INTERNATIONAL HANDBOOK ON THE ECONOMICS OF MIGRATION Acclaim for the International Handbook on the Economics of Migration ‘Constant and Zimmermann have assembled a collection of essays that is remarkable in one extremely important way: it integrates many novel research topics into the mainstream immigration literature, including ethnic hiring patterns, obesity, the economic consequences of interethnic marriages, the link between natural disasters and migration, immigrant time use, and the relationship between migration and happiness These survey papers are destined to become beacons for future researchers as each of these topics will inevitably receive much more attention in future research.’ – George Borjas, Harvard University, USA ‘This is an extremely impressive volume which guides readers into thinking about migration in new ways. In its various chapters, international experts examine contemporary migration issues through a multitude of lenses ranging from child labor, human trafficking and jobs to the political economy of migration and refugees. The result is a fascinating assessment of the role of migration in driving population change in the modern age This will surely serve as a reference volume for those interested in migration for years to come.’ – Deborah Cobb-Clark, Melbourne Institute of Applied Economic and Social Research, Australia ‘A comprehensive, truly encyclopedic collection of original surveys and essays discussing migration and topics related to the movement of people among countries and areas The studies both present and review the literature critically and in many cases offer new results The basic theory is laid out right from the start, providing a nice introduction and framework for the other 27 chapters While most are interesting and worth reading, as a novice in the field of migration I found the essays on human smuggling and natural disasters to be particularly enlightening and important I can recommend this Handbook to any labor economist or sociologist with a scholarly interest, either for research or for instruction, in this general area The volume is definitive.’ – Daniel S Hamermesh, University of Texas at Austin, USA and Royal Holloway, University of London, UK ‘As immigration has spread from traditional receiving nations to developed countries throughout the world, the economics of migration has become a burgeoning field of research Amelie Constant and Klaus Zimmermann’s International Handbook offers an excellent, state-of-the-art guide to the rapidly changing intellectual terrain, providing comprehensive coverage of the topics necessary to comprehend patterns and processes of migration in the world today It will be an indispensable guide to scholars and policy-makers for years to come.’ – Douglas S Massey, Princeton University, USA ‘The International Handbook on the Economics of Migration is an excellent book that broadens our understanding of the economics of migration It covers classic issues related to immigration such as labor market integration and wages as well as much newer and less explored aspects of it, such as happiness, religiosity and crime I commend Constant and Zimmermann for gathering an excellent team of young and more experienced scholars, and for producing a book that will become an important reference in teaching and learning about immigration.’ – Giovanni Peri, University of California, Davis, USA International Handbook on the Economics of Migration Edited by Amelie F Constant George Washington University and Temple University, USA and IZA, Bonn, Germany Klaus F Zimmermann IZA and Bonn University, Bonn, Germany Edward Elgar Cheltenham, UK Northampton, MA, USA â Amelie F Constant and Klaus F Zimmermann 2013 All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc William Pratt House Dewey Court Northampton Massachusetts 01060 USA A catalogue record for this book is available from the British Library Library of Congress Control Number: 2012955223 This book is available electronically in the ElgarOnline.com Economics Subject Collection, E-ISBN 978 78254 607 ISBN 978 84542 629 01 Typeset by Servis Filmsetting Ltd, Stockport, Cheshire Printed and bound in Great Britain by T.J International Ltd, Padstow Contents vii List of contributors Frontier issues in migration research Amelie F Constant and Klaus F Zimmermann PART I  INTRODUCTION   Migration and ethnicity: an introduction Amelie F Constant and Klaus F Zimmermann 13 PART II  THE MOVE           Modeling individual migration decisions John Kennan and James R Walker The economics of circular migration Amelie F Constant, Olga Nottmeyer and Klaus F Zimmermann The international migration of health professionals Michel Grignon, Yaw Owusu and Arthur Sweetman Independent child labor migrants Eric V Edmonds and Maheshwor Shrestha Human smuggling Guido Friebel and Sergei Guriev 39 55 75 98 121 PART III  PERFORMANCE AND THE LABOR MARKET       10 11 12 13 Labor mobility in an enlarged European Union Martin Kahanec Minority and immigrant entrepreneurs: access to financial capital Robert W Fairlie Migrant educational mismatch and the labor market Matloob Piracha and Florin Vadean Ethnic hiring David Neumark Immigrants in risky occupations Pia M Orrenius and Madeline Zavodny Occupational sorting of ethnic groups Krishna Patel, Yevgeniya Savchenko and Francis Vella Immigrants, wages and obesity: the weight of the evidence Susan L Averett, Laura M Argys and Jennifer L Kohn v 137 153 176 193 214 227 242 vi   Contents PART IV  NEW LINES OF RESEARCH 14 Immigrants, ethnic identities and the nation-­state Amelie F Constant and Klaus F Zimmermann 15 Interethnic marriages and their economic effects Delia Furtado and Stephen J Trejo 16 The impact of migration on family left behind Francisca M Antman 17 Natural disasters and migration Ariel R Belasen and Solomon W Polachek 18 Immigration–religiosity intersections at the two sides of the Atlantic: Europe and the United States Teresa García-­Moz and Shoshana Neuman 19 Immigration and crime Brian Bell and Stephen Machin 20 Immigrants’ time use: a survey of methods and evidence David C Ribar 21 Happiness and migration Nicole B Simpson 259 276 293 309 331 353 373 393 PART V  POLICY ISSUES 22 23 24 25 26 27 28 Frontier issues of the political economy of migration Gil S Epstein Skill-­based immigrant selection and labor market outcomes by visa category Abdurrahman Aydemir Refugee and asylum migration Timothy J Hatton The economics of immigrant citizenship ascension Don J DeVoretz Welfare migration Corrado Giulietti and Jackline Wahba Diaspora resources and policies Sonia Plaza The evaluation of immigration policies Ulf Rinne Name index Subject index 411 432 453 470 489 505 530 553 561 Contributors Francisca M Antman  University of Colorado at Boulder, USA and Institute for the Study of Labor (IZA), Bonn, Germany Laura M Argys  University of Colorado Denver, USA and Institute for the Study of Labor (IZA), Bonn, Germany Susan L Averett  Lafayette College, USA and Institute for the Study of Labor (IZA), Bonn, Germany Abdurrahman Aydemir  Sabanci University, Turkey and Institute for the Study of Labor (IZA), Bonn, Germany Ariel R Belasen  Southern Illinois University Edwardsville, USA Brian Bell  Department of Economics, University of Oxford and Centre for Economic Performance, London School of Economics, UK Amelie F Constant  George Washington University and Temple University, USA and Institute for the Study of Labor (IZA), Bonn, Germany Don J DeVoretz  Simon Fraser University, Canada and Institute for the Study of Labor (IZA), Bonn, Germany Eric V Edmonds  Dartmouth College, USA and Institute for the Study of Labor (IZA), Bonn, Germany Gil S Epstein  Department of Economics, Bar-­Ilan University, Israel and Institute for the Study of Labor (IZA), Bonn, Germany Robert W Fairlie  University of California, Santa Cruz, USA and Institute for the Study of Labor (IZA), Bonn, Germany Guido Friebel  Goethe University Frankfurt, Germany, Centre for Economic Policy Research (CEPR), UK and Institute for the Study of Labor (IZA), Bonn, Germany Delia Furtado  University of Connecticut, USA and Institute for the Study of Labor (IZA), Bonn, Germany Teresa García-­Moz  University of Granada, Spain Corrado Giulietti  Institute for the Study of Labor (IZA), Bonn, Germany Michel Grignon  McMaster University, Hamilton, ON, Canada Sergei Guriev  New Economic School Moscow, Russia and Centre for Economic Policy Research (CEPR), UK Timothy J Hatton  University of Essex, UK and Australian National University, Australia and Institute for the Study of Labor (IZA), Bonn, Germany vii viii   Contributors Martin Kahanec  Central European University, Hungary, Institute for the Study of Labor (IZA), Bonn, Germany and Central European Labour Studies Institute (CELSI), Slovakia John Kennan  University of Wisconsin-­Madison, USA, NBER, USA and Institute for the Study of Labor (IZA), Bonn, Germany Jennifer L Kohn  Drew University, USA Stephen Machin  Department of Economics, University College London and Centre for Economic Performance, London School of Economics, UK and Institute for the Study of Labor (IZA), Bonn, Germany Shoshana Neuman  Bar-­Ilan University, Israel, Institute for the Study of Labor (IZA), Bonn, Germany and Centre for Economic Policy Research (CEPR), UK David Neumark  UCI Department of Economics and Center for Economics and Public Policy, NBER, USA and Institute for the Study of Labor (IZA), Bonn, Germany Olga Nottmeyer  Institute for the Study of Labor (IZA), Bonn, Germany Pia M Orrenius  Federal Reserve Bank of Dallas, USA and Institute for the Study of Labor (IZA), Bonn, Germany Yaw Owusu  McMaster University, Hamilton, ON, Canada Krishna Patel  Federal Deposit Insurance Corporation, Division of Insurance and Research, USA Matloob Piracha  University of Kent, UK and Institute for the Study of Labor (IZA), Bonn, Germany Sonia Plaza  World Bank and Institute for the Study of Labor (IZA), Bonn, Germany Solomon W Polachek  State University of New York at Binghamton, USA and Institute for the Study of Labor (IZA), Bonn, Germany David C Ribar  University of North Carolina at Greensboro, USA and Institute for the Study of Labor (IZA), Bonn, Germany Ulf Rinne  Institute for the Study of Labor (IZA), Bonn, Germany Yevgeniya Savchenko  Georgetown University, USA Maheshwor Shrestha  Massachusetts Institute of Technology, USA Nicole B Simpson  Colgate University, USA and Institute for the Study of Labor (IZA), Bonn, Germany Arthur Sweetman  McMaster University, Hamilton, ON, Canada and Institute for the Study of Labor (IZA), Bonn, Germany Stephen J Trejo  The University of Texas at Austin, USA and Institute for the Study of Labor (IZA), Bonn, Germany Florin Vadean  University of Kent, UK Contributors  ­ix Francis Vella  Georgetown University, USA and Institute for the Study of Labor (IZA), Bonn, Germany Jackline Wahba  University of Southampton, UK and Institute for the Study of Labor (IZA), Bonn, Germany James R Walker  University of Wisconsin-­Madison, NBER, USA and Institute for the Study of Labor (IZA), Bonn, Germany Madeline Zavodny  Agnes Scott College, USA and Institute for the Study of Labor (IZA), Bonn, Germany Klaus F Zimmermann  Institute for the Study of Labor (IZA) and Bonn University, Bonn, Germany Frontier issues in migration research Amelie F Constant and Klaus F Zimmermann With the inescapable progress of globalization, labor markets are bound to become more integrated The impending demographic disruptions will set in with full force in many countries within the coming years Climate change, natural disasters and the rise of the BIC countries (Brazil, India, China) will pose additional labor market challenges Ethnic diversity will continue to gain importance – as both an opportunity and a threat All of these will eventually require a global reallocation of resources, which will force international and domestic labor markets to undergo major adjustment processes The strong demand for skilled workers along with the fight against extreme economic inequality, the creation of ‘good’ jobs, and the increased employment of specific groups such as the young, older, female, low-­skilled and ethnic minority workers will need scientific monitoring and evaluation, in order to initiate necessary adjustment processes and labor market programs in time Therefore, migration economics is a fast-­growing and exciting research area with very significant and rising policy relevance While its scope is extending persistently, there is no adequate authoritative treatment of its various branches in one volume The new International Handbook on the Economics of Migration (IHEM) goes beyond providing basic information on migration It offers the latest experiences on migration research and tackles frontier issues in the field It provides comprehensive guidance to economics scholars, inquiring researchers, students of migration and policy advisers This handbook is a carefully commissioned and refereed compilation of 28 state-­of-­the-­art chapters of research in the economics of migration written by 44 leading experts in the field together with this introduction Well-­written and simply explained, each chapter comprises a critical assessment of the status quo and provides challenges to the traditional economics of migration by dealing also with taboo topics The IHEM systematically and tactically covers all relevant frontier issues on migration It deals with innovations in the modeling of migration, with the determinants of migration such as natural disasters, refugee and asylum seeking, and the welfare magnet, including child labor migration, human smuggling, the international move of health professionals and labor mobility in the enlarged European Union Other chapters study the consequences of migration for happiness, obesity, religiosity, crime, citizenship ascension, ethnic hiring, employment in risky occupations, occupational sorting and migrant educational mismatch The IHEM also covers the economic reflections and empirical findings on ethnicity and integration, such as immigrant entrepreneurship, inter-­ethnic marriages and immigrants’ time use Lastly, the IHEM tackles specific issues of policy relevance such as the impact of migration on the family left behind, immigrant selection by visa category, circular migration, diaspora policies, evaluation techniques for migration policies and the political economy of migration The IHEM is structured in five parts: ‘Part I: Introduction’, ‘Part II: The move’, ‘Part III: Performance and the labor market’, ‘Part IV: New lines of research’ and, finally, 2   International handbook on the economics of migration ‘Part V: Policy issues’ Following this introduction, some core knowledge of migration research is presented in the chapter by Amelie F Constant and Klaus F Zimmermann on ‘Migration and ethnicity: an introduction’ This chapter deals with the economic and ethnic diversity caused by international labor migration, and their economic integration possibilities It brings together three strands of literature dealing with the neoclassical economic assimilation, ethnic identities and attitudes towards immigrants and the natives, and provides analysis to understand their interactions The issue of how immigrants fare in the host country, especially in terms of their labor force participation and remuneration, has been the core of research in the labor migration literature If immigrants fare as well as the natives, then they are economically assimilated While some immigrant groups do, most not, especially in Europe Of equal importance is how immigrants identify with the culture of their home and host countries, and if natives and immigrants have the right attitudes about each other Ethnic identities and attitudes seem to be less affected by the economic environment but have implications for economic performance ‘Part II: The move’ deals with the migration decision and migratory flows The first chapter by John Kennan and James R Walker on ‘Modeling individual migration decisions’ sets the stage for modeling the migration decision It summarizes recent research that formulates life-­cycle models of migration which are estimated using longitudinal data These models consider multiple destinations and multiple periods The framework offers a unified view applicable to internal and international migration flows However, data limitations severely hinder studies of international migration As is common in modeling life-­cycle decision-making, strong assumptions are imposed Yet, most critical assumptions are empirically testable The primary advantage is that these models offer an interpretable economic framework for evaluating policy alternatives and other counterfactual thought–experiments that offer insight on behavioral determinants and tools for improved policy-­making The second chapter by Amelie F Constant, Olga Nottmeyer and Klaus F Zimmermann deals with ‘The economics of circular migration’, an issue that has generated keen interest by researchers and policy-­makers alike For too long, migratory movements have been considered to be mostly permanent, an evaluation that has never been right and is increasingly accepted as an incorrect description of labor migration Temporary, return, repeat or circular migration have become the keywords of the new migration research This type of migration presents more challenges for modeling and predicting migration patterns, as well as for migration policies The chapter presents a review of the empirical evidence, outlines implications for policy and summarizes the policies to manage circular migration Given the rising scarcity of skilled workers, skilled migration receives much more attention The chapter by Michel Grignon, Yaw Owusu and Arthur Sweetman on ‘The international migration of health professionals’ is, therefore, particularly timely Health workforce shortages in developed countries are perceived to be central drivers of the health professionals’ international migration, one ramification being negative impacts on developing nations’ health-­care delivery After a descriptive international overview, the authors discuss selected economic issues for both developed and developing countries Health labor markets’ unique characteristics imply great complexity in developed economies involving government intervention, licensure, regulation and (quasi-­)union activity These features affect migrants’ decisions and their economic integration, and Frontier issues in migration research  ­3 impact on the receiving nations’ health workforce and society Developing countries sometimes educate citizens in expectation of emigration, while others pursue international treaties in attempts to manage migrant flows The next two chapters consider the dark side of migratory moves and deal with child labor migration and human smuggling The chapter by Eric V Edmonds and Maheshwor Shrestha investigates the situation of ‘Independent child labor migrants’ Children living and working away from home are the most vulnerable in our societies Parents, family, friends and home communities offer protection that can reduce a child’s susceptibility to abuse and exploitation, as well as alleviate the consequences of bad or poorly informed decisions This chapter reviews the nascent literature on the prevalence, causes and consequences of independent child labor migration Measurement challenges have constrained progress in understanding this phenomenon There is considerable scope for future research to transform how we think about issues related to the millions of children living and working away from their parents Guido Friebel and Sergei Guriev undertake the complex and thorny case of migration, that of ‘Human smuggling’ Despite its importance and prevalence in global illegal migration, there is little – and mostly theoretical – research on human smuggling The authors suggest an analytical framework to understand the micro structure of the human smuggling market Migrants interact with smuggling and financing intermediaries, who may or may not be integrated with each other, and with the migrants’ employers Migration policies in the receiving countries such as border controls, employer sanctions, deportation policies and sales of visas strongly affect the interactions in the smuggling market and, hence, migration flows and the surfacing of illegal immigrants The chapter reviews the theoretical work, points to the scarce empirical evidence, and identifies challenges for future theoretical, empirical work and policy advice ‘Part III: Performance and the labor market’ contains seven chapters covering migrant and minority performance and the labor market consequences of mobility In the first chapter, Martin Kahanec presents a landmark labor migration in the European history ‘Labor mobility in an enlarged European Union’ is about the 2004 and 2007 enlargements of the European Union (EU) that extended the freedom of movement to workers from the 12 new member states mainly from central eastern Europe This chapter summarizes and comparatively evaluates what we know about mobility in an enlarged Europe to date The pre-­enlargement fears of free labor mobility proved to be unjustified No significant detrimental effects on the receiving countries’ labor markets have been documented, nor has there been any welfare shopping Rather, there appear to be positive effects on the EU’s productivity While the sending countries face some risks of losing their young and skilled labor force, they have also been relieved of some redundant or idle labor and associated fiscal burdens, as well as having profited from remittances sent back by migrants Of key importance for the sending countries is to reap the benefits from brain gain and brain circulation in the enlarged EU For the migrants the benefits in terms of better career prospects have, with little doubt, exceeded any pecuniary and non-­pecuniary costs of migration Consequently, the freedom of labor movement in the EU provided for a triple-­win situation for the receiving and sending countries as well as for migrants themselves Self-­employment is viewed as a key strategy to survive economically and even 4   International handbook on the economics of migration f­ lourish for migrants and minorities Robert W Fairlie’s chapter deals with ‘Minority and immigrant entrepreneurs: access to financial capital’ Reviewing existing research, the author indicates that inadequate access to financial capital, partly owing to wealth inequality, restricts the creation and growth of minority-­owned businesses Access to financial capital is thus essential for entrepreneurial success There is less evidence on access to financial capital among immigrant-­owned businesses New estimates from the US Census Bureau indicate that immigrant-­owned businesses start with higher levels of capital than non-­immigrant owned businesses The most common source of start-­up capital for immigrant firms is from personal or family savings, which is similar for ­non-­immigrant firms Immigrants have relatively low rates of home ownership, however, which may partly limit business formation The next chapter, by Matloob Piracha and Florin Vadean, investigates ‘Migrant educational mismatch and the labor market’ This chapter reviews the literature on the educational mismatch of immigrants in the host country labor market It draws on the theoretical arguments postulated in the labor economics literature and discusses their extension in the analysis of the causes and effects of immigrants’ educational mismatch in the destination country The authors also present relevant empirical approaches, which show that immigrants are in general more over-­educated than natives and the reasons for these findings range from imperfect transferability of human capital to discrimination to perhaps lack of innate ability Lastly, they assess the state of current literature and propose an agenda for further research The chapter on ‘Ethnic hiring’ by David Neumark deals with discrimination, spatial mismatch and networks which may pose barriers to employment Widespread evidence of ethnic discrimination from audit or correspondence studies may be questionable because these studies may not identify discrimination Application of a new method that ­identifies discrimination is needed to reassess this evidence Recent evidence discounts spatial mismatch as an important contributor to the low employment of minorities in the USA; living in an area with many jobs does not help minorities if these jobs are held by other groups Ethnically stratified networks may explain this evidence, although ethnic networks may also help minorities connect to labor markets Pia M Orrenius and Madeline Zavodny follow with their study ‘Immigrants in risky occupations’ The chapter reviews the economics literature on immigrant–native differentials in occupational risk It begins by briefly explaining the theory of compensating wage differentials, and then provides a more detailed discussion of the empirical evidence on the subject, which reaches several conclusions First, immigrants are overrepresented in occupations and industries with higher injury and fatality rates Second, immigrants have higher work-­related injury and fatality rates in some advanced economies, but not in all Finally, most, but not all, immigrants appear to earn risk premiums similar to natives for working in risky jobs The chapter closes with a discussion of areas where additional research is needed ‘Occupational sorting of ethnic groups’ is the next chapter, by Krishna Patel, Yevgeniya Savchenko and Francis Vella The chapter discusses research on immigrant occupational sorting in the destination country, and how immigrant occupational outcomes depend on both the demand for skill and the supply of immigrant skill On the demand side, immigration policies in the destination countries affect the degree to which immigrants are suitably matched in their occupation On the supply side, immigrant occupational sorting depends on factors such as experience in the Frontier issues in migration research  ­5 home country and the skill transferability in the country of relocation Social networks also play an important role in the job search and matching process for immigrants, and ­influence their occupational placement The final chapter in Part III is on ‘Immigrants, wages and obesity: the weight of the evidence’ written by Susan L Averett, Laura M Argys and Jennifer L Kohn In this novel study the authors integrate disparate literatures on the effect of immigration on obesity and the effect of obesity on labor market outcomes Their review finds support for the ‘healthy immigrant’ hypothesis: immigrants are less likely to be obese, but obesity increases with duration in their new home There is conflicting evidence on the causal effect of obesity on labor market outcomes for immigrants and non-­immigrants alike Only two existing studies examine the dual effects of immigration and obesity Researchers need more complete data to address endogeneity concerns and assess the causal effects of immigration and obesity on labor market outcomes Part IV of the handbook deals with ‘New lines of research’ The first chapter by Amelie F Constant and Klaus F Zimmermann deals with ‘Immigrants, ethnic identities and the nation-­state’ Concepts of individual and group identities have become increasingly relevant in economics following the pace of other disciplines Migrants, minorities and natives have their own identity which differs from their national identities The chapter outlines the non-­economic roots of ethnic and national identities, and discusses the relationship with religious and social identities The authors introduce a model of identity formation and review the empirical findings concerning ethnic identity formation They then present and discuss the available data and the results of the relevant literature for several countries The second chapter by Delia Furtado and Stephen J Trejo reviews ‘Interethnic marriages and their economic effects’ Immigrants who marry outside of their ethnicity tend to have better economic outcomes than those who marry within their ethnicity It is difficult, however, to interpret this relationship because individuals with stronger preferences for ethnic endogamy are likely to differ in unobserved ways from those with weaker preferences To clarify some of the complex issues surrounding inter-­ethnic marriages and assimilation, this chapter starts by considering the determinants of intermarriage It proceeds with an examination of the economic consequences of intermarriage, and ends with a discussion of the links between intermarriage, ethnic identification and measurement of long-­term socio-­economic integration Francisca M Antman undertakes the study of the often forgotten family of the migrant in the home country ‘The impact of migration on family left behind’ addresses the effects of migration on families left behind and offers new evidence on the impact of migration on elderly parents After discussing the identification issues involved in the estimation, the chapter reviews the literature on the effects of migration on the education and health of non-­migrant children as well as the labor supply of non-­migrant spouses Finally, it discusses the impact of adult child migration on contributions toward non-­migrant parents as well as on the effects on parental health Results show that elderly parents receive lower time contributions from all of their children when one child migrates In their chapter, ‘Natural disasters and migration’, Ariel R Belasen and Solomon W Polachek make a case about the intrinsic link between man and the environment Since the dawn of civilization man has battled with environmental disasters, from massive hurricanes and tsunamis to slow, yet persistent, soil erosion and climate change When the environment wins, thousands are displaced and forced to emigrate from their 6   International handbook on the economics of migration homes The chapter presents a three-­pronged approach examining the impact of environmental disasters on migration: first, a literature survey; second, a meta-­analysis based on this literature, and third, it puts forward new techniques isolating the marginal impact of disasters on migration The chapter finds stronger impacts in developing countries, particularly contingent upon whether the affected populace is in an urban or rural setting The chapter by Teresa García-­Moz and Shoshana Neuman investigates ‘Immigration–religiosity intersections at the two sides of the Atlantic: Europe and the United States’ In this avant-­garde chapter, they explore the intertwined relationship between immigration and religiosity in Europe and the USA Starting with (1) the current religious landscape and projections for the future, they continue with (2) the religiosity of immigrants compared to natives, and they move on with (3) the religiosity of immigrants and their integration; the relevant question being, is religiosity a ‘bridge’ or a ‘buffer’? The authors lastly compare the two continents of Europe and the USA The main conclusions are that: immigrants are indeed more religious than the local ­populations, leading to major changes in the future religious landscapes; and while in the USA the religiosity of immigrants serves as a ‘bridge’, in Europe it has mainly the function of a ‘buffer’ Brian Bell and Stephen Machin provide in the chapter on ‘Immigration and crime’ a highly politicized link The authors examine first the economic literature on the links between immigration and crime In spite of popular concern, there is only a sparse literature on the topic After discussing some simple predictions from an economics of crime model, they review the extant empirical evidence While causal effects are difficult to identify, the evidence points to the importance of focusing on the labor market attachment and earnings opportunities of different immigrant groups Those groups that are disadvantaged across this dimension tend to be associated with rises in property crime There appears to be no significant links between immigrants and violent crime David C Ribar authors another frontier chapter about ‘Immigrants’ time use: a survey of methods and evidence’ This chapter discusses research questions related to immigrants’ time use, reviews conceptual and methodological approaches to examining time allocations, and reviews evidence from previous studies Using time-­diary data from the American Time Use Survey, the chapter also provides new descriptive evidence While results vary with the country of origin, immigrant men in the USA tend to devote more time to market work and sleeping; they allocate less time to housework, community activities and leisure than native men Immigrant women tend to devote more time to housework, care-­giving and sleep, but less time to market work, community activities and leisure than native women The last of the cutting-­edge chapters in this part is ‘Happiness and migration’ by Nicole B Simpson This chapter explores the various channels in which happiness and migration are related Happiness may be important in the decision to migrate, but migration may also affect happiness, and specifically the happiness of the migrants, the natives in the destination and non-­migrants back home Existing literature indicates that migration increases the happiness of the migrants but migrants are generally less happy than natives in the destination There is considerable heterogeneity documented in the happiness of migrants across origin and destination countries and in migration duration Despite a recent surge in work on the topic, several unexplored areas of research remain ‘Part V: Policy issues’ of the handbook starts with the chapter on ‘Frontier issues of Frontier issues in migration research  ­7 the political economy of migration’ written by Gil S Epstein Migration has a strong economic impact on the sending and host countries Since individuals and groups not benefit equally from migration, interest groups emerge to protect and take care of their narrow self-­interests and compete for rents generated by migration Narrow self-­ interests may be present not only for interest groups but also for ruling politicians and civil servants This chapter considers how political culture is important for determining policy and how interest groups affect, via a lobbying process, the choice of public policy The chapter lastly analyzes how interest groups and lobbying activities affect assimilation and attitudes towards migrants and international trade The narrow interests of the different groups may cause a decrease in social welfare, in some cases, and may enhance welfare in other situations Immigrant selection, political migration and citizenship ascension are the topics of the next three chapters dealing with significant policy issues Attracting skilled immigrants is emerging as an important policy goal for immigrant receiving countries In his chapter ‘Skill-­based immigrant selection and labor market outcomes by visa category’ Abdurrahman Aydemir first discusses the economic rationale for immigrant selection The author reviews selection mechanisms of the receiving countries in the context of deteriorating labor market outcomes for immigrants across destination countries which fuels the debate on selection Next, he discusses the variation in immigrant characteristics across countries and visa types Lastly, he reviews the evidence on labor market outcomes of immigrants by visa category that portrays the experiences of countries with different selection mechanisms He concludes by underlining the challenges for realizing aimed benefits of a skill-­based immigrant selection policy Timothy J Hatton deals with another hot migration topic, ‘Refugee and asylum migration’ He provides an overview of asylum migration from poor strife-­prone countries to the Organisation for Economic Co-­operation and Development (OECD) since the 1950s and examines the political and economic factors in source countries that generate refugees and asylum seekers Particular attention is given to the rising trend of asylum applications up to the 1990s, and the policy backlash that followed The chapter then considers the political economy of restrictive asylum policies, especially in EU countries, as well as the effectiveness of those policies in deterring asylum seekers It concludes with an outline of the assimilation of refugees in host country labor markets ‘The economics of immigrant citizenship ascension’ by Don J DeVoretz observes that naturalized immigrants often receive an earnings premium after obtaining citizenship It is argued that the size of this ‘citizenship premium’ varies across immigrant receiving countries and the immigrants themselves; in conjunction with the cost of obtaining citizenship this premium determines the differential rates of citizenship ascension The size of the premium obtained by ‘Old World and New World’ naturalized immigrants is a consequence of positive discrimination in the labor market for naturalized immigrants and a by-­product of their human capital accumulation prior to citizenship ascension The largest economic premium from naturalization accrues under a ‘triple selection’ regime where economic immigrants self-­select on an economic basis to migrate to a country with stringent economic entry and ­citizenship acquisition criteria The chapter on ‘Welfare migration’ by Corrado Giulietti and Jackline Wahba reviews and discusses major theories and empirical studies about the welfare magnet hypothesis, that is, whether immigrants are more likely to move to countries with 8   International handbook on the economics of migration generous welfare systems Although economic theory predicts that welfare generosity affects the number, composition and location of immigrants, the empirical evidence is rather mixed The chapter offers explanations for the existence of such mixed evidence and highlights that the literature so far has overlooked the presence of different migration regimes, as well as the possibility of reverse causality between welfare spending and immigration Sonia Plaza further studies ‘Diaspora resources and policies’ suggesting that migration presents significant untapped potential for development Globalization makes it possible for immigrants to remain connected with their native countries while residing abroad, thus diminishing their loss of identity and separation from their countries of origin The contribution of the diaspora goes beyond remittances and includes promotion of trade, investments, knowledge and technology transfers Diasporas facilitate bilateral trade and investment flows between their country of residence and their home country Diaspora members can also act as catalysts for the development of capital markets in their countries of origin by diversifying the investor base, by introducing new financial products and by providing reliable sources of funding, such as diaspora bonds Diasporas my also provide origin-­country firms access to technology and skills In recent years there has been a shift in the analysis of high-­skilled migration Instead of viewing the emigration of skilled people as a loss, many economists view it as an opportunity to increase trade, investment and technology flows This chapter covers a diverse range of diaspora issues and provides a number of analytical and empirical results that are relevant for policy-­makers in both developed and developing countries Ulf Rinne provides a chapter on the under researched area ‘The evaluation of immigration policies’ summarizing the literature on the evaluation of immigration policies The chapter brings together two strands of the literature dealing with the evaluation of labor market programs and with the economic integration of immigrants Next to immigrant selection and settlement policies, there are four types of interventions that aim at improving the economic and social outcomes of immigrants: (1) introduction programs, (2) language training, (3) labor market programs, and (4) anti-­discrimination policies The chapter discusses problems associated with the evaluation of such programs, presents methodological approaches to circumvent these problems, and surveys empirical results and findings It concludes with lessons from previous research and identifies avenues for future research An endeavor such as a handbook cannot be successfully undertaken without the devoted support of many people This includes the 44 authors of the chapters and the many experts who have provided excellent anonymous referee reports as well as editorial support: Olivier Bargain, Brittany Bauer, Costanza Biavaschi, Marco Caliendo, John Cawley, Nancy H Chau, Deborah A Cobb-­Clark, Amelie F Constant, Horst Entorf, René Fahr, Denis Fougère, Martin Guzi, Dan Hamermesh, Gaby Herbrig, Jasmin Kantarevic, Annabelle Krause, Steffen Künn, Evelyn L Lehrer, Marco Manacorda, Kostas Mavromaras, David McKenzie, Olga Nottmeyer, Margard Ody, Ulf Rinne, Regina T Riphahn, Ralph Rotte, Sabrina Pabilonia, Maurice Schiff, Zahra Siddique, Erdal Tekin, Bienvenue Tien, Marie-­Anne Valfort, Nicolas R Ziebarth and Klaus F Zimmermann The editorial work has been done at the various stages together and alone at the premises of IZA, Bonn and DIW DC, Washington The perfect working conditions in both institutions provided us with the necessary environment and support Frontier issues in migration research  ­9 to foster this project Finally, we gratefully acknowledge all the encouragement and support provided by the publisher, Edward Elgar, and his staff, including Alex Pettifer, Alexandra O’Connell, Laura Seward and Caroline Cornish PART I INTRODUCTION 1  Migration and ethnicity: an introduction* Amelie F Constant and Klaus F Zimmermann 1  INTRODUCTION Migration as ‘factor mobility’ and migrants as a ‘factor of production’ are of paramount importance in economics The different skills and education that are embodied in immigrants, while valuable in the production process, may not be appreciated by all members of the host country In addition, migrants as human beings are an integral part of the human development in a society and a country Yet, resistance to the spreading of diversity and concerns about the growth of the immigrant population from several groups make immigrants feel unwanted The imbroglio of migration touches and raises problems in the social, economic, political, cultural and religious spheres not only domestically, but also internationally Migration scholars, pundits and policymakers alike are deeply divided over the responsibilities and the best concepts for analyzing or solving the issue of international migration The issue of how immigrants fare in the host country especially in terms of their labor force participation and remuneration occupies the minds of social scientists, politicians and the general public Using the natives as the gold standard, immigrants have been compared to them If immigrants fare as well as the natives, then they are economically assimilated Of equal importance is the question of whether immigrants socialize and mingle with the natives, if they feel comfortable in their new country or they create parallel societies, and if natives and immigrants have the right attitudes about each other Terms such as cultural or social assimilation, acculturation, integration, and so on, have been used to capture and describe these concerns This chapter focuses on economic migrants, that is, individuals who leave their country and loved ones to go abroad to a new country in search of job and other economic opportunities to better their and their children’s lives We first review the economic status quo theories on immigrant performance dealing with what is sometimes called economic assimilation: how migrants become like natives in economic terms? We then present recent advances in economics about the formation of ethnic identity and its role in the economic and social spheres: how identities shape and how are they related to economic success? We finally discuss the importance of attitudes and perceptions in the integration process: are they affected by economic conditions and they influence economic performance? The chapter is designed as an introduction to the core issues of migration research We neither attempt to cover all relevant basic knowledge nor we discuss most of the recent advances in the field, which is the purpose of the other chapters in this volume 13 14   International handbook on the economics of migration 2  THE ECONOMICS OF ASSIMILATION Starting with the pioneering work of Chiswick (1978) on the assimilation of immigrant men in the United States (US), the overarching research that has preoccupied the literature deals with the economic performance of immigrants relative to that of comparable natives The literature is set within the Mincerian human capital framework,1 whereby immigration is perceived as an investment in human capital (Sjaastad, 1962), the young and the better educated are more likely to migrate and migration yields higher returns to the more able and the more highly motivated; assimilation is a labor market phenomenon.2 The conjecture is that immigrants are rational individuals who want to maximize their lifetime utility; they are a self-­selected group of individuals characterized by a strong incentive to invest in human capital and have a ferocious drive to succeed in the host country’s labor market They have set preferences that they reveal in a rational ranking order Migrants with higher levels of human capital will command higher wages in the labor market since investment in human capital raises productivity Chiswick’s (1978) hypothesis, as well as that of many others who followed his lead, was that the earnings of newly arrived immigrants are significantly lower than those of natives with the same observed socioeconomic characteristics, mainly because immigrants’ skills are not always or perfectly transferable to the host country’s labor market However, as immigrants gain information about the functioning of the new labor market and invest in human capital in the new country, their earnings increase rapidly and can reach and even exceed the earnings of natives When the catching up of earnings occurs, then economic assimilation is achieved, meaning that immigrants and natives are ­indistinguishable in terms of their earnings Therefore, assimilation is the rate at which the earnings of immigrants converge to the earnings of comparable natives due to their accumulation of human capital in the host country’s labor market with additional years of residence (Chiswick, 1978).3 Assimilation is attributed to the positive selection of immigrants, that is, their innate ability, their high motivation for labor market success and their higher incentives to invest in host country’s specific human capital Indeed, this generation of studies4 found that immigrant earnings reach parity with native earnings within 10 years of residence, and after 10 years, immigrant earnings exceed the earnings of natives The main drawback of these studies was that the models were estimated based on a single cross-­section of data that includes individuals from all ages A new generation of studies was ignited by Borjas’s (1985)5 seminal paper which questioned the empirical validity of the above results from cross-­section data on the grounds that the assimilation effects were confounded with cohort effects That is, based on one cross-­section, the estimated earnings of immigrants of different ages are overstated if the quality of more recent immigrant cohorts is lower than that of older cohorts Borjas (1985) attempted to estimate the selection bias which may contaminate cross-­sectional comparisons and to establish a relationship between cohort quality and immigrant self-­selection Borjas’s contribution was to track the progress of a particular cohort over successive waves of cross-­sectional data and to identify cohort and assimilation effects by creating synthetic cohorts.6 Borjas and subsequent research suggested that immigrants in the US were not necessarily positively selected As a result, and despite the fact that earnings increase with additional years of residence, immigrants may not assimilate as rapidly as Migration and ethnicity: an introduction  ­15 the traditional view hypothesized, and the earnings of more recent cohorts may never reach parity with the earnings of natives Meanwhile, other researchers demonstrated that the age of immigrants at the time of arrival in the new host country plays a decisive role in their earnings assimilation Indeed, the profiles of those migrating as children resemble the profiles of the native-­born rather than the profiles of immigrants migrating later in life (Friedberg, 1992; Kossoudji, 1989) Assimilation for these immigrants is, therefore, not a labor market phenomenon but the result of acculturation.7 Afterwards, most studies agreed that the assimilation process is very slow and the earnings of male immigrants will probably never reach parity with natives The declining skills of more recent immigrant cohorts (within cohort differences), as well as the changing national origin composition of immigrants (across cohort differences), hold back assimilation (Borjas, 1992; Chiswick, 1986) Some disparity in these findings was documented in other studies LaLonde and Topel’s (1991, 1992) different results are due, however, to the different variables chosen and to the different comparison group – whether is intra-­ethnic or ethnic-­native They found that the assimilation of immigrants is mainly intergenerational and that estimates are sensitive to the choice of the base group Yet, all these studies are subject to additional biases related to the comparability of the samples gathered across decennial censuses Extra selection biases exist because of the highly selective return migration, which was overlooked in the estimation of earnings assimilation In theory, return migration is non-­ random and depends on immigrants’ performance in the host country’s labor market, whether successful or failing Assimilation estimates based on the pool of stayers will be under-­ or over-­estimated depending on whether or not the successful immigrants emigrate Empirical studies can answer these possibilities Some find higher return migration by skilled immigrants (Jasso and Rosenzweig, 1988), others by less successful immigrants (Borjas, 1989), while others find little evidence of any selectivity with respect to schooling (Chiswick, 1986) Constant and Massey (2003) in their 14-­year longitudinal study on immigrants in Germany find that emigrants are negatively selected with respect to occupational prestige and to stable full-­time employment, but no selectivity with respect to human capital or gender Return migration is strongly determined by the range and nature of social attachments to Germany and origin countries It is also bimodal, that is, very high during the first five years from arrival, and grows higher again toward retirement Selective emigration, however, does not appear to distort cross-­sectional estimates of earnings assimilation in a relevant way.8 Finally, selection with respect to labor force participation, occupational attainment, labor market success by female immigrants and the performance of the children of immigrants are some aspects neglected by the literature A study on the relative earnings of native-­and foreign-­born women in turn-­of-­the-­century America revealed that immigrant women ‘fared somewhat better relative to the native-­born than men did’, earning from 102.2 percent to 113.2 percent of the native women’s wages (Fraundorf, 1978, p. 213) Long (1980), among the first to study female immigrants, found that the earnings of recent female immigrants were higher than those of natives, but this advantage declined over time In particular, married female immigrants increased their labor force participation initially to subsidize their husbands’ investments in human capital, but, later, 16   International handbook on the economics of migration as their husbands earnings increased, they switched to nonmarket activities and their ­earnings declined In contrast, other studies on female immigrants found strong evidence of assimilation, which varied considerably across countries of origin (Field-­Hendrey and Balkan, 1991) Studying life-­cycle patterns of immigrant women’s labor force participation in the US, Schoeni (1998) finds that the cross-­sectional approach significantly overestimates assimilation Nonetheless, he finds that immigrant women’s assimilation measured with cohort effects is still sizable and occurs within 10 years of arrival Japanese, Korean and Chinese women have the highest degree of assimilation in the labor market Parallel research in Canada presented evidence that the initial earnings differential for Canadian immigrant women is likely permanent and may be even worse for highly educated women (Beach and Worswick, 1993) Many researchers try to explain the earnings disparity between immigrants and natives by adding more characteristics to the theoretical and empirical estimation Others like Piore (1979) argue that labor market performance is not a function of the duration of residence in the host country, but a function of when an individual came For example, immigrants who arrived in Germany during the prosperous years of the mid-­1960s until the first economic recession of the early 1970s should fare better than more recent immigrants Miller and Chiswick (2002) corroborate this by showing that the business cycle of the host country plays an important role in the assimilation process More refined studies on earnings assimilation control for additional characteristics of the host country labor market, institutional variables, network effects and demographics in their quest to solve the earnings assimilation debate Even after adding ethnicity and legal status, results show that earnings assimilation is a rather elusive realization and varies widely by nationality; immigrants earn less than comparable natives when they work as employees In some countries like Germany and France, for example, earnings assimilation does not take place at all Still, immigrants who are self-­employed not only exhibit higher earnings than comparable immigrants in the paid employment sector, but they earn substantially more than comparable natives (Borjas, 1986) A study in Germany shows that the earnings of self-­employed Germans are not much different from the earnings of the self-­employed immigrants However, immigrants suffer a strong earnings penalty if they feel discriminated against, while they receive a premium if they are German educated (Constant and Zimmermann, 2006) New facets of immigrant performance are important and can offer key insights to an operative migration policy We refer to, for example, immigrant performance with regard to housing, wealth, education, even crime, as well as ­intergenerational assimilation 3 THE ROLE OF ETHNIC IDENTITY IN ECONOMIC INTEGRATION 3.1  The Identity-­Based Theory of Utility Maximization Personal identity is what makes individuals unique and different from others, including the self-­definition of one’s self How identity forms and manifests is a dynamic process Migration and ethnicity: an introduction  ­17 linked to social interactions Norms, values and rules binding members of a social group are inherent in the formation of social identities When conflicts arise, identities may result in suboptimal behavior Sociologists are well aware of these issues Massey and Denton (1993, p. 8) suggest in American Apartheid: Segregation and the Making of the Underclass, that segregated neighborhoods can create the structural conditions for some individuals to develop ‘an oppositional culture that devalues work, schooling and ­marriage’ and impedes success in the larger economy While identity has occupied a central role in other social sciences, such as psychology, sociology and anthropology, it has not been fully incorporated in economic theory and empirics Sen (1977), in his avant-­garde piece about the rational egoistic man of Edgeworth, talked about psychological issues that underlie choice and relate to consumer decisions and production activities He introduced the concepts of sympathy and commitment as part of the utility maximizing function, arguing that commitment as part of behavior can result in nongains-­maximizing answers, even when answers are truthful Economic theory should therefore accommodate commitment as part of behavior While commitment does not presuppose reasoning, it does not exclude it either In the 2000s, economists started looking at the concept of identity as a determinant of labor market attachment, performance and earnings This is along the strand of literature that places identity, behavior and personality traits in the heart of labor markets and the performance of individuals The quest is to explain schooling performance and economic labor market integration and unexplained wage differentials Some researchers have considered personality and behavior traits as part of the individual human capital, which counts differentially for men and women and for different ethnic groups (Bowles et al., 2001) In another empirical work that tries to improve human capital models and gain a greater understanding of the behavioral determinants of occupational success, Groves (2005) finds that traits such as locus of control, aggression and withdrawal are all statistically significant factors in the wage determination models of white women Akerlof and Kranton (2000) offer a novel theoretical framework of the utility maximization function by incorporating an individual’s self-­identification as powerful motivation for behavior They imply that if individuals achieve their ‘ideal self’ and are comfortable with their identity then their utility increases, otherwise, their utility decreases In this framework, it is then possible that even rational individuals choose nonoptimal occupations because of identity considerations For instance, a rational individual’s decision may very well be influenced by other social considerations as this person chooses a social category or affiliation, or a group to belong to or an occupation to self-­identify with As an example, suppose that someone identifies with and emulates being part of the armed forces If this person fails to so, then his or her utility decreases This in turn may affect the identity and behavior of others around him or her, and so on and so forth The choice of an individual to be a particular type of person then becomes a powerful economic decision with substantial changes in the conclusions in comparison with traditional economic analysis Bénabou and Tirole (2011) model a broad class of beliefs of individuals including their identity, which people value and invest in They also study endogenously arising self-­serving beliefs linked to pride, dignity or wishful thinking Norms about ‘fitting in’ or not, differ across time and space (Akerlof and Kranton, 2005) Modeling identity 18   International handbook on the economics of migration and work incentives, Akerlof and Kranton (2005) envisage corporate culture as the division of the workers into different groups, the prescribed behavior for each group and the extent to which workers identify with the organization or with the workgroup and adopt their respective goals They argue that identity is an important supplement to monetary compensation and enterprises that inculcate in employees a sense of identity and ­attachment to an organization are well-­functioning These emerging important contributions can very well explain labor market integration and wage differentials Accordingly, while some individuals have the drive and human capital to integrate and succeed in the labor market, they may not reach their goal because of behavioral norms and unfulfilled or confused self-­identity images In an empirical setting, Russo and van Hooft (2011) link identities, conflicting behavioral norms and job attributes They find that because individuals can adhere to multiple identities, when they experience conflicting norms in the labor market, they tend to value and choose job characteristics that can reduce the degree of conflict (that is, favorable working hours and good relationships with colleagues and managers) An interesting gender split shows that while men usually resolve any conflict between career and leisure by favoring a career, for women the presence of role conflict is not associated with the importance of a career While there is a large potential to use these frameworks for the analysis of ethnic, racial and immigrant identity along with the quest for economic inequality explanations, they have not been applied further 3.2  Ethnic, Racial and Cultural Identity Ethnic identity is ‘developed, displayed, manipulated, or ignored in accordance with the demands of a particular situation’ (Royce, quoted in Ruble, 1989, p. 401) It is whatever makes individuals the same or different in comparison to other ethnic groups But, it may also encompass a network of strong beliefs, values and what people hold dear; it builds and shapes peoples’ lives Fearon and Laitin (2000) argue that ethnic identities are socially constructed, either by individual actions or by supra-­individual discourses of ethnicity Some studies develop economic theories of ethnic identity and explicitly explore their implications for economic behavior Kuran (1998) has created a theory of reputational cascades that explains the evolution of behavioral ethnic codes that individuals follow to preserve social acceptance The speed of acting ethnic is chosen under the influences of social pressures that the individuals themselves create and sustain It is fostered by interdependencies among individual incentives that crucially affect personal choices This theory can explain why similar societies may show very different levels of ethnic activity Darity et al (2006) provide a long-­term theory of racial (or ethnic) identification formation Their evolutionary game theory model may result in equilibrium where all persons follow an individualist identity strategy, another where all persons pursue a racialist (or ethnic) identity strategy, or a mixture of both Consequently, race or ethnicity may be more or less significant for both market and non-­market social interactions A positive impact of racial identity on economic outcomes, that is, the productivity of social interactions, is the cornerstone of the theory This also explains the persistence of racial or ethnic privileges in market economies In sum, if there is a dominant or majority group or culture and a subordinate or Migration and ethnicity: an introduction  ­19 minority group or culture in a country, individuals in the minority group will either identify with the majority (in the hope that they will be recognized and accepted by the majority) or they will develop what is called oppositional identities and fight the majority culture because they know they will not be accepted by the majority anyway Sociologists and anthropologists know this all too well Ogbu (1999) argues that nonimmigrant minorities in the US constructed an oppositional collective identity after white Americans forced them into minority status and mistreatment He finds that a Black speech community in Oakland, California, faces a dilemma in learning and using proper English because of their incompatible beliefs about standard English However, since identity is multidimensional, science should allow for more than ‘either with them or against them’ identities It is also possible as Anderson (1999) shows in Code of the Street that some residents of segregated communities develop the capacity of ‘code switching’, which enables them to go back and forth between the predominantly white mainstream culture and the culture of their neighborhoods in order to navigate neighborhood perils Levels of attachment to, or detachment from, the dominant culture of the country of residence can therefore be extremely pertinent and crucial for policy design In the Battu et al (2007) model – where nonwhites identify with their social environment, their culture of origin, and where social networks can find them jobs – they find that individuals, who are otherwise identical, end up with totally different choices Depending on how strong peer pressures are, nonwhites choose to adopt ‘oppositional’ identities because some individuals may identify with the dominant culture and others may reject that culture, even if it implies adverse labor market outcomes In another empirical study, Battu and Zenou (2010) investigate the relationship between ethnic identity and employment They find that in the United Kingdom (UK) individuals’ identity choice is very much influenced by their social environment, that there is considerable heterogeneity in the nonwhite population in terms of preferences and that those nonwhites who develop and manifest oppositional and extreme identities are penalized in the labor market, experiencing a percent to percent lower probability of being in employment Mason (2004) establishes a stable identity formation among Mexican-­Americans and other Hispanics He shows that these ethnicities are able to increase their income substantially through acculturating into a non-­Hispanic white racial identity Bisin et al (2006) find that, in line with their theoretical analysis, identity with and socialization to an ethnic minority are more pronounced in mixed than in segregated neighborhoods The strength of identification with the majority culture regardless of strength of (ethnic) minority identity is important for labor market outcomes (Nekby and Rödin, 2010) Aguilera and Massey (2003) provide a better understanding of societal and economic behavior Expanding on the concept of ethnic human capital, Chiswick (2009) shows that economic determinants of ‘successful’ and ‘disadvantaged’ group outcomes are sensitive to the relationship between ethnic and general human capital, especially with regard to externalities in the processes by which they are formed Policies that welcome ethnic diversity within the larger society without encouraging separation would be desirable A genuinely inclusive policy of multiculturalism would also be beneficial Notice, however, that while there is a general understanding of flexible ethnic identity 20   International handbook on the economics of migration among many social scientists, there is still no consensus on all the elements that compose ethnic identity In the aforementioned studies, some use a self-­reported identification question, others use religion and language, and so on Reviewing the relevant literature outside economics, we find that among the suggested and widely used key elements of ethnic identity are the subjective expression of one’s commitment to, sense of belonging to or self-­identification with the culture, values and beliefs of a specific ethnic group and social life (Makabe, 1979; Masuda et al., 1970; Unger et al., 2002) Most frequently employed are cultural elements such as language, religion, media and food preferences, celebrated holidays and behavior (Phinney, 1990, 1992; Unger et al., 2002) 3.3  A Theory of Ethnic Identity While ethnic identity exists even when migrants are in their home country,9 it surfaces and manifests when they arrive in a host country that is dominated by a different ethnicity, culture, language and so on Typically, immigrants come from countries where they are part of the majority and become part of the minority in the host country.10 Ethnic identity is then like an attribute that an individual can have for some time, he or she can lose it and acquire a new one, or lose it and never take on or assume another one While it is unique to the individual – in the sense that even people from the same country of origin can have different ethnic identities – ethnic identity can create feedback loops as individuals interact with other or the same ethnicities In contrast, ethnicity is what people are born with, is static as well as permanent and usually denotes segments of the host country population with economic and social inequality between the dominant and minority groups, with political and social repercussions As the United Nations Economic Commission for Europe (UNECE, 2006, p. 100) put it, ‘ethnicity is based on a shared understanding of the history and territorial origins (regional, national) of an ethnic group or community as well as on particular cultural characteristics: language and/or religion and/or specific customs and ways of life’ Ethnicity is thus more related to the roots of peoples, their ancestry, the actual territory and physical boundaries of a country Here the reference is the group, a shared sense of peoplehood and not the individual Ethnic identity, ethnicity and culture are very much related, but they designate different things While the role of ethnicity or country of origin is documented to be a significant determinant of labor force participation and earnings as well as other socioeconomic areas concerned with integration (for example, homeownership, citizenship, voting and entrepreneurship) the role of culture and ethnic identity on economic ­outcomes is less widely acceptable There is a growing literature on the effects of culture on economic outcomes Guiso et al (2006) (using beliefs about trust) show a pervasive impact of culture in many economic choices The value of cultural diversity is evidenced in US cities through its net positive effect on the productivity of natives (Ottaviano and Peri, 2006) Bellini et al (2009) confirm that diversity is positively correlated with productivity in 12 of the EU15 European countries and causation runs from the former to the latter In Germany, the cultural diversity of people fosters the recognition, absorption and realization of entrepreneurial opportunities and has a positive impact on new firm formation, even more than the diversity of firms (Audretsch et al., 2010) Zimmermann (2007a), special issues Migration and ethnicity: an introduction  ­21 of the Journal of Population Economics (volume 20, issue 3, 2007), International Journal of Manpower (volume 30, issue 1–2, 2009) and Research in Labor Economics (volume 29, 2009) have documented the rising interest of economists in the field of ethnicity and identity In 2006, Constant et al were the first to introduce the multidimensional concept of ethnic identity in economics, by borrowing literature from social psychology and other social sciences Following the original work of Berry et al (1989),11 they developed a framework of ethnic identity and tested it empirically with German data Specifically, they created a two-­dimensional quantitative index – the ethnosizer – that measures the degree of the ethnic identity of immigrants Ethnic identity is how individuals perceive themselves within an environment as they categorize and compare themselves to others of the same or a different ethnicity It is the closeness or distance immigrants feel from their own ethnicity or from other ethnicities, as they try to fit into the host society; it can differ among migrants of the same origin, or be comparable among migrants of different ethnic backgrounds In stark distinction to ethnicity, ethnic identity attempts to measure how people perceive themselves rather than their ancestors The authors allow for the individuality, personality, distinctiveness and character of a person in an ethnic group to prevail, to differ from one person to another, and to alter and evolve in different directions They define ethnic identity to be the balance between commitment to, affinity to or self-­identification with the culture, norms and society of origin, and commitment to or self-­identification with the host culture and society Constant et al (2009a) propose that an immigrant moves along a plane formed by two positive vectors normalized from to 1, with representing maximum commitment The horizontal axis measures commitment to and self-­identification with the country of origin, and the vertical axis commitment to and self-­identification with the host country The origin of the Cartesian co-­ordinates (0,0) shows that an immigrant has no commitment to either the home or host country Point (0,1) exhibits maximum commitment to the culture of origin and no identification with the host country Diametrically opposite is point (1,0) that indicates immigrants who achieve full adaptation of the new culture and norms while they deny their own heritage If commitments to the home and host countries are linearly dependent and mutually exclusive and they sum up to one, then immigrants move along the diagonal (1,0) to (0,1).12 This is the case of the one-­dimensional ethnosizer That is, if immigrants retain their ethnic culture and norms they must not identify with the host country, and vice versa if they adopt the persona of the host country they must shed their ethnic and ­cultural identity related to the home country Confronted with both cultures, which combination of commitments migrants choose to uphold? The two-­dimensional ethnosizer of Figure 1.1 answers this question and shows where exactly migrants are in the positive quadrant As illustrated in Figure 1.1, the ethnosizer contains four states or regimes of ethnic identity differentiated by the strength of cultural and social commitments Quadrants A, I, M and S correspond to: assimilation (A), a pronounced identification with the host culture and society, coupled with a firm conformity to the norms, values and codes of conduct, and a weak identification with the ancestry; Integration (I), an achieved amalgam of both dedication to and identification with the origin and commitment and conformity to the host society This is the case of a perfect bicultural state; marginalization (M), a strong detachment 22   International handbook on the economics of migration Commitment to host country (0,1) (0,0) (1,1) A I M S Commitment to origin (1,0) Source:  Constant et al (2009a) Figure 1.1  The two-­dimensional non-­negative ethnosizer from either the dominant culture or the culture of origin; and, separation (S), an exclusive commitment to the culture of origin even after years of emigration, paired with weak involvement in the host culture and country realities Starting at point (1,0), a migrant can undergo a more complicated journey through the various states, leaving separation towards integration, assimilation or marginalization, or remaining separated Constant and Zimmermann (2008) augment the theoretical possibilities of the formation and manifestation of ethnic identity to include negative commitments Assuming a plane formed by two axes representing commitment to the home and host countries, an immigrant has four quadrants to express his or her ethic identity Commitment to and self-­identification with the country of origin is measured along the horizontal axis, and commitment to and self-­identification with the host country along the vertical axis Figure 1.2 illustrates the theoretical model of a complete multidimensional ethnic identity of positive, fanatical, and subvert ethnic identity Point (0,0) represents the stance of immigrants who have lost all ethnic identity related to the country of origin A movement to the right along the positive part of the horizontal axis (or in the northeast quadrant) indicates ethnic retention and increasing commitment to the country of origin Moving beyond point (1,0), suggests that immigrants not only identify with the country of origin but they more fanatically so practicing extreme views Going in the other direction along the negative part of the horizontal vector indicates immigrants who can turn against their own heritage and culture With respect to commitment to the host country, point (0,0) exhibits no identification with the host country either Immigrants going north on the vertical axis to point (0,1) evince increasing identification with the host country Moving beyond point (0,1) indicates the case of overzealous migrants, who over-­identify with the host country Migration and ethnicity: an introduction  ­23 (–1,0) Ethnic subversion towards the origin Ethnic subversion towards the host country Ethnic loss and commitment to host country (0,1) (0,0) Ethnic retention and commitment to origin (1,0) (0,–1) Source:  Constant and Zimmermann (2008) Figure 1.2 Complete illustration of ethnic identity; retention, relinquishment and subversion Going south on the negative part of the vertical axis, shows dissatisfied and disgruntled immigrants with the host country who can develop a subverted self-­identification towards it Note that, when migrants move along the negative part of the vertical axis, they can be either in the southeast or the southwest quadrant The southeast quadrant represents immigrants who keep the ethnic identity of the home country and oppose the host country While being in the southwest quadrant is a valid theoretical possibility of individuals turning against both countries, it is rather unlikely to happen in the real world (if we assume rational and mentally sound individuals) In reality, individuals may exhibit strong association with, commitment to and malcontent with either or both the culture of ancestry and the host culture The two-­dimensional model of the measurement of ethnic identity suggests that commitments to two different societies can coexist and influence each other in several ways In other words, the level of dedication to the original society does not preclude the degree of the commitment to the host society This assumption recognizes that a migrant, who strongly identifies with the culture and values of his or her ancestry, may or may not have a strong involvement with the dominant culture Similarly, a migrant with a strong affinity to the values and beliefs of the host country may or may not totally identify with the culture of ancestry At the same time, migrants may also be completely detached from the home or host country The two-­dimensional ethnosizer of Constant et al (2009a) allows for this case as well While they are theoretically possible, the negative parts of Figure 1.2 are impossible 24   International handbook on the economics of migration to be examined empirically No survey to our knowledge so far has any questions on negativity towards either the home or host country culture 3.4  Ethnosizing Migrants and Economic Integration To empirically test the ethnic identity of immigrants, Constant et al (2009a) developed an index, the ethnosizer They define the verb ethnosize13 to quantify how ethnic is an individual Based on data from the German Socio-­Economic Panel (GSOEP) they construct the four states or regimes of the two-­dimensional ethnosizer by identifying pairs of questions that transmit information on personal devotion and commitment to both the German culture and society and to the culture and society of origin They choose five essential elements of cultural and societal commitment that compose the ethnic identity, as they are widely accepted in social psychology These elements pertain to both the country of origin and the host country and give us a multidimensional view They are: (1) language; (2) visible cultural elements; (3) ethnic self-­identification; (4) ethnic interactions with natives; and (5) future citizenship and locational plans In some cases, individuals may be classified clearly with one concept, in other cases not at all In most cases, people will fall in several different regimes at the same time For example, with respect to element 3, immigrants who answered that they self-­ identify with Germany but not with the country of origin are considered assimilated Immigrants who self-­identify with both the country of origin and the host country are classified as integrated Those with total identification with the country of origin and little or no identification with Germany are labeled separated, and those who cannot self-­identify with either country are classified as marginalized The same classification is applied with respect to the other four elements of ethnic identity Providing equal weights to the five elements, each of the four measures or regimes of the ethnosizer can take a value between zero and five, and add up to five for each individual The idea of the index of ethnic identity is that it can be used to test the performance of immigrants in the host country’s labor market and possibly explain unexplained differences and deficiencies Potentially it can also be used to compare immigrants with natives and revisit the earnings assimilation literature Zimmermann et al (2008) find that human capital acquired in the host country does not affect the attachment and affinity to the receiving country Instead, it is pre-­ migration characteristics that dominate ethnic self-­identification In particular, human capital acquired in the home country leads to lower identification with the host country for both men and women immigrants, while men only have a higher affiliation with the original ethnicity and culture However, Aspachs-­Bracons et al (2008) have shown that ­compulsory language policy implemented in Catalonia have an effect on identity Constant and Zimmermann (2008) argue that while ethnic identity should affect work participation and cultural activities as human capital formation does, the ethnic identity of those working should not be influenced by work intensity and education from the receiving country Applying the ethnosizer on a sample of working men, they find that the ethnosizer mainly depends on pre-­migration characteristics, suggesting that ethnic identity is predictable through characteristics measured at the time of entry in the host country They also find that the ethnosizer is de facto independent of measured economic activity and significantly affects economic outcomes Migration and ethnicity: an introduction  ­25 Zimmermann (2007b) deals with the role of ethnic identity in earnings Adding the two-­dimensional ethnosizer to standard tobit regressions to examine the particular contribution of ethnic identity, he finds that ethnic identity matters significantly and that the findings are very robust with respect to the concrete model specification That is, the inclusion of the ethnosizer does not change the parameter estimates of the standard variables in any relevant way Nevertheless, the parameter estimates of the ethnic identity have a strong impact on economic behavior Constant and Zimmermann (2009) extend this framework to model the labor force participation and earnings of both men and women immigrants, because men and women may have completely different understanding and expression of their ethnic identity.14 This is based on the idea that immigrants are mostly useful in the host country when they bring different talents and skills than natives possess If the resulting diversity reflects ethnic characteristics that are relatively scarce, the labor market functions smoothly In the case of a homogeneous population, there is always the risk of lost creativity ‘Successful migration implies integration, assimilation, loyalty and good citizenship but also diversity and multiple identities’ (Hieronymi, 2005, p. 132) There are costs and benefits associated with this cultural capital embodied in immigrants When immigrants and natives complement each other, there can be a win-­win situation; immigrants and natives can profit and the economy and society can benefit from creativity, dynamism and greater prosperity Constant and Zimmermann (2009) find that the ethnic identity of immigrants is a strong determinant of their labor force participation They also find interesting gender dynamics, whereby separated or marginalized men have a much lower probability to work when compared with immigrants who totally identify with natives and demonstrate a strong commitment to German society However, being assimilated does not offer a particular advantage to the working probabilities of men compared with the identity state of being integrated In contrast, women who identify with both cultures (are integrated) have a much higher probability to work than women who only identify with natives (are assimilated) Separated or marginalized women have lower chances of joining the labor force than those who are assimilated Unexpectedly, the authors find that once immigrants start working, ethnic identity does not affect their earnings in a significant way This is consistent with other studies on the effect of identity and personality on occupations and earnings Therefore, the findings reported in Zimmermann (2007b) on earnings using tobit regressions are driven by the decision to work Dealing with other forms of economic integration, Constant et al (2009c) study the home ownership and wealth of immigrants They find that immigrants with a stronger commitment to the host country are more likely to achieve home ownership for a given set of socioeconomic and demographic characteristics, regardless of their level of attachment to their home country That is, assimilated and integrated immigrants move up to home ownership Other forms of immigrant integration can also offer valuable insights into their economic integration For example, immigrants who integrate in the political arena by naturalizing may follow different paths of economic integration in the host country Naturalization, in turn, may very well be influenced by ethnic identity Zimmermann et al (2009) study how ethnic identity can affect the probabilities of actual naturalization, future naturalization and refusal of naturalization They find that integration in 26   International handbook on the economics of migration German society has a stronger effect on naturalization than ethnic origin and religion, and women immigrant household heads are more likely to want to acquire or to already have acquired German citizenship The risk proclivity of immigrants and individuals in general in a society is important to study as it affects many socioeconomic facets, from gambling to obesity, to crime, to labor market performance Bonin et al (2012) explore the role of ethnic identity in the risk proclivity of immigrant and native Germans Specifically, they use measures of immigrants’ ethnic persistence and assimilation They find that assimilation or adaptation to the attitudes of the majority population closes the immigrant–native gap in risk proclivity, while stronger commitment to the home country or ethnic persistence preserves it As risk attitudes are behaviorally relevant, and vary by ethnic origin, these results could also help explain differences in the economic assimilation of immigrants 3.5  Empirical Ethnic Identity Issues This section provides two empirical ethnic self-­identification examinations to support the usefulness of the ethnic identity approach The first is, to what extent does the ethnosizer differ from the direct measure of15 provided by survey data? The self-­ identification question is subjective, and hence open to debate People are asked, for example, how native or foreign they feel and how much they identify with one or the other country The ethnosizer, however, uses another four elements besides the self-­ identification question; elements that are objective, such as what people did or are actually doing This can balance the judgement the self-­identification question provides Table 1.1 uses data from the GSOEP optimized for the purpose of comparison of the ethnosizer with the direct measure of ethnic self-­identification We observe 1339 individual migrants and generate 6695 observations that are cross-­classified according to the four regimes (assimilation, integration, marginalization and separation) The cells on the main diagonal of the contingency table contain the cases where self-­classification coincides with the judgement of the ethnosizer The agreement is, in general, small: 45.9 percent for integration, 53.6 percent for assimilation, 54.9 percent for separation and 31.9 percent for marginalization (percentages from the column totals) From those who consider themselves to be marginalized, in 23.7 percent of the cases we find evidence of integration In 32.5 percent of the cases for those who self-­report integration, we find evidence of assimilation Self-­classified assimilation goes with 12.6 percent cases of marginalization, and self-­reported separation coincides with 21 percent cases of integration This provides support for the attempt to balance the self-­evaluation question out through the ethnosizer The second example demonstrates the differences the ethnic identity regimes have for economic performance Here, we choose data from a new frontier survey, the German IZA Evaluation Dataset (Caliendo et al., 2011), that collects data on ethnic self-­identity of immigrants and natives who are unemployed and who receive unemployment benefits The assumption is that for immigrants the alternative to the home culture is German, and for the native Germans the alternative culture is international Table 1.2 contains raw data on the net hourly reservation wages for natives and immigrants and the reservation wage ratio (reservation wage divided by the respective wage in the last job) for all four ethnic identity regimes For immigrants, reservation wages are the highest if they Migration and ethnicity: an introduction  ­27 Table 1.1  Direct measure of ethnic self-­identification and the ethnosizer Ethnosizer: four regimes Self-­identification Total Integration Assimilation Separation Marginalization 143 32.50 2.14 202 45.91 3.02 33 7.50 0.49 62 14.09 0.93 440 100.00 6.57 488 53.62 7.29 219 24.07 3.27 115 12.64 1.72 88 9.67 1.31 910 100.00 13.59 435 12.52 6.50 729 20.98 10.89 403 11.60 6.02 1908 54.90 28.50 3475 100.00 51.90 378 20.21 5.65 444 23.74 6.63 597 31.93 8.92 451 24.12 6.74 1870 100.00 27.93 Assimilation Integration Marginalization Separation Total 1444 21.58 1594 23.81 1148 17.15 2509 37.48 6695 100.00 Notes: Own calculations on the basis of the GSOEP (Zimmermann, 2007b) Number of individuals: 1339 Bold numbers are cell counts, followed by percentages of the column totals (italic) and the relative frequencies of the total sample size Table 1.2  Ethnic identity regimes and reservation wages Ethnosizer: four regimes Assimilation Integration Marginalization Separation Reservation wages Reservation wage ratio Immigrants Natives Immigrants Natives 7.29 7.61 7.16 6.98 7.10 7.73 6.72 7.15 1.10 1.12 1.19 1.15 1.08 1.12 1.12 1.20 Notes:  Sample sizes are 1515 migrants and 5975 natives Net hourly reservation wage in euros measured about three months after unemployment entry Reservation wage ratio is defined as the reservation wage divided by the last net wage from (self-)employment before entering unemployment Source:  IZA Evaluation Dataset (wave 1: 2007−08), own calculations are integrated, followed by those assimilated, marginalized and separated Assimilated and integrated immigrants report reservation wages which are roughly 10 percent higher than their previous hourly wages However, separated and marginalized immigrants’ reservation wages exceed their previous hourly wages by 15 percent and 18 percent, respectively 28   International handbook on the economics of migration This picture is different for natives, where those integrated have the highest reservation wages, those marginalized have the lowest and the others rank in between Assimilated, integrated and marginalized natives report reservation wages which are between percent and 12 percent higher than their previous hourly wages However, the reservation wages of separated natives exceed their previous hourly wages by 20 percent This suggests that identity matters for labor market behavior 4 ATTITUDES FROM AND ABOUT IMMIGRANTS IN THE INTEGRATION PROCESS Attitudes are extremely relevant in the integration setting as they represent the position a person has towards others Attitudes could be feelings or emotions towards a fact, a negative mindset, or the way people respond to a stimulus They reflect complex historical, psychological and social processes, can change according to experience and stimuli, and can be positive, negative or neutral ranging from xenophilia or allophilia to xenophobia Attitudes and sentiments towards migrants, foreigners or ethnic minorities vary widely across countries They may arise from ethnic or racial antipathy and xenophobia, or may be based on economic fears regarding the labor market and the welfare state, and one’s own economic outlook Since immigration is the consequence of policy, migration policy is partly responsible for the types of immigrants a country receives, their economic performance, the functioning of the economy and, hence, natives’ perceptions towards immigrants While attitudes and perceptions form or influence our behavior, they are also the outcome of a complex social, political and economic process, shaped through the engagement of individuals in social and working life and influenced by public discourse and the media This suggests that attitudes are only partly predetermined, and are also the outcome of a complex economic, political and social reality In the scientific literature, the concepts of ethnicity, ethnic identity, multiculturalism, social exclusion and xenophobia are relatively well researched by sociologists, social psychologist, cultural anthropologists and political theorists Phinney et al (2001) argue that ethnic and national identity are inter-­related and play a role in the psychological well-­being of immigrants They perceive this as an interaction between the attitudes and characteristics of immigrants and how the host society responds to them That is, ethnic identity strongly interacts and changes with the immigrant policy a country has and with the attitudes of natives The strengths of ethnic identity vary according to the support immigrants receive to keep their ethnic culture and the pressure immigrants receive to assimilate and relinquish their ethnic values and norms The authors find that the best adaptation is achieved by a combination of a strong ethnic identity and a strong national identity When the host society accepts multiculturalism and immigrants want to keep their ethnic identity, then ethnic identity is strong When immigrants are pressured to assimilate but they are accepted, then national identity is strong When immigrants face real or perceived hostility towards them, then some may reject their ethnic identity while others may over-­exhibit their ethnic identity For example, they find that immigrants in Finland have largely marginalized identities, in the Netherlands have overly separated identities and in Israel half of the immigrants were assimilated and half integrated Migration and ethnicity: an introduction  ­29 Overall, integrated ethnic identities are associated with higher levels of well-­being than are other ethnic identity states There is also a growing literature in economics on the attitudes towards immigrants and ethnic minorities Bauer et al (2000) study the effect of different immigration policies in OECD countries on attitudes towards immigrants and document the relevance of economically motivated migration policy for the social acceptance of immigrants Analyzing the role of labor market competition, immigrant concentration, racial/ethnic bias, educational attainment and a set of other variables that potentially determine attitudes towards immigrants, Gang et al (2010) find that negative attitudes towards foreigners have increased and those who directly compete with immigrants have stronger negative attitudes towards foreigners About 12 percent of the increased anti-­foreigner attitudes are explained by differences in people’s characteristics and 88 percent of the rising anti-­foreigner sentiment is related to behavioral changes among the population that has strengthened the impact of various individual characteristics on negative attitudes towards foreigners Key among these behavioral changes is the fact that the strength of the ameliorating impact of education on anti-­foreigner attitudes diminished over time Overall, people with higher levels of education and occupational skills are more likely to favor immigration and cultural diversity regardless of the skill attributes of the immigrants in question, and they are also more likely to believe that immigration generates benefits for the host economy as a whole (Hainmueller and Hiscox, 2007) Dustmann and Preston (2004) show that in the UK, attitudes towards foreigners depend on where immigrants come from Those from other European countries face more favorable attitudes than those from Asia or the West Indies Looking at the formation of attitudes towards future immigrants they find evidence that economic matters such as welfare and labor market performance contribute to negative perceptions However, it is striking that the most important determinant is non-­economic; it is racial intolerance Moreover, high concentrations of ethnic minorities are associated with more hostile attitudes towards immigrants in Germany (Gang and Rivera-­Batiz, 1994) Continuing on the quest for the determinants of attitudes, Card et al (2005) show that attitudes toward immigration vary systematically with age, education and urban/rural location, and that there is substantial variation in the strength of anti-­immigrant opinion across European countries If immigrants are to switch identities in light of different attitudes from natives this may very well result in different economic integration patterns Manning and Roy (2010), in a theoretical and empirical exercise, discuss the cultural assimilation of immigrants in the UK, the British identity and the views on rights and responsibilities in societies They find that almost all UK-­born immigrants see themselves as British and others feel more British the longer they stay in the UK However, not all of the white UK-­born population thinks of these immigrants as British, because they are more concerned about values than national identity For example, they are worried that Pakistanis who feel British are causing problems but are not worried about Italians who not feel British and cause problems How people see others and how they see themselves is the interesting question Epstein and Gang (2009) acknowledge the three elements required to bring minorities into line with the majority: assimilation efforts, time and the degree to which the majority welcomes the minority They set up a theoretical model to examine the consequences for assimilation and harassment of growth in the minority population, time and the role 30   International handbook on the economics of migration of political institutions They find that as the size of the minority increases, assimilation and the effort to assimilate also increases But growth in the minority also increases harassment by the majority If the groups are very asymmetric, the best thing to is for the minority to fight harassment and continue with assimilation If the asymmetry between the abilities of both groups to affect the minority’s productivity is less than the ratio between the effects of the marginal efficiency of their investments on their rents, then the minority will give up on assimilation Lastly, the minority will give up on its assimilation efforts if the majority is strong and united against the minority The role of culture and family attitudes towards employment rates in OECD countries is studied by Cahuc and Algan (2007) They argue that family labor supply interactions and cross-­country heterogeneity in family culture are key to explaining divergent employment rates and employment gaps The emphasis is on employment disparities that mostly affect specific demographic groups like women In particular, they show that people facing a priori the same economic environment by living in the same country – but who differ by the national origin of their ancestors – have significantly different family attitudes, even after controlling for all their relevant socioeconomic individual characteristics That is, they confirm cultural foundations of family attitudes In addition, their family attitudes are perfectly in line with those currently expressed in their country of origin They show that the stronger preferences for family activities in European countries may explain both their lower female employment rate and the fall in the employment rates of younger and older people As valuable as explaining the facts is, it is a different matter than implementing recommendations such as the Lisbon agenda Cahuc and Algan (2007) wonder if this implementation can be reached and if it can be welfare improving, given the cultural forces that reign in some segments of the population Constant et al (2009b) study opinions and attitudes towards immigrants and minorities and their interactions with other barriers to minorities’ economic integration They use a unique dataset that gauges the perspectives of expert stakeholders and of ethnic minorities on their integration situation and the main barriers that hinder it In this survey both immigrants and natives were asked about attitudes and perceptions towards others and about themselves They find that ethnic minorities face integration problems and the natives’ general negative attitudes are a key factor of the challenging situation of minorities While discrimination is acknowledged as the single most important integration barrier, low education and self-­confidence as well as cultural differences also hinder integration Lastly, minorities want change as long as it comes about by policies based on the principle of equal treatment 5  CONCLUSIONS AND OUTLOOK This chapter has reviewed the economic assimilation of labor migrants, the evolution of bi-­ethnic identities and the value and relevance of perceptions and attitudes, within the broader framework of the economic integration of immigrants There is evidence that ‘soft’ factors such as attitudes, perceptions and identities affect economic behavior more than they are driven by them Some of these issues are further studied in Chapter 14, in this volume, by Constant and Zimmermann However, most of the findings are still based on cross-­sectional evidence and available Migration and ethnicity: an introduction  ­31 only for a few countries We need to expand our analysis to the available panel datasets and to employ cross-­national comparisons especially of the performance and adaptation of specific ethnic groups in different cultural settings A major difficulty here that future research should try to tackle is to model the endogeneity of the processes of economic performance and social and cultural interactions The often expressed societal norms, and even political innuendos, that migrants should assimilate is not a conclusion that can be derived from economic reasoning First, migrant groups hardly ever reach economic assimilation, at least not the first-­generation migrants If people with a migration background, the second generation, are performing like natives in an economic sense, they are often also ethnically assimilated Second, labor migrants pulled by the host country are requested because they exhibit scarce characteristics – in the short or the long run Hence, they are wanted because they are different, either because they have talent or skills which are not sufficiently produced at a certain time or because they bring with them ethnic capital that is valuable for the global competitiveness of the receiving economy Research is needed to better understand and empirically validate ethnic capital and its potential use in the economy Globalization and demographic changes will lead to a higher level of permanent and temporary labor migration around the world Circular migration, the move of workers back and forth as well as onwards will become even more regular and standard than it currently is (See Chapter on circular migration by Constant et al in this volume for further details.) This increases the demand for individuals with multiethnic identities and generates more diversity within migrant-­receiving countries Coping with increasing ethnic, cultural and religious diversity, especially in societies with either a history of conflicts between certain groups, or a strong tradition of cultural homogeneity, is not an easy task Hence, this will also raise the importance of attitudes Observing and understanding future labor migration trends and their interactions with cultural and societal conditions is therefore a future research agenda of great importance This volume, in particular in Parts III and IV, offers a number of chapters that address these challenges, including Chapter 10 on ethnic hiring by Neumark, Chapter 12 on occupational sorting of ethnic groups by Patel et al and Chapter 18 on immigration–­religiosity intersections by García-­Moz and Neuman NOTES   * Financial support from the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) for the project on ‘Ethnic Diversity and Labor Market Success’ in the DFG-­Priority Program ‘Flexibility in Heterogeneous Labor Markets’ (Flexibilisierungspotenziale bei heterogenen Arbeitsmärkten) is gratefully acknowledged This is an updated version of Amelie F Constant and Klaus F Zimmermann (2011), ‘Migration, ethnicity and economic integration’, in M.N Jovanovic (ed.), International Handbook of Economic Integration, Edward Elgar Publishing, Cheltenham, UK and Northampton, MA, USA, pp. 145–68   Human capital theory, as formulated by Becker (1991) and Mincer (1974), evaluates how improvement in the skills and talents of workers influences future real income Investment in human capital includes education, labor market experience (with specific or general on-­the-­job training), health (both mental and physical), and knowledge about the labor market An increase in skill increases productivity and earnings, but at the cost of foregone income Human capital theory posits that wage differentials result from differences in human capital accumulation and specialization Investment in human capital in the host country should help disadvantaged groups (immigrants) increase their earnings and approach the earnings of natives 32   International handbook on the economics of migration   Another theory that can explain the economic assimilation of immigrants is the segmented labor market theory, as first formulated by Doeringer and Piore (1971) Here, the labor markets are divided into a primary or capital-­intensive sector with skilled workers and secondary or labor-­intensive sector with mostly unskilled workers assigned menial jobs Earnings differentials across the primary and secondary sectors are significant, and market forces are unable to erode these differentials Consequently, immigrants who are in the second tier, earn lower wages than natives, their wages increase slowly over time, wages are rigid upwards but flexible downwards and can fall if supply increases (Piore, 1979) The wage gap between immigrants and natives is expected to widen over time and additional years of residence in the host country not affect the economic process of the assimilation of immigrants   Alternatively, some researchers define assimilation as the increase in the immigrants earnings brought about by additional years of residence in the host country Others measure assimilation as the rate at which the earnings of newly arrived immigrants converge to the earnings of other ethnically similar immigrants residing in the host country for more than 25 years   These studies include Carliner (1980), Borjas (1982) and Abbott and Beach (1993), to name a few of the early contributors   Jasso and Rosenzweig (1986, 1990), Bloom and Gunderson (1991) and Schoeni (1998), just to name a few of the first longitudinal studies   This is an alternative method to longitudinal data analysis and is typically used with decennial censuses or CPS data   In contrast to the first- and second-­generation immigrants, these immigrant children are often called the half generation   Allowing for other theories, such as the new economics of labor migration, studies on return migration there may not be a unitary process of return migration, but several (Constant and Massey, 2002) The authors also caution against an over-­reliance on single theories in understanding and explaining international migration   Constant and Zimmermann (2007) consider the ethnic identity of natives in Germany Inevitably, natives are also affected by the incoming migrants in several dimensions For example, they can become more cosmopolitan and open to new cultures, stay locked in their own ethnic identity or even develop an identity opposing immigrants, or their own culture 10 An exception is the case of Jews who are usually a minority in the home country and become part of the majority when they migrate to Israel 11 See also Berry (1980) and Phinney (1990, 1992) 12 See Figure 1.2 for this illustration 13 The word comes from the combination of the terms ‘ethnic’ and ‘size’ (ethno/size), where ‘size’ indicates its status as a form of measurement of the ethnic identity 14 Ethnic identity, much like personality and other individual characteristics, influences labor market outcomes We know, for example, that preferences affecting earnings, efficacy and other psychological aspects of individuals are significant influencers of earnings (Bowles et al., 2001) Moreover, cultural hypotheses are economically important for fundamental economic issues like national rates of saving (Guiso et al., 2006) Beliefs that people value and invest in have important economic implications (Bénabou and Tirole, 2011) 15 See Zimmermann et al (2007, 2008) for an econometric analysis of ethnic self-­identification using GSOEP data REFERENCES Abbott, M.G and C.M Beach (1993), ‘Immigrant earnings differentials and birth-­year effects for men in Canada: post-­war-­1972’, Canadian Journal of Economics, 26 (3), 505–24 Aguilera, M.B and D.S Massey (2003), ‘Social capital and wages of Mexican migrants: new hypotheses and tests’, Social Forces, 82 (2), 671–701 Akerlof, G.A and R.E Kranton (2000), ‘Economics and identity’, Quarterly Journal of Economics, 115 (3), 715–53 Akerlof, G.A and R.E Kranton (2005), ‘Identity and the economics of organizations’, Journal of Economic Perspectives, 19 (1), 9–32 Anderson, E (1999), Code of the Street: Decency, Violence, and the Moral Life of the Inner City, New York: Norton Aspachs-­Bracons, O., I Clots-­Figueras, J Costa-­Font and P Masella (2008), ‘Compulsory language educational policies and identity formation’, Journal of the European Economic Association, (2–3), 434–44 Migration and ethnicity: an introduction  ­33 Audretsch, D.B., D Dohse and A Niebuhr (2010), ‘Cultural diversity and entrepreneurship: a regional analysis for Germany’, The Annals of Regional Science, 45 (1), 55–85 Battu, H and Y Zenou (2010), ‘Oppositional identities and employment for ethnic minorities: evidence from England’, Economic Journal, 524 (120), F52–F71 Battu, H., M McDonald and Y Zenou (2007), ‘Oppositional identities and the labor market’, Journal of Population Economics, 20 (3), 643–67 Bauer, T.K., M Lofstrom and K.F Zimmermann (2000), ‘Immigration policy, assimilation of immigrants and natives’ sentiments towards immigrants: evidence from 12 OECD-­countries’, Swedish Economic Policy Review, (2), 11–53 Beach, C.M and C Worswick (1993), ‘Is there a double-­negative effect on the earnings of immigrant women?’, Canadian Public Policy, 19 (1), 36–53 Becker, G.S (1991), ‘Investment in human capital: a theoretical analysis’, in R.E Kuenne (ed.), Microeconomics: Theoretical and Applied 1, International Library of Critical Writings in economics 11, Aldershot, UK and Brookfield, VT, USA: Edward Elgar Bellini, E., G.I.P Ottaviano, D Pinelli and G Prarolo (2009), ‘Cultural diversity and economic performance: evidence from European regions’, Nota di Lavori 632009, Milano: Fondazione Eni Enrico Mattei (FEEM) Bénabou, R and J Tirole (2011), ‘Identity, morals, and taboos: beliefs as assets’, Quarterly Journal of Economics, 126 (2), 805–55 Berry, J (1980), ‘Acculturation as varieties of adaptation.’, in A.M Padilla (ed.), Acculturation: Theory, Models and Some New Findings, Boulder, CO: Westview, pp. 9–25 Berry, J., U Kim, S Power, M Young and M Bujaki (1989), ‘Acculturation attitudes in plural societies’, Applied Psychology: An International Review, 38 (2), 185–206 Bisin, A., E Patacchini, T Verdier and Y Zenou (2006), ‘Bend it like Beckham’: identity, socialization and assimilation’, CEPR Discussion Paper No 5662, Centre for Economic Policy Research (CEPR) London Bloom, D.E and M Gunderson (1991), ‘An analysis of the earnings of Canadian immigrants’, in J.M Abowd and R.B Freeman (eds), Immigration, Trade, and the Labor Market, Chicago, IL: University of Chicago Press, pp. 321–42 Bonin, H., A Constant, K Tatsiramos and K.F Zimmermann (2012), ‘Ethnic persistence, assimilation and risk proclivity’, IZA Journal of Migration, 1(5), available at: http://www.izajom/content/1/1/5 Borjas, G.J (1982), ‘The earnings of male hispanic immigrants in the United States’, Industrial Labor Relations Review, 35 (3), 343–53 Borjas, G.J (1985), ‘Assimilation, changes in cohort quality, and the earnings of immigrants’, Journal of Labor Economics, (4), 463–89 Borjas, G.J (1986), ‘The self-­employment experience of immigrants’, The Journal of Human Resources, 21 (4), 485–506 Borjas, G.J (1989), ‘Immigrant and emigrant earnings: a longitudinal study’, Economic Inquiry, 27 (1), 21–37 Borjas, G.J (1992), ‘National origin and the skills of immigrants in the postwar period’, in G.J Borjas and R.B Freeman (eds), Immigration and the Work Force, Chicago, IL: University of Chicago Press for the National Bureau of Economic Research, pp. 17–48 Bowles, S., H Gintis and M Osborne (2001), ‘Incentive-­enhancing preferences: personality, behavior, and earnings’, American Economic Review, 91 (2), 155–8 Cahuc, P and Y Algan (2007), ‘The roots of low European employment: family culture?’, in J Frankel and C Pissarides (eds), NBER International Seminar on Macroeconomics 2005, Cambridge, MA: MIT Press, pp. 65–109 Caliendo, M., A Falk, L Kaiser, H Schneider, A Uhlendorff, G van den Berg and K.F Zimmermann (2011), ‘The IZA evaluation dataset: towards evidence-­based labor policy making’, International Journal of Manpower, 32 (7), 731–52 Card, D., C Dustmann and I Preston (2005), ‘Understanding attitudes to immigration: the migration and minority module of the first European social survey’, CReAM Discussion Paper No 0305, University College Centre, Department of Economics, Centre for Research and Analysis of Migration (CReAM), London Carliner, G (1980), ‘Wages, earnings and hours of first, second, and third generation American males’, Economic Inquiry, 18 (1), 87–102 Chiswick, B.R (1978), ‘The effect of Americanization on the earnings of foreign-­born men’, Journal of Political Economy, 86 (5), 897–922 Chiswick, B.R (1986), ‘Is the new immigration less skilled than the old?’, Journal of Labor Economics, (2), 168–92 Chiswick, C.U (2009), ‘The economic determinants of ethnic assimilation’, Journal of Population Economics, 22 (4), 859–80 Constant A and D.S Massey (2002), ‘Return migration by German guestworkers: neoclassical versus new economic theories’, International Migration, 40 (4), 5–38 34   International handbook on the economics of migration Constant, A and D.S Massey (2003), ‘Self-­selection, earnings, and out-­migration: a longitudinal study of immigrants to Germany’, Journal of Population Economics, 16 (4), 631–53 Constant, A and K.F Zimmermann (2006), ‘The making of entrepreneurs in Germany: are native men and immigrants alike?’, Small Business Economics, 26 (3), 279–300 Constant, A.F and K.F Zimmermann (2007), ‘Does ethnosizing pay? Gender and native–migrant differences’, mimeo, presented at the 4th Annual Migration Meeting, Institute for the Study of Labor (IZA), Bonn Constant, A.F and K.F Zimmermann (2008), ‘Measuring ethnic identity and its impact on economic behavior’, Journal of the European Economic Association, (2–3), 424–33 Constant, A.F and K.F Zimmermann (2009), ‘Work and money: payoffs by ethnic identity and gender’, Research in Labor Economics, 29, 3–30 Constant, A.F., L Gataullina and K.F Zimmermann (2006), ‘Ethnosizing immigrants’, IZA Discussion Paper No 2040, Institute for the Study of Labor (IZA), Bonn Constant, A.F., L Gataullina and K.F Zimmermann (2009a), ‘Ethnosizing immigrants’, Journal of Economic Behavior and Organization, 69 (3), 274–87 Constant, A.F., M Kahanec and K.F Zimmermann (2009b), ‘Attitudes towards immigrants, other integration barriers and their veracity’, International Journal of Manpower, 30 (1/2), 5–14 Constant, A., R Roberts and K.F Zimmermann (2009c), ‘Ethnic identity and immigrant homeownership’, Urban Studies, 46 (9), 1879–98 Darity, W.A., P.L Mason and J.B Stewart (2006), ‘The economics of identity: the origin and persistence of racial identity norms’, Journal of Economic Behavior and Organization, 60 (3), 283–305 Doeringer, P.B and M.J Piore (1971), Internal Labor Markets and Manpower Analysis, Lexington, MA: D.C. Heath Dustmann, C and I Preston (2004), ‘Is immigration good or bad for the economy? Analysis of attitudinal responses’, Research in Labor Economics, 24, 3–34 Epstein, G.S and I.N Gang (2009), ‘Ethnicity, assimilation and harassment in the labor market’, Research in Labor Economics, 29, 67–88 Fearon, J.D and D.D Laitin (2000), ‘Violence and the social construction of ethnic identity’, International Organization, 54 (4), 845–77 Field-­Hendrey, E and E Balkan (1991), ‘Earnings and assimilation of female immigrants’, Applied Economics, 23 (10), 1665–72 Fraundorf, M.N (1978), ‘Relative earnings of native- and foreign-­born women’, Explorations in Economic History, 15 (2), 211–20 Friedberg, R.M (1992), ‘The labor market assimilation of immigrants in the United States: the role of age at arrival’, working paper, Providence, RI: Brown University Gang, I.N and F.L Rivera-­Batiz (1994), ‘Unemployment and attitudes towards foreigners in Germany’, in G Steinmann and R Ulrich (eds), Economic Consequences of Immigration to Germany, Heidelberg: Physica-­ Verlag, pp. 121–54 Gang, I.N., F.L Rivera-­Batiz and M.-­S Yun (2010), ‘Changes in attitudes toward immigrants in Europe: before and after the fall of the Berlin Wall’, in G.S Epstein and I.N Gang (eds), Migration and Culture, Bingley: Emerald Publishing, pp. 649–76 Groves, M.O (2005), ‘How important is your personality? Labor market returns to personality for women in the US and UK’, Journal of Economic Psychology, 26 (6), 827–41 Guiso, L., P Sapienza and L Zingales (2006), ‘Does culture affect economic outcomes?’, Journal of Economic Perspectives, 20 (2), 23–­48 Hainmueller, J and M.J Hiscox (2007), ‘Educated preferences: explaining attitudes toward immigration in Europe’, International Organization, 61 (2), 399–442 Hieronymi, O (2005), ‘Identity, integration and assimilation: factors of success and failure of migration’, Refugee Survey Quarterly, 24 (4), 132–50 Jasso, G and M.R Rosenzweig (1986), ‘What’s in a name? Country-­of-­origin influences on the earnings of immigrants in the United States’, in O Stark and I Sirageldin (eds), Research in Human Capital and Development 4, A Research Annual, Greenwich, CT: JAI Press, pp. 75–106 Jasso, G and M.R Rosenzweig (1988), ‘How well U.S immigrants do? Vintage effects, emigration selectivity, and occupational mobility of immigrants’, in P.T Schulz (ed.), Research of Population Economics 6, A Research Annual, Greenwich, CT: JAI Press, pp. 229–53 Jasso, G and M.R Rosenzweig (1990), The New Chosen People: Immigrants in the United States, New York: Russel Sage Foundation Kossoudji, S.A (1989), ‘Immigrant worker assimilation: is it a labor market phenomenon?’, The Journal of Human Resources, 24 (3), 494–526 Kuran, T (1998), ‘Ethnic norms and their transformation through reputational cascades’, Journal of Legal Studies, 27 (2), 623–59 Migration and ethnicity: an introduction  ­35 LaLonde, R.J and R.H Topel (1991), ‘Immigrants in the American labor market: quality, assimilation, and distributional effects’, American Economic Review, 81 (2), 297–302 LaLonde, R.J and R.H Topel (1992), ‘The assimilation of immigrants in the U.S labor market’, in G Borjas and R Freeman (eds), Immigration and the Work Force, Chicago,IL: University of Chicago Press, pp. 67–92 Long, J.E (1980), ‘The effect of Americanization on earnings: some evidence for women’, Journal of Political Economy, 88 (3), 620–29 Makabe, T (1979), ‘Ethnic identity scale and social mobility: the case of Nisei in Toronto’, The Canadian Review of Sociology and Anthropology, 16 (2), 136–45 Manning, A and S Roy (2010), ‘Culture clash or culture club: national identity in Britain’, Economic Journal, 120 (542), F72 –F100 Mason, P.L (2004), ‘Annual income, hourly wages, and identity among Mexican-­Americans and other Latinos’, Industrial Relations, 43 (4), 817–34 Massey, D.S and N.A Denton (1993), American Apartheid: Segregation and the Making of the Underclass, Cambridge, MA: Harvard University Press Masuda, M., G Matsumoto and G Meredith (1970), ‘Ethnic identity in three generations of Japanese Americans’, Journal of Social Psychology, 81 (2), 199–207 Miller, P.W and B.R Chiswick (2002), ‘Immigrant earnings: language skills, linguistic concentrations and the business cycle’, Journal of Population Economics, 15 (1), 31–57 Mincer, J (1974) ‘Family investments in human capital: earnings of women’, Journal of Political Economy, 82 (2), Pt2, S76–S108 Nekby, L and M Rödin (2010), ‘Acculturation identity and employment among second and middle generation immigrants’, Journal of Economic Psychology, 31 (1), 35–50 Ogbu, J (1999), ‘Beyond language: ebonics, proper English, and identity in a Black-­American speech community’, American Educational Research Journal, 36 (2), 147–84 Ottaviano, G.I.P and G Peri (2006), ‘The economic value of cultural diversity: evidence from US cities’, Journal of Economic Geography, (1), 9–44 Phinney, J.S (1990), ‘Ethnic identity in adolescents and adults: review of research’, Psychological Bulletin, 180 (3), 499–514 Phinney, J.S (1992), ‘The multigroup ethnic identity measure: a new scale for use with diverse groups’, Journal of Adolescent Research, (2), 156–76 Phinney, J.S., G Horenczyk, K Liebkind and P Vedder (2001), ‘Ethnic identity, immigration, and well-­being: an international perspective’, Journal of Social Issues, 57 (3), 493–510 Piore, M.J (1979), Birds of Passage: Migrant Labor and Industrial Societies, Cambridge: Cambridge University Press Ruble, B (1989), ‘Ethnicity and Soviet cities’, Soviet Studies, 41 (3), 401–14 Russo, G and E van Hooft (2011), ‘Identities, conflicting behavioural norms and the importance of job attributes’, Journal of Economic Psychology, 32 (1), 103–19 Schoeni, R.F (1998), ‘Labor market assimilation of immigrant women’, Industrial and Labor Relations Review, 51 (3), 483–500 Sen, A (1977), ‘Rational fools: a critique of the behavioral foundations of economic theory’, Philosophy and Public Affairs, (4), 317–44 Sjaastad, L.A (1962), ‘The costs and returns of human migration’, Journal of Political Economy, 70 (5), pt 2, 80–93 Unger, J., P Gallaher, S Shakib, A Ritt-­Olson, P Palmer and C Johnson (2002), ‘The AHIMSA Acculturation Scale: a new measure of acculturation for adolescents in a multicultural society’, Journal of Early Adolescence, 22 (3), 225–51 United Nations Economic Commission for Europe (UNECE) (2006), Conference of European Statisticians Recommendations for the 2010 Censuses of Population and Housing, New York: United Nations Zimmermann, K.F (2007a), ‘The economics of migrant ethnicity’, Journal of Population Economics, 20 (3), 487–94 Zimmermann, K.F (2007b), ‘Migrant ethnic identity: concept and policy implications’, Ekonomia, 10 (1), 1–17 Zimmermann, K.F., A.F Constant and L Gataullina (2009), ‘Naturalization proclivities, ethnicity and integration’, International Journal of Manpower, 30 (1–2), 70–82 Zimmermann, L., L Gataullina, A Constant and K.F Zimmermann (2008), ‘Human capital and ethnic self-­ identification of migrants’, Economics Letters, 98 (3), 235–39 Zimmermann, L., K.F Zimmermann and A Constant (2007), ‘Ethnic self-­identification of first-­generation immigrants’, International Migration Review, 41 (3), 769–81 PART II THE MOVE 2  Modeling individual migration decisions* John Kennan and James R Walker 1  INTRODUCTION We review empirical analyses of migration decisions, using life-­cycle models to interpret migration histories The starting points are Schultz (1961), who considered migration as a form of investment in human capital, and DaVanzo (1983), who documented the richness of individual migration histories, pointing out that although most individuals never move, those who are likely to move again, often returning to a home location.1 This means that migration decisions should be viewed as a sequence of location choices, where the individual knows there will be opportunities to modify or reverse moves that not work out well Sjaastad’s (1962) treatment of migration as an investment emphasizes the dynamic aspect of migration − expected costs and payoffs to migration change over time Viewed within a life-­cycle perspective, individuals (or families) decide whether and when to move Allowing households to make multiple migration decisions substantially increases the model’s complexity Decisions made in previous periods (for example, savings, education, marriage and fertility) determine choices available in the current period, and expectations of future events also influence current decisions Within this perspective migration and fertility choices are connected through the ‘primitives’ of the decision-­ making process: current opportunities (determined in part by decisions made in the past), expectations (anticipated future events and outcomes) and preferences (values assigned to different ­outcomes) The economic perspective thus provides a unified framework connecting several important demographic behaviors Any of a large number of factors may influence an individual to move from their current location The special nature of the ‘home’ location suggests that home will be a favored destination of natives who have moved away A common explanation for a sequence of moves builds on the idea of differences between expected and realized outcomes and the role of learning An initial move is motivated by anticipated outcomes The move is made and the realized outcome is experienced If the realized outcome is greatly different than expected, a second, corrective, move may be required (which may be to home, given its the special nature) or learning may take place and the updated assessment of the costs and payoffs (at all locations) may induce onward migration.2 This perspective also provides an explanation for the commonly observed pattern that places of high in-­migration also experience high out-­migration Prospective migrants may be uncertain about payoffs and costs, and indeed, prospective migrants may differ in their attitudes toward risk, liquidity to finance a move and insurance against poor labor market outcomes following a move To the extent that moving entails a large monetary cost, financing a move becomes important Families can play a central role in providing information and in helping to finance a move Families can also provide an important form of insurance Social insurance programs provide 39 40   International handbook on the economics of migration a safety net for some families, while family members provide an important backstop These types of intra-­family exchanges have been extensively modeled and investigated in developing countries In particular the role of information, credit constraints and intergenerational transfers have received the most attention As in all forward-­looking models of behavior, specification of how expectations are formed plays a critical role in the model specification Most commonly, individuals are assumed to have rational expectations (and thus use all available information) When agents are not well informed, a model of learning is necessary to determine subjective probabilities of events and payoffs The optimal policy may involve experimentation early in the process to gain information relevant for subsequent payoffs Such models can help rationalize chain and return migration Selection is an important element in every empirical study of migration.3 As is well understood, the fundamental problem is that many patterns may stem from true causal forces or may arise from unobserved individual characteristics For example, if we observe large positive economic gains accruing to migrants, is this evidence of the gain anticipated by the migrant or does the gain reflect characteristics of the migrant that might have also led to above average payoffs even if the individual had not moved? A theoretical framework is needed to control for selection, which is much more difficult than in many other applications, where panel data can generate observations on the same person in different states (albeit at different points in time) Here, most individuals are seen only in a single location, and even the movers move only a few times; no one is observed in all locations Descriptive studies are useful for highlighting important flows and temporal patterns, but the ubiquity of selection means that an analytic framework is needed to unravel behavioral mechanisms In this chapter we describe recent research modeling of individual migration decisions These models consider multiple destinations and multiple periods: individuals may move more than once, may move to several distinct locations or may return to a previously visited location In addition, the models recognize the special status of the home location We proceed by first outlining the main elements of the models and then present empirical applications 2  COMPONENTS OF A MODEL OF MIGRATION The basic idea of a (dynamic) migration model is that there are a number of alternative (mutually exclusive) payoff flows, associated with different locations, with options to switch between these alternatives, subject to a moving cost At any given time, the individual is in a particular location, and must choose whether to stay there or go somewhere else, exchanging the current payoff flow for one of the alternatives The future payoffs in the current location are generally uncertain, and the alternative payoffs even more so The individual acts so as to maximize the expected present value of the realized payoffs, net of moving costs The model is most conveniently formulated in the language of dynamic programming, since it involves solving variants of the same choice problem over and over again Suppose there are J locations, and the payoff flow in each location is a random variable with a known distribution Let x be the state vector (which includes current location and Modeling individual migration decisions  ­41 age, as well as all currently available information that helps to predict future payoffs, as discussed below) The utility flow for someone who chooses location j is specified as u (x, j) zj where zj is a random variable that is assumed to be independent and identically distributed across locations and across periods, and independent of the state vector, representing influences on migration decisions that are not included in the model Let p (xr x, j) be the transition probability from state x to state xr, if location j is chosen The decision problem can be written in recursive form as: where and V (x, z) max (v (x, j) zj) j v (x, j) u (x, j) b a p (xr x, j) v (xr) xr v (x) EzV (x, z) and where b is the discount factor, and Ez denotes the expectation with respect to the distribution of the J-­vector z with components zj The interpretation is as follows The payoff shocks (for all locations) are realized before the location decision is made, and the distinction between n (x, j) and v (x) is that one represents the continuation value for each alternative choice, while the other represents the expectation of the optimized continuation value, taken before the payoff shocks have been realized The above formulation can be used to describe both finite-­horizon and infinite-­ horizon problems In the finite-­horizon case, it is convenient to remove age from the state vector, and write the model as: Vs (x, z) max vs (x, j) zj) vs (x, j) us (x, j) b a ps (xr x) vs21 (xr) j xr vs (x) Ez (x, z) where s is the number of periods remaining, with v0 (x) V0 (x, z) Thus the decision for s is a simple static choice problem, and once this has been solved, the decision problem for s can be specified explicitly, and the solution of this can be used to specify the decision problem for s 3, and so on In other words the general solution can be obtained by backward induction The model is implemented by specifying the function u (x, j) as well as the transition probabilities, up to a vector of unknown parameters q Ideally, the specification should parsimoniously capture the most important features of the choice problem, giving plausible interpretations of the main features of the data, while also facilitating the prediction of behavior in situations not actually seen in the data (such as alternative policy environments) A major limitation of models of this kind is that the solution cannot be computed 42   International handbook on the economics of migration unless special assumptions are made For one thing, computation of the function v involves an integral overall possible realization of the payoff shocks, and if there are many locations this is infeasible unless the payoff shocks are drawn from a distribution for which a closed-­form solution for this integral is known (which in practice means the generalized extreme value distribution) We assume that zj is drawn from the Type I extreme value distribution In this case, following McFadden (1974) and Rust (1987), we have: exp (vs (x)) exp (g) a exp (vs (x, k)) J k 51 where g is the Euler constant Let rs (x, j) be the probability of choosing location j, when the state is x with s periods remaining Then: rs (x, j) exp (g vs (x, j) vs (x)) exp (vs (x, j)) a exp (vs (x, k)) J k 51 A more general computational issue is that solution of the dynamic programming problem requires computation of the continuation values from all decision nodes that might possibly be reached, and the number of such nodes explodes as the number of possible states increases.4 This severely limits the number of locations that can be considered For instance, the natural specification of a location is a local labor market, but it is impossible to compute such a model for a large economy; in applications to the United States economy, locations must be aggregated to the level of states, or census regions Even then, there is an outrageous number of decision nodes, but since the vast majority of these are almost never reached, the solution can be well approximated by just ignoring most of them Thus we assume that the individual retains information about wage draws in at most two locations (even if more locations have actually been visited) Another important general consideration is that the initial conditions of the decision problem are typically the result of some previous decisions, which means that even if the stochastic components of payoffs are randomly assigned ex ante, the distribution of these components in the data is contaminated by selection bias This is of course especially true for models of migration by older people But it may be reasonable to assume away the initial conditions problem in models that begin at the point of entry to the labor force 3  AN EMPIRICAL MODEL In Kennan and Walker (2011) we develop a dynamic model of migration decisions, and estimate it using data on young white male high school graduates in the United States We follow respondents of the National Longitudinal Survey of Youth, Cohort 1979 (NLSY79) from age 20 until their mid-­thirties We define locations as states We include a utility premium for workers residing in their ‘home’ location, defined as Modeling individual migration decisions  ­43 the state of residence at age 14 The details of the model are outlined below, followed by a summary of the empirical results We also present results for white male college graduates 3.1  Payoff Flows Let O (O0, O1) be a vector recording the current and previous locations (with the convention that O1 if there is no previous location), and let w (w0, w1) be a vector recording wage information at these locations The state vector x consists of O, w and age The flow payoff for someone whose ‘home’ location is h is specified as: where uh (x, j) uh (x, j) zj uh (x, j) a0w (O0, w0) a akYk (O0) aHc (O0 h) Dt (x, j) Here the first term refers to wage income in the current location This is augmented by the non-­pecuniary variables, Yk (O0) , representing amenity values The parameter aH represents a premium that allows each individual to have a preference for his native location (c (A) denotes an indicator meaning that A is true) The cost of moving from O0 to j for a person of type t is represented by Dt (x, j) The unexplained part of the utility flow, zj, may be viewed as either a preference shock or a shock to the cost of moving, with no way to distinguish between the two 3.2  Wages It is assumed that workers know their wage in the current location, are assumed to have rational expectations and to know the distribution of offered wages at all other locations The wage of individual i in location j at age a in year t is specified as: wij (a) mj ui G (Xi, a, t) hi eij (a) where mj is the mean wage in location j, u is a permanent location match effect, G (Xi, a, t) represents a (linear) time effect and the effects of observed individual characteristics, hi is an individual effect that is fixed across locations, and e is a transient effect We assume that h, u and e are independent random variables that are identically distributed across individuals and locations We also assume that the realizations of h and u are seen by the individual The age component and the fixed effect are common to all locations and consequently not influence migration decisions Thus since the current realization of the transient wage component is known only after the current location has been chosen, migration decisions are driven exclusively by the State means (mj) and the match specific component (uij) between worker i and location j For computational reasons, we model the worker-­location component as a discrete distribution with three points of support (low, middle and high) Even this simple model gives workers two motivations to migrate: to leave a bad local labor market (a low mj) or 44   International handbook on the economics of migration a bad location match, (a low uij) The incentives to migrate are strong For example, the 90-­10 differential across state means is about $4700 a year (in 2010 dollars) and the value of replacing a bad location match with a good one is about $17 000 a year 3.3  Moving Costs Let D (O0, j) be the distance from the current location to location, j and let A (O0) be the set of locations adjacent to O0 (where states are adjacent if they share a border) The moving cost is specified as: Dt (x, j) (g0t g1D (O0, j) g2c ( j [ A (O0)) g3c ( j O1) g4a g5nj) c ( j O0) We allow for unobserved heterogeneity in the cost of moving: there are several types, indexed by t with differing values of the intercept g0 In particular, there may be a ‘stayer’ type, meaning that there may be people who regard the cost of moving as prohibitive The moving cost is an affine function of distance, but moves to a previous location may be less costly, and moves to an adjacent location may also be less costly (because it is possible to change states while remaining in the same general area) In addition, the cost of moving is allowed to depend on age, a Finally, we allow for the possibility that it is cheaper to move to a large location, as measured by population size nj It has long been recognized that location size matters in migration models (see, for example, Schultz, 1982) For example, a person who moves to be close to a relative is more likely to have relatives in California than in Wyoming One way to model this in our framework is to allow for more than one draw from the distribution of payoff shocks in each location Alternatively, location size may affect moving costs − for example, relatives might help reduce the cost of the move In practice, both versions give similar results 3.4  Transition Probabilities The state transition probabilities ps (xr x, j) for this model are straightforward First, if no migration occurs this period, then the state remains the same except for the age component If there is a move to a previous location, the current and previous locations are interchanged And if there is a move to a new location, the current location becomes the previous location, and the new location match component of wages is drawn at random In all cases, age is incremented by one period (equivalently, s is decremented by one period) 3.5  Results We show the basic estimation results from Kennan and Walker (2011), along with estimates of the same model using data for college graduates The estimates in Table 2.1 show that expected income is an important determinant of migration decisions, for both education groups Even though the overall migration rate is much higher for college graduates, the parameter estimates are quite similar for the two samples, aside from a substantially lower estimated migration cost for college graduates Modeling individual migration decisions  ­45 Table 2.1  Interstate migration, white male high school and college graduates High School Utility and cost Disutility of moving (g0) Distance (g1) (1000 miles) Adjacent location) (g2) Home premium (a H) Previous location (g3) Age (g4) Population (g5) (millions) Stayer probability Cooling (a1) (1000 degree days) Income (a0) Wages Wage intercept Time trend Age effect (linear) Age effect (quadratic) Ability (AFQT) Interaction (Age,AFQT) Transient s.d Transient s.d Transient s.d Transient s.d Fixed effect Fixed effect Fixed effect Location match (tu) Loglikelihood   No of person-year observations   No of Respondents/no of moves College q^ s^ q q^ s^ q 4.794 0.267 0.807 0.331 2.757 0.055 0.654 0.510 0.055 0.314 0.565 0.181 0.214 0.041 0.357 0.020 0.179 0.019 0.019 0.100 3.583 0.483 0.852 0.168 2.374 0.084 0.679 0.221 0.001 0.172 0.686 0.131 0.130 0.019 0.178 0.024 0.116 0.058 0.011 0.031 −5.133 0.245 −0.034 0.008 7.841 0.356 −2.362 0.129 0.011 0.065 0.144 0.040 0.217 0.007 0.375 0.015 0.546 0.017 1.306 0.028 0.113 0.036 0.296 0.035 0.933 0.016 0.384 0.017 −4214.160 4274 423/124 −6.019 0.496 0.065 0.008 7.585 0.649 −2.545 0.216 −0.045 0.158 0.382 0.111 0.212 0.007 0.395 0.017 0.828 0.026 3.031 0.037 0.214 0.024 0.660 0.024 1.020 0.024 0.627 0.016 −4902.453 3114 440/267 Sources:  Estimates for high school graduates are from Table of Kennan and Walker (2011) Estimates of college graduates are calculated from Kennan (2011) The importance of the home location is clearly shown in Table 2.1, especially for the high school sample This attachment to home reduces out migration and induces return migration It helps to explain why most people never move, despite large spatial wage differences; it also implies that the losses from forced migration (such as the migration due to hurricane Katrina) are very large.5 3.6  How Big Are the Moving Costs? There are big differences in wages across states, and the estimated dispersion of the worker-­location component of wages is also quite large Yet migration rates are low: the 46   International handbook on the economics of migration Table 2.2  Average moving costs Move origin and destination (2010 $) Previous location None Home Other Total From home To home Other Total −147 619 – −150 110 −147 930 138 095 18 686 113 447 25 871 −39 677 −124 360 −67 443 −97 656 −139 118 −9924 −87 413 −80 768 Source:  Table of Kennan and Walker (2011) interstate migration rate for white men in the NLSY is 2.9 percent for high school graduates, and while the rate for college graduates is much higher (8.6 percent), this still seems low in relation to the estimated wage gains A natural reaction is to infer that moving costs must be very high Indeed, if it is assumed that the moving cost is the same for everyone, the estimated model indicates that the cost is on the order of $300 000 (for high school graduates) Yet people move, and those who move tend to move again, and it is hard to believe that they are paying costs of this magnitude every time they move The answer to this riddle is that people are heterogeneous For some people, at some times, the moving cost is very high; for some people, at some times, the cost is quite low The model allows us to quantify the extent of this heterogeneity In particular, we can estimate the average moving costs for those who actually move The estimates for the high school sample are given in Table 2.2 There is considerable variation in these costs, but for a typical move the cost is negative The interpretation of this is that the typical move is not motivated by the prospect of a higher future utility flow in the destination location, but rather by unobserved factors yielding a higher current payoff in the destination location, compared with the current location That is, the most important part of the estimated moving cost is the difference in the payoff shocks In the case of moves to the home location, on the other hand, the estimated cost is positive; most of these moves are return moves, but where the home location is not the previous location the cost is large, reflecting a large gain in expected future payoffs due to the move 3.7  Why Do College Graduates Move So Much? It is well known that the migration rate for skilled workers is much higher than the rate for unskilled workers; in particular the migration rate for college graduates is much higher than the rate for high school graduates (see, for example, Bound and Holzer, 2000; Greenwood, 1997; Molloy et al., 2011; Topel, 1986; Wozniak, 2010) Malamud and Wozniak (2008), using draft risk as an instrument for education, find that an increase in education causes an increase in migration rates (the alternative being that people who go to college have lower moving costs, so that they would have higher migration rates even if they did not go to college).6 The model described in Table 2.1 can be used to simulate the extent to which the differences in migration rates for college graduates can be explained by differences in expected incomes, as opposed to differences in moving costs This distinction affects the interpretation of measured rates of return on Modeling individual migration decisions  ­47 Table 2.3  Interstate migration, white male high school and college graduates High school Utility and cost Disutility of moving (g0) Distance (g1) (1000 miles) Adjacent location) (g2) Home premium (a H) Previous location (g3) Age (g4) Population (g5) (millions) Stayer probability Cooling (a1) (1000 degree days) Income (a0) Wages Location match (tu) NLSY data Observations Migration rate Simulated Observations Migration rate College wages College q^ q^ 4.794 0.267 0.807 0.331 2.757 0.055 0.654 0.510 0.055 0.314 3.570 0.482 0.852 0.167 2.382 0.085 0.678 0.227 0.001 0.172 0.384 0.634 4274 2.90% 3114 8.57% 427 429 3.16% 427 421 3.96% 311 571 8.60% High school wages 311 428 8.23% Source:  Estimates of utility and cost parameters from Table 2.1 investments in college education For example, if college graduates move more because the college labor market has higher geographical wage differentials, then a substantial part of the measured return to college is spurious, because it is achieved only by paying large moving costs Table 2.3 shows the observed annual migration rates for the high school and college graduate samples along with the migration rates predicted by the estimated model, where these rates are computed by using the model to simulate the migration decisions of 100 replicas of each person in the data The extent to which the large observed difference in migration rates can be attributed to differences in geographical wage dispersion can be measured by simulating the migration decisions that would be made by one group if they faced the same wage dispersion as the other group Thus, according to the model, the migration rate of high school graduates would increase considerably if they faced the higher wage dispersion seen by college graduates, but the migration rate in this simulation is still only about percent per year, compared with 8.6 percent for the college sample The reverse experiment gives a similar result: the migration rate for college graduates facing the high school wage process would still be more than twice the observed rate for high school graduates Thus although geographical wage dispersion can explain a nontrivial part of the difference in the explained migration rates, the model attributes the bulk of this difference to other factors, such as differences in moving costs 48   International handbook on the economics of migration 3.8  Spatial Labor Supply Elasticities The estimated model can be used to analyze labor supply responses to geographical wage differentials We are interested in both the magnitude and the timing of these responses For example, Blanchard and Katz (1992) found that the half-­life to a unit shock to the relative wage is more than a decade Studies by Barro and Sala-­i-Martin (1991) and Topel (1986) report similar findings Given that college graduates move more often than high school graduates, it is also interesting to ask whether the greater mobility of college graduates is associated with a more elastic response to geographic wage differences Since the model assumes that the wage components relevant to migration decisions are permanent, it cannot be used to predict responses to wage innovations in an environment in which wages are generated by a stochastic process Instead, it is used to answer comparative dynamics questions: the estimated parameters are used to predict responses in a different environment The first step is to take a set of young white males who are distributed over states as in the 1990 census data, and allow the population distribution to evolve, by iterating the estimated transition probability matrix (given the observed wages) The transition matrix is then recomputed to reflect wage increases and decreases representing a 10 change in the mean wage of an average 30-­year-­old, for selected states, and the population changes in this scenario are compared with the baseline simulation Supply elasticities are measured relative to the supply of labor in the baseline calculation For example, the elasticity of the response to a wage increase in California after five years is computed w as, DL Dw L , where L is the number of people in California after five years in the baseline calculation, and DL is the difference between this and the number of people in California after five years in the counterfactual calculation Figure 2.1 shows the results for three large states that are near the middle of the one-­ period utility flow distribution The high school results show substantial responses to spatial wage differences, occurring gradually over a period of about 10 years The wage responses for college graduates are larger (with a supply elasticity around unity), and the length of the adjustment period is longer This is consistent with the hypothesis that college graduates face substantially lower moving costs than high school graduates The reason for the long adjustment period is that wage differences are just one of many influences on migration decisions Tilting the wage difference in favor of a particular location therefore has a relatively small effect on migration probabilities, but since this effect is permanent, while the payoff shocks are transient, the cumulative effect of wage differences is substantial in the longer run 4  WELFARE MIGRATION At various times during the past 30 years, the existence of ‘welfare magnets’ has surfaced in public policy debates and particularly during welfare reform in the mid-­1990s in the US Prior to the reform, Aid to Families with Dependent Children (AFDC) was the primary source of income support States followed federal guidelines but were otherwise free to determine benefits schedules A state offering relatively high income support Modeling individual migration decisions  ­49 Responses to 10% wage changes Proportional population change White male high school graduates 0.12 0.1 0.05 –0.05 –0.1 10 11 12 13 14 15 16 17 Year Responses to 10% wage changes Proportional population change White male college graduates 0.12 0.1 0.05 –0.05 –0.1 Year 10 11 CA, decrease IL, decrease NY, decrease CA, increase IL, increase NY, increase 12 13 Figure 2.1  Geographical labor supply elasticities could function as a ‘magnet’ for low-­income population; retaining those already living in the state and attracting poor from relatively low-­benefits states Indeed, during the welfare reform debates there was concern that competition among states would become a ‘race to the bottom’ in setting benefit levels.7 In Kennan and Walker (2010) we develop a dynamic model to investigate the migration decision making of welfare-­eligible mothers.8 Our motivation for estimating the model was to investigate migration flows that might be induced by alternative welfare benefit policy regimes not seen in the data Specifically, we investigated the implied migration flows if all states set benefits equal to Mississippi’s (the lowest benefit) or California’s (the highest benefit) Our estimates suggest that income has a significant but quantitatively weak influence on migration 50   International handbook on the economics of migration We find that migration adjusts slowly, but with slightly greater responsiveness to earnings than to benefits This is to be expected as everyone is affected by wages, but high-­wage women are not much affected by welfare benefits In particular, women with favorable individual fixed effects are unlikely to be on benefits In the second set of counterfactual experiments we investigate migration responses to uniform benefit levels for all states Differences in AFDC benefits are seen as the driving force behind welfare-­induced migration Investigation of a national benefit level is also interesting because the result is a priori ambiguous − implementing a national welfare benefit may serve to increase or decrease migration rates.9 One striking feature of our results is the insensitivity of migration to substantial changes in either benefits or wages Another rather surprising result is that uniform benefits increase migration A natural intuition is that state variation in benefit levels should increase migration relative to a uniform benefit regime The intuition is correct if the state benefit levels are independent of other influences on migration However, benefits are in fact negatively correlated with these other influences, and thus serve to dampen migration flows Women are eligible for welfare benefits only if they are single, with dependent children, and if their earnings are low Thus a complete specification of the value function would require a model of marriage and divorce, including a theory of how the marital surplus is divided, and of how likely it is that the surplus disappears, so that the marriage breaks up This is a tall order Moreover, to incorporate the temporal change in benefits requires a significant extension to our model − we must model beliefs about future benefits For a discussion and an application of such forward-­looking behavior that does not consider migration see Keane and Wolpin (2002a, 2002b) In addition, a woman who is out of the labor force (either because she is collecting welfare or because she is married and doing non-­market work) forgoes the human capital accumulation associated with labor market experience Thus a fully specified model should encompass the relationship between current work and future wages, as in Shaw (1989) and Imai and Keane (2004) In particular, the opportunity cost of being on welfare may be considerably higher than the current wage Thus a more complete model would require a much larger state space than that used here, with marital status, number of children, and accumulated market work experience treated as state variables Our model can be viewed as a simplification based on two approximations First, when a welfare-­eligible woman marries, she receives no surplus, either because the surplus is negligible or because her share is negligible Second, the experience associated with non-­market work yields the same increment of human capital as the same amount of market work experience 5  HOUSEHOLD MIGRATION DECISIONS Mincer (1978) recognized that when married individuals have distinct preferences and different opportunities across locations, the most preferred location for a couple may differ from the locations preferred by each individual Mincer assumed that couples maximize the sum of their incomes The couple has to decide where to live and whether to stay together This gives rise to notions of ‘tied-­movers’ and ‘tied-­stayers’ along with predictions on who should remain married and who should divorce One interesting prediction is that migration is more likely after a divorce (as the newly unwed individuals Modeling individual migration decisions  ­51 move from their ‘tied’ locations) Mincer also noted that these forces become stronger as women’s labor force participation and earnings increase Gemici (2011) extends Mincer’s static formulation with a model of household decision-­making that considers multiple locations over multiple periods Gemici seeks to quantify the inhibiting effect of location ties on labor mobility, wage growth and marital stability The couple acts so as to maximize the expected present value of the sum of their consumption levels, allowing the possibility that this might imply that divorce is optimal Consumption is a linear combination of a private good, a public good that is produced by the marriage, and leisure, with weights that depend on the duration of the marriage and the presence of children Location is defined as one of the nine census regions in the US Each period each member of a household residing in location , receives a wage offer from an alternative location with some probability The household must then decide on where to live and whether to work, allowing for the possibility that the spouses might live apart to take advantage of attractive wage offers in distinct locations The model is estimated by the method of simulated moments, using data from the Panel Study of Income Dynamics 6  IMMIGRATION The analysis of individual immigration decisions is in principle no different from the analysis of internal migration, but in practice empirical work in this area is severely limited by the scarcity of panel data sets spanning national borders Thom (2010) and Lessem (2011) model migration between Mexico and the US using one of the few available data sets, from the Mexican Migration Project Thom (2010) uses a two-­location model with borrowing constraints and a concave utility of consumption.10 We focus on Lessem’s model, which emphasizes income differences and family links as the main variables of interest, while allowing for repeat and return migration, and allowing for other idiosyncratic influences on migration choices Most of the migrants in the data set are illegal, and so the cost of migrating depends on border enforcement activity Variations in border patrol man-­hours along different sections of the border are used to estimate the relationship between cost and migration decisions The richness of the model can be illustrated by considering the effect of increased enforcement on the length of spells in the US For someone who is already in the US illegally, tougher border enforcement means that it will be more difficult to return to the US after returning to Mexico Thus one paradoxical effect of increased enforcement is that it tends to keep illegal immigrants in the country for longer periods of time Although Lessem’s empirical results indicate that the magnitude of this effect is not large, the estimates illustrate the ability of the model to quantify behavioral responses that are just not accessible without a well-­specified model of how migration decisions are made Such a model of course entails making various restrictive assumptions, but the assumptions can be examined and modified, enabling the analysis to move beyond mere qualitative descriptions Lessem’s model is estimated by maximum likelihood.11 As in Gemici’s model, when spouses have the option of living apart in order to take advantage of income opportunities in different locations, the number of relevant contingencies in the decision problem becomes unmanageable Lessem assumes that one spouse (generally the husband) is the 52   International handbook on the economics of migration ‘primary’ mover, and restricts the choice set so that the other spouse cannot choose to live in one of the US locations unless the primary mover is also there; in addition, she assumes that the decisions of the two spouses are made in sequence, thereby restricting the number of choices that must be considered at each node of the decision problem In the data it is rare to find the wife working in the US while the husband stays in Mexico Thus by excluding this option Lessem makes the required computations feasible without losing much in terms of realism As a result, she can measure the extent to which migration is influenced by family ties, and in particular the relevance of family ties for return migration decisions In terms of substantive empirical results, the most important feature of Lessem’s model is that it gives a coherent analysis of the extent to which increases in Mexican wages relative to US wages leads to a reduction in the flow of immigrants to the US At this stage, the model does not deal with general equilibrium effects of migration, but a well-­specified model of the supply side of the labor market is an essential first step toward a full-­blown general equilibrium analysis An important feature of Lessem’s model is that it allows for uneven development across regions within Mexico, so that emigration to the US and internal migration within Mexico can be analyzed within the same decision problem Lessem’s estimates imply that an increase in wages in Mexico reduces migration to the US and increases return migration from the US A 10 percent increase in Mexican wages reduces the amount of time that individuals spend in the US by approximately percent Increased border enforcement decreases not only immigration from Mexico, but also return migration from the US Simulations indicate that a 50 percent increase in enforcement would reduce the amount of time that migrants spend in the US by up to percent, depending on the allocation of additional enforcement at the border 7  CONCLUSION We have presented an analytical framework capable of modeling individual migration decisions over the life cycle Through our choice of examples we show that the framework can include family and household determinants as well as standard economic factors The causal linkages among core demographic processes (such as migration, child bearing, household formation, marriage and divorce) can potentially be recovered assuming appropriate longitudinal (event history) data are available The framework requires parsimonious empirical specifications which frequently mandates strong functional form and or distributional assumptions Yet, these assumptions can and should be empirically investigated The most demanding challenge imposed by the framework is the specification of a parsimonious yet flexible specification that recovers the primary features of the data The payoff from this analytical approach is the ability to trace out life-­cycle and distributional consequences of policies or (social or economic) environments that are conjectured and not estimable using existing data The increasing availability of longitudinal household microdata suggest a wealth of future research opportunities to investigate Even though the work to date is quite limited, it shows that this approach is empirically fruitful and adds to our understanding of individual life-­cycle decision-­making Modeling individual migration decisions  ­53 NOTES   *                 10 11 We thank an anonymous referee and the editors, Amelie F Constant and Klaus F Zimmermann, for helpful comments See Dierx (1988) for what we believe is the earliest model of return migration See Pessino (1991) for a model of learning applied to data on migration in Peru For example, see Bayer and Juessen (2012) This is Bellman’s ‘Curse of Dimensionality’ Interestingly, recent work by Gregory (2011) suggests that rebuilding subsidies offered post-­Katrina by the US federal government incurred relatively small deadweight losses due to the low-­income elasticity of return migrants to New Orleans Notowidigdo (2011) interprets the difference in migration rates between skilled and unskilled workers in terms of differential responses to local demand shocks When there is an adverse local shock, house prices decline Low-­wage workers spend a large fraction of their income on housing, so the decline in the price of housing substantially reduces the incentive to migrate, while this effect is less important for high-­ wage workers At the same time, public assistance programs respond to local shocks, and these programs benefit low-­wage workers (although the relevance of this in explaining the differential migration rates for high-­school and college graduates is doubtful, especially for men) See Gelbach (2004) for evidence of welfare magnets The effect of welfare benefits and particularly AFDC on migration decisions of poor women with children has a long history The common perception, dating back to the English Poor Laws of the nineteenth century, is that high welfare benefits attract poor (welfare-­eligible) individuals from other locations, and retain poor among the local population However, benefits may help finance a move for liquidity-­ constrained households (that is, access to welfare benefits may lessen poverty traps) The empirical evidence on the influence of AFDC benefits is mixed See Brueckner (2000) for a summary of the literature In principle, the same issue faces countries within the European Union; there however, language differences and cultural differences may serve as barriers to migration It might seem that a national welfare benefit must reduce migration But in the model, income is either the wage or the welfare benefit, whichever is higher In the absence of any welfare benefits, people tend to move to high-­wage locations Introducing a (uniform) national welfare scheme reduces the incentive to move towards high-­wage locations, since even in such a location, the wage might turn out to be lower than the welfare benefit And if the benefit level is higher, migration due to spatial wage differences is reduced On the other hand a welfare scheme that pays different benefits in different locations (like the AFDC scheme) introduces additional migration incentives, especially for low-­wage workers Standardizing the benefit across locations eliminates the incentive to move towards high-­benefit locations, but if the standardized benefit is low, the incentive to move towards high-­wage locations is enhanced The main state variable in Thom’s model is the level of assets Initially, young people in Mexico not have enough money to get across the border They save enough to pay the cost, and then they migrate In the US they build up assets to the point where they would rather live in Mexico, because they can consume at a reasonably high level, and staying in the US to build up a higher asset level is not worth it, because they have diminishing marginal utility, and a preference for living in Mexico So they go back Then after returning to Mexico, it may happen that the asset level falls to a point where it is optimal to return to the US Thom (2010) uses the method of simulated moments REFERENCES Barro, R.J and X Sala-­i-Martin (1991), ‘Convergence across states and regions’, Brookings Papers on Economic Activity, 1, 107–58 Bayer, C and F Juessen (2012), ‘On the dynamics of interstate migration: migration costs and self-­selection’, Review of Economic Dynamics, 15 (3), 377–401 Blanchard, O.J and L.F Katz (1992), ‘Regional evolutions’, Brookings Papers on Economic Activity, 1, 1–37 Bound, J and H.J Holzer (2000), ‘Demand shifts, population adjustments, and labor market outcomes during the 1980s’, Journal of Labor Economics, 18 (1), 20–54 Brueckner, J.K (2000), ‘Welfare reform and the race to the bottom: theory and evidence’, Southern Economic Journal, 66 (3), 505–25 54   International handbook on the economics of migration DaVanzo, J (1983), ‘Repeat migration in the United States: who moves back and who moves on?’, Review of Economics and Statistics, 65 (4), 552–9 Dierx, A.H (1988), ‘A life-­cycle model of repeat migration’, Regional Science and Urban Economics, 18 (3), 383–97 Gelbach, J.B (2004), ‘Migration, the lifecycle and state benefits: how low is the bottom?’, Journal of Political Economy, 112 (5), 1091–130 Gemici, A (2011), ‘Family migration and labor market outcomes’, working paper, New York University, Department of Economics Greenwood, M.J (1997), ‘Internal migration in developed countries’, in M.R Rosenzweig and O Stark (eds), Handbook of Population and Family Economics, 1B, Amsterdam: North Holland, pp. 851–87 Gregory, J (2011), ‘The impact of rebuilding grants and wage subsidies on the resettlement choices of Hurricane Katrina victims’, unpublished paper, University of Michigan, Department of Economics Imai, S and M.P Keane (2004), ‘Intertemporal labor supply and human capital accumulation’, International Economic Review, 45 (2), 601–41 Keane, M.P and K.I Wolpin (2002a), ‘Estimating welfare effects consistent with forward-­looking behavior Part I: Lessons from a simulation exercise’, Journal of Human Resources, 37 (3), 570–99 Keane, M P and K.I Wolpin (2002b), ‘Estimating welfare effects consistent with forward-­looking behavior Part II: Empirical results’, Journal of Human Resources, 37 (3), 600–622 Kennan, J (2011), ‘Higher education subsidies and human capital mobility’, unpublished manuscript, University of Wisconsin-­Madison, Department of Economics Kennan, J and J.R Walker (2010), ‘Wages, welfare benefits and migration’, Journal of Econometrics, 156 (1), 229–38 Kennan, J and J.R Walker (2011), ‘The effect of expected income on individual migration decisions’, Econometrica, 79 (1), 251–91 Lessem, R (2011), Mexico–U.S immigration: effects of wages and border enforcement, working paper, Carnegie Mellon University, Tepper School of Business, Pittsburgh Malamud, O and A Wozniak (2008), ‘The impact of college graduation on geographic mobility: identifying education using multiple components of Vietnam draft risk’, Harris School Working Paper No 08.11, University of Chicago, The Harris School, Chicago, IL McFadden, D.D (1974), ‘Conditional logit analysis of qualitative choice behavior’, in P Zarembka (ed.), Frontiers in Econometrics, New York: Academic Press, pp. 105–42 Mincer, J (1978), ‘Family migration decisions’, Journal of Political Economy, 86 (5), 749–73 Molloy, R., C.L Smith and A Wozniak, A (2011), ‘International migration in the United States’, Journal of Economic Perspectives, 25 (3), 173–96 Notowidigdo, M.J (2011), ‘The incidence of local labor demand shocks’, NBER Working Paper No 17167, National Bureau of Economic Research (NBER), Cambridge, MA Pessino, C (1991), ‘Sequential migration theory and evidence from Peru’, Journal of Development Economics, 36 (1), 55–87 Rust, J.P (1987), ‘Optimal replacement of gmc bus engines: an empirical model of Harold Zurcher’, Econometrica, 55 (5), 999–1033 Schultz, T.P (1982), ‘Lifetime migration within educational strata in Venezuela: estimates of a logistic model’, Economic Development and Cultural Change, 30 (3), 559–93 Schultz, T.W (1961), ‘Investment in human capital’, American Economic Review, 51 (1), 1–17 Shaw, K (1989), ‘Life cycle labor supply with human capital accumulation’, International Economic Review, 30 (2), 431–56 Sjaastad, L.A (1962), ‘The costs and returns of human migration’, Journal of Political Economy, 70 (5, pt 2), 80–89 Thom, K (2010), ‘Repeated circular migration: theory and evidence from undocumented migrants’, working paper, New York University, Department of Economics Topel, R.H (1986), ‘Local labor markets’, Journal of Political Economy, 94 (3, pt 2), S111–S143 Wozniak, A (2010), ‘Are college graduates more responsive to distant labor market opportunities?’, Journal of Human Resources, 45 (4), 944–70 3  The economics of circular migration* Amelie F Constant, Olga Nottmeyer and Klaus F Zimmermann 1  INTRODUCTION The economics of migration literature saw an increasing interest in circular or repeat migration in the past two decades At the same time, more and more programs concerning circular, revolving-­door or ‘va-­et-­vient’ migration started sprouting By the year 2003 about 176 bilateral labor agreements had been signed among the Organisation for Economic Co-­operation and Development (OECD) countries alone (Newland et al., 2008) The importance of circular migration is underscored in the September 2005 communiqué on migration and development of the European Commission The benefits of this labor movement ‘back and forth’ between the home and host countries as well as labor movements involving third countries are listed in the Global Commission on International Migration (GCIM), the International Organization for Migration (IOM), the World Bank, the European Commission and the House of Commons International Development Committee The emerging importance of circular migration comes from its potential benefits to all countries involved in migration – be it the home or sending country, the host or receiving country or a third country, as well as to the migrants themselves Circular migration is not a new phenomenon or a new form of migration that suddenly appeared in the late twentieth century Looking back in time even before current national borders existed, many migration moves were de facto circular Nomads, traders and other seasonal laborers have been pursuing livelihood strategies by consciously taking advantage of opportunities in time and space in order to meet their needs under scarce resources These needs include food security and access to institutions such as welfare, insurance, access to credit markets and risk diversification In developing ­countries, circular migration has also been the road to eliminate poverty In neoclassical economic theory1 in which rational labor migrants (the agents) have complete information, maximize utility under income constraints, make decisions at the margin, are optimizers and pursue their own self-­interest, it is natural that labor migrants move multiple times taking advantage of each situation in each country Migrants optimize their income, savings and investment strategies according to employment options and possibilities in both the home and host country and therefore improve their economic, social, and personal situation in every move and reach equilibrium (Constant and Zimmermann, 2011) With technology allowing relative inexpensive travels and with communications allowing easier information access, the past two decades have witnessed circular migration in a larger scale The creation of de facto migration markets ensures the efficient allocation of the scarce resource labor In addition, circular migrants can be the buffer 55 56   International handbook on the economics of migration during periods of economic volatility and provide employers with a flexible labor force Agreements about free labor mobility among countries can also foster circular migration We find such examples in the European Union of 27 and between Australia and New Zealand.2 Besides the receiving and sending countries, the migrants themselves can profit from repeated or circular moves Just as there are benefits, so there are disadvantages and costs associated with circular migration that we will cite and discuss in a later section In addition to theoretical arguments empirical studies can enlighten our understanding of circular migration and produce sound evidence-­based policy recommendations Unfortunately, empirical evidence about circular migration is scarce and empirical analyses are limited owing to missing or problematic data A thorny issue in the data collection is whether a free democratic country should prohibit its residents from going abroad and/or gather information about their exit and re-­entry.3 Even if this information could be obtainable, reliable data often not exist because there is no matching information in home and source countries A notable exception is the Mexican Migration Project (MMP), a bi-­national study that surveys Mexicans on both sides of the US–Mexican border since 1983 The MMP allows for unique insights into the characteristics and determinants of circular migration For example, these data show that Mexicans are indeed practicing circular migration; they are going to the United States (US) to work for a couple of months, and then return to Mexico to live Other limitations arise because surveys may follow migrants only for a limited period of time and report observations of only parts of migrants’ lives, or they not ask about migration history at all The availability of better datasets, such as the German Socio-Economic Panel (GSOEP) that commenced in 1984 and has rich information on before and after migration experiences, make empirical studies on circular migration possible According to the analysis of Constant and Zimmermann (2011), male and single migrants, and those who have lower levels of education and are closely attached to the labor market, are more likely to be circular migrants in their GSOEP sample The New Immigrant Survey (NIS) is another exceptional migration dataset combining information from US administrative data as well as survey data The NIS contributes to our knowledge about circular migration by containing information about previous migration trips, their frequency and duration; it goes a step further to provide a glimpse into the behavior of migrants by offering information about possible previous repeated ‘illegal’ trips to the US (Jasso et al., 2008) One stunning finding about unintended consequences in policymaking became apparent from research on circular migration Namely, restricting free circular mobility between home and host countries, such as by introducing immigration quotas, increasing border protection or enforcing return migration, has often backfired Flagrant examples are the increasing number of undocumented migrants living in the US and the permanence of the Turkish population in Germany Instead of reducing labor migration, restrictions simply shift migration routes, change the type of migrants entering the country and render sojourner migrants permanent Entry without inspection and overstaying of legal entry visas increase, increasing the number of migrants staying in the host country Lost options for legal re-­entry makes families unify and procreate in the host country The organization of this chapter is as follows: the next section sets the stage for the The economics of circular migration  ­57 economics of circular migration starting with a short presentation of definitions (2.1) and continues with a detailed analysis of the benefits and costs of circular migration (2.2) Section 3.1 presents examples or case studies of the – often negative and counter-­ productive – consequences of restrictive migration policies; it also reviews empirical studies that analyze the characteristics of circular movers worldwide (3.2) The final section concludes and offers policy recommendations as it also lists remaining gaps in the literature of circular migration 2  THE ECONOMICS OF CIRCULAR MIGRATION 2.1  The Circular Migration Nomenclature Following Constant and Zimmermann (2011), we define circular migration as the systematic and regular movement of migrants between their homelands and foreign countries typically seeking work.4 Both skilled and unskilled workers practice circular migration In the literature we find the following synonyms of circular migration: repeat, shuttling, rotating, multiple, cyclical, or circuit migration For migration between the US and Puerto Rico, where there is free border mobility, the terms commuter or revolving-­door migration have been used Circular migration should be differentiated from the one-­ time emigration or out-­migration and the eventual permanent return migration Out-­ migration denotes a single move out of the home country and into the host country with no prospects of return – as was, for instance, the case for many Europeans immigrating to the US between the fifteenth and seventeenth centuries Similarly, return migration describes the only and final move back to the home country after a single migration trip Within the umbrella of circular migration there is seasonal and non-­seasonal migration, mobility of professionals or brain circulation, and transnational entrepreneurs, to name a few The seasonal type of migration is the most popular and probably the most often occurring form of circular migration; it takes place between high-­income and low-­income countries or regions during certain seasons such as during the summer to harvest tomatoes Unregulated systems and spontaneous circulation should also be ­differentiated from regulated systems and managed circulation (Newland et al., 2008) Unregulated systems are established by the migrants themselves Examples of this type are the nomads and traders Regulated systems on the other hand are based on collaborations and diplomatic agreements between states and/or employers Examples of this type are the ‘bracero’ program in the US in the 1940s, 1950s and early 1960s and ‘guestworker’ recruitment in many European countries in the 1950s, 1960s and 1970s.5 The type of circulating migration one observes among the European Union (EU) member states and also among the former Soviet Union countries can be labeled transnational or commuter migrants Zapata-­Barrero et al (2012) use the term ‘circular temporary labor migration’ to focus on temporary workers who travel back and forth between countries over years for a substantial period of time of the year with no permanent nature They discuss the Unió de Pagesos Programe of the largest agricultural business association in Catalonia/Spain with, in particular, Colombia as a role model for deregulated liberalization of mobility It is seen as an innovative way to prevent illegal migration of low-­skilled workers by regulating the flow 58   International handbook on the economics of migration However, the most important restricted circular migration in the world are the less well studied and not yet well understood Chinese internal ‘floating populations’ regulated by the Hukou system (Chan and Zhang, 1999; Wang, 2004) An example for free circular migration is probably the new East European labor mobility generated by EU enlargement (Kahanec and Zimmermann, 2009; Zaiceva and Zimmermann, 2008) Circular migration has also been found to the South and East of the Mediterranean (Fargues, 2008) A specific and distinct feature of circular migration, especially in its unregulated form, is its self-­perpetuating nature (Constant and Zimmermann 2011; Massey and Espinosa, 1997; Massey et al., 2002) Circular or repeat migration spawns migration-­specific capital that causes and encourages the continuity of future migration Like human capital and social capital, it is this migration-­specific capital that ensures the exchange of information about the labor markets in the destination and the source countries More than that, migrant networks ‘grease the wheels’ so migrants have current information about where to go, how to find jobs, what the market wage is, how to maneuver in the system, how to find places to live and shop for groceries, where to send their children to school, and so on These migrant networks are genuine sources of reliable information at any time in either country In the neoclassical microeconomic setting this can be viewed as the ‘migration market’ that reaches equilibrium wages and employment 2.2  Costs and Benefits of Circular Migration 2.2.1  Triple win scenario During the past decade, circular migration became very popular not as a description of reality but as a policy concept, particularly among government institutions Besides the associated benefits for both sending and receiving countries, regulated circular migration programs offer the only legal loophole to labor migration Labor migration is only a small fraction of immigrants to the US, who are mostly tied to families and kin Likewise, in Europe, labor migration to the EU from outside the EU countries is only possible under certain exceptions Circular migration offers a way out of the stringent regulations, pacifies anti-­migrant public opinion and satisfies short-­term excess labor demand With regard to the receiving countries, circular migration fills labor market mismatches and gaps in a highly flexible way, provided that the right to enter and the requirement to leave are linked to the availability of work and the filter for migration is the labor market (Zimmermann, 2009) A historic example of regulated circular migration is the bracero program in the US that started in the 1940s and continued until the 1960s to ease the labor shortages of the Second World War (Massey, 2011) According to the agreements between the US and Mexico, a Mexican laborer was allowed to come and work in the US for a limited period of time, especially in agriculture and railroads In principle, temporary and short-­term agreements like this, aimed to decompress the labor markets and increase competitiveness Another famous example of controlled circular migration is the guestworker scheme that took place in Europe after the Second World War.6 Several European countries, such as France, the Netherlands, Denmark, Germany and Switzerland, delved into migrant recruitment of mostly low-­skilled laborers Agreements with Southern European countries, such as Spain, Portugal, Greece, Italy, the former Yugoslavia and The economics of circular migration  ­59 Turkey as well as Morocco, specified the spell of migration and the sector in the economy Guestworkers were earmarked for low-­level jobs in the industrial and manufacturing sectors Guestworkers contributed to the health of the labor markets and the economy of the host countries; the ‘German Economic Miracle’ (Deutsches Wirtschaftswunder) in the 1960s is one example Lastly, a lesser known circular labor migration scheme exists in oil-­rich Arab countries Massive guestworker migration to these countries occurred in the 1970s with the oil crises and has continued to this day Guestworkers in these countries played a major role in the economic development and structural changes in oil-­rich Arab countries The rational of the guestworker migration is that: (1) employers can satisfy their short-­ term excess demand for low-skilled workers without having to increase wages Hiring low-­wage unskilled workers also guarantees a bottom in the hierarchy of jobs and wages and thus eliminates the need to increase anybody else’s wages This scheme ensures high profits for the employers; (2) employers can recruit young, healthy, and brawny men for these jobs from a known and reliable pool of workers with the blessing of the workers’ home countries; (3) unskilled guestworkers are almost always substitutes of machines and complements of native workers, therefore they create no friction to the skilled native manpower; (4) native workers are free to move up on the professional and socioeconomic ladder and receive extra training; (5) guestworkers contribute to the low-­cost increased production of consumer goods, at relatively low prices; (6) guestworkers as consumers increase demand for goods and can introduce new ethnic goods to the receiving country, enriching the gamut of available goods; (7) guestworkers pay taxes and contribute to the public coffers; (8) employers, under the auspices of their governments can always send these guestworkers back to their home when they not need them, after all, these workers were meant to be guests and not to stay forever, and (9) based on the temporary and circular character of these schemes politicians can pacify xenophobic interest groups and concerned laborers who fear being replaced by guestworkers without the risk of losing their electorate The formation of migration networks and migration-­specific capital helps to establish optimal matches between workers and employers Besides economic benefits, another advantage from migration comes from the influx of mostly young, healthy and ­productive workers who have a ‘rejuvenating effect’ on aging societies From the sending countries’ point of view, circular migration programs are beneficial for the following reasons The sending countries (1) are relieved from any unemployment frictions and labor market imbalances, (2) benefit tremendously from the remittances that their emigrants send back, which can be monetary or in-­kind, (3) gain from the new skills and knowledge that return migrants acquired abroad and bring with them upon return, and (4) through bilateral or multilateral agreements in migration, the sending countries may also strike additional agreements in trade and development and benefit even further from cooperation It should be noted that these countries have usually lower levels of economic development and are characterized by masses of able and available laborers, low wages, scarce capital and high interest rates Remittances are an integral part of migration and are resilient to any economic recession or the business cycle This financial aid helps individual recipients at the micro ­level, as well as governments at the macro level Remittances reduce poverty of families left at home, stimulate markets by increasing the demand for local goods and services (Bird and 60   International handbook on the economics of migration Deshingkar, 2009), reduce child labor,7 and contribute to investments in human capital, development and other productive assets, such as enhanced infrastructure Returnees invest and build up new enterprises in their home communities while maintaining non-­ family networks and highly productive linkages for trade and investment, as is the case for highly skilled returnees Circular migration returnees bring with them knowhow and a different way of doing things that can foster innovative ideas While their visits in each country are for a shorter time, they are offering updated information on the latest happenings in each country Circular migrants are often conduits of technology, fashion and news by their mere presence in the country All these mechanisms are often cited as counter examples of ‘brain drain’ Shrewd migrants delve into circular movements because they can take advantage of better employment and payment opportunities in different countries over time They are able to optimize and re-­optimize their income, savings and assets strategies, and improve their economic, social and personal situation in every period (Constant and Zimmermann, 2011) They minimize their search, relocation and psychic costs while generating a comparable advantage over nonmigrants and one-­time movers as they build up migration-­specific knowledge and local-­specific capital in both countries Migrants are willing to accept low-­paid jobs because they believe that these jobs are only temporary and for the short-­term On the other hand, these jobs pay more than what they would have earned in their home countries and, thus, they can remit more and accumulate more wealth in a short time Here we should note that there are a few programs of short-­term repeated migration that are similar to guestworker programs, but for highly skilled information technology (IT) and research and development (R&D) personnel In the US these are H1-­A and H1-­B visa programs, and specific visiting programs in academia 2.2.2  Shortcomings In reality, the circular migration model has practical disadvantages and caveats First, receiving countries may face grave compliance problems and illegal ‘overstaying’ if migrants not conform to the terms of the circular migration program and not return to their homelands However, even if guestworkers overstay under a legal status, the receiving countries suddenly face increased migration and have to provide for these individuals and their families In fact, the circularity and short-­term migration program prevents the host countries from planning the long-­term settlement of these workers or their families In the event that migrant workers stay permanently in the host country, there are often no provisions for their socioeconomic or educational integration The worst scenario is that migrants stay in the host country – usually without the proper documentation – and they manage to bring their families in for long-­term residence This puts a huge strain on the cities and localities where immigrants settle Primarily, overstayers and their families can cause financial burden on local schools and hospitals Germany is among the prominent examples of what happens when efforts to encourage return migration fail After the recruitment ban in 1973 most of the guestworkers from non-­EU countries did not return to their home countries as intended.8 Instead, they stayed in Germany, became permanent residents, and brought their family members to Germany in the course of family reunification and family building In contrast, guestworkers from EU countries left Germany knowing that they could go back to Germany to work any time they needed to The economics of circular migration  ­61 While in Germany and other continental EU countries illegal or undocumented migrants are not a big problem, they are a problem in the US and have immense ramifications in everyday life and politics Clandestine migration and smuggling are very much affected by government interventions9 and often in a counter-­intuitive way Intensifying efforts to stop illegal entries, increasing raids and stepping-­up deportation policies may make illegal migrants more likely to stay as exploited workers ‘underground’ rather than ‘surfacing’ and facing deportation From the sending countries’ perspective, a mass out-­migration of the working population can create grave labor market shortages In the case of an exodus of highly educated individuals, the sending countries can experience brain drain, at least in the short run until the ‘brains’ return In addition, some authors refer to ‘reverse remittances’, to describe the situation that remittances trigger in the home country Remittances can have a negative impact on labor force participation and productivity of migrants’ relatives and friends who receive them Reverse remittances include services such as ‘raising the migrant’s children and managing housing construction and business, as well as services related to obtaining documents for regularizing a migrant’s stay abroad and for ensuring their social security’ (Mazzucato, 2009) Lastly, repeat migration can be seen as a way of comfortable living and can trigger further out-­migration that may not be ­desirable for the sending or receiving countries As circular migrants often move on their own and leave their children at home with their families,10 reverse remittances can be expected to be higher for them than for migrants who move with the entire family Children living in ‘split families’ where one or both parents migrate regularly may suffer from a lack of parental care which, in turn, may reduce their well-­being and have negative consequences for their economic performance later in life Poor governance of migration institutions may reduce the potential benefits of circular migration, such as financial support, even further Regarding the Chinese migration experience to North America and Australasia, one encounters the ‘astronaut’ syndrome and the ‘parachute kid’ syndrome (Skeldon, 1998) as vivid illustrations of transnationalism The first syndrome conveys the case of the typical male astronaut migrant who returns home for business while leaving his family in the host country In the parachute kid syndrome, the migrant parents return home while leaving their kids with relatives established in the host countries Probably even more severe are the costs to the circular migrants themselves, who can be victims of discrimination and xenophobic attitudes, as well as bearing the risk of exploitation due to lack of employment protection, and lack of integration opportunities There is the possibility that circular migrants may remain trapped in low-­paid menial jobs and need to continue their participation in guestworker programs if they target a certain amount of money and are used to a certain lifestyle; Mexican workers coming in as part of a Canadian guestworker program have displayed this feature (Basok, 2003) Temporary migrants caught in unskilled and precarious jobs may also be exposed to various forms of abuse, discrimination and exploitation which, as Amnesty International (AI) claims, can lead to a ‘modern form of slavery’ (Schöni, 2000) Deshingkar and Start (2003) show that even though circulation is extremely prevalent in India, with millions of (poor) laborers migrating for the best part of the year, Indian policymakers refuse to improve migrants’ situation Accordingly, migrants have no entitlement to livelihood support systems or formal welfare schemes, and they are also refused full payment which 62   International handbook on the economics of migration is reduced even further because contractors make deductions as well.11 It is therefore the case that improved monetary prosperity comes at a high psychological cost born by the migrants The psychological toll of being separated from their households, communities and cultures, can offset some of the material benefits accrued to the sending country (Basok, 2003) Migrants, in general, can jeopardize their health through migration Working hard in precarious jobs and without health insurance can erode the health of workers It is even worse when circular migrants are parents who have left their children behind.12 Their well-­being may decline in the course of their absence from home with serious ramifications for their children’s welfare Circular migrants, whether legal or illegal, tend to avoid hospitals and doctors If they are working legally they may not have health insurance and may not be able to afford to go to the doctor until the very last minute, when they go to the emergency room If they are undocumented, they never go to the doctor because they fear they might be arrested or deported They only seek help when their illnesses become emergencies Even when they seek health care, it is not always easy to be treated because of communication barriers as well as cultural barriers A largely unstudied circular labor migration takes place in Saudi Arabia, which receives many guestworkers from poor non-­oil-­rich countries as well as Thailand or Bangladesh Labor migration to oil-­rich Arab countries boomed in the 1970s with the oil crisis and has been continuing since then Unfortunately, there are practically no data available to study this type of migration Many Southern Arabian economies rely heavily on a ‘sponsor system’ that strips migrants of most of their rights; payment is low, migrants are unable to leave their work place without permission, and they are completely at the mercy of their employer who collects the migrant’s passport upon arrival (Schöni, 2000) Migrants are forbidden to marry natives, to bring over their family or to naturalize Even in the best scenarios, circular migrants risk being seen as transient, as ‘different’ from the locals, and as outsiders who cannot belong to one community This can start a downward spiral of alienation of migrants The pervasiveness of the temporary character of regulated circular migration can harm the migrants in the following way: if migrants are convinced that they are in the host country temporarily they will most likely not devote any efforts to become part of the new country’s life, learn the language, invest in it and/or naturalize Hence it is of utmost importance for programs managing circular migration to achieve the best economically, while guaranteeing humane livelihood conditions for people moving back and forth The following section provides the empirical evidence of circular migration schemes 3  EVIDENCE 3.1  Experiences from Labor Migration Restrictions As discussed in the previous section, circular migration can create a positive stimulation of labor markets in both the sending and the receiving countries, but it can also lead to dependencies, discrimination and alienation Social acceptance by the receiving The economics of circular migration  ­63 country is a serious issue Many societies are worried about ‘foreign infiltration’ owing to increased immigration However, simply restricting circular migration is not a solution and often has quite the opposite effect, as the following examples will illustrate Before the elimination of the bracero migration in 1964, Mexican workers were free to move between Mexico and the US, in three states in particular: California, Illinois, and Texas.13 Mexican laborers commuting for employment purposes to the US on a regular basis used to enjoy free labor mobility until an (initially benevolent) change in legislation ended this privilege This change in legislation resulted in severe restrictions, such as immigration quotas and enhanced border control for laborers coming from Mexico who sought work in the US This led to major changes in the structure and pattern of ­migration, as Massey (2011) discusses comprehensively The change in American legislation and the emphasis on the US–Mexican border resulted in changing the geography of migration, and in more Mexican families settling permanently in the US The militarization of the border induced enormous costs on American society and increased the costs and risks for Mexicans Rendering border crossing so much more difficult, dangerous and expensive has not stopped unauthorized Mexicans from coming to the US, but has made ‘coyotes’ richer, resulted in thousands of deaths and kept Mexicans inside the wall, as back-­and-­forth migration was no longer an option From his 30-­year experience in studying Mexican migrants in the US, Massey (2011) has shown that before any US intervention Mexican migrants would return to Mexico after the harvest in the US Currently, they stay and settle in the US in poverty, which also put a strain on the cities they live in Instead of returning home to ‘live’ and going back to the US to ‘work’, Mexicans reunite with their families in the US where many live for years in the shadows, while they raise American children In the early 1980s the average Mexican worker (usually undocumented) was staying in the US for about three years, but by the late 1990s, the average stay was nine years Responding to raids and militarization of the US–Mexican border, Mexican migrants stopped going home and brought their families to the US, where there has been a ­tremendous growth of the Mexican population, with about 12 million undocumented Since the 1990s, and especially after 9/11, anti-­Hispanic hate crimes have increased ‘The demonization of Mexican immigrants set off a chain reaction that ultimately yielded a massive increase in both border and internal enforcement, which transformed the circularity, demography, and geography of Mexico-­U.S Migration’ (Massey, 2011) While the de jure labor system in the US was based on the annual circulation of legal temporary workers and a small number of legal permanent residents, the recently introduced migration cap changed the system to a de facto system of yearly circulation of undocumented workers and a larger number of permanent residents (Massey, 2011) German history offers two more examples of failed attempts to manage circular migration and to redirect migration flows The first is the guestworker program during Germany’s extraordinarily fast economic growth (Deutsches Wirtschaftswunder) in the late 1950s and in particular in the early 1960s that made the need for imported cheap labor imperative In response, Germany signed bilateral treaties for recruitment of blue-­collar workers in low-­qualified sectors with several Southern European states, such as Spain, Italy, Greece, Turkey and Yugoslavia This was a demand-­driven and project-­tied immigration system The system was successful in recruiting the desperately needed laborers, who worked hard and contributed to Germany’s economic boom 64   International handbook on the economics of migration However, the first oil crisis in 1973 prompted the German government to stop its active recruitment of low-­skilled workers (see Constant et al., 2012; Zimmermann, 1996) Subsequent efforts to encourage and plan to achieve return migration failed, and quite the contrary occurred Germany experienced an increase of its guestworker population owing to family reunification in Germany and high fertility rates of these families in Germany Migration may erode institutional constraints (Zimmermann, 1996) Even though many European countries that once actively recruited labor migrants turned to restrictive policies afterwards, they were unable to find successful policy instruments and incentives that materialize and enforce return migration Migrants from non-­EU member countries, such as Turkey and the former Yugoslavia, who did not have the option to easily return to Germany for work in the event of an upswing of the business cycle, often stayed and became permanent residents Interestingly and paradoxically, labor migrants from EU member states, such as Italy or Spain who could re-­enter and work in Germany at any desired period, returned to their homelands more often (Constant and Zimmermann, 2011).14 The increase of the Turkish population in the 1980s after four rounds of incentives to leave is a testimony to backfired policymaking with ensuing increased social tension between native and foreign population groups The US and German paradigms concur that, as odd as it sounds, free labor mobility decreases migration and constrained or forbidden migration results in increased numbers of permanent migrants Based on negative experiences, after the recruitment ban Germany’s government became very cautious regarding the EU East Enlargements in 2004 and 2007 Like Austria, Germany opted against free labor mobility from the new member states for a seven-­year transition period Fears of mass migration, of ‘welfare tourism’ and of displacement effects in the labor market kept the German doors closed, but not without repercussions.15 As a result, the best qualified workers from the new states went to other EU countries For example, skilled workers from Poland preferred to go to the UK or Ireland – states that did not restrict immigration at all It was the less qualified from the EU-­8 who entered Germany under legal exceptions, counteracting Germany’s protective immigration politics.16 These less-skilled workers compete with the other non-­EU migrants who live in Germany Once more, restrictive immigration policies backfired and led to the opposite of the desired outcome In fact, the closed-­door policy not only failed to attract the much needed high-­skilled workers, but it was also unable to avoid the attraction of low-­skilled immigrants 3.2  Empirical Findings: Who Are the Circular Migrants? Knowing the characteristics of circular migrants is a powerful and successful tool of migration policy Following their self-­interests, sovereign governments can decide which migrants are allowed entry Empirical evidence on circular migrants is scarce since not many countries have records on the entry and exit of their residents It is even rarer to record temporary moves such as those in circular movements Even when reliable data on entry and exit exist, migration studies suffer from selection biases Some knowledge exists regarding internal migration, that is, migration within a country or close region For instance, there are available studies for the US (DaVanzo, 1983), Africa (Beguy The economics of circular migration  ­65 et al., 2012; Bigsten, 199617), India (Deshingkar and Start, 2003) and the Asia-­Pacific region (Hugo, 2008, 2009; Lidgard and Gilson, 2002) Massey and his collaborators have been studying the behavior of Mexican migrants to the US for many decades and have offered tremendous insights in circular or repeat migration The series of trips to the US is determined by the migrant experience, while social networks also play an important role in undertaking an additional trip (Massey, 1987) Conditional on having one trip already, the odds of taking an additional trip to the US for both documented and undocumented migrants increase with experience, occupational achievement and social capital (Massey and Espinosa, 1997) The leitmotif of these studies is that the characteristics for circular migration are different from the characteristics of an initial migration A study on the migration past of new permanent residents in the US showed that 32 percent of them had a previous illegal experience (Jasso et al., 2008) Out of these, 19 percent had entered without inspection, 12 percent had visa overstay experience and 11 percent have had unauthorized jobs The authors find that in the movement from illegal to legal some immigrants experienced substantial upward social mobility There are nascent empirical studies in Asia, where circular migration is of a much higher order A recent study about Thailand and its circular or repeat migrants to six major destination countries (Taiwan, Hong Kong, Israel, Brunei, Singapore and South Korea) concurs that circular migrants are a highly select group (Lee et al., 2011) These circular Thai migrants are more likely to be males, less likely to send remittances and more likely to save – compared with first-­time migrants Circular migration exhibits a strong correlation with age and follows an inverted U-­shape Interestingly, Hugo (2009) finds that circular migrants in Asia remit more money than permanent emigrants For the Pacific region, Gibson and McKenzie (2011) are able to study emigration and return migration among the very highly skilled Based on a unique survey about the best and brightest academic performers from Tonga, Papua New Guinea and New Zealand, the authors find that emigration is most strongly linked with risk aversion, patience and choice of subjects in secondary school Paradoxically, the gain in income from migrating is not a significant determinant The decision to return is strongly associated with family and lifestyle reasons, rather than income opportunities in different countries The authors conclude that income maximization has a very limited role in migration decisions among the very highly skilled For Europe little is known about circular movements per se, except for pioneer work by Constant and Zimmermann (2003a, 2003b, 2011, 2012) on circular immigrants in Germany The existing related literature studies return and out-­migration for Germany (Constant and Massey, 2003), the Netherlands (Bijwaard, 2010), Denmark (Jensen and Pedersen, 2007) and Sweden (Nekby, 2006) All studies confirm the highly selective nature of circular migration From a migrant-­sending country’s perspective, Vadean and Piracha (2010) as well as de Coulon and Piracha (2003) study out-­migration and return patterns of Albanian laborers going to geographically close regions in Greece and Italy Most of these studies actively support the assumption that circular migration is primarily labor migration.18 Taking advantage of particularly rich, representative and longitudinal data from the German Socio-­Economic Panel (GSOEP), Constant and Zimmermann (2011) study circular migration patterns of former guestworkers in Germany They analyze the 66   International handbook on the economics of migration number of exits and the total number of years away from the host country and find notably ­circular movements with about 60 percent of Germany’s immigrant population being circular movers The authors identify some interesting characteristics of circular migrants Migrants from other EU member countries who usually face fewer, if any, institutional hurdles when they want to return to Germany are more frequent circular migrants Those who not own a dwelling in Germany, the younger and the older individuals (excluding the middle-­aged), are significantly more likely to engage in repeat migration and to stay out of Germany for longer Males are more likely to engage in exit and entry than females Migrants with German passports exit Germany more frequently, demonstrating that possession of the German citizenship induces out-­migration Specifically, migrants from Italy, Greece or Spain – that is, migrants coming from one of the three former guestworker countries that are now EU member states – and migrants who have gained German citizenship exit Germany more frequently than migrants coming from Turkey or the former Yugoslavia who face more legal restrictions upon returning to Germany Turks and Yugoslavs are thus less mobile; they exit fewer times and spend substantially fewer years out of Germany The same study finds that, as expected, migrants who are closely attached to the labor market in Germany remain outside for less time and leave (exit) the country less frequently In contrast, when family is left back in the home country this definitely elongates the time circular migrants spend out of Germany Migrants with higher education exit less often and there are no differences with time spent out of Germany Interestingly, vocational training – an important feature of the German educational system – does not play a significant role in the international mobility decision This is in contrast with findings from New Zealand (Lidgard and Gilson, 2002) and the Netherlands (Bijwaard, 2010) In fact, labor migrants from Western countries, who are usually better educated, show higher mobility rates than labor migrants from countries that are more distant both geographically and culturally (Bijwaard, 2010).19 Another study on migrants in Germany, using a dynamic Markov modeling approach, identifies factors generating single moves, circular migration and absorption states from the first 14 individual years of GSOEP (Constant and Zimmermann, 2012) Newcomers are more likely to leave shortly after their arrival and when they have social and familial bonds with the home country Conversely, migrants are less likely to leave Germany when they have a job and speak the German language well After exiting Germany the probability of returning is influenced by remittances and family considerations Circular migration, in particular, is significantly fostered by vocational training acquired in Germany and by older age While men are more likely to return to the home country, gender is not significant in predicting the re-­return to Germany after an exit According to the Constant and Zimmermann (2012) study, the annual probability of immigrants leaving Germany is low, about 10 percent, but once they are in their home country the likelihood of undertaking a repeat move and returning back to Germany is high, about 80 percent of the observed transitions They further report life-­time simulations with the estimated models to study the dynamic process These exercises show that while the probability of returning to the home country remains low as time elapses, the probability of returning back to Germany from the home country is high and approaches 1, the older the immigrants are and the earlier they have migrated to Germany for the The economics of circular migration  ­67 first time This suggests that the remaining repeat migrants are indeed migrant workers, who come to Germany to earn money, but there is no evidence that they finally attempt to return to the home country On the contrary, Germany remains the magnet and will eventually become their real home country These findings are confirmed by various studies for other countries such as Sweden, where repeat migrants are mostly of Nordic origin, male, single, between the ages of 36 and 55, and migrate to Sweden for work-­related reasons (Nekby, 2006) Similarly, Lidgard and Gilson (2002) report that circular migration of New Zealand nationals is most common among single individuals and those aged around 30 For New Zealand nationals, however, non-­economic reasons are as important as economic factors for their return from Australia and the US Other studies show similar migration patterns in Denmark, a country that ­experienced analogous migration to Germany in the past The Danish government recruited many guestworkers in the 1950s and 1960s, mainly from Yugoslavia, Pakistan and Turkey, to satisfy labor market demand (Jensen and Pedersen, 2007) Comparable to ­situations in Germany and the Netherlands, coming from a high-­income country increased the ­likelihood of migrants leaving Denmark Those coming from less developed ­countries, in particular from Pakistan and Turkey, exhibited much lower levels of return This creates a paradox where immigrants who are assimilable or who for all practical ­purposes are integrated in the labor market are more likely to leave the host country, but those who face the most challenges to integration not leave (Jensen and Pedersen, 2007) Still, economic success and attachment to the labor market, as indicated by labor market experience in Denmark, have a positive impact on the propensity to stay Concurring findings exist for Albanian returnees, although Albania is a migrant-­ sending and not a migrant-­receiving country like Germany Vadean and Piracha (2010) find that circular migrants are negatively selected from the pool of returning migrants with regard to education Accordingly, among those who return to Albania, the least educated engage in repeat migration while the better educated re-­enter the Albanian labor market As for other countries, circular migrants are mostly men and, on average, younger than nonmigrants; they are often members of poor and relatively large families, have only primary education, and come from rural and less developed areas close to Albania’s neighboring countries of Greece and Italy However, better educated Albanians are more likely to migrate to other Western European countries, the US and Canada These better skilled individuals often migrate permanently, and not return to Albania Employing MMP data, Reyes (1997) sketches the portrait of return Mexican migrants and comes to similar conclusions; return rates are especially high for males, Mexicans with low education, low-­wage earners and undocumented immigrants Even more importantly, she finds that those who move once are very likely to move again and, thus, engage in repeat migration In summary, studies show that circular migration is mostly seasonal labor migration It is most prevalent among the young, among men and among the single Attachment to the host country, as indicated by owning a dwelling or being married and having family in the host country, reduces the likelihood to circulate In contrast, the acquisition of the host country’s citizenship fosters circularity Hence, the likelihood of a circular 68   International handbook on the economics of migration ­ igration pattern arises with the freedom to leave and with the right to return (Constant m and Zimmermann, 2011, p. 511) 4  RECENT CIRCULAR PROGRAMS AND POLICY ADVICE Many governments and governmental institutions increasingly consider setting up programs that aim to manage circular movement of laborers in order to take advantage of the benefits associated with circular movement The Commission of the European Communities (COM, 2007) for instance, has listed numerous ongoing and planned circular migration and mobility partnerships between the EU and third countries Complementing this list, Newland et al (2008) provide a broad overview of initiatives worldwide differentiated by type of migration: seasonal, non-­seasonal and circular movement of professionals, academics and entrepreneurs Among the successful programs listed by the European Commission are partnerships signed by Spain with Morocco, Colombia and Romania While other countries were touched by mild recessions in the early 2000s, Spain followed a remarkable economic boom Spain needed migrants for many sectors and different periods and delved into circular migration agreements, cautiously designed to have a carrot and a stick The ‘Programme de gestion integral de l’immigration saisonnière’, for example, aims at implementing a system for the management of seasonal migration of Moroccan workers to Spain for strawberry and citrus fruit cultivation Its objectives are, among others, to develop legal immigration for temporary jobs, to prevent illegal practices, and to guarantee the return after the season Preventing irregular migration is also one of the aims of the ‘Temporary and circular labour migration’ (TCLM) project between Colombia and Spain, a program implemented by the International Organization for Migration (IOM) The remarkable success of circular migration programs in Spain20 ended however with the worldwide recession of 2008 The programs will now have to demonstrate that they also operate as expected on the downside of the business cycle There are also programs focusing on South–South migration rather than migration into the European Union One such program, the ‘Management of labour migration as an instrument of development’, is implemented in Africa by the International Labour Organization (ILO) The ILO also promotes the ‘Asian Programme on the Governance of Labour Migration’ that targets countries such as China, Korea and Japan One of its aims is to ‘minimize exploitation and abusive treatment by encouraging active dialog and cooperation among countries in the Asian region’ (COM, 2007, p. 26) Protecting migrant workers’ rights and reducing bureaucratic obstacles to recruitment are also among the aims of the ‘Towards sustainable partnerships for the effective governance of labour migration in the Russian Federation, the Caucasus and Central Asia’ project As the name indicates, this is a cooperation agreement between the governments of the Russian Federation, Armenia, Kazakhstan, Kyrgyzstan and Tajikistan.21 Newland et al (2008) provide a detailed list of: seasonal worker programs between Canada and Mexico, the ‘Seasonal Agricultural Workers Program’ (SAWP) and Canada’s new initiative, the ‘Low-­Skilled Pilot Project’; agreements between Spain and ‘foreign workers who not enjoy free circulation within the EU labor market’ The economics of circular migration  ­69 in addition to Spain’s bilateral agreements with Colombia, Ecuador and Morocco; the Dominican Republic, Poland and Romania; Bulgaria; Mauretania and Senegal; Cape Verde; with Gambia, and Guinea and Mali They describe New Zealand’s new ‘program for seasonal workers from the Pacific islands’, the UK’s ‘Seasonal Agricultural Workers Scheme’ and its ‘Sector Based (point) Scheme’ (SBS), as well as the H-­2B and H1-­B visa program of the US This exemplary documentation demonstrates that there is a huge variety of projects and programs aiming to manage circular labor movements in various countries worldwide The complexity in setting up such programs lies in the different requirements and priorities of sending and receiving countries, as well as in migrants’ needs, anxieties of members of the host society and worries of family members staying back home Designing a system that takes into consideration all these needs and necessities at the same time is a major undertaking Nonetheless it is useful to promote a catalog of measures that should be considered when setting up programs to encourage circularity (Constant and Zimmermann, 2011; Newland et al., 2008; Zimmermann, 2009) The successful policy agenda of circular migration should include, for example, the free access of immigrants to the global labor market However, this policy agenda should connect a migratory move to a job generated from the market system, thus rendering the labor market the filter for migration The right to enter a country and the requirement to leave it should be linked to the availability of work The basic principle of circular migration should be the right or the chance to return back to the host country and should even offer rewards to those return migrants who honored the return migration code International standard settings should include giving minimum work contract standards, providing the means to preserve pension rights, facilitating the free circulation of remittances and enabling the reunion of family members Further, governments (1) need to find instruments that improve the fit between employers and migrants; (2) need to provide channels through which entitlements, such as pensions, that migrants build up during their stay and work abroad can be transferred easily, and (3) need to offer possibilities for upward mobility by providing training for skill upgrading Angenendt (2009) draws four conclusions from the circular migration debate which could also be important elements in a circular migration policy bill First, circular migration programs should be realistic and transparent, meaning that they should contain provisions for every contingency even for illegal or irregular migrants, which are bound to happen Above all, circular policies should secure the human rights of migrants Second, a successful policy bill should have clear and unambiguous goals, and should, third, be tailored to skill levels Lastly, agents involved in this should be aware that the success of these programs requires substantial governance efforts Furthermore, migrants should be able to easily gather information about possibilities and risks of working abroad as well as about safe travel routes and migration channels Conditions that try to ‘enforce’ circularity, for instance by offering only very short-­term contracts and non-­renewable visas or visas that are tied to particular employers without the option to switch to other admission categories, will increase the risk that migrants refuse to conform to the terms of circular migration programs As a consequence, they are more likely to engage in illegal and unauthorized migration Hence, it is most ­important to remove obstacles and to encourage circular movement by simplifying bureaucracy and 70   International handbook on the economics of migration red tape and by making programs more flexible Often programs are too slow to respond in a timely manner to employers’ needs and market conditions (Newland et al., 2008) To promote high-­skilled circular migration – as is often desired by governments – the receiving and sending countries have to work together and for the long term Migration and circulation of people is more complex than international trade for goods Examples of bilateral cooperation are the setting up of service centers that offer housing assistance to returnees, as is the case in China, or the loosening of their foreign-­currency exchange controls, as India does Agreements between Germany and Turkey include educational support to the children of returnees who not speak the Turkish language and need to integrate in Turkey The IOM implemented a ‘return-­of-­talents’ program in Africa, Latin America, Afghanistan and Sudan that offers recruitment, job placement, transport and some employment support Dual citizenship and voting rights are complementary instruments that can bind expatriates to their home country Through these channels, migrants maintain personal and political attachment with both the sending and receiving country and keep in touch with their home communities To intensify attachment with the origin, the UN subsidizes volunteer professionals who return to the home country on a short-­term basis to pass on skills and knowledge that they acquired abroad The most successful programs balance push and pull factors, and gently navigate circular movements 5  CONCLUSIONS The ‘once and forever’ migration moving strategy is passé; circular migration is considerable and highly selective Circular migration is pivotal in assessing the relative success of immigrants in the host country labor market as well as in measuring the economic impact of immigration, with serious fiscal implications Understanding circular migration patterns and gaining knowledge about circular migrants is very important Research on circular/repeat migration can improve the ability to forecast trends in immigration and to design a better migration policy The increasing number of studies dealing with circular migration and burgeoning circular migration agreements among countries are indicative of this Circular migration is widely perceived as making it possible to generate benefits for sending and receiving countries as well as for the migrants themselves Circular labor migration is the best tool for labor economists who contemplate successful labor market and migration policies By its mere definition it is not permanent and, if managed properly, it can accommodate volatile labor markets best Circular migration is the ideal solution for filling labor market shortages over time Even though empirical evidence is rare, existing studies show that circular migration is most common among the young, among men and among those who not show a strong commitment to the host country Contradicting common beliefs and intuition, naturalization and acquisition of the host country’s citizenship does not bind or immobilize immigrants Instead, it provides immigrants with freedom of re-­entry and thus encourages out-­migration In fact, the easier mobility is, the more likely are migrants to engage in and practice repeat and circular migration (Constant and Zimmermann, 2011, p. 513) For governments planning to establish circular migration systems it is important to The economics of circular migration  ­71 find ways to bring out the benefits while mitigating negative side effects such as exploitation and disadvantageous dependences However, intervening in labor mobility, restricting back-­and-­forth movement and coercing immigrants to return to their homelands will not serve the purpose as various historical experiments have demonstrated Migration restrictions often backfired, and tight and restrictive policy measures appeared to be counter-­productive From that experience, circular migration resulting from free labor mobility filtered by the availability of jobs in the host countries could work better Governments can encourage circularity by offering incentives such as possibilities for upward mobility and training as well as portability of benefits across countries Furthermore, repeat migration can be fostered by incentives given in the source countries ‘Diaspora’ policies linking the ethnic networks in receiving countries to the countries of origin is another approach with potential.22 Still, circular migration may not be ‘a silver bullet’ to every problem; for example, skills learned in developed economies may not be transferable to circumstances and realities in the developing world (Skeldon, 2010) It is clear that migration policies cannot be successful when applied unilaterally Receiving and sending countries need to liaise if they want to reach a beneficial goal for all parties Migration in general and circular migration in particular cannot be viewed in isolation but only as part of a broader foreign policy International trade, economic development, environmental changes, technological advances and any kind of abrupt shocks are all interrelated and linked to migration Despite the gain in insight as discussed in this chapter, many questions are still unanswered For instance, the effects of circular migration on economic factors such as output, wages or unemployment rates, or on ‘soft’ factors such as integration, ethnic identification or individual well-­being are still unclear Consequently, more research is needed on this topic in order to completely understand driving factors, benefits and shortcomings of circular movement Only when determinants and consequences are much better understood can regulated systems be established to try to manage circular migration more successfully NOTES   * We thank the anonymous referee for many helpful comments on earlier drafts   In this chapter we not discuss a formal theoretical model of circular migration For relevant theoretical contributions we refer to Dierx (1988), Constant and Zimmermann (2012), Thom (2010) and Chapter by Kennan and Walker on modeling individual migration decisions, in this volume   However, unlike what intuition dictates, circular movements among the members of the EU states in practice are rather limited; it is as if freedom to move immobilizes individuals For an overview of the European determinants of immobility see Bonin et al (2008) The migrant workers from the new member states after EU enlargement are an exception so far, see Zaiceva and Zimmermann (2008) and various chapters in Kahanec and Zimmermann (2009)   We find an exception in the Scandinavian countries where some information about exits is potentially available through the public registers   This could be either frequent moves from the country of origin or a particular host country as the basis Some studies investigate situations where migrants frequently move home from one host country (Germany; Constant and Zimmermann, 2011, 2012) or to one home country (New Zealand; McCann et al., 2010)   More information about the ‘bracero’ program can be found in Massey (2011) The European ‘guestworker’ programs are studied by Castles (2006) and Zimmermann (1996) A comprehensive list of 72   International handbook on the economics of migration         10 11 12 13 14 15 16 17 18 19 20 21 22 regulated programs between the European Union and third countries can be found in COM (2007), for example An exemplary abstract of these projects is also discussed in section See again Castles (2006) and Zimmermann (1996) for further information See Chapter 5, by Edmonds and Shrestha, on independent child labor migrants, in this volume The myth that most of the guestworkers did not go home other than planned is, however, not generally true Most of the guestworkers went home as originally planned However, some groups stayed and brought in their family Zimmermann (1996) has studied the broader picture See Chapter 6, by Friebel and Guriev on human smuggling, in this volume Chapter 16 by Antman in this volume provides an elaborated survey of the new research on the impact of migration on family left behind See also Newland et al (2008) for a detailed description of the Canadian Seasonal Agricultural Workers Program (SWAP), as well as work programs in Spain, New Zealand, Germany, the UK and the USA See Chapter 16 by Antman in this volume for further discussion For further information see Massey (2011), Massey and Espinoza (1997), and Massey et al (2002) Other European countries like France, Denmark and the Netherlands experienced similar changes in their immigrant population whereby those leaving had the option to re-­enter easily See, for example, Jensen and Pedersen (2007) for out-­migration of immigrants from Denmark, and Bijwaard (2010) who studies immigration migration dynamics models for the Netherlands Such fears were unjustified, as research on the effects of enlargement on the EU labor markets (see Kahanec and Zimmermann, 2009) and on the European welfare magnet (Giulietti et al., 2013) has shown See Brenke et al (2009) for a detailed discussion of Germany’s reaction to EU east-­enlargement Bigsten (1996) is interesting since it treats not the individual but the household as the decision unit and investigates circular migration of farm household members in Kenya Except for New Zealand nationals for whom non-­economic reasons play a similarly important role as economic issues (see Lidgard and Gilson, 2002), employment purposes are among the most pressing reasons for migration New Zealand nationals predominantly migrate for educational reasons In contrast, Vadean and Piracha (2010, p 473) show explicitly that employment purpose is the main reason for migration of Albanian migrants Depending on the country of destination, circular migrants seek for seasonal employment in construction, farming and tourism in Greece and temporary employment in manufacturing, construction and services in Italy Nekby (2006) also shows that better education increases the probability to leave Sweden in order to migrate to a third country Using quasi-­exogenous staged Norwegian school reform data, Machin et al (2012) have demonstrated that compulsory education has a causal impact on regional labor mobility They suggest that part of the US–Europe difference in labor mobility, as well as the European North– South difference, is likely to be related to different levels of education across those regions See Zapata-­Barrero et al (2012) for a first analysis of this success For a more detailed description of all those and many other projects funded by the European Commission please consider the corresponding communication (COM, 2007) See Chapter 27, by Plaza on diaspora resources and policies, in this volume REFERENCES Angenendt, S (2009), ‘Labor migration management in times of recession: is circular migration a solution’, Transatlantic Academy Paper Series, The German Marshall Fund of the United States, Washington, DC Basok, T (2003), ‘Mexican seasonal migration to Canada and development: a community-­based comparison’, International Migration, 41 (2), 3–26 Beguy, D., P Bocquier and E.M Zulu (2012), ‘Circular migration patterns and determinants in Nairobi slum settlements’, Demographic Research, 23 (20), 549–86 Bigsten, A (1996), ‘The circular migration of smallholders in Kenya’, Journal of African Economics, (1), 1–20 Bijwaard, G.E (2010), ‘Immigrant migration dynamics model for the Netherlands’, Journal of Population Economics, 23 (4), 1213–47 Bird, K and P Deshingkar (2009), ‘Circular migration in India’, Policy Brief No 4, prepared for the World Bank’s World Development Report 2009 Bonin, H., W Eichhorst, C Florman, M.O Hansen, L Skiöld, J Stuhler, K Tatsiramos, H Thomasen and K.F Zimmermann (2008), ‘Geographic mobility in the European Union: optimising its economic and social The economics of circular migration  ­73 benefits’, joint expertise with NIRAS Consultants and AMS for the European Commission, IZA Report No 19, Institute for the Study of Labor (IZA), Bonn Brenke, K., M Yuksel and K F Zimmermann (2009), ‘EU enlargement under continued mobility restrictions: consequences for the German labor market’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin: Springer-­Verlag, pp. 111–29 Castles, S (2006), ‘Guestworkers in Europe: a resurrection?’, International Migration Review, 40 (4), 741–66 Chan, K.W and L Zhang (1999), ‘The Hukou system and rural–urban migration in China: processes and changes’, The China Quarterly, 160, 818–55 COM (2007), ‘On circular migration and mobility partnership between the European Union and third countries’, communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the regions, Brussels, COM (2007) 248 final Constant, A and D.S Massey (2003), ‘Self-­selection, earnings, and out-­migration: a longitudinal study of immigrants to Germany’, Journal of Population Economics, 16 (4), 631–53 Constant, A.F and K.F Zimmermann (2003a), ‘The dynamics of repeat migration: a Markov chain analysis’, IZA Discussion Paper 885, Institute for the Study of Labor (IZA), Bonn Constant, A.F and K.F Zimmermann (2003b), ‘Circular movements and time away from the host country’, IZA Discussion Paper 960, Institute for the Study of Labor (IZA), Bonn Constant, A.F and K.F Zimmermann (2011), ‘Circular and repeat migration: counts of exits and years away from the host country’, Population Research and Policy Review, 30 (4), 495–515 Constant, A.F and K.F Zimmermann (2012), ‘The dynamics of repeat migration: a Markov chain analysis’, International Migration Review, 46 (2), 361–87 Constant, A.F., O Nottmeyer and K.F Zimmermann (2012), ‘Cultural integration in Germany’, in Y Algan, A Bisin, A Manning and T Verdier (eds), Cultural Integration in Europe, Oxford: Oxford University Press, pp. 69–124 DaVanzo, J (1983), ‘Repeat migration in the United States: who moves back and who moves on?’, The Review of Economics and Statistics, 65, 552–9 De Coulon, A and M Piracha (2003), ‘Self-­selection and the performance of return migrants: the source country perspective’, Journal of Population Economics, 18 (4), 779–807 Deshingkar, P and D Start (2003), ‘Seasonal migration for livelihoods in India: coping, accumulation and exclusion’, ODI Working Paper 220, Overseas Development Institute, London Dierx, A (1988), ‘A life-­cycle model of repeat migration’, Regional Science and Urban Economics, 18, 383–97 Fargues, P (2008), ‘Circular migration: is it relevant for the south and east of the Mediterranean?’, CARIM Analytic and Synthetic Notes 2008/40, Circular Migration Series, European University Institute, Robert Schuman Centre for Advanced Studies Gibson, J and D McKenzie (2011), ‘The microeconomic determinants of emigration and return migration of the best and brightest: evidence from the Pacific’, Journal of Development Economics, 95 (1), 18–29 Giulietti, C., M Guzi, M Kahanec and K.F Zimmermann (2013), ‘Unemployment benefits and immigration: evidence from the EU’, International Journal of Manpower, 34 (1), 24–38 Hugo, G.J (2008), ‘In and out of Australia: rethinking Chinese and Indian skilled migration to Australia’, Asian Population Studies, (3), 267–92 Hugo, G.J (2009), ‘Circular migration and development: an Asia-­Pacific perspective’, in O Hofirek, R.  Klvanova and M Nekorjak (eds), Boundaries in Motion: Rethinking Contemporary Migration Events, Brno: Centre for the Study of Democracy and Culture (CDK), pp. 165–90 Jasso, G., D.S Massey, M.R Rosenzweig and J.P Smith (2008), ‘From illegal to legal: estimating previous illegal experience among new legal immigrants to the United States’, The International Migration Review, 42 (4), 803–43 Jensen, P and P.J Pedersen (2007), ‘To stay or not to stay? Out-­migration of immigrants from Denmark’, International Migration, 45 (5), 87–113 Kahanec, M and K.F Zimmermann (eds) (2009), EU Labor Markets after Post-­Enlargement Migration, Berlin: Springer-­Verlag Lee, S.-­H., N Sukrakarn and J.-­Y Choi (2011), ‘Repeat migration and remittances: evidence from Thai migrant workers’, Journal of Asian Economics, 22 (1), 142–51 Lidgard, J and C Gilson (2002), ‘Return migration of New Zealanders: shuttle and circular migrants’, New Zealand Population Review, 28 (1), 99–128 Machin, S., P Pelkonen and K.G Salvanes (2012), ‘Education and mobility’, Journal of the European Economic Association, 10 (2), 417–50 Massey, D.S (1987), ‘Understanding Mexican migration to the United States’, American Journal of Sociology, 92 (6), 1372–403 Massey, D.S (2011), ‘Chain reaction: the causes and consequences of America’s war on immigrants’, keynote paper presented at the 8th IZA AM2 and 3rd Migration Topic Week, Washington, DC, 13 May 74   International handbook on the economics of migration Massey, D.S and K.E Espinosa (1997), ‘What’s driving Mexico–U.S migration? A theoretical, empirical, and policy analysis’, American Journal of Sociology, 102 (4), 939–99 Massey, D.S., J Durand and N.J Malone (2002), Beyond Smoke and Mirrors: Mexican Immigration in an Era of Economic Integration, New York: Russell Sage Foundation Mazzucato, V (2009), ‘The development potential of circular migration: can circular migration serve the interests of countries of origin and destination?’, in Labor Migration and its Development Potential in the Age of Mobility, Round table theme 2: Circular migration, se2009.eu McCann, P., J Poot and L Sanderson (2010), ‘Migration, relationship capital and international travel: theory and evidence’, Journal of Economic Geography, 10 (3), 361–87 Nekby, L (2006), ‘The emigration of immigrants, return vs onward migration: evidence from Sweden’, Journal of Population Economics, 19 (2), 197–226 Newland, K.K., D.R Agunias and A Terrazas (2008), ‘Learning by doing: experiences of circular migration’, mpi-­Insight, Washington, DC: Migration Policy Institute (MPI) Reyes, B (1997), ‘Dynamics of immigration: return migration to western Mexico’, Public Policy Institute of California, San Francisco, CA Schöni, D (2000), ‘Moderne Sklaven’, AI Journal, available at: http://www.amnesty.de/umleitung/2000/ deu05/312?lang5de&mimetype5text/html (accessed 12 January 2012) Skeldon, R (1998), ‘Migration from China’, Journal of International Affairs, 46 (2), 434–55 Skeldon, R (2010), ‘Managing migration for development: is circular migration the answer?’, The Whitehead Journal of Diplomacy and International Relations, 11 (1), 21–33 Thom, K (2010), ‘Repeated circular migration: theory and evidence from undocumented migrants’, mimeo, New York University Vadean, F and M Piracha (2010), ‘Circular migration or permanent return: what determines different forms of migration?’, in G.S Epstein and I.N Gang (eds), Migration and Culture, Frontiers of Economics and Globalization, 8, Bingley: Emerald Group pp. 467–95 Wang, F.L (2004), ‘Reformed migration control and new targeted people: China’s Hukou system in the 2000s’, The China Quarterly, 177, 115–32 Zaiceva, A and K.F Zimmermann (2008), ‘Scale, diversity, and determinants of labour migration in Europe’, Oxford Review of Economic Policy, 24 (3), 428–52 Zapata-­Barrero, R., R.F Garcia and E Sánchez-­Montijano (2012), ‘Circular temporary labour migration: reassessing established public policies’, International Journal of Population Research, available at doi: 10.1155/2012/498158 Zimmermann, K.F (1996), ‘European migration: push and pull’, Supplement to The World Bank Economic Review and The World Bank Research Observer, 10 (1995), 313–42 Reprinted in: International Regional Science Review, 19 (1996), 95–128; K.F Zimmermann and T Bauer (eds) (2002), The Economics of Migration, vol I, pt I, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 70–99 Zimmermann, K.F (2009), ‘Towards a circular migration regime’, in Labor Migration and its Development Potential in the Age of Mobility, Round table theme 2: Circular migration, se2009.eu 4  The international migration of health professionals* Michel Grignon, Yaw Owusu and Arthur Sweetman 1  INTRODUCTION International migration by health professionals is an area of increasing policy interest Shortages of medical personnel in several developed countries are perceived to be central drivers of this phenomenon, and there are critical ramifications for developing countries (for example, the World Health Organization – WHO, 2006) After a period of perceived excess supply in many developed countries during the 1990s, more recent years have seen an increased demand for health professionals, a growing concern about the need to provide healthcare services to aging populations, and an increasing focus on health human resources more generally The International Migration Outlook (Organisation for Economic Co-­operation and Development – OECD, 2007a) identifies ‘several international initiatives  .  ­formulating policy recommendations to overcome the global health workforce crisis’ (p. 162).1 In response to these flows, in 2010 the WHO adopted a global code of practice on the international recruitment of health personnel with a focus on ethics and protecting less-­ developed sending countries (WHO, 2010) Aligned with this initiative, several developed countries have devised their own protocols regarding the ethics of international health professional migration (for example, Canadian Federal/Provincial/Territorial ACHDHR, 2009; Norwegian Directorate of Health and Social Affairs, 2007; UK Department of Health, 2011) Moreover, migration among more developed countries is also an issue In particular, the enlargement of the European Union (EU) has heightened concerns regarding an exodus of health workers from accession countries towards Western Europe (Wismar et al., 2011) A substantial amount of academic research by health clinicians, health services researchers, migration scholars and those from other backgrounds, together with significant contributions from international organizations, advocacy groups, and others has explored the broad issue of health workforce migration Economists are adding a particular perspective to this interdisciplinary discussion, but much work remains to be done This chapter surveys the topic from an economic perspective while also being situated within the broader interdisciplinary literature Moreover, this survey employs selected well-­known theoretical tools from economics in an effort to conceptualize the substantive phenomenon and to point to areas for future research After this introduction, the second section provides an empirical overview to elucidate the magnitudes of international flows of health professionals and set the context for the analysis that follows In the third section the focus is on developed countries, which implies simultaneously addressing issues regarding both receiving and sending health workforce migrants Less-­developed, primarily sending, countries are addressed in the fourth section The final section concludes 75 76   International handbook on the economics of migration 2 DESCRIPTIVE STATISTICS OF THE INTERNATIONAL CONTEXT International comparisons and analyses of health human resources have become much easier in recent years because of the work of the WHO and the OECD Building on this, Figure 4.1 depicts the percentage of physicians that are foreign born in a wide variety of (mostly) OECD countries circa 2000 and compares it to the percentage of the entire population that is foreign born It is clear that in almost all countries immigrants are much more likely to be physicians than are native-­born persons, Germany and Greece being notable exceptions Most OECD countries import a very substantial share of their physicians In contrast, in Figure 4.2 nurses are seen to be, on average, roughly proportionately drawn from each country’s immigrant and native-born populations, although there is also greater variation across nations for nurses, with some nations drawing a disproportionate share of their nursing workforce from the immigrant population and others doing the reverse Clearly, there are important differences in the migration of physicians and nurses Figure 4.3 displays a similar plot for the aggregation of a variety of other health professions On the whole, these other professions look more like physicians than nurses, with a disproportionate share of these health workers being drawn from the migrant population Of course, being foreign born does not imply being foreign educated since immigrants may arrive in their host country prior to completing their education Also, the aptly named 1.5 generation – those who immigrate as children – may have a different probability of obtaining health professional certification than does the average native-born person This difference is addressed in Table 4.1 The traditional immigrant receiving countries of Australia, Canada and New Zealand clearly have a substantial number of immigrants educated in the domestic school system; surprisingly, the US does not Also, in France and Portugal a substantial proportion of the foreign born are educated domestically, although, as noted in OECD (2007a), this may reflect the repatriate community Conversely, being native born does not imply receiving a domestic health professional education An increasing phenomenon is the emergence of medical and other health professional schools offering education for export Such institutions provide medical and other health professional education to an international clientele of foreign students as an export industry The best known of these medical schools are probably in the Caribbean, Ireland and, more recently, Australia This implies that there are increasingly two major types of international medical and other health professional graduates: residents/citizens of the country providing their education who emigrate post-­graduation, and international students who obtain their medical/health professional education at an international ‘destination’ school with the primary intention of departing that country to return home or to practice in some third nation This is particularly the case for the US where, for physicians, the American Medical Association (AMA, 2010) reports that in 2008 about 15 percent of international medical graduates (IMGs) among those in Accreditation Council for Graduate Medical Education programs are native US citizens.2 Of course, as discussed by Hawthorne and Hamilton (2010), countries housing such medical schools may also come to view these foreign students as a source of potential domestic supply In fact, it could be argued that from some perspectives this is an 77 ic o e rk Tu y nd H y ar g un 3.9 2.6 la n Fi n K U 33.7 nd la e Ir 7.9 N ce an Fr e ec um gi l Be 8.6 10.3 re G ia tr us A 11.8 10.3 % population foreign born s nd la er h et D EC O % MDs foreign born l y k ga ar rwa tu m r o n e N Po D 7.5 4.9 Sp 14.6 10.4 U S er itz Sw en ed Sw G nd la N ew y an nd la a Ze 10.1 12.5 m er 46.8 g a ia ur al ad bo an ustr C m A xe Lu 17.2 Figure 4.1 Share of physicians and total population that is foreign born, OECD countries circa 2000 (percentage) Sources:  OECD (2007a, 2007b, 2011) Notes:  Poland and Greece’s foreign-born population is from OECD (2007b); all others are from OECD (2011) d 3.2 1.6 an l Po 1.5 0.5 ex M 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.3 8.7 35.0 24.4 11.0 40.0 19.7 5.1 22.8 11.3 35.1 17.4 45.0 6.2 1.9 11.0 2.9 10.9 5.8 16.6 6.8 18.0 9.9 16.6 10.1 16.9 10.1 27.9 11.3 42.4 23.0 50.0 30.2 33.2 78 M o d 6.1 6.8 3.4 4.9 3.1 2.9 0.8 2.6 K U nd la e Ir O D EC B m iu g el G e ec re tr us A 9.7 10.3 ia % Population foreign born h s nd la er S U Sw 11.9 11.0 r ze 24.5 23.0 17.2 17.4 rg lia da nd any and la l ou na t a m b a s r u e C Ze it em A G ux ew Sw L N en ed Figure 4.2 Share of nurses and total population that is foreign born, OECD countries circa 2000 (percentage) N et e nc a Fr 10.6 9.9 % Nurses foreign born y k al in ry nd ar rwa la pa rtug ga m S n o u en N Po H D n Fi 1.9 y ke r Tu 0.4 1.6 an l Po 0.2 0.5 ic ex 4.1 5.8 Sources:  OECD (2007a, 2007b) and OECD (2011) 0.0 5.0 10.0 15.0 20.0 13.9 5.1 15.2 7.9 14.3 8.7 25.0 10.1 5.5 30.0 10.1 6.8 14.5 10.4 28.4 11.3 35.0 10.3 6.6 11.3 8.8 12.5 9.6 23.1 17.2 33.2 25.8 79 M o ic d y ke r Tu 2.7 1.6 an l Po 1.2 0.5 ex F H d an g un 3.4 2.6 l in y ar 29.2 d an el Ir 7.9 K U 28.1 N s nd la r he et D EC O 8.7 ce an Fr % Other health professionals foreign born y k al ar rwa ug t m r o n o e N P D n 6.4 4.9 Sp um ia tr us A 10.2 10.3 gi l Be 14.4 10.4 U S en e itz Sw ed Sw % Population foreign born e ec re G 38.5 Figure 4.3  Share of other health professionals and total population that is foreign born, OECD countries circa 2000 (percentage) Sources:  OECD (2007a, 2007b) and OECD (2011) 17.2 ia rg da nd any and al ou na rla rm tr al b a s e u e C m Z A G xe ew Lu N Note:  Other professions defined as ISCO 222: dentists, pharmacists, veterinarians and other health professionals not elsewhere classified 0.0 5.0 10.0 15.0 20.0 25.0 17.2 5.1 20.9 11.0 30.0 15.7 9.9 26.7 11.3 35.0 14.5 6.8 14.5 10.1 14.7 10.1 20.0 11.3 32.0 17.4 40.0 5.4 1.9 9.3 2.9 9.3 5.8 10.3 7.4 12.5 8.8 33.6 23 45.0 30.5 33.2 80   International handbook on the economics of migration Table 4.1 Foreign born and foreign trained in selected OECD countries circa 2000 (percentage) Country Australia Austria Canada Switzerland Denmark Finland France UK Ireland Netherlands New Zealand Sweden US Physicians Nurses Foreign trained Foreign born Foreign trained Foreign born 25.0 1.8 23.1 11.8 7.7 3.6 3.9 33.1 10.3 6.2 34.5 4.3 25.5 42.4 14.6 35.1 27.9 10.9 3.9 16.9 33.7 35.3 16.6 46.8 22.8 24.4 12.1 – 6.4 – 6.0 0.2 – 8.0 14.0 1.4 19.3 2.5 3.5 24.5 – 17.2 – 4.1 0.8 – 15.2 14.3 6.8 23.1 8.8 11.9 Notes:  Foreign-­trained MDs for Australia, the UK and the Netherlands, and foreign-­trained nurses for Australia, Ireland, the Netherlands and US, are 2005 data Sources:  OECD (2007a, 2007b) and OECD (2011) ideal supply since it is inexpensive, with the students paying international student fees, and the graduates are already fully acclimatized to local medical practice This (increasingly common) revenue-­generating education for non-­residents is distinct from the well-­known practice of some countries training their citizens with the explicit intent of serving international markets with migrants – such as some Philippine and Indian medical and nursing schools (Masselink and Lee, 2010; Nullis-­Kapp, 2005) Cycling and permanent migration among developed countries is also quite common, as discussed more broadly in Chapter in this volume For example, Mullan (2005) provides a useful description of the flows of domestic physicians among Australia, Canada, the UK and the US, summary statistics from which are presented in Table 4.2 At the top is the UK and Canada, both of which are net exporters in this quadruple, and at the bottom are Australia and the US, which are net recipients.3 Turning to nurses, the OECD (2007a) takes a somewhat broader view looking at all intra-­OECD migration and characterizing it as having a cascading pattern Australia and Canada, and to a lesser extent the UK, New Zealand and Switzerland, are positioned near the bottom of the chain and receive nurses from other OECD nations However, similar to the situation for physicians, the US is at the very bottom of this migratory flow and is by a substantial amount the largest net recipient of immigrant nurses Given the enlargement of the EU  – and associated treaties and regulations regarding the mutual recognition of ­credentials – ­substantial geographic mobility within that area is observed and more is expected, although to date there has perhaps been less mobility than some feared (Wismar et al., 2011) The international migration of health professionals  ­81 Table 4.2 Emigration and immigration of physicians among Australia, Canada, the UK and the US Immigration Emigration Australia Canada UK US Total in % Australia Canada 54 212 65 4664 73 4802 8.9 247 68 096 2735 519 3501 5.1 UK US 872 1144 50 8990 3439 138 667 79 836 036 1001 13 573 0.7 1.6 Total out % 2263 9105 10 838 671 – – 4.2 13.4 7.8 0.1 – – Source:  Mullan (2005) 3  A DEVELOPED COUNTRY PERSPECTIVE This section addresses three questions Why rich countries import health professionals? How these immigrant health professionals fare in the healthcare market of recipient countries? What is the impact of licensure and other specific institutions of the healthcare market in recipient countries on the integration of foreign-­trained professionals? Health Labor and Product Markets: A Brief Overview Before focusing on the international migration of the health workforce it is worth briefly considering relevant economic aspects of health product and factor markets since these are crucial for interpretation As is well-­known, the product market for health services is nonstandard, and asymmetric information problems are extremely serious Most health service provision is ‘custom work’ employing specialized knowledge that makes it very costly to judge provider effort and the quality of the work performed Principal–agent models are often employed to characterize the context, as are models of monopolistic competition (see, for example, McGuire, 2000) It is not even clear that the patients/­customers have sufficient information to know the optimal quantity of services to purchase leading to (somewhat controversial) models of supplier induced demand whereby if providers (usually physicians) perceive that they not have sufficient business/income they can influence the demand for additional services (McGuire, 2011) One ramification of professional knowledge and other sources of asymmetric information is licensing and/or regulation by government, which is frequently delegated to a professional college or similar organization of experts Licensing bodies become important hurdles for the international migration of health professionals, and it is often argued that in addition to simply verifying ability/credentials these organizations serve to create provider ­monopoly power, as is discussed below Another important determinant of the structure of healthcare product markets is the potential occurrence of infrequent, unanticipated, large-­scale, expensive negative health 82   International handbook on the economics of migration shocks in contexts where borrowing is not normally feasible One response to this is private health insurance, which is ubiquitous in developed countries Health insurance can be viewed as a pay-­in-­advance approach to funding healthcare service purchases that spreads risk across the population However, ‘pure’ insurance is not sufficient in healthcare markets since there are individuals whose health status (sometimes from birth) is uninsurable in an actuarially fair system This would lead to a profound – and morally/ ethically unacceptable – market failure However, society in the form of government intervenes to provide (or at least pay for and/or subsidize) healthcare services In practice, governments for various reasons normally provide far more healthcare services than those medically necessary, because of this example of a market failure Although the structure of healthcare systems and the nature and extent of governments’ roles vary dramatically across nations, in most developed countries governments play a very significant role in the operation of the healthcare system In terms of international migration, this institutional structure frequently serves to separate healthcare recipients (that is, patients) from those who in the first instance pay for those services (that is, governments, insurance companies, patients and others) This separation of the ‘payer’ from the recipient adds an important characteristic to the marketplace The aggregate-­level payers are concerned about costs, but akin to patients, they face asymmetric information challenges regarding courses of treatment for individuals In an effort to control costs governments in particular sometimes seek to control the number of providers, which has a variety of quite complex ramifications for the international migration of health professionals On the factor market side it is important to recognize the enormous time lags in training health professionals, especially physicians, which can lead to a type of ‘overshooting’ whereby a perception of a current surplus generates a cap on or a reduction in enrollments, which seems frequently to lead to a shortage in the medium to long term (see, for example, Bärnighausen and Bloom, 2011, for a survey of the literature) In large part these caps are motivated by concerns about cost containment For example, there are caps on both undergraduate (MD) and graduate (specialty/residency) positions in the US (AMA, 2010),4 and the central governments of Australia, France, Sweden and the UK control medical school enrollment through university funding Some countries, such as Canada, have a more decentralized though still government mandated approach, with each province setting local medical enrollment.5 Mis-­targeting can lead to oscillating periods of (perceived) surplus and shortage, as have been observed in recent decades (Bloor and Maynard, 2003) Given the extended training periods of physicians and other health professionals, almost by default international migration is used as a short-­term quantity adjustment mechanism From a different perspective, employers/payers in healthcare are frequently viewed as monopsonistic, especially for nurses (for example, Hirsch and Schumacher, 2005; Machin and Manning, 2004; Staiger et al., 2010), although there is some controversy Monopsony explains the ongoing perception of shortages concurrent with unemployment or underemployment in the sector and/or a queue of potential immigrant workers willing to work at the going wage who are prevented from practicing in the relevant profession Simultaneously, public-­sector unionization also plays an extremely important role, with some professional associations/unions in healthcare having enormous influence (see, for example, Drexler, 2008) In fact, in some jurisdictions health human The international migration of health professionals  ­83 resource decisions might best be viewed as the outcomes of a bilateral monopoly with governments on one side and professional unions/associations on the other Of course, there are usually also many influential economic actors, such as associations of hospital administrators, medical schools and accreditation bodies as well as patient advocacy groups and policy think tanks, among others All of these combine to determine the ‘pull’ factors that attract health workers to developed countries or act to limit their entry into national labor markets Why Do Developed Countries Import Health Professionals? It is not immediately obvious why developed countries would want to import health professionals en masse Although a small number may be involved in entrepreneurial health research or delivery with substantial spillovers for the domestic economy and that provides an economic justification for the receiving country, this is unlikely to represent the majority of migrants In many countries these are ‘good’ jobs that are rationed among a domestic population eager to become health professionals, especially physicians, but prevented from doing so via quotas on, in particular, medical school entry.6 We, therefore, next discuss a few potential explanations for the acceptance of health professional migration by many developed countries – though certainly not all as evidenced by Germany A key rationale for the international migration of health professionals is a response to perceptions of short-­term shortages – although these can alternate with perceptions of surpluses that can induce barriers to entry.7 As mentioned, governments have multiple incentives for managing the size of the workforce for physicians and other health providers, especially cost containment Other payers as well as provider associations have similar incentives, though different objective functions, but ultimately governments have the authority to act whereas the others only have influence Further, governments can and overshoot – more frequently toward shortages than surpluses Given the extended education/training durations (planning horizons) involved, international migration is frequently used by wealthy countries as a tool to manage short-­term shortages An interesting variation on this theme is put forward by Rutten (2009a) who, using a computable general equilibrium model for the UK, and very specific assumptions, argues that for some situations importing immigrant health providers has superior welfare implications to increasing the health service’s budget since the former avoids pay increases for inframarginal (existing) workers A related but distinct rationale involves local or specialty-­specific shortages (Zurn et al., 2004) Rural, remote and underprivileged areas are underserviced by domestic health professionals in many developing (for example, Kanchanachitra et al., 2011) and developed (for example, Rabinowitz et al., 2008) countries Surprisingly, for an important issue that has been a focus of discussion for decades, there are essentially no studies that credibly estimate medium- and long-­term causal impacts of programs seeking to address health services provision in underserviced areas; systematic reviews that have found no such studies are by Grobler et al (2009) and Wilson et al (2009) Nevertheless, there is a large research literature looking at non-­causal relationships, and some jurisdictions, particularly Australia, Canada, New Zealand and the US, clearly and explicitly employ immigration as a tool to service rural and remote areas (for a survey of one aspect of this 84   International handbook on the economics of migration literature see Bärnighausen and Bloom, 2009) To tie them to appropriate geographical areas, at least in the short run, such workers are frequently admitted under limited visas, for example J-­1 visas in the US (for example, Hagopian et al., 2003; US GAO, 2006) or given limited/provisional licenses to practice in Australia and Canada (for example, Auda et al., 2005) Secondary migration by international migrants, and internal migration more generally, is a key mechanism explaining why underserviced areas have been the focus of an ongoing policy issue for decades as described by McDonald and Worswick (2010) for Canada They focus on out-­migration from underserviced rural areas that actively recruit IMGs and, in the Canadian case, physicians practicing in these regions also typically receive additional financial and non-­financial benefits Their research observes low retention rates, which are not entirely associated with pecuniary factors but have substantial relationships with marital status and spouse’s level of education Further, the migrants tend to move not to other underserviced areas, nor to regions with intermediate levels of service, but to the very largest cities with the highest physician densities Overall, regions with the greatest need and lowest entry hurdles are effectively stepping stones to other locations within the country for many IMGs Although the international migration of health professionals clearly provides practitioners to underserviced areas, it appears that a steady flow of such practitioners is required Closely related to finding workers for geographic locations that are (by some) deemed less desirable, is the need to find health workers for less desirable work shifts (times of the day/week) and other practice characteristics (for example, Denour and Junker, 1995; Drexler, 2008) Again, immigrants tend to be found working in these situations – at least initially upon arrival in a new country Lastly, it has been suggested that importing countries should try to match their increasingly diverse population through importing physicians of the same ethnic backgrounds who would be more culturally sensitive than native-­born and locally educated ones There is not, however, clear evidence that sharing the same ethnic background makes healthcare professionals more effective A fuller exposition of migration, ethnicity and economic integration can be found in Chapter in this volume Since the education and training of health professionals is both subsidized and costly to governments in many countries (for example, McGuire, 2000), savings on these fronts is sometimes suggested as motivating immigration While this may indeed prompt some short-­term decisions, it is hard to believe that governments would be sufficiently myopic (and insensitive to the demands of prospective domestic health/medical students) as to drive policy based on these upfront costs that are relatively modest as a percentage of total healthcare expenditures Population aging is an often mentioned rationale for health worker migration However, the link to the demand for immigrant health professionals is more tenuous than it might at first appear It is not a short-­term problem but one of the few public policy issues that can be forecast decades in advance Also, while the associated increased demand for health professionals cannot be ignored, it is more modest than popular perceptions allow, although there is a need to alter the composition of the health professional workforce since demand will shift from, for example, pediatrics, obstetrics and gynecology to ophthalmology and chronic diseases (for example, Denton et al., 2009) On the other hand the story is quite different for low-­skilled healthcare providers since, The international migration of health professionals  ­85 as Haberkern et al (2012) point out, the need for eldercare will likely increase substantially in rich countries Because home and long-­term care are labor intensive, with limited possibilities for automation, lower skilled healthcare workers from developing or less rich countries will probably supply it in wealthier nations Canada’s Live-­in Caregiver Program is a formalized immigration stream that is increasingly used for elder care, whereas it was formerly primarily used for childcare The Labor Market Integration of Health Providers Unlike much of the immigration literature, there has been relatively little analysis of economic integration that is specific to health professionals However, recent work using US data by Schumacher (2011) has begun to address this issue for nurses as a precursor to looking at the impact of immigration on domestic workers in that occupation, and it is worth reviewing this article at length Using the US National Council Licensure Examination where candidates are required to pass qualifying exam(s) that are set by a third party rather than an educational institution where the conflict of interest between pedagogical delivery and assessment is obvious, Schumacher initially observes that first time internationally educated test-­takers have pass rates roughly 20 percent below those of US-­educated test-­takers This influences the interpretation Schumacher employs two datasets in his analysis: the US Current Population Survey (CPS), and the National Survey of Registered Nurses (NSRN) Only country of birth is identified in the CPS, whereas the NSRN records country of education The CPS descriptive statistics show that while the foreign born have hourly earnings slightly below that of the native born, the average wage of those with foreign education is higher However, controlling for observed characteristics, the foreign born have earnings about to percent less than the US born, although there is no difference for those from Canada, and the gap is less than percent (and not statistically significant) for those from the Philippines – the second largest and largest source countries respectively Also, the years since migration profile suggests that full earnings equality occurs within about six years Data from the NSRN show effectively no gap in average wages between the domestic and foreign educated, and the years since migration profile indicates even swifter integration One aspect of the specifications employed is that they may ‘over-­control’ for background variables For example, some models control for union membership and others for hospital employment While these variables are undoubtedly important, it is entirely possible that they are determined simultaneously with wages and may mask relevant wage effects If immigrant nurses are, for example, less likely to obtain ‘good’ union jobs or are differentially likely to be employed in hospitals because of their immigrant status, then controlling for these factors may answer an economic/policy question different than the one being posed Future work could include specifications with and without these potentially endogenous regressors to see how the coefficient on the variable of interest changes Similarly, a specification excluding those who immigrated younger than age 25 could differentiate between individuals most likely to have been educated in the US and those educated elsewhere As seen in Schaafsma and Sweetman (2001) for example, there is a noticeable difference in labor market outcomes between those who immigrate as children and are educated domestically and those who immigrate later in life Nevertheless, 86   International handbook on the economics of migration overall the evidence indicates that practicing immigrant nurses have hourly earnings that are extremely similar to those of the domestic born Of course, these results are for practicing nurses and say nothing about outcomes for those who immigrated intending to practice, but who are not doing so Schumacher next attempts to estimate the impact of immigrant nurses on the wages of domestic nurses, which is very similar in intent to work by Kaestner and Kaushal (2012) who use the NSRN However, the two take distinct methodological approaches Schumacher employs a ‘factor proportions’ methodology that exploits (potential) changes in wage gaps between nurses and other occupations not affected by nurse migration, while Kaestner and Kaushal pursue an instrumental variables strategy Both draw similar conclusions These results are interesting in their own right, but it is difficult to extrapolate to other occupations since nursing is primarily female, highly unionized, regulated and potentially monopsonistic (this last is very controversial but, minimally, perennial labor shortages are reported) For both studies, a credible identification strategy is crucial since nurses are not randomly assigned employment but end up in particular locations because, for example, opportunities are greatest, social networks exist and/ or information is available While estimating causal impacts is extremely difficult and it is not clear that either of these strategies identifies the desired parameter perfectly, they are useful exercises Comparisons across these identification strategies, and across specifications for each, are informative.8 Altogether, the regression results suggest either no effect, or an extremely small negative effect, on domestic nurses’ wages as a result of immigration Huang (2011) pursues a similar strategy to the first portion of Schumacher’s paper using the NSRN data She finds those nurses licensed to practice are able to transfer their foreign human capital with relative ease and, indeed, obtain a premium in the labor market The premium is driven almost entirely by international nursing graduates from English-­speaking countries working in hospitals Even within a narrowly defined occupation, she finds substantial earnings heterogeneity across source country and work environment Kalist et al (2010) undertake research similar to the second part of Schumacher’s paper, but using an instrumental variables approach, and find largely consistent results An issue raised but not analyzed by Schumacher is host country language ability Although this is much discussed in the literature and is clearly extremely important in labor markets, relatively little is known about the effect of language facility on earnings and employment outcomes of international immigrants in health professions Unlike the general labor market, these regulated professions frequently require language proficiency to be tested prior to licensure A small piece of evidence is provided by Sharieff and Zakus (2006) who interview 21 non-­randomly selected IMGs in Canada Although one of the study’s selection criteria was self-­assessed English language fluency, some had appreciable difficulty with the Test of English as a Foreign Language (TOEFL) and the Test of Spoken English (TSE) Although all passed the TOEFL, it took some four attempts to so and of the 15 who took the TSE after three attempts only 83 percent had passed While not directly relevant to health professionals, a similar problem is noted by Ferrer et al (2006) among Canadian immigrants in general There appears sometimes to be a gap between literacy/language skills as measured by these tests and individuals’ self-­perception This is an important area for future research.9 The international migration of health professionals  ­87 Turning to physicians, recent papers focus on the impact of being granted a license to practice on labor market outcomes Kugler and Sauer (2005) take advantage of a discontinuity in the relicensing requirements for physicians leaving the former Soviet Union for Israel Crucially, the Israeli administration assigned immigrant physicians with 20 or more years of experience to an observation track, whereas those with fewer years had to undertake a relicensing exam The observation track was both immediate and associated with a high probability of licensure, whereas the examination track delayed entry to practice and had a lower probability of licensure While ordinary least squares (OLS) results show a very substantial, in the area of 100 percent, return to licensure on monthly earnings, the instrumental variable estimates around the discontinuity imply that the causal impact is actually in the range of 180 to 340 percent Clearly, the forgone earnings of an immigrant physician not, or being delayed in, receiving a license to practice are enormous Given that the instrumental variable results are larger, the authors impute the presence of rents accruing to practitioners and negative selection into licensing This implies that higher-­quality physicians – those with valuable outside options – are unwilling to bear the cost of acquiring a license It is not clear how generalizable these results are since the migration of Jews from the former Soviet Union to Israel was part of a large and unusual event, but it does point to a methodology that could easily be used in other countries to evaluate the value of licensure to immigrant practitioners It also highlights the importance of both licensing criteria and the cost of the licensing process for health professionals, which in most countries is complicated, time-­consuming and expensive This is stressed by Sharieff and Zakus (2006) Lesky (2011) outlines the administrative process for physicians entering Canada and the US, and Forcier et al (2004) is an overview for OECD countries In a set of papers, McDonald et al (2011a, 2011b) compare immigrants with medical degrees in Canada and the US Interestingly, although some observers in both countries argue that there is a shortage of physicians, both countries control the number of IMGs who can enter each year and facilitate immigration for physicians willing to work in underserviced areas One interpretation for this follows from the distinction discussed above between the perspective of patients and payers where cost control is important Of course, quality control is also important and the two issues frequently intermingle, as is discussed in a subsequent section Both papers exploit differences in the immigration processes in Canada and the US: Canada has a points system and the US uses employer or state nomination (through the H1-­B and J-­1 visas) Points systems may be designed in a variety of manners, and may admit new immigrants with medical degrees who have a low probability of being licensed to practice, whereas employer nomination improves the probability of professional employment.10 Of course, immigrant physicians may enter both countries through other administrative channels, such as spousal or refugee streams Nevertheless, this is an important specific case of the general issue regarding the importance of the details of the administrative implementation of points systems There may be a gap between the allocation of educational points for immigration and the value of those same credentials for occupational licensing – and Kugler and Sauer’s results suggest that the value of that gap is enormous for applicants if not necessarily for the receiving society Comparing an earlier Canadian regime, when it more closely resembled that in the 88   International handbook on the economics of migration US, to the one following 2002 when the Canadian points system facilitated the entry of highly educated individuals, including physicians, without regard for the likelihood of occupational recognition for regulated/licensed professions, McDonald et al (2011a, 2011b) observe the emergence of a substantial gap in the probability of practicing Additionally, immigrants reporting medical education from non-­English-­speaking countries are more likely to be either not employed or employed in a low-­skilled occupation Clearly, the structure of the administrative process adjudicating entry (immigrant selection) into developed countries can have important implications for the ability to practice post-­immigration in regulated health professions (and plausibly in other regulated and/ or licensed occupations), which have domestic language as well as occupational skill requirements It appears likely that the Canadian government did not appreciate these ramifications of the legislation, and associated regulations, in shifting the system to more general and generic entrance criteria in 2002 Of course, the licensure process may have biases, as discussed below A distinct, but related and understudied issue involves practice styles and treatment delivery standards of immigrant practitioners following migration Chalkley et al (2011) compare the practice patterns of comparable in-­service foreign and domestically trained dentists in Scotland using administrative data Given the discretion in treatment protocols exercised by health professionals, this is a particularly important issue Using a difference-­in-­differences approach, they observe that in particular situations and for male patients, foreign and domestically trained dentists have modestly dissimilar treatment patterns, but these attenuate within two years of post-­immigration practice Licensing/Registration, Regulation and Unionization Although not a problem exclusive to health professions, immigration for health professionals, as became clear in the previous section, is far more costly, administratively complex and fraught with risk than is the case for the typical worker These are important elements in the calculus of migration, a general discussion of which can be found in Chapter in this volume on modeling individual migration decisions For expositional purposes we categorize the institutional hurdles in health into three ‘pure’ concepts: licensing/registration (with licensure being more common in North America and the approximately synonymous term registration used in Europe), regulation and unionization However, we recognize that these terms not have standardized definitions in the literature, with regulation and licensure/registration sometimes, but not here, being close to interchangeable Also, in practice the agencies involved frequently implement versions of more than one of these concepts simultaneously Nevertheless, for our purposes licensure refers to the objective adjudication of skills, knowledge and abilities It is an expert assessment of an individual’s ability to perform his or her profession safely and in accord with the standards in place in the relevant jurisdiction We will define regulation to embody a broader set of goals reflective of the objective function of society in a particular jurisdiction and typically overseen by government Hence, for example, an individual might satisfy a jurisdiction’s licensure requirements in that the person has the appropriate qualifications to practice but still might not be granted permission to practice because of, for example, issues regarding the optimal allocation of resources within society Regulation may be socially beneficial; however, in the The international migration of health professionals  ­89 extreme, regulation may also reflect an objective that is deemed unethical by some For example, a regulatory body may exclude or restrict a particular gender or racial group from practicing despite individuals from that group being appropriately qualified (see, for example, Forcier et al., 2004; OECD, 2002) Regulatory bodies may also intervene by limiting access to educational and training programs providing occupation specific skills For example, a regulatory body might limit the number of entry positions available in health professional education programs; this is frequently associated with controlling costs A trade union or a professional association in this context is an organization that facilitates collective action by the members of a health profession.11 Its objective function is not that of society but of its membership There is a very large literature discussing the economic implications of unionization (see, for example, Addison and Schnabel, 2003) and a small literature looking at the economic operation of medical associations and other provider unions/interest groups that frequently act to protect their members from increased competition that might reduce incomes (for example, Drexler, 2008; McGuire, 2000) Internationally, some jurisdictions believe that the conflict between the alternative functions and perspectives requires distinct organizations, whereas others not.12 At issue in this context is that migrating health professionals need to navigate a series of competency tests, credential verifications and location-­specific hurdles motivated by social and/or professional objective functions Having appropriate qualifications is not sufficient to practice in many jurisdictions A very long and at times controversial economic literature looks at licensure/­ regulation/unionization in health professions (Arrow, 1963; Friedman, 1962; Inoue, 2010; Kleiner and Kudrle, 2000), and most studies conclude that while a socially beneficial function is served on average, on the margin licensing practices satisfy a union/­ association motive more than a socially beneficial regulatory one Overall, more so than for most occupations, institutions are crucial for migrants who are health professionals and for consumers of immigrant health services For instance, Noether (1986) shows that the ease with which foreign medical graduates can practice in the US is a key element of the degree of competition in the health labor market in that country Credentialing and licensing bodies, regulatory agencies and professional associations/trade unions all play important roles serving distinct and sometimes conflicting purposes In many countries the role and function of these institutions is not well studied from an economics perspective, and international comparisons are likely to yield fruitful insights in the future regarding the strengths and weaknesses of alternative structures In this respect, country-­to-­country bilateral agreements can play a very important role in regulating the migration of health professionals Inoue (2010) shows these are common, and he also points to the legal and practical conflicts in some jurisdictions, especially the EU, around language requirements From an industrial relations perspective and with reference to the international migration of nurses to the UK, Bach (2007) similarly notes the importance of bilateral agreements that allow the tailoring of professional education in one country to the requirements of another He views the institutional and regulatory framework as crucial to the short-­term operation of the immigration process for health professionals and describes current trends being towards 90   International handbook on the economics of migration internationalization rather than globalization Manning and Sidorenko (2007) also look at the importance of regulatory frameworks but focus on migration within a trading arrangement among 10 Southeast Asian nations with an emphasis on the supranational influence of mode under the General Agreement on Trade in Services (GATS).13 They observe that more developed economies have more liberal regulations regarding worker mobility, and comparing the information technology and healthcare sectors they find healthcare to be more highly regulated given both social considerations and professional interests 4  A DEVELOPING COUNTRY PERSPECTIVE Many developing countries face a crisis in healthcare owing to a crippling shortage of health professionals While emigration to developed countries is not the only source of the problem, it is an important aspect of the issue In sub-­Saharan Africa the shortage is particularly acute, and Dal Poz and Gupta (2009) estimate that 36 countries not have sufficient workers to provide minimal services for maternal, newborn and child health Important externalities arise from the health of the population, which is influenced by the size and quality of the health workforce In particular, although causality goes in multiple directions, improved population health can lead to a higher growth rate of gross domestic product (GDP) and more rapid economic development Many observers, therefore, see this as not only affecting the current, but also the future, well-­being of a nation’s population A very large research and advocacy literature exists discussing the migration of health professionals from less-­developed to developed countries Although somewhat arbitrary, research can be classified as falling within academic disciplines, primarily: medical, although typically not clinical (for example, Mills et al., 2011; Mullan, 2005); health services (for example, Parkash et al., 2006; Zurn et al., 2004, which is a review); development (for example, Beine et al., 2001); or health economics (for overviews see Rutten, 2009b; Bärnighausen and Bloom, 2011, s 21.4.3; for examples see Hagopian et al., 2005; Kangasniemi et al., 2007) Data It is very difficult to obtain information globally, but the WHO, the OECD and the World Bank have led in developing consistent and internationally comparable data on the migration of health professionals looking both at developed and developing countries (for example, OECD, 2007a; Diallo, 2004; Docquier and Bhargava, 2007) Despite this, a lack of detailed comparable data tends to cause research projects to be narrowly defined on one dimension or another Most empirical studies either focus on the flow or stock of (immigrant) health professionals in individual or a small set of source or receiving countries – frequently with a focus on physicians or nurses Some, such as Mills et al (2011), calculate the cost of the lost human capital and/or productivity to the source country; others measure the value of the benefit to the receiving country Another stream, such as Goldfarb et al (1984), focus on remittances with some evidence that this revenue on average more than makes up for the loss to source countries and explains why The international migration of health professionals  ­91 some developing countries and/or medical schools in those countries have effectively set themselves up as exporters However, even if some countries benefit from the export of health professionals, this does not prevent the healthcare crisis that they and other developing countries face Push and Pull Factors, and ‘Beneficial’ Migration A number of studies address the ‘push and pull’ factors that motivate health professionals to migrate (Rutten, 2009b, and Bärnighausen and Bloom, 2011, provide an extensive catalog of studies; Astor et al., 2005, provide an extensive opinion survey) One particularly interesting observation by Vujicic et al (2004) is that the gap in earnings between source and destination countries is so enormous that no plausible increase in source country wages will have an effect on emigration rates All told, there is a substantial list of financial and nonfinancial reasons One central conclusion from this stream of research is that there is no single ‘magic bullet’ that will stop health professional migration, nor is it obvious that stopping migration altogether would be beneficial Rather, managing the migration flow is a more realistic approach to protecting developing nations’ investments and serving their populations’ healthcare needs Developing countries that manage their health professional education systems with an eye on graduates’ migration decisions will be able to extract greater benefits for their populations; if nothing else, active management may increase remittances and/or obtain offsets for the costs of education The concept of managed migration coincides with the development of bilateral and multilateral national agreements as discussed above regarding licensure Although this will clearly not solve all the relevant problems, formalizing some of the key relationships may provide management tools to both sending and receiving governments One trend that is evident is the movement of migrant health professionals within or across sets of countries with common or closely related languages Even more so than with the average immigrant, practicing in a health profession requires advanced language skills – and these skills are frequently tested This appears to be a key motivating factor for relevant existing international agreements One particularly important hypothesis justifying the receipt of economic migrants from less-developed countries by developed countries is that the source countries benefit from the exchange One aspect of this idea, by Stark and Wang (2002), envisions migration as a type of ‘prize’ that may motivate human capital development in source countries and, despite the loss of a limited number of workers, raise the overall quality and level of training in those countries While there is some appeal to the theoretical possibility of emigration that is ‘beneficial’ to the source country, empirical research is mixed Kangasniemi et al (2007) seek to determine whether the conditions required for immigration to be beneficial exist for a particular set of countries, and conclude that they not – although these authors observe that return migration and/or remittances have beneficial effects for the sending country In contrast, looking at national immigrant populations as a whole rather than focusing on health professionals, Beine et al (2008) find some support for the hypothesis in countries with both low levels of human capital and low emigration rates It may well be that the beneficial effect turns on being able to control ­emigration; 92   International handbook on the economics of migration also, education programs in healthcare typically have excess demand and limited enrollment, and this may differentiate these fields of study Bhargava et al (2011), using an extremely large dataset of countries from 1991 to 2004 and appropriate panel data techniques, find physician migration indeed induces increased levels of medical education, but the magnitude is too small for a beneficial effect See Docquier and Bhargava (2007) for an introduction to the useful and interesting dataset employed However, they also observed that child mortality rates decrease as physician density increases but only when aggregate literacy rates are sufficiently high They conclude that reducing medical brain drain will only have modest benefits Clearly, medical workforce availability is only one element of the healthcare system and social structure required for improved health outcomes 5  CONCLUSION International immigration of health professionals has substantial externalities for both developed and developing nations Developing nations are frequently negatively affected by the phenomenon, but for these nations it appears to be but one element of a larger systemic health challenge Mobility is primarily towards higher-­paying, more prestigious and more amenity-­rich areas Health professionals migrate from rural to urban areas, from low- to middle- to high-­income countries, from developed countries with lower wages to those with higher ones Less-developed countries are most likely to be net suppliers, and although other developed countries are also net recipients, the main receiver appears to be the US Extensive multinational datasets have recently been developed that can extend the scope of future analyses addressing many of these issues Shortages of domestically trained health professionals in developed nations are a key driver affecting the human resource losses of developing nations Since the health workforce supply in wealthy countries is, by one means or another, influenced and/or controlled by governments, many of the perceived shortages driving the observed migration appear to result from planning errors or historical decisions putting substantial weight on short-­term concerns and/or relying on immigration A better understanding of these dynamics could improve planning to alleviate future shortages and thereby assist less-developed countries Quantifying and better understanding both the stocks and flows of health human resources, especially in developing countries but also in developed ones, seems extremely worthwhile Much of the research thus far is foundational and primarily descriptive, but a few studies are more analytical, and some seek not just correlations but causal impacts A few papers looking at the economic integration, and causal impacts resulting from, the arrival of international health professionals have been written – mostly for the US But the issues of asymmetric information and moral hazard in healthcare delivery that are reflected in health labor markets in the form of, for example, government regulation, licensure and other forms of practice standards have not been yet integrated into the immigration literature to any great degree To this point it appears that among new immigrants employed in a health profession, economic integration occurs relatively swiftly and their arrival causes a relatively small or negligible negative impact on the domestic labor force However, most of these studies tend to focus on immigrants who The international migration of health professionals  ­93 find employment, ignoring those unable to work in the profession in which they were trained In the case of physicians, one study suggests that the opportunity cost of not practicing is enormous, suggesting substantial rents Management tools that governments may potentially employ, for example, bilateral and multilateral treaties facilitating recognition and occupational integration, also seem particularly interesting and worthy of study both to provide feedback regarding their design and implementation, and as sources of exogenous variation to allow underlying economic parameters to be estimated In sum, there is enormous scope for future work, both in establishing the basic stylized facts and in understanding and economically ­modeling the underlying market structures NOTES   *                 We would like to thank the editors, Amelie F Constant and Klaus F Zimmermann, and an anonymous referee for helpful comments that have improved the manuscript Any views expressed in this paper not necessarily reflect those of the government of Ontario A useful summary of the shortages and maldistribution of the health workforce is the OECD’s (2008) report, The Looming Crisis in the Health Workforce This can be thought of as the flow of new MD graduates Given the integration of Canadian and American medical education, graduates of Canadian medical schools are not classified as IMGs in the US Of course, when all nations are included then all of these four countries are net recipients In 1997 the Balanced Budget Act froze the number of resident physicians at 1996 levels Currently, a 2009 bill to increase resident physician positions by 15 percent is pending in the US Congress Not all countries follow this approach In Germany, for example, there is no centralized control over the size of the medical workforce This is, in part, evident in the increasing number of individuals from developed countries, especially the US and Canada as reported by McAvinue et al (2005), who pursue their education in ‘destination’ medical schools See, for example, St George’s school of medicine in Grenada (http://www.sgu.edu/ about-­sgu/medical-­students-­demographics.html, accessed December 2011), which in 2011 had 3272 US citizens (68 percent of its total) as students in its medical program Canadians were the next largest group at 15 percent (665 students), while native Grenadians accounted for only percent of total MD enrollment Hawthorne and Hamilton (2010) provide an accounting of foreign student enrollment in Australian medical schools There is a substantial number of individuals who meet the requirements for entrance to medical school, are willing to pay very substantial tuition, and end up practicing in their native or some other developed country via this circuitous international route It is worth noting that payer, provider, and patient perceptions of the adequacy of the supply of health professionals are not necessarily in accord An overview of the history of forecasting in this area is by Bärnighausen and Bloom (2011) Identification for Schumacher is threatened if, for example, similar to the inclusion of endogenous variables in the economic integration regression in the first half of the paper, the comparison group ‘over controls’ for any wage impact on domestic workers correlated with the local nursing immigration rate His factor proportion – the location-­specific percent foreign born (or foreign educated) in nursing – could be picking up general local labor market effects, and might be correlated with similar immigrant percentages in the comparison occupations, with both occupational wage structures responding not only to the common location-­specific economic shock (as desired for identification), but also the common effect of immigrants on domestic wages (thereby removing too much of the variance in nursing wages across cities) Inference in Schumacher’s paper is also an issue since the inclusion of the percent of nurses who are immigrants implies including a macro variable in the micro regression, which requires appropriate estimation of the standard errors reflecting the ‘clustered’ nature of the data It is not clear if this was undertaken or not A controversial assumption that nursing labor markets are not monopsonistic is important for Kaestner and Kaushal’s approach, and identification in their data is also threatened by weak instruments Despite these technical challenges, this is an important area in which additional work would be worthwhile 94   International handbook on the economics of migration   Other recent work looking at language and immigration includes Chiswick and Miller (2010), Goldmann et al (2011) and the references cited therein 10 At times Canada’s points system was structured in such a way that physicians without prearranged employment were inadmissible, but since 2002, years of education regardless of field of study/­occupation has been a key determinant of points awarded This permitted physicians to enter with neither ­pre-­arranged employment nor any guarantee of licensure 11 In some jurisdictions certain health professions, for example physicians, have tended to shy away from the term ‘union’, whereas others, for example nurses, have used this nomenclature 12 Perhaps most commonly there is a distinction between the union/association that serves the professional membership and a regulatory college (at arm’s length from, or a branch of, government) comprised primarily of experts from within the profession but whose duty is to serve the broader social perspective In this context the college is sometimes charged with licensure, although that is sometimes administered directly by government 13 However, GATS’ application to the healthcare sector has limitations since it does not cover public services, that is, services provided under government authority where there is neither a commercial nor competitive basis REFERENCES Addison, John T and Claus Schnabel (2003), International Handbook of Trade Unions, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing American Medical Association (AMA) (2010), International Medical Graduates in American Medicine: Contemporary Challenges and Opportunities, Chicago, IL: International Medical Graduates Section Governing Council Arrow, Kenneth J (1963), ‘Uncertainty and the welfare economics of medical care’, American Economic Review, 53 (5), 941–73 Astor, Avraham, Tasleem Akhtarb, Maria Alexandra Mattallanac, Vasantha Muthuswamyd, Folarin A Olowue, Veronica Tallof and Reidar K Liea (2005), ‘Physician migration: views from professionals in Colombia, Nigeria, India, Pakistan and the Philippines’, Social Science & Medicine, 61 (12), 2492–500 Auda, Rick, Amanda Ross and David Vardy (2005), ‘The use of provisionally licensed international medical Graduates in Canada’, Canadian Medical Association Journal, 173 (11), 1315–16 Bach, Stephen (2007), ‘Going global? The regulation of nurse migration in the UK’, British Journal of Industrial Relations, 45 (2), 383–403 Bärnighausen, Till and David E Bloom (2009), ‘Financial incentives for return of service in underserved areas: a systematic review’, BMC Health Services Research, (86), 1–17 Bärnighausen, Till and David E Bloom (2011), ‘The global health workforce’, in Sherry Glied and Peter C Smith (eds), Oxford Handbook of Health Economics, Oxford: Oxford University Press, pp. 486–519 Beine, Michael, Frédéric Docquier and Hillel Rapoport (2001), ‘Brain drain and economic growth: theory and evidence’, Journal of Development Economics, 64 (1), 275–89 Beine, Michael, Fréderic Docquier and Hillel Rapoport (2008), ‘Brain drain and human capital formation in developing countries: winners and losers’, The Economic Journal, 118 (528), 631–52 Bhargava, Alok, Frédéric Docquier and Yasser Moullan (2011) ‘Modeling the effects of physician emigration on human development’, Economics and Human Biology, (2), 172–83 Bloor, Karen and Alan Maynard (2003), ‘Planning human resources in health care: towards an economic approach – an international comparative review’, Canadian Health Services Research Foundation, Ottawa, available at http://www.chsrf.ca (accessed 10 October 2011) Canadian Federal/Provincial/Territorial Advisory Committee on Health Delivery and Human Resources (ACHDHR) (2009), ‘How many are enough? Redefining self-­sufficiency for the health workforce’, discussion paper, Health Canada, Ottawa, available at: http://www.hc-­sc.gc.ca/hcs-­sss/pubs/hhrhs/2009-­self-­ sufficiency-­autosuffisance/index-­eng.php (accessed July 2010) Chalkley, Martin, Colin Tilley and Shaolin Wang (2011), ‘Comparing the treatment provided by migrant and non-­migrant health professionals: dentists in Scotland’, Working Paper No 249, Department of Economic Studies, University of Dundee, Scotland Chiswick, Barry R and Paul W Miller (2010), ‘Occupational language requirements and the value of English in the United States labor market’, Journal of Population Economics, 23 (1), 353–72 Dal Poz, Mario R and Neeru Gupta (eds) (2009), Handbook on Monitoring and Evaluation of Human Resources for Health with Special Applications for Low- and Middle-­Income Countries, Geneva: World Health Organization The international migration of health professionals  ­95 Denour, Linda and Rémi Junker (1995), Les mộdecins ộtrangers dans les hụpitaux franỗais, Revue Européenne des Migrations Internationales, 11 (3), 145–6 Denton, F.T., A Gafni, and B.G Spencer (2009), ‘Users and suppliers of physician services: a tale of two populations’, International Journal of Health Services, 39 (1), 189–218 Diallo, K (2004), ‘Data on the migration of health-­care workers: sources, uses, and challenges’, Bulletin of the World Health Organization, 82 (8), 601–7 Docquier, Frédéric and Alok Bhargava (2007), ‘A new panel data set on physicians’ emigration rates (1991– 2004)’, working paper, Catholic University of Louvain, available at: http://perso.uclouvain.be/frederic docquier/filePDF/MBD1_Description.pdf (accessed 30 December 2012) Drexler, Armelle (2008), ‘Le défi du recrutement des médecins diplômes étrangers dans les hôpitaux publics’, memoire de L’École des Hautes Etudes en Santé Publique (EHESP), Rennes Ferrer, Ana, David A Green and Craig W Riddell (2006), ‘The effect of literacy on immigrant earnings’, Journal of Human Resources, 41 (2), 380–410 Forcier, M.B., S Simoens and A Giuffrida (2004), ‘Impact, regulation and health policy implications of physician migration in OECD countries’, Human Resources for Health, (12), 1–11 Friedman, Milton (1962), Capitalism and Freedom, Chicago, IL and London: University of Chicago Press Goldfarb, Robert, Oli Havrylyshyn and Stephen Mangum (1984), ‘Can remittances compensate for manpower outflows: the case of Philippine physicians’, Journal of Development Economics, 15 (1–3), 1–17 Goldmann, Gustave, Arthur Sweetman and Casey Warman (2011), ‘The portability of new immigrants’ human capital: language, education and occupational matching’, IZA Discussion Paper No 5851, Institute for the Study of Labor, Bonn Grobler, L., B.J Marais, S.A Mabunda, P.N Marindi, H Reuter and J Volmink (2009), ‘Interventions for increasing the proportion of health professionals practising in rural and other underserved areas’, Cochrane Database of Systematic Reviews, issue 1, art no CD005314 Haberkern, K., T Schmid, F Neuberger and M Grignon (2012), ‘The role of the elderly as providers and recipients of care’, in The Future of Families to 2030, Paris: OECD, pp. 189–257 Hagopian, Amy, Anthony Ofosu, Adesegun Fatusi, Richard Biritwum, Ama Essel, L Gary Hart and Carolyn Watts (2005), ‘The flight of physicians from West Africa: Views of African physicians and implications for policy’, Social Science and Medicine, 61 (8), 1750–60 Hagopian, Amy, Matthew J Thompson, Emily Kaltenbach and Gary L Hart (2003), ‘Health departments’ use of international medical graduates in physician shortage areas’, Health Affairs, 22 (5), 241–9 Hawthorne, Lesleyanne and Jan Hamilton (2010), ‘International medical students and migration: the missing dimension in Australian workforce planning?’, Medical Journal of Australia, 193(5), 262–5 Hirsch, Barry T and Edward J Schumacher (2005), ‘Classic or new monopsony? Searching for evidence in nursing labor markets’, Journal of Health Economics, 24 (5), 969–9 Huang, Serana H (2011), ‘The international transferability of human capital in nursing’, International Journal of Health Care Finance and Economics, 11 (3), 145–63 Inoue, Jun (2010), ‘Migration of nurses in the EU, the UK, and Japan: regulatory bodies and push-­pull factors in the international mobility of skilled practitioners’, Discussion Paper Series A No 526, Institute of Economic Research, Tokyo: Hitotsubashi University Kaestner, R and N Kaushal (2012), ‘Effects of immigrant nurses on labor market outcomes of US nurses’, Journal of Urban Economics, 71 (2), 219–29 Kalist, David, Stephen Spurr and Tatsuma Wada (2010), ‘Immigration of nurses’, Industrial Relations, 49 (3), 406–27 Kanchanachitra, C., M Lindelow, et al (2011), ‘Human resources for health in southeast Asia: shortages, distributional challenges, and international trade in health services’, Lancet, 377 (9767), 769–81 Kangasniemi, Mari, Alan L Winters and Simon Commander (2007), ‘Is the medical brain drain beneficial? Evidence from overseas doctors in the UK’, Social Science and Medicine, 65 (5), 915–23 Kleiner, Morris M and Robert T Kudrle (2000), ‘Does regulation affect economic outcomes? The case of dentistry’, Journal of Law and Economics, 43 (2), 547–82 Kugler, Adriana, D and Robert M Sauer (2005), ‘Doctors without borders? Relicensing requirements and negative selection in the market for physicians’, Journal of Labor Economics, 23 (3), 437–65 Lesky, Linda G (2011), ‘Physician migration to the United States and Canada: criteria for admission’, in Barry R Chiswick (ed.), High-­Skilled Immigration in a Global Labor Market, Washington, DC: American Enterprise Institute Press, pp. 155–64 Machin, Stephen and Alan Manning (2004), ‘A test of competitive labor market theory: the wage structure among elder care assistants in the South of England’, Industrial and Labor Relations Review, 57 (3), 370–85 Manning, Chris and Alexandra Sidorenko (2007), ‘The regulation of professional migration: insights from the health and IT sectors in ASEAN’, The World Economy, 30 (7), 1084–113 Masselink, Leah E and Shoou-­Yih Daniel Lee (2010), ‘Nurses, Inc.: expansion and commercialization of nursing education in the Philippines’, Social Science and Medicine, 71(1), 166–72 96   International handbook on the economics of migration McAvinue, Mary B., John R Boulet, William C Kelly, Stephen S Seeling and Amy Opalek (2005), ‘U.S citizens who graduated from medical schools outside the United States and Canada and received certification from the educational commission for foreign medical graduates, 1983–2002’, Academic Medicine, 80 (5), 473–8 McDonald, James Ted and Christopher Worswick (2010), ‘The determinants of the migration decisions of immigrant and non-­immigrant physicians in Canada’, SEDAP Research Paper No 282, Hamilton, ON: McMaster University, Program for Research on Social and Economic Dimensions of an Aging Population (SEDAP) McDonald, James Ted, Casey Warman and Christopher Worswick (2011a), ‘Earnings, occupation and schooling decisions of immigrants with medical degrees: evidence for Canada and the US’, in Barry R Chiswick (ed.), High-­Skilled Immigration in a Global Labor Market, Washington, DC: American Enterprise Institute Press, pp. 165–97 McDonald, James Ted, Casey Warman and Christopher Worswick (2011b), ‘Immigrant selection systems and occupational outcomes of international medical graduates in Canada and the United States’, Queen’s Economics Department, Working Paper No 1285, Kingston, ON: Queen’s University, Department of Economics McGuire, Thomas G (2000), ‘Physician agency’, in Anthony J Culyer and Joseph P Newhouse (eds), Handbook of Health Economics, Amsterdam: North-­Holland and Elsevier, pp. 461–536 McGuire, Thomas G (2011), ‘Physician agency and payment for primary medical care’, in Sherry Glied and Peter C Smith (eds), Oxford Handbook of Health Economics, Oxford: Oxford University Press, pp. 602–23 Mills, Edward J., Steve Kanters, Army Hagopian, Nick Bansback, Jean Nachega, Mark Alberton, Christopher G Au-­Yeung, Andy Mtambo, Ivy, L Bourgeault, Samuel Luboga, Robert S Hogg and Nathan Ford (2011), ‘The financial cost of doctors emigrating from sub-­Saharan Africa: human capital analysis’, British Medical Journal, 343, 1–13 Mullan, Fitzhugh (2005), ‘The metrics of the physician brain drain’, New England Journal of Medicine, 353, 1810–18 Noether, Monica (1986), ‘The growing supply of physicians: has the market become more competitive?’ Journal of Labor Economics, (4), 503–37 Norwegian Directorate of Health and Social Affairs (2007), ‘Recruitment of health workers: towards global solidarity’, Department of Health and Social Services Personnel/Secretariat for International Work, Rapport IS-­1490 E, Sosial-­og Helsedirektorate, Oslo, available at http://www.helsedirektoratet.no/publi​ kasjoner/recruitment-­of-­health-­workers-­towards-­global-­solidarity/Sider/default.aspx (accessed 12 October 2011) Nullis-­Kapp, C (2005), ‘Efforts under way to stem “brain drain” of doctors and nurses’, Bulletin of the World Health Organization, 83 (2), 84–5 Organisation for Economic Co-­operation and Development (OECD) (2002), International Migration of Physicians and Nurses: Causes, Consequences and Health Policy Implications, Paris: OECD Organisation for Economic Co-­operation and Development (OECD) (2007a), International Migration Outlook 2007, Paris: OECD Organisation for Economic Co-­operation and Development (OECD) (2007b), OECD Factbook 2007: Economic, Environmental and Social Statistics, Paris: OECD Organisation for Economic Co-­operation and Development (OECD) (2008), The Looming Crisis in the Health Workforce: How Can OECD Countries Respond? Paris: OECD Organisation for Economic Co-­operation and Development (OECD) (2011), OECD Factbook 2011: Economic, Environmental and Social Statistics, Paris: OECD Parkash, Hari, V.P Mathur, R Duggal and B Jhuraney (2006), ‘Dental workforce issues: a global concern’, Journal of Dental Education, 70 (11), S22–S26 Rabinowitz, Howard K., James J Diamond, Fred W Markham and Jeremy R Wortman (2008), ‘Medical school programs to increase the rural physician supply: a systematic review and projected impact of widespread replication’, Academic Medicine, 83 (3), 235–43 Rutten, Martine (2009a), ‘The economic impact of medical migration: a receiving country’s perspective’, Review of International Economics, 17 (1), 156–71 Rutten, Martine (2009b), ‘The economic impact of medical migration: an overview of the literature’, The World Economy, 32 (2), 291–325 Schaafsma, Joseph and Arthur Sweetman (2001), ‘Immigrant earnings: age at immigration matters’, Canadian Journal of Economics, 34 (4), 1066–99 Schumacher, Edward, J (2011), ‘Foreign-­born nurses in the US labor market’, Health Economics, 20 (3), 362–78 Sharieff, Waseem and David Zakus (2006), ‘Resource utilization and costs borne by international medical graduates in their pursuit for practice license in Ontario, Canada’, Pakistan Journal of Medical Sciences, 22 (2), 110–15 The international migration of health professionals  ­97 Staiger, Douglas O., Joanne Spetz and Ciaran S Phibbs (2010), ‘Is there monopsony in the labor market? Evidence from a natural experiment’, Journal of Labor Economics, 28 (2), 211–36 Stark, Oded and Yong Wang (2002), ‘Inducing human capital formation: migration as a substitute for subsidies’, Journal of Public Economics, 86 (1), 29–46 UK Department of Health (2011), ‘International recruitment – code of practice’, available at: http:// www.nhsemployers.org/RecruitmentAndRetention/Internationalrecruitment/code-­of-­practice/Pages/Code-­ practice-­international-­recruitment.aspx (accessed 14 December 2011) United States Government Accountability Office (US GAO) (2006), Foreign Physicians Data on Use of J-­1 Visa Waivers Needed to Better Address Physician Shortages, highlights of GAO-­07-­52, a report to congressional requesters, Washington, DC: Government Accountability Office Vujicic, Marko, Pascal Zurn, Khassoum Diallo, Orvill Adams and Mario R Dal Poz (2004), ‘The role of wages in the migration of health care professionals from developing countries’, Human Resources for Health, (3), 1–14 Wilson, Nathan W., Ian D Couper, Elma De Vries, Steve Reid, Therese Fish and Ben J Marais (2009), ‘A critical review of interventions to redress the inequitable distribution of healthcare professionals to rural and remote areas’, Rural Remote Health, (1060), 1–21 Wismar, Mathias, Claudia B Maier, Irene A Glinos, Gilles Dussault and Josep Figueras (eds) (2011), Health Professional Mobility and Health Systems: Evidence from 17 European Countries, Copenhagen: World Health Organization, on behalf of the European Observatory on Health Systems and Policies World Health Organization (WHO) (2006), Working Together for Health, the World Health Report 2006, Geneva: World Health Organization World Health Organization (WHO) (2010), Global Code of Practice on the International Recruitment of Health Personnel, Sixty-­third World Health Assembly-­WHA63.16, Geneva: World Health Organization Zurn, Pascal, Mario R Dal Poz, Barbara Stilwell and Orvill Adams (2004), ‘Imbalance in the health workforce’, Human Resources for Health, (13), 1–21 5  Independent child labor migrants* Eric V Edmonds and Maheshwor Shrestha 1  Introduction Independent child labor migrants are working children who have migrated to their current location of employment and not cohabitate with a parent We not know how many children are independent child labor migrants Yaqub (2009) tallies counts from case studies He concludes the number of independent child labor migrants must be in the tens of millions Gurung (2004) documents 121 000 from Nepal Kielland (2008) identifies 100 000 from Benin Kielland and Sanogo (2002) estimate 330 000 from Burkina Faso Camacho (2006) measures 400 000 in the Philippines Independent child labor migrants can be either international or domestic migrants We suspect that the latter is more prevalent Independent child labor migrants are an extremely vulnerable population They are often found in the worst forms of child labor The International Labour Office (IPEC, 2003) estimates 1.8 million children in prostitution and pornography Qualitative work with children in the commercial sexual exploitation sector typically finds that most participants started as independent child labor migrants Edmonds and Salinger (2008) tabulate a nongovernmental organization’s (NGO’s) records that document more than 5000 child labor migrants confined to their work sites in Mumbai slums Fifty-­four percent of these children were under 12 All lived away from their parents Independent children are more easily abused and exploited as they may lack the information and capacity to identify dangerous conditions, exploitation and abuse The purpose of this chapter is to examine the current state of research in economics on independent child labor migrants Children may live away from their parents for a number of reasons other than child labor In this chapter, we omit discussion of issues specific to refugees, orphans, child soldiers and human smuggling Friebel and Guriev discuss this last topic in detail in Chapter in this volume Instead, we focus on research where parent and child separation and child labor migration is a choice There are few studies that directly examine the decision that leads to the child living independently We suspect that the paucity of research in part owes to sampling difficulties Many independent children are in institutional settings outside the frame of traditional household surveys Some surveys specifically sample migrant or independent children but research on the consequences of independence require a counterfactual of what the child would be doing without independence Forming a counterfactual requires information from the independent child’s home location in the case of a migrant child Such matched data typically not exist This chapter is structured as follows Section reviews the measurement issues salient in empirical discussions of independent child labor migration Section reviews research on why employers might prefer independent child laborers and summarizes the findings of several case studies on that topic Section considers the evidence on the factors 98 Independent child labor migrants  ­99 that influence the decision for the child to live independently Sections and draw extensively from Edmonds and Shrestha’s (2009) review of the implications of the child labor literature for understanding child migration Section examines the research on the consequences of child independence for child welfare Section concludes with ­recommendations for future research priorities An obvious, important initial question is whether there is anything in the economics of independent child labor migration that differentiates it from migration research in general We not think the literature is sufficiently well developed to answer this question There are two obvious issues that differentiate independent child labor migration First, children are more vulnerable to abuse and exploitation than adults Hence, the scope of harms from independent child labor migration is broader than adult migration Second, children may not have agency in their migration and employment situation Hence, issues related to who is making migration and employment decisions and the information that agent possesses are more central in the discussion of independent child labor migration compared with adult migration 2  Measurement There are three types of survey data used to identify independent child labor migrants: specialized surveys, representative surveys using roster methods and representative surveys using fertility methods All have significant disadvantages for estimating the incidence of independent child labor migrants and for economic analysis of the causes or consequences of independent child labor migration Specialized surveys often employ opportunistic sampling to directly count contacts of independent child labor migrants in specific sectors These data are difficult to use to create estimates of the number of independent child labor migrants because of the lack of a sample design Specialized surveys also not contact children outside of the targeted sector Hence, they generally cannot be used to infer counterfactuals of what independent migrants would be doing without migration or independence Representative surveys, usually household based, can be used in two ways to learn about independent child labor migrants The roster method identifies children outside of parental care by responses to questions on whether the parents are present for an individual listed on the household roster The fertility method asks a sampled woman about her fertility history and compares her fertility to the roster A variant on the fertility methods asks household members about any children who are living away from the household Typically, questions on children living away focus on children who are temporarily away Hence, these questions on children living away can lead to a different count of children than would follow from a fertility survey The roster method is typically used by studies that examine independent children in the location where they live The fertility method is useful for studying from where independent children come Because the roster method is based on children who have been identified in the roster, it is less subject to errors of omission than the fertility method and typically contains a large amount of information on the children The fertility method is less subject to bias from children located outside of the sampling frame but is unlikely to produce an accurate picture of the time allocation of children living elsewhere 100   International handbook on the economics of migration Table 5.1 Number of children aged 5–17 years in 2002/03 Bangladesh living independently No parent or grandparent present   Married   Of non-­married:    Economically active    Of economically active:     Growing cereals     Farming of animals     Mixed farming     Textiles     Furniture     Retail trade     Other Total Male Female 1 092 927 104 371 504 014 39 216 588 911 65 154 135 270 87 174 48 096 39 717 11 617 7 624 6 825 17 228 18 101 34 158 32 892 3 993 2 832 799 8 358 15 120 23 180 6 825 7 624 4 792 6 027 8 869 2 981 10 978 Note:  Authors’ calculations from the 2002/03 Bangladesh Child Labor Survey Mixed farming includes both the growing of cereals or other crops and the farming of animals Edmonds and Shrestha (2009) use the roster method to tabulate and examine the incidence of children living independently of their parents in the 32 countries covered by the 2005/06 UNICEF Multiple Indicator Cluster Survey (MICS 3) project data The MICS3 data are representative of an estimated 160 million children Ten percent are independent The phenomenon of children living without any co-­resident parent present is most pronounced in the sub-­Saharan African countries covered by the MICS project More than in children under 17 live without a parent present in Cameron, Côte d’Ivoire, Gambia, Guinea Bissau, Malawi, Sierra Leone and Togo Less than in 100 children under 17 live independently in countries such as Albania, Bosnia, Macedonia, Montenegro and Syria In every MICS3 country, apart from Macedonia, girls are more apt to live away from a biological parent than boys Fifty-­nine percent of those who are independent attend school compared with 77 percent who live with a parent Sixteen percent of independent children participate in paid employment compared with 14 percent of children that live with a parent Child independence occurs for many reasons beyond labor and marriage When detailed data on time allocation and living arrangements are available, it is possible to estimate the prevalence of independent child laborers among the population of independent children Table 5.1 contains tabulations of the activity status of independent children using the 2002/03 Bangladesh Child Labor Survey In 2002/03, over a million children aged 5–17 years in Bangladesh live without either a co-­resident parent or grandparent We exclude children living with grandparents, because we cannot identify whether the child lives with a parent present if the grandparent is present in this specific questionnaire As is typical of most household survey rosters, we cannot identify whether the child or the parent migrated Hence, Table 5.1 (like the MICS tabulation above) is not restricted to migrant children Less than a tenth of the resident, independent children are married Of those who are not married, Independent child labor migrants  ­101 Prevalence of working independent children 15 SLE GNB CIV 10 MWI TG GMB CMR THA BDI BGD TJK SYR UKR TTO SRB 10 Log GDP per capita Sources:  Estimates of the prevalence of economically active independent children are authors’ calculation from the MICS-­3 data Population estimates are from the UN population database for 2005 for 5–19 GDP per capita is from the World Development Indicators, 2005 PPP Series Prevalence estimates are reported in Edmonds and Shrestha (2009) Figure 5.1 Economically active, independent children aged 5–17 years and GDP per capita 135 thousand are economically active Of those, agriculture is the dominant sector of employment Boys are also substantively involved in furniture and retail trade Girls are in textiles and furniture Thus, in the Bangladesh example, most independent children identified with the roster method are not economically active Economically active, independent children are more prevalent in poor countries than in rich in the MICS data Figure 5.1 contains the plot of economically active independent children against GDP per capita for the 32 countries in the MICS3 project Ten percent means that 10 percent of children aged 5–17 years are economically active and independent (not living with either parent) Each country is marked with a circle, the size of which is proportional to the country’s population of children The prevalence of economically active independent children declines rapidly with gross domestic product (GDP) per capita and is unusual in the MICS3 countries with incomes above $3000 per person per year A comparison of roster method and fertility method calculations from within one survey illustrates that estimates of the prevalence of independent children can vary substantively across approaches The 2010 Nepal Living Standards Survey contains 102   International handbook on the economics of migration Table 5.2  Different types of independent children in Nepal, 2010 Independent children (roster) Independent migrants Migrants Independent children (fertility) Absentee Non-­migrants Prevalence Age Female 4.9 3.0 9.9 8.6 11.5 90.1 12.7 12.4 12.0 13.5 11.6 10.8 67.9 48.5 51.0 41.0 34.7 50.9 Currently Working Currently schooled for wage abroad 65.7 89.4 93.8 n.a 79.4 88.0 9.3 10.0 7.0 n.a 13.0 5.4 0 n.a 19.5 Notes:  For absentees, the work categories are based on parents’ report of the absentee’s primary occupation, not based on all the activities that an individual does Roster matches table in definition ‘Independent migrants’ restricts the ‘roster’ sample to children who have migrated ‘Migrants’ refers to children who have migrated regardless of residency with parents ‘Fertility’ refers to children of a resident female who are not co-­resident ‘Absentee’ refers to children that the household reports as living elsewhere N.a indicates that the data is not available Source:  Authors’ calculations from Nepal Living Standards Survey III Methods are described in the text a detailed roster, a fertility survey, and a report on the activities of temporary out migrants Hence, it can be used to contrast the roster method with two versions of the fertility method to identify independent child labor migrants These tabulations are in Table 5.2 The first three rows of Table 5.2 are based on the roster method The fourth and fifth rows use the fertility method to identify independent children The first row, ‘Independent children (roster)’, uses the roster method to identify independent children identically to Edmonds and Shrestha (2009) ‘Independent migrants’ defines migrants as those independent children who also migrated to their current household from somewhere else In 2010 Nepal, 61 percent of independent children are also migrants Independent migrants are more male, more likely to be in school and more likely to work for pay than independent children in general The data imply that 0.3 percent of children in Nepal are independent migrants who work in paid employment The third row of Table 5.2 includes children who have migrated regardless of their co-­residency with parents Comparing row and row indicates that independent child migrants are more likely to be male, less likely to attend school and more likely to work for pay than children who migrate with their parents Most migrating children so with their parents The fertility method is used to tabulate independent children in the fourth row of Table 5.2 The prevalence of independent children is 75 percent greater with the fertility method compared with the roster The fertility method is not precisely comparable with the roster method as roster-­based estimates of independent children are conditional on the children living away from the father and mother The fertility method identifies children living away from their mother They may be with their father A typical shortcoming of the fertility method is that we know nothing about their current schooling and work status The fifth row of Table 5.2 tabulates responses to a survey question about absent Independent child labor migrants  ­103 family members that the respondent expects to return According to that instrument, nearly 12 percent of children are living away from home We were surprised to see this above both the roster and fertility survey estimates of independent children under 18, although these children may be absent with a parent and thus not included in the independent children tabulations We suspect that some children living away are included in the roster and that some children in the absentee responses are not biological children of mothers present Absent children are less likely to attend school than independent migrants and more likely to work for pay Nearly one in five absent children are abroad and completely outside the frame for a roster-­based method for estimating independent child labor migrants If all absentee children are considered independent and all children abroad are assumed to be in child labor (some are students), then the fertility method from the fifth row implies that at most 2.6 percent of Nepali children are independent child labor migrants or 331 000 children Unfortunately, the outcome from Table 5.2 is not constructive Measures of independent children and child migrants are sensitive to the approach to measurement and the conditions of those children depend on what group is identified A thorough census is clearly an alternative that should resolve the sampling frame problems associated with the roster method and the selection problems intrinsic to the fertility survey approach (although out-­of-­country children will be missed) To our knowledge most census questionnaires are ill suited to measure whether a child is independent or a migrant More research on measuring and identifying the status and conditions of independent children and child labor migrants in particular is needed for this literature to progress 3  Demand for Independent Migrant Child Labor Why employers hire independent child labor migrants? Independent child labor migrants might be a perfect substitute for other types of child labor, their independence and migration status reflecting the spatial distribution of child labor supply and demand Alternatively, their independence might make them less costly and easier to exploit Their status as migrants might make them less expensive, because employers can offer services that it would be more costly for the child to acquire on their own (security, housing) and the migrant values location-­specific amenities In this section we review several cases where there appears to be a concentration of independent, usually migrant, child labor in a sector Our purpose is to review several cases in order to inform a more general discussion of why employers hire independent children 3.1  Domestics A child domestic laborer is a child under 18 who performs domestic chores in his/her employer’s household with or without remuneration Domestics can be boys and girls, although there is substantive sex typing of tasks For example, studies observe male domestics tending gardens or livestock, with girls focused inside the home The use of children as domestic workers is common practice in the developing world, especially in South Asia (see, for example, ACPR, 2006; KC et al., 2002) We are not aware of global estimates of the prevalence of child domestics, but there are several 104   International handbook on the economics of migration country specific estimates For example, the Associates for Community and Population Research (ACPR, 2006) estimates 421 426 in Bangladesh ‘Encuesta de Condiciones de Vida’ (ENCOVI 2000) estimates 17 350 in Guatemala Sharma et al (2001) estimate 21 191 in Nepal The survey of activities of young people in South Africa 1999 (Statistics South Africa, 2001) estimates 53 942 in South Africa Domestics often live in their employer’s house and work within the premises of the house Child residency in their place of employment is especially common for independent child migrant domestics Employers commonly reference the need to shelter the domestic from the dangers of urban life as a reason for confining the domestic to their work site Reports of physical abuse, violence and sexual abuse are not uncommon Children report feeling ‘threatened’ in some ways and that they are not free to leave their current work at their will Many countries consider domestic service one of the worst form of child labor Domestics are generally sent by their parents or relatives (see, for example, ACPR, 2006) A recent study from Ethiopia documented extensive informal networks to match rural children to households in Addis Ababa (Kifle, 2002) In Thailand, the ILO/IPEC Rapid Assessment (Phlainoi, 2002) found that communities of origin have developed mechanisms and social networks to ensure confidence in the recruitment and conditions of their children Sometimes agents are involved in matching children and jobs, but often relatives help place children It is not unusual to find domestics working in the homes of distant relatives In a study of domestics in Phnom Penh, the National Institute of Statistics (NIS, 2004) found that 60 percent of domestics reported that their employer was a relative In terms of motive for becoming a domestic, studies typically report that the child became a domestic at a parent’s request Sharma et al (2001) report that 82 percent of domestics in Nepal answer that the decision to become a domestic was made by their parents When asked about motives, respondents usually mention the primacy of poverty-­related concerns However, a significant proportion of child domestic labor mention the possibility of better schooling as one of the reasons for their decision to work as a domestic worker (for example, KC et al., 2002) Why are children employed as domestics? The literature is unclear on this point Domestic tasks are not ones where there is clearly a ‘nimble fingers’ comparative advantage story for child labor In a study in Bangladesh, ACPR (2006) reported that 80 percent of the employers of child domestic workers in Bangladesh indicated that domestics were easier to deal than adults Thirteen percent reported that children were less expensive The fact that domestics are often independent child migrants as well, suggests that something about the provision of food, shelter and the employer’s location in urban areas might be valuable to the person deciding to send the child to work as a domestic 3.2  Mining Mining is another sector where independent migrant children have been documented Mining is considered by many to be an easy way to make quick money Most children are likely to work in informal small-­scale and nonskill-­intensive mining rather than in large-­scale mining where much of the processes are highly mechanized Often, small-­ Independent child labor migrants  ­105 scale mining sites are surrounded by a hub of temporary households full of migrants looking for jobs Global estimates of children in mining could not be located In some countries the involvement of children can be substantial An ILO (2006) study documented 200 000 children in mines in Burkina Faso A recent Human Rights Watch (2011) report estimated that 20 000 children work in artisanal gold mines in Mali We cannot find estimates of what share of these child labor migrants are living independent of their parents Children can be involved in different activities directly related to mining Children work in above-­ground activities like crushing rocks, drilling rocks, washing rock dusts, collecting and carrying pieces of crushed rocks or heaps of mud, or under the ground in tunnels and mine shafts Child labor in mining is usually viewed as one of the worst forms of child labor Children might also be involved in other activities not related to mining For example, they might work in restaurants, bars and shops in temporary settlements around the mining sites Are children in mines used differently than other types of labor? This has not been examined scientifically as far as we can identify There are anecdotes that children benefit from their smaller frames and lack of awareness of risks For many tasks, independent children are likely just another type of labor that is available to mines However, it seems plausible that independent children might be more easily induced into dangerous tasks that parents would not permit 3.3  Agriculture Agriculture is the largest employer in low-­income countries Most estimates suggest that agriculture is the largest employer of independent child labor, but there is little evidence that this is for any reason other than that agriculture is the largest employer of all types of labor In our tabulations from Bangladesh in Table 5.1, we found that 44 percent of economically active independent children work in agriculture Based on the case study literature, evidence exists of migrant child labor in sugarcane, cocoa and cottonseed At the time of writing, there is a scandal involving the use of migrant child labor in cotton farms in Burkina Faso that source cotton for Victoria’s Secret products Sugarcane, cocoa and cottonseed are prevalent in the sector studies, because plantation production will often require a seasonal, migrant population for harvesting It is natural that migrant child labor will be drawn into migrant farm labor in general Estimates of the volume of migrant child labor can be very large in agriculture Venkateswarlu (2007) reports that 416 460 migrant children work in hybrid cottonseed farms in India The ILO (IPEC, 2003) estimates as many as 240 000 migrant child laborers in seasonal cotton harvesting in Turkey Few studies document how many of these child laborers are migrant Dàvalos (2002) found that 18 percent of boys and percent of girls were independent migrants in sugarcane farms in two districts of Bolivia It is not unusual for seasonal agriculture labor to be managed by intermediaries who help bring together labor and employers Both the India and the Turkey study document reliance on intermediaries An International Institute of Tropical Agriculture (IITA, 2002) study of cocoa farms in Côte d’Ivoire documented that 41 percent of child labor 106   International handbook on the economics of migration found their job using an intermediary Stories of deceptive recruitment permeate the case study literature Albertine De Lange (2007) reports on the types of deceptions used by recruiters to draw independent children to cotton farms in Burkina Faso While children might be easier to recruit, especially in ways that allow recruiters to capture significant rents, it does not appear to be the case that independent children work in significantly different tasks 3.4  Handcrafts Such as Handmade Carpets Carpet production is associated with migrant child labor in South Asia, especially in Nepal and India An ILO/IPEC rapid assessment in the Nepalese carpet sector (KC et al., 2002) estimated that 7689 children worked in the sector and that 96 percent were migrants The rapid assessment did not separately identify independent child migrants but, given the low average age of workers in the sector reported in the rapid assessment, it is plausible to assume that independence was prevalent in the project In handmade carpets as well as many handcrafts, one often hears that there is a value for little fingers While the ‘nimble fingers’ story for child labor is not as compelling as popularly believed (Edmonds, 2007, has a discussion), it may be the case that employers of migrant children are able to work them in conditions that would not be feasible with an adult While we are not aware of any formal studies that meet modern standards of evidence, there are many anecdotes in the press of employers working children in tasks where small fingers might be at a comparative advantage in conditions that adults would not tolerate It is unclear how systematic or widespread this is 3.5  Street Children, Beggers and Ragpickers Street children are children under 18 who are living in the streets and detached from their families They usually have no fixed place to stay and are highly mobile Giani (2006) describes them as ‘run away’ children who migrate to the streets because they feel emotionally, physically and sexually vulnerable at home There are two broad definitions of street children in the studies below: those who are on the street during the day but return home to sleep at night and those who work and sleep on the streets Children in the latter category can be considered as independent children Even when these children originate from the same locality as the streets they inhabit, they have moved out of their homes without their parents or adult guardians Street children often petty jobs available in the streets Those mostly include street hawking, begging, ragpicking, selling goods, and so forth Since street children not have a particular job, the use of recruiting networks is absent On the other hand, children who have migrated with false hopes and promises to work in various other sectors often end up in the streets Estimates of the prevalence of street children, beggars and ragpickers vary wildly Given the low barriers to entry into the job, it is possible that the number of independent children in these activities vary substantively from season to season and year to year The Bangladesh Bureau of Statistics conducted a quick count survey and estimated about 2500 street children in Bangladesh (FREPD, 2003) Most (58 percent) of these children have very weak links with their parents The street children’s survey in Ghana (Ghana Independent child labor migrants  ­107 Statistical Service, 2003) identified 2314 street children under the age of 17 Most of them (53.2 percent) lived outside their parents’ district of residence, and some had traveled considerable distances to become street children In most countries, boys are more likely to be street children than girls The question of why employers use children as ragpickers is difficult as it is not known what fraction of ragpickers and street children are employed by a third party or are self-­employed Given the extremely low costs to entering the industry, it seems likely that much of the work that beggars and ragpickers should be motivated by the labor supply concerns discussed below One hears anecdotes of organized crime creating non-­ market barriers to entry for beggars and ragpickers, but it is not clear how widespread these anecdotes really are 3.6  Summary We know very little about why employers might prefer independent child migrants The popular press and some rapid assessment work document stories built around the idea that children are easier to exploit There are competing explanations In the case of domestics, there is some evidence that employers can provide amenities and in-­kind services that are of value to independent child migrants at lower costs than the migrant could provide on their own In mining and handcrafts, there is some evidence that children might have an advantage in some types of tasks, although that evidence is far from conclusive Children might also have an advantage in ragpicking and begging In all of these types of child labor that employ independent child migrants, it is possible that independent child labor migrants are not distinct from other forms of child labor This seems most likely to be the case with agriculture, where there is less evidence of physical segmentation of migrant labor from non-­migrant labor 4  Entry into Independence We focus our discussion on whether a single child is sent away by an agent The agent is able to control the child’s activity at home and to decide what the child will if the child is sent away The agent’s problem is to maximize household welfare, and the agent chooses the combination of activity and location that does so Because the agent decides location and activity simultaneously, it is not possible to predict the next best activity to the one the child is observed in For example, suppose a child stays home and works in a slaughterhouse If the slaughterhouse activity is removed from the agent’s choice set, the agent could be best off moving the child to a different job away from home When a child participates in multiple activities in a given location, the agent’s payoff must be equalized across activities The agent’s payoff for having a child participate in a given activity at some location depends on the marginal utility of income, the effect of participating in the activity on agent’s income and relative prices Thus, factors that influence the child labor decision then influence the decision to migrate Anything that improves the economic return to keeping the child at home will increase the likelihood the child stays Anything that raises the net economic return to leaving will lead to more independent child migrants 108   International handbook on the economics of migration Edmonds (2007) is a dedicated review of the child labor literature In what follows, we highlight some of the findings from that literature that have specific relevance to the context of child labor migrants 4.1  Poverty Motives There is broad statistical support for the qualitative evidence that links work and living standards Two-­thirds of the cross-­country variation in economic activity rates can be explained by differences in gross domestic product per capita (Edmonds 2010a) Causal evidence comes from many sources including Edmonds and Schady (2012) who document substantial declines in paid employment in reaction to a randomized cash transfer valued at 20 percent of foregone earnings Rigorous statistical studies on the independent child labor migration–poverty connection are rare The narrative evidence from working, independent child migrants appears to put a lot of emphasis on poverty at home as a motive for migration Roe’s (1999) study of street children in Bangladesh is an excellent example They report that migration to the street improves their access to income, food, clothing and other necessities that their parents cannot adequately provide Sheikh Hasina’s (1989) discussion of street children in Bangladesh emphasizes that in addition to supporting themselves, there is hope that the child will contribute financially to their home family’s welfare as well Sometimes this contribution comes simply by relieving their family of the need to care for the child Other times, the support comes from remittances or advanced payments on the child’s earnings See Chapter 16 in this volume for further discussion of how migrants might support their sending families The link between child independent migration and poverty is a bit subtler than some case studies emphasize For example, Kielland and Sanogo (2002) study Burkinabe children and argue that poverty is a weaker determinant of migration than one would expect from studies that focus on the responses of child migrants They observe that the challenge of meeting basic needs is more influential for girl migration than boys and that poverty seems more influential in rural to urban migration than in rural to rural migration In a study from the Indian states of Bihar and Uttar Pradesh, Edmonds and Salinger (2008) observe that child independent migration is more likely from poorer households but in remote locations the poverty factor plays a lesser role Their explanation is that the costs of migration become a larger influence on migration decisions in more remote areas of India Djebbari and Mayrand (2011) provide a rare, direct study of the relationship between income and child migration They find that the child support grant in Kwa Zulu-­Natal reduces the prevalence of child independence The grant is only available if the parent is present, so it is somewhat complicated to interpret whether the impact of the grant is through the impact of income or the relative cost of child–parent separation However, income issues could be extremely important and it is not obvious whether income itself would lead to more or less child independence in this case where families may have difficulty affording to migrate The credit–migration connection is similarly complex Although credit constraints force families to make child labor and schooling decisions based on immediate concerns, improved incomes and access to credit not necessarily ameliorate child labor Independent child labor migrants  ­109 School Here, RH+WH Should the child attend school, work, migrate? School Away, RA+WA No School, WO • Idle, WHI • Domestic Work WHD • Work in family enterprise, WHF • Work outside of family enterprise, WHO • Idle, WAI • Domestic Work for host, WAD • Work in host family enterprise, WAF • Work away from host, WAO • Live independently, WAW • Idle, WOI • Domestic Work, WOD • Work in family enterprise, WOF • Work outside of family enterprise, WOO • Leave home for work with family member, WOF • Leave home for work independently, WOW Source:  Edmonds and Shrestha (2009) Figure 5.2  The child time allocation problem or reduce child migration Migration is costly, and Edmonds and Salinger (2008) emphasize that wealthier households will be better able to afford to migrate Kielland and Sanogo (2002) explicitly emphasize this in their discussion of child migration in rural Burkina Faso Many families cite an inability to finance migration as a major barrier to migration This point is salient in the analytical framework of Figure 5.2 At times, child independent migration for work might be perceived as the best option available to the child Migration is costly Sometimes, poverty might limit a child’s ability to migrate Edmonds et al (2005) document the diverse and complicated relationship between income and the migration of children, young adults and others They look at how household composition responds to eligibility for the Old Age Pension in South Africa They find evidence of some household members moving away from the income and others moving into the money Migration becomes less costly with the transfer so individuals with large returns to migrating can so But, the return to moving in with the elder pensioner may increase the return to migrating towards the pensioner The relative import of such influences varies across the population 110   International handbook on the economics of migration 4.2  Insurance Failures Children may migrate both as a way to diversify risk that would maximize the net expected return to migrate or to cope with a realized shock that either lowers income or reduces the net expected return to staying home Migration serves as a way to cope with shocks and crises Migration of household members and remittances from migrants help families cope with difficult times regardless of whether the origins of the difficulties rest with financial crises (Yang, 2008), productivity shocks and natural disasters (Halliday, 2008) or weather shocks (Rosenzweig and Stark, 1989) Akresh (2009) gives rare evidence of a link between economic shocks and child out-­ migration In data from rural Burkina Faso, Akresh finds that household agricultural shocks induce families to send children away through fostering networks Health shocks can also induce migration Sometimes that migration will be away from the shock Edmonds (2010b), for example, finds that paternal disability is a strong predictor of entry into ragpicking in Nepal Other times, health shocks might induce migration towards the shock in order to provide care Young and Ansell (2003) document the migration of Southern African children to households with a sick member to provide care One consequence of an economic shock in the context of credit constraints is that it may create a short-­term liquidity crisis that has substantial implications for child migrants Srivastava (2005), for example, documents that laborers in India are often bonded, because their parents received an advance on their labor in exchange for migration and the job It is difficult to know how widespread this phenomena is, but it implies that short-­term needs for medicine, seeds or fertilizer may induce families to choose to send children away into bonded labor settings 4.3  Sending-­Area Opportunities Children are more likely to work when the relative return to work is higher Children will be drawn into migrating for that work if the returns in receiving labor markets are high relative to the opportunities at home and the costs of migrating This section focuses on evidence on the influence of sending-­area employment opportunities Most children work inside their home The availability of productive assets can thus be an important influence on whether children work (for example, Mueller, 1984) Generally, it appears that children migrate more from households with fewer productive assets Punch (2002) characterizes the migration of rural Bolivian youths to urban areas or abroad to Argentina as a strategy to cope with lack of access to land and limited economic opportunities Ford and Hosegood (2005), in their study of child migration from a rural district of a province in South Africa, also find that children in households with more assets are less likely to migrate On the other hand, Edmonds and Salinger (2008) and Iversen (2002) find little clear association between household working capital (household farm ownership) and child migration One explanation for their finding is that the potential returns to time spent in household-­based activities are small relative to the anticipated returns to migrating When children work outside of their home in sending areas, opportunities outside the home in sending areas should deter migration in the same way that household-­based Independent child labor migrants  ­111 opportunities might be expected to Child migration studies and reports mention the lack of employment opportunities and lower wages in origins as one of the main reasons for child migration for work Punch’s (2002) study of youth migration in rural Bolivia, and Erulkar et al.’s (2006) study of adolescents in slum areas of Addis Ababa are but a few examples of such studies The challenge in these analyses is that more remote areas often have fewer employment opportunities, but they are also more expensive to migrate from Kielland’s (2008) study of child migrants in Benin overcomes this problem by looking at differential employment opportunities by gender in the same localities She observes that agriculturally intensive localities leave girls with fewer independent opportunities than for boys and hence female migration is greater than male migration from those places Her findings are consistent with the view that boys and girls have similar employment opportunities in destination areas We might see no such patterns in countries where girls have few employment opportunities in destination areas as well The relationship between seasonal patterns of labor demand and child migration is complicated because there is seasonal variation in employment opportunities as well as incomes Households that depend highly on income from agricultural labor would also face seasonality in household income When agricultural labor demand is low, households will suffer from lower income This situation is aggravated by lack of credit markets To cope with this seasonality of income variation, family members, including children are likely to migrate temporarily in search of work opportunities during lower labor demand periods, assuming that higher labor demand areas are accessible They often return to their origins during harvest or sowing times to help their families as labor demand in origin areas tend to be high Giani (2006) and Baas (2008) are but a few studies that document the increase in child migration in the lean season and their return during harvest times Quiroz (2008) further documents that entire families, including children, migrate to the coffee plantations in Guatemala for seasonal work Seasonal migration is not limited to farm work – Bastia (2005) also finds seasonal migration to be customary in urban–rural migration of Bolivian migrants Schooling must be an important component of the return to time in the sending area When schools are far away or unavailable, children have less alternative uses of their time Typically, children either migrate to other places with educational opportunities (most relevant for boys) or work The effect of school access on child migration is, however, not well documented One exception is from Kielland and Sanogo’s study of child migration in Burkina Faso They observe no effect of presence of a primary school in the village on child labor migration overall However, they find that the presence of a primary school in the village reduces girls overall labor migration (within and outside the country) and reduces child migration to work as a domestic If schooling is available but of low quality, it may induce migration in the same way that a lack of access does Giani (2006) studies rural–urban migration of children who have moved to Dhaka from various parts of rural Bangladesh through case studies and child interviews She finds that migrant children take migration as an alternative to poor schooling at their origins She argues that poor quality of schooling, coupled with poor performance, lack of interest and abusive behavior from teachers trigger child migration to urban centers 112   International handbook on the economics of migration 4.4  Destination Employment Opportunities Higher wages and therefore income lures many children to migrate to work in urban areas and abroad Several case studies and interviews with child migrant labor document higher expected wages in urban centers and more employment opportunities in cities as one of the main reasons for child migration Punch’s (2002) study of youth migration in rural Bolivia and the Erulkar et al (2006) study of adolescents in slum areas of Addis Ababa are but two examples of such studies While evidence of an urban wage premium is widely documented for adults, we not know of similar evidence for children It is not clear if migrant children achieve higher wages than their counterfactual if they had stayed home However, it is clear that hope for higher wages and a brighter future is important in the decision to leave home Bastia (2005) documents the use of lies and deceit by recruiters to rural Bolivian children in order to persuade them to migrate to urban centers or to Argentina Pearson et al (2006) also document that children often migrate to urban Thailand in the hope of better jobs but often end up with worse jobs than in their origins For further discussion in the context of migration in general, see the ­discussion in Chapter in this volume Traineeships and apprenticeships are two formalized institutions to which children migrate in the hope of a better life Kok (1997) examines historic youth migration in the Netherlands She observes that in cities, children whose fathers were skilled workers were most prone to migrate She argues that these parents had the necessary contacts, information and money to find useful and interesting jobs or apprenticeships in another town Kok observes similar patterns among merchant and elite families Empirical evidence on the impact of transport and search costs on migration varies based on whether researchers control for other correlates of remoteness Edmonds and Salinger (2008), for example, control for individual family living standards and local employment opportunities in areas of migrant origin in Bihar and Uttar Pradesh With these controls, they observe that more remote communities are less likely to have children move away This finding that migrants are less likely to come from more  remote locations, everything else equal, is consistent with the large historical literature on the migration of Americans out of the South at the start of the twentieth century Additional costs come into play when one considers international migration Legal cross-­border migration often requires lengthy bureaucratic process and is often costly McKenzie (2007) finds that countries with high passport costs, indicative of poor bureaucracy, have lower migration rates Sending a child legally across borders then could turn out to be prohibitively expensive in a poor, developing country context Child migration, then, could take the form of illegal migration across borders These illegal children are most likely to be trafficked and to be working in exploitative situations Lending support to this hypothesis, Caouette (2001) finds that there are significant proportions of undocumented women, and children as young as 13, along the borders of China, Myanmar and Thailand Those children and migrants, she posits, are likely to suffer from extensive debt bondage, arrests and extortion, forced labor and poor living arrangements Independent child labor migrants  ­113 4.5  Information Migration is a selective process Individuals rarely migrate without having some form of network already present in the destination A social network in destination can be an important factor because of several reasons First and foremost, a possible migrant gets detailed information about the conditions at the destinations through his social network This information is usually more valuable to the migrant than that available through media or otherwise Secondly, a migrant gets more support in the destination after he or she migrates, which makes transition to the destination easier Therefore, the propensity of migration of an individual to a particular destination depends upon the migration experience of his or her social network at that destination Empirical studies support the idea that existing social networks in a destination promote migration McKenzie and Rapoport (2007), in their study of international migration from Mexico, find that migration networks increase the likelihood of migration by spreading the benefits to poor members of the network Similarly, Curran et al (2005) also find that migration experience in a destination increases the propensity of migration to that destination significantly in Thailand They also observe that female migration experience at a destination has a stronger impact than male migration experience in all (individuals, household, community) levels of observations Networks are not just important for migrants to find work, it is also important for employers to find employees Employers use their own network through intermediaries, recruiters, relatives, friends and previous employers to find workers Rigorous empirical evidence on how employers use networks to identify and recruit workers does not appear to exist However, various reports of sector studies reveal that employers use their network or hire recruiters in order to find workers, including children We will review the use of employers’ network under recruiting sections when we discuss the sectors in which child independent migrants are most likely 4.6  Sibling Interactions Sibling interactions arise with great frequency in discussions of independent child migration Siblings influence the marginal utility of income, the return to providing services to the household, and the relative cost of different types of spending and investment in the household Parish and Willis (1993) find that, more important than the caring and support she provides to her siblings, the biggest contribution of the eldest daughter comes through marriage and out-­migration from the family Their finding in Taiwan is consistent with Vogl’s (2011) finding in contemporary Nepal and India Sibling composition, especially birth order and spacing, can have an important role to play in a child migration decision beyond their influence on marriage Edmonds and Salinger (2008) observe that migrant children tend to be older among siblings Punch (2002) also notes that older siblings are much more likely to migrate at a young age compared with their siblings Conditional on an elder sibling moving away from home, parents will likely keep younger siblings at home until they reach an appropriate age However, a very young sibling at home reduces the propensity to migrate, as the older sibling is likely to assume an important caretaking role Punch’s study nicely illustrates that the 114   International handbook on the economics of migration r­ elationship between siblings and migration is complex and will vary with sibling cohort characteristics 4.7  Agency Conventional wisdom suggests that parents decide whether a child should migrate to work or not Many reports on studies of domestics report that children are rarely consulted before they are sent to work, indicating no autonomy (for example, Brown, 2007) However, studies of street children show a great extent of child autonomy (for example, Conticini, 2005; Giani, 2006) In one of very few studies that focus on the autonomy of general child migration, Iversen (2002) finds autonomous behaviors among migrant children in his study in rural South India Boys outnumber girls and exhibit greater autonomy His finding is consistent with Keilland and Sanogo’s (2002) observations that girls migrate with their families and boys migrate with friends and other relatives in rural Burkina Faso Iversen finds that older children and children from higher caste families exhibit greater autonomy compared to other migrants He also finds that peer group autonomy greatly enhances a child’s migration decision, whereas household wealth and household social network does not Child abuse and neglect also cause children to behave autonomously and ‘run away’ from their homes In her study of children living in the streets of Bangladesh, Giani (2006) argues that abuse and neglect are primary reasons for children living in the streets Similarly, Conticini and Hulme (2007) argue that children move out to the streets because of excessive control and abuse at home and of gradual breakdown of trust within the households They emphasize the role of poverty in increasing stress and tension within the households This discussion of the causes of demand for agency among children is a good illustration of how important multiple factors can be in the decision to migrate No single factor can be the cause of a child’s migration and work status, because a child’s status depends on its payoff relative to all of the other possible solutions to the child’s problem With a wide variety of causes, there is then a wide variety of influences and policies that might impact a child’s migration status 5  The Impact of Independence There is little evidence on the impact of child independence outside of the literature on orphans The generalizability of evidence from orphans to the topic of this chapter is suspect The loss of both parents is a trauma whose impact may differ substantively from the consequences of independence We are not aware of any scientific study that identifies the impact of independent migrant child labor The problem in this literature is that identifying the impact of being an independent child migrant requires establishing the counterfactual of what the child would be doing without independence and migration Identifying this counterfactual requires knowledge of the child in her destination and at her source area in addition to quasi-­random variation in entry to independence The related research on the impact of independence that we could identify comes from the fostering literature Independent child labor migrants  ­115 Fostering is typically mentioned in the sub-­Saharan Africa context where children move between connected households for work, support and schooling Akresh (2008) examines the impact of child fostering on school enrollment At the core of his study is an impressive data collection effort that matched fostered children in their destination to their source families Akresh compares fostered children to non-­fostered children in the same location as well as fostered children to their biological siblings residing elsewhere Akresh documents that after fostering, young fostered children are more likely to be in school than either their hosts or siblings Fostered children themselves are less likely to be enrolled in school, but once they are fostered, their schooling increases substantively for children aged 5–7 years The opposite appears true for children aged 12–15 years, who attend school less after being fostered as well as before fostering Thus, for at least the youngest children in Burkina Faso, fostering seems to be important in helping them enroll in school While no other studies that we know of can compare fostered children to their hosts and siblings elsewhere, there are several cross-­sectional studies that document that fostered children receive schooling Zimmerman (2003) for example documents that fostered children in South Africa are more likely to attend school Similar evidence is in Beck et al (2011) for Senegal, although they point out that there is enormous heterogeneity in fostering situations The original models of fostering from Ainsworth (1996) focused on child labor demand as a determinant of the decision to foster in and Akresh (2009) documents poverty motives for sending children Beck et al emphasize that some children are fostered for work, some for school, some to protect the child’s food intake, and these different motives will have different implications for the impact of fostering Serra (2009) formalizes these ideas in a theoretical model of fostering with heterogeneous treatment effects Even with heterogeneous impacts of fostering, it is feasible to estimate the average consequences of fostering Coppoletta et al (2011) consider the long-­run consequences of fostering in Senegal by looking at the adult outcomes of individuals fostered in their youth They rely on self-­reports about whether an adult was fostered in youth The authors note that there are substantive swings in the prevalence of fostering across cohorts and years They argue that these large fluctuations in fostering rates imply that the unobserved characteristics associated with selection into fostering should average out across cohorts and years It appears that men who were fostered in their youth wind up with better education, job market outcomes and earlier marriage than men who were not fostered The long-­term consequences of fostering on average are less clear for women in Senegal They posit that women fostered in traditional ways marry early and are more apt to be in a polygamous union Less traditional cases of fostering may be associated with better education and reduced polygamy, although more research is necessary to understand selection into different types of fostering relationships The ability to send and receive child labor between households may also reduce distortions in human capital decisions Akresh and Edmonds (2011) argue that sibling influences on time allocation stem from labor market imperfections that families can overcome if fostering allows households to move child labor between residences In the study’s rural Burkina Faso setting, households are more comfortable sending children away when households can send or receive children within fostering networks Thus, 116   International handbook on the economics of migration the availability of fostering networks determines the ability of households to adjust ­composition When households can import and export child labor, the value of child labor in the fostering network determines the opportunity costs of schooling When households are constrained to use the labor to hand because of the opposite of fostering opportunities, Akresh and Edmonds find household composition influences school enrollment Thus, the ability to move children between families can moderate the impact of sibling composition on human capital accumulation There are good reasons why the evidence from fostering might extend to some types of independent child migration, including that associated with worst forms of child labor However, there are many reasons why fostering might be different than many types of child independence as there is explicitly an agent responsible for the child’s welfare in fostering exchanges That sense of responsibility might be important and could differentiate fostering exchanges that are for child labor from the typical domestic servant relationship In the end, we are left with very little sense about the consequences of independent child labor migration Anecdotes of child abuse and exploitation raise reasons for concern, but scientific evidence about the scope and scale of such abuse compared to the counterfactual for the child does not exist Even the fostering studies suffer from concerns about self-­selection and problems of omitted variables Perhaps the best hope for stronger evidence on the impacts of child independence come from future, yet to be conducted field experiments, where the treatment effects come from the treatment’s impact on child out-­migration 6  Conclusion The literature on independent child labor migrants is in its infancy This chapter documents major research needs in measurement, causes and consequences of independent child labor migration Studies of the prevalence of independent child labor migration typically use a roster method that identifies independent children by their current location The roster method suffers from a lack of knowledge of what the child’s environment was before coming to its current location Studies of the prevalence of independent children also sometimes use a fertility survey method that asks adult women or other household members about their children and compares the stated fertility history against the roster list of who is present This fertility method suffers from a lack of information on the absent child’s current environment Dedicated questions about out-­migrant children should be included in multi-­purpose surveys to help future research We not know whether employers view migrant child laborers differently from other child laborers Migrant child laborers may be easier to control, manipulate and exploit, but they also might value amenities that the employer can offer by virtue of his location or industry In some of the most common forms of migrant child labor such as in agriculture, we found very little evidence to suggest that employers view migrant child laborers differently than any other type of child labor Children become independent for a variety of reasons, and the literature strongly emphasizes that it is unrealistic to expect one cause of independence or to expect the Independent child labor migrants  ­117 cause to be consistent across space and time Based on anecdotes from the field, it seems that poverty and economic opportunities are two issues that must be central in any discussion of why children become independent child labor migrants Few studies consider the consequences of independence because of the data demands required to establish the counterfactual of what children would be doing without independence or migration The best evidence that exists on this topic comes from fostering studies that seem to establish some benefit to fostered children from fostering This general finding differs starkly from press accounts of horrific working conditions and abuse of child labor migrants and independent children Understanding the relationship between the circumstances of migration and the consequences of those migrations seems a priority for formulating and improving policy aimed at helping independent child migrants NOTE * We are grateful to an anonymous referee, Klaus F Zimmermann, and Amelie F Constant for helpful and constructive comments and suggestions Much of our thinking on this topic developed in the context of preparing Edmonds and Shrestha (2009), and we are grateful to Shahin Yaqub, Cinzia Tusco Bruschi, Eva Jespersen, David Parker, Anne Kielland, David McKenzie, Richard Akresh, Sylvie Lambert and Furio Rosati for helpful discussions and input We appreciate the research assistance of Dana Niu References Associates for Community and Population Research (ACPR) (2006), ‘Baseline survey on child domestic labour in Bangladesh’, International Labour Office, Dhaka, available at: http://www.ilo.org/ipecinfo/ product/download.do?type5document&id54647 (accessed 30 September 2008) Ainsworth, Martha (1996), ‘Economic aspects of child fostering in Cote d’Ivoire’, in T Paul Schultz (ed.), Research in Population Economics, vol 8, Greenwich: JAI Press, pp. 25–62 Akresh, Richard (2008), ‘School enrollment impacts of non-­traditional household structures’, BREAD Working Paper 89, Bureau for Research and Economic Analysis of Development (BREAD), Cambridge, MA Akresh, R (2009), ‘Flexibility in household structure: child fostering decisions in Burkina Faso’, Journal of Human Resources, 44 (4), 976–97 Akresh, Richard and Eric Edmonds (2011), ‘Residential rivalry and constraints on the availability of child labor’, NBER Working Paper No 17165, National Bureau of Economic Research (NBER), Combridge, MA Baas, Laura (2008), ‘Child labour in the sugarcane harvest in Bolivia’, The IREWOC Research Project on the Worst Forms of Child Labour in Latin America, Amsterdam: International Research on Working Children (IREWOC) Bastia, Tanja (2005), ‘Child trafficking or teenage migration? Bolivian migrants in Argentina’, International Migration, 43 (4), 58–89 Beck, Simon, Philippe de Vreyer, Sylvie Lambert, Karine Marazyan and Alba Safir (2011), ‘Child fostering in Senegal’, mimeo, Paris School of Economics, available at: http://www.parisschoolofeconomics.eu/docs/ lambert-­sylvie/confiage7-­1.pdf (accessed 17 January 2012) Brown, Eleanor (2007), ‘Out of sight, out of mind? Child domestic workers and patterns of trafficking in Cambodia’, International Organization for Migration, Geneva Camacho, Agnes (2006), ‘Children and migration: understanding the migration experiences of child domestic workers in the Philippines’, MA thesis, Institute of Social Studies, The Hague Caouette, Therese M (2001), ‘Small dreams beyond reach: the lives of migrant children and youth along the borders of China, Myanmar and Thailand’, a participatory action research project of Save the Children (UK), available at: http://www.savethechildren.org.uk/sites/default/files/docs/small_dreams_1.pdf (accessed 14 October 2008) 118   International handbook on the economics of migration Conticini, Alessandro (2005), ‘Urban livelihoods from children’s perspectives: protecting and promoting assets on the streets of Dhaka’, Environment and Urbanization, 17 (2), 69–81 Conticini, Alessandro and David Hulme (2007), ‘Escaping violence, seeking freedom: why children in Bangladesh migrate to the street’, Development and Change, 38 (2), 201–27 Coppoletta, Rosalinda, Philippe de Vreyer, Sylvie Lambert and Alba Safir (2011), ‘The long-­term impact of child fostering in Senegal: adults fostered in their childhood’, mimeo, Paris School of Economics, available at: http://www.parisschoolofeconomics.eu/docs/lambert-­sylvie/lotif_ver22.pdf (accessed 17 January 2012) Curran, Sara, Filiz Chung and Kanchana Tangchonlatip (2005), ‘Gendered migrant social capital: evidence from Thailand’, Social Forces, 84 (1), 225–55 Dávalos, Guillermo (2002), ‘Bolivia – child labour in sugarcane: a rapid assessment’, ILO/IPEC, Geneva, available at: www.ilo.org/ipecinfo/product/download.do?type5document&id5380 (accessed October 2008) De Lange, A (2007), ‘Child labour migration and trafficking in rural Burkina Faso’, International Migration, 45 (2), 147–67 Djebbari, Habiba and Helene Mayrand (2011), ‘Cash transfers and children’s living arrangements in South Africa’, mimeo, Laval University, Quebec, available at: http://www1.carleton.ca/economics/ccms/wp-­ content/ccms-­files/seminar-­paper-­110325.pdf (accessed 24 September 2008) Edmonds, Eric (2007), ‘Child labor’, in T Paul Schultz and John Strauss (eds), Handbook of Development Economics, vol 4, Amsterdam: North Holland, pp. 3607–710 Edmonds, Eric (2010a), ‘Trade, child labor, and schooling in poor countries’, in Guido Porto and Bernard Hoekman (eds), Trade Adjustment Costs in Developing Countries: Impacts, Determinants, and Policy Responses, Washington, DC: World Bank Press, pp. 179–96 Edmonds, Eric (2010b), ‘Selection into worst forms of child labor’, Research in Labor Economics, 31, 1–33 Edmonds, E and P Salinger (2008), ‘Economic influences on child migration decisions: evidence from Bihar and Uttar Pradesh’, Indian Growth and Development Review, (1), 32–56 Edmonds, E and N Schady (2012), ‘Poverty alleviation and child labor’, American Economic Journal: Economic Policy, (4), 100–124 Edmonds, Eric and Maheshwor Shrestha (2009), ‘Children’s work and independent child migration: a critical review’, Innocenti Working Paper No 2009-­19, UNICEF Innocenti Research Centre, Florence Edmonds, E., D Miller and K Mammen (2005), ‘Rearranging the family? Household composition responses to large pension receipts’, Journal of Human Resources, 40 (1), 186–207 ENCOVI (2000), ‘Encuesta de Condiciones de Vida (ENCOVI) 2000’, Instituto National de Estadisticas (INE), Guatemala Erulkar, Annabel, Tekle-­Ab Mekbib, Negussie Simie and Tsehai Gulema (2006), ‘Migration and vulnerability among adolescents in slum areas of Addis Ababa, Ethiopia’, Journal of Youth Studies, (3), 361–74 Ford, Kathleen and Victoria Hosegood (2005), ‘AIDS mortality and the mobility of children in KwaZulu Natal, South Africa’, Demography, 42 (4), 757–68 Foundation for Research on Educational Planning and Development (FREPD) (2003), ‘A baseline survey of street children in Bangladesh’, Bangladesh Bureau of Statistics, National Child-­Labour Survey, available at: http://www.ilo.org/ipecinfo/product/download.do?type5document&id5287 (accessed February 2012) Ghana Statistical Service (2003), ‘Ghana child labour survey’, available at: http://www.ilo.org/ipecinfo/ product/download.do?type5document&id5690 (accessed 24 September 2008) Giani, Laura (2006), ‘Migration and education: child migrants in Bangladesh’, Sussex Migration Working Paper 33, University of Sussex, Brighton Gurung, Yogendara Bahadur (2004), ‘Nature, extent and forms of child labour in Nepal’, Nepal Population Journal, 11 (10), available at: http//www.childmigration.net/Gurung_04 (accessed 26 September 2008) Halliday, Timothy J (2008), ‘Migration, risk and the intra-­household allocation of labor in El Salvador’, IZA Discussion Paper No 3322, Institute for the Study of Labor (IZA), Bonn Hasina, Sheikh (1989), Ora Tokai Keno, Dhaka: United Nations Children’s Fund Human Rights Watch (2011), ‘A poisonous mix: child labor, mercury, and artisanal gold mining in Mali’, Report, New York: Human Rights Watch International Institute of Tropical Agriculture (IITA) (2002), ‘Child labour in the cocoa sector of West Africa: a synthesis of findings in Cameroon, Côte d’Ivoire, Ghana, and Nigeria’, Sustainable Tree Crops (STCP)/ International Institute of Tropical Agriculture, Ibadan, Nigeria, available at: www.globalexchange.org/sites/ default/files/IITACocoaResearch.pdf (accessed 25 October 2008) International Labour Organization (ILO) (2006), ‘International Programme on the Elimination of Child Labour (IPEC)’, routes des Morillons, CH-­1211, Geneva International Programme on the Elimination of Child Labour (IPEC) (2003), Facts on Commercial Sexual Exploitation of Children, Geneva: International Labor Office, available at http://www.wotclef.org/ documents/fs_sexualexploit_0303.pdf (accessed 25 September 2008) Iversen, Vegard (2002), ‘Autonomy in child labor migrants’, World Development, 30 (5), 817–34 KC, Bal Kumar, G Subedi, Y.B Gurung and K.P Adhikari (2002), ‘Child labour in the Nepalese carpet Independent child labor migrants  ­119 sector: a rapid assessment’, Investigating the Worst Forms of Child Labour, Kathmandu, Nepal: International Labour Office, International Programme on the Elimination of Child Labour, available at: http:// www.childtrafficking.com/Docs/ilo_2002_child_labour_in_the_nepalese_carpet_sector_7.pdf, (accessed 25 September 2008) Kielland, Anne (2008), Child Labor Migration in Benin: Incentive, Constraint, or Agency? A Multinomial Logistic Regression, Saarbrücken: VDM Verlag Dr Müller Aktiengesellschaft & Co KG Kielland, Anne and Ibrahim Sanogo (2002), ‘Burkina Faso: child labor migration from rural areas – the magnitude and the determinants’, Analysis of Finding, with Incorporated Comments, from the Workshop of Interpretation and Validation in Ouagadougou, July 16–17, 2002, Washington, DC: World Bank, available at: http://home.online.no/~annekie/Africa_docs/BFEnglish.pdf (accessed 30 September 2008) Kifle, Abiy (2002), ‘Ethiopia: child domestic workers in Addis Ababa: a rapid assessment’, Investigating the Worst Forms of Child Labour No 38, International Labour Office, International Programme on the Elimination of Child Labour (IPEC), Geneva Kok, Jan (1997), ‘Youth labour migration and its family setting, the Netherlands 1850–1940’, The History of the Family, (4), 507–26 McKenzie, D (2007), ‘Paper walls are easier to tear down: passport costs and legal barriers to emigration’, World Development, 35 (11), 2026–39 McKenzie, David and Hillel Rapoport (2007), ‘Network effects and the dynamics of migration and inequality: theory and evidence from Mexico’, Journal of Development Economics, 84 (1), 1–24 Mueller, Eva (1984), ‘The value and allocation of time in rural Botswana’, Journal of Development Economics, 15 (1–3), 329–60 National Institute of Statistics (NIS) (2004), ‘Child domestic workers survey in Phnom Penh, survey report, Ministry of Planning, National Institute of Statistics/ILO, Phnom Penh, available at: http://www.ilo.org/ wcmsp5/groups/public/ -asia/ -ro-­bangkok/documents/publication/wcms_bk_pb_27_en.pdf (accessed 23 October 2008) Parish, William and Robert Willis (1993), ‘Daughters, education, and family budgets: Taiwan experiences’, Journal of Human Resources, 28 (4), 863–98 Pearson, Elaine, Sureeporn Punpuing, Aree Jampaklay, Sirinan Kittisuksathit and Aree Prohmmo (2006), ‘The Mekong challenge: underpaid, overworker, and overlooked: the realities of young migrant workers in Thailand’, International Labour Organization, International Program on the Elimination of Child Labour, Geneva Phlainoi, Nawarat (2002), ‘Thailand: child domestic workers: a rapid assessment’, Investigating the Worst Forms of Child Labour No 23, International Labour Office, International Programme on the Elimination of Child Labour, Geneva Punch, S (2002), ‘Youth transitions and interdependent adult-­child relations in rural Bolivia’, Journal of Rural Studies, 18 (2), 123–33 Quiroz, Carolina S (2008), ‘Importancia de Contar una Politica Migratoria Centroamericana’, MA thesis, Universidad de San Carlos de Guatemala, Escuela de Ciencia Politica Roe, Emery (1999), Except Africa: Remaking Development, Rethinking Power, New Brunswick, NJ: Transaction Rosenzweig, M.R and O Stark (1989), ‘Consumption smoothing, migration, and marriage: evidence from rural India’, Journal of Political Economy, 97 (4), 905–26 Serra, Renata (2009), ‘Child fostering in Africa: when labor and schooling motives may coexist’, Journal of Development Economics, 88, 157–70 Sharma, Shiva, M Thakurathi, K Sapkota, B Devkota and B Rimal (2001), ‘Nepal: situation of domestic child labourers in Kathmandu: a rapid assessment’, International Labour Office, International Programme on the Elimination of Child Labour, Geneva Srivastava, Ravi S (2005), ‘Bonded labor in India: its incidence and patterns’, Working Paper 43, International Labour Office, Special Action Programme to Combat Forced Labour, Geneva Statistics South Africa (2001), ‘Survey of activities of young people in South Africa 1999: country report on children’s work-­related activities’, Statistics South Africa and Department of Labour, Pretoria, and International Labour Office, Geneva, available at: http://www.ilo.org/ipecinfo/product/viewProduct.do? productId55013 (accessed 26 September 2008) Venkateswarlu, Davuluri (2007), ‘Child bondage continues in Indian cotton supply chain’, study jointly commissioned by OECD Wartch, Deutsche Welthungerhilfe (DWHH), India Committee of the Netherlands (ICN), Eine Welt Netz NRW (EWN NRW), International Labor Rights Forum (ILRF), available at: http:// www.indianet.nl/pdf/childbondagecotton.pdf (accessed 15 October 2008) Vogl, Tom (2011), ‘Sisters, schooling, and spousal search: evidence from South Asia’, mimeo, Princeton University, available at: http://www.princeton.edu/~tvogl/vogl_sisters.pdf (accessed 17 January 2012) Yang, Dean (2008), ‘International migration, remittances and household investment: evidence from Philippine migrants’ exchange rate shocks’, Economic Journal, 118 (528), 591–630 120   International handbook on the economics of migration Yaqub, Shahin (2009), ‘Independent child migrants in developing countries: unexplored links in migration and development’, Innocenti Working Paper No 2009-­01, UNICEF Innocenti Research Centre, Florence Young, Lorraine and Nicola Ansell (2003), ‘Young AIDS migrants in Southern Africa: policy implication for empowering children’, AIDS Care, 15 (3), 337–45 Zimmerman, F (2003), ‘Cinderella goes to school: the effects of child fostering on school enrollment in South Africa’, Journal of Human Resources, 38 (3), 557–90 6  Human smuggling* Guido Friebel and Sergei Guriev 1  INTRODUCTION While of increasing empirical importance, the phenomenon of human smuggling in illegal migration has only recently received notable attention in academics A search in scholar.google.com reveals that among the more than 20 000 hits for ‘human smuggling’ almost all stem from the late 1990s or later Economics has been quiet about smuggling and illegal migration until recently, while scholars in law, criminology, sociology, political science and demography have discovered the topic somewhat earlier (for example, Salt, 2000; Salt and Stein, 1997) Illegal migration, in contrast, has been an important topic in economics for quite a while, at least since Ethier’s (1986) analysis of the host country’s problem and Djajic’s (1987) two-­country model of illegal migration These models are very informative on a macrolevel and they point to important institutional determinants of illegal migration such as the structure of the host country’s labor market However, they not look at the micro structure of the market for migration, which is increasingly determined by the relationship between intermediaries that finance and organize illegal migration, and their ‘customers’, the potential migrants As human smuggling is a multibillion global business, it lends itself readily to microeconomic analysis.1 We will hence use the neutral language of microeconomics, although we are quite aware that illegal migration and human smuggling is a dangerous, violent and often humiliating business We discuss the small but growing microeconomic literature on the human smuggling business, and will show that the predictions about the effects of policies such as border controls, employer sanctions, deportation policies and amnesties depend quite crucially on whether or not the microstructure of illegal migration, that is, the relationship between migrants and intermediaries is considered In most of this chapter, we will maintain a positive perspective The normative aspects will be discussed later, albeit in a nonexhaustive way A first important step to integrating intermediaries in the economic analysis of illegal migration is to clearly identify the reason for their existence First, international migration is a costly activity, not only in terms of the risk for health and life involved, but also in terms of the financial means needed During the first wave of modern international migration during the seventeenth century, an estimated 60 percent of the migrants from the British Isles to the North American colonies overcame financing constraints by selling themselves into indentured servitude for a limited period of time A first striking difference to modern human smuggling is that in the times of indentured servitude, transportation costs were very high compared to wages Nowadays, transportation per se is cheap but the legal barriers to migrations are high This drives migration costs up and creates the business for intermediaries The second difference is that, at the time, indentured servitude was legal; people had the right to ‘sell’ themselves and the 121 122   International handbook on the economics of migration c­ ontracts between the ‘owner’ and the migrant were enforced Contemporaneous societies, however, are quite concerned about the moral and economic repercussions of such intermediaries, and servitude contracts are outlawed The second important reason for the existence of intermediaries in illicit migration are economies of scales in technology and access to networks Intermediaries have the know-­how to make migration happen, in terms of getting migrants from the home to the host country, potentially avoiding border controls and internal enforcement activities They provide housing, food and work, which although often of low quality, may not be available for illegal migrants who cannot operate in the legal sector Intermediaries have information that migrants may not have, they have access to employer and social networks, and can hide the migrants from law enforcement agencies seeking to deport the migrants Economies of scale in transportation and in the provision of work provide further rationales for the existence of intermediaries Because of its illegal character, the relationship between intermediaries and migrants is fraught with problems of potential abuse The contracts between migrants, smugglers and employers are not enforceable, therefore each party in these relationships is subject to the risk of opportunistic behavior of counterparties To overcome the contractual imperfections, smugglers often vertically integrated with employers of migrants However, as they cannot vertically integrate with the migrant, the moral hazard problems – both on the side of the intermediary and on the migrant’s side – remain important features of the relationship Intermediaries may not only have substantial bargaining power vis-­à-­vis the migrants, they may also behave in opportunistic ways ex post, for instance, by charging the migrants higher prices than agreed upon, or paying lower wages They may also be prone to ex ante opportunistic behavior The most extreme form of such ex ante opportunistic behavior is coercion, and the presence or absence of coercion can be used to define the limits between smuggling and trafficking While the use of smuggling involves that migrant and intermediary agree to enter a contractual relationship, trafficking means that the migrant is coerced into the relationship This becomes evident when one compares the definitions of smuggling, and trafficking provided by the UN Convention on transnational organized crime (United Nations Office on Drugs and Crime, 2004) Trafficking is here defined as: the recruitment, transportation, transfer, harboring or receipt of persons, by means of the threat or use of force or other forms of coercion, of abduction, of fraud, of deception, of the abuse of power or of a position of vulnerability or of the giving or receiving of payments or benefits to achieve the consent of a person having control over another person, for the purpose of exploitation (Article 3) Smuggling is ‘the procurement, in order to obtain, directly or indirectly, a financial or other material benefit, of the illegal entry of a person into a State Party of which the person is not a national or a permanent resident’ (Article 3) In most of this chapter, we look at smuggling, that is, we assume that migrants enter the relationship with intermediaries without being coerced, at least ex ante.2 We are aware that this distinction is a fuzzy one, and discuss this problem in the last section We here review the literature we are aware of, without claiming to be exhaustive, neither in terms of theory, nor in terms of empirical literature we may have overlooked Human smuggling  ­123 In what follows we suggests frameworks to think about many of the above issues, in terms of the reasons of existence, the potential for abuse and the consequences of the microstructure of illegal migration on the effect of polices We review the small number of available models to better understand the relationship between intermediaries and potential migrants We also highlight how different strategies of the receiving countries, such as border enforcement, internal enforcement and deportation policies are likely to affect the relationship, the division of rents and the outcomes in terms of quantities and types of equilibrium migration We conclude by discussing a list of challenges in terms of theory, empirical work and policy advice 2  THEORETICAL FRAMEWORKS FOR HUMAN SMUGGLING The model of Friebel and Guriev (2006) is the first to consider the microstructure of illegal migration It adds to the existing analyses of migration behavior by taking into account that migration involves costs that most migrants cannot finance themselves, and that the policies of host governments – by affecting the relationship between migrants and intermediaries – change migration outcomes both in terms of quantities and skill composition The effects of policies targeted against illegal migration in a model with a smuggling intermediary are indeed quite different from the ones of traditional models that not consider the microstructure of the market for illegal migration In Friebel and Guriev’s model, migrants need to enter a host country in which expected wages are higher than in the home country Intermediaries are integrated structures that provide the smuggling services, as well as financing and jobs A (potentially small) proportion of high-­skilled, wealthy migrants hire the intermediary to provide the smuggling services and pay them cash Low-­skilled potential migrants are wealth constrained.3 They can enter a contract with the intermediary in which the intermediary provides the smuggling services and finances the illegal migration The migrant in turn promises to pay back his debt through the labor income after successful migration, in the host country This model is inspired by the incomplete contracting literature in corporate finance (see Tirole, 2006), inasmuch as it assumes that, in a way similar to a wealth-­constrained entrepreneur vis-­à-­vis his investors, a potential migrant cannot commit himself to paying back the debt to the intermediary, rather than defaulting strategically Consequently, the main source of ex post opportunistic behavior or moral hazard is on the side of the migrant The model thus reflects the reality of modern illegal migration as the contract between the intermediary and the migrant cannot be enforced in the legal sector of the host country (in contrast to the indentured servitude contracts of the seventeenth century) However, as long as the migrant stays in the illegal sector of the host country, the intermediary can enforce the contract through the threat of violence, by excluding the migrant from a network, or by keeping the migrant in a safe house When the migrant exits the relationship by moving to the legal sector, he defaults on the repayment of the debt to the intermediary Upon doing so, however, he takes the risk of being deported to the home country, which occurs with probability D The larger D, that is, the stricter a country’s deportation and legalization policies, the smaller a migrant’s default probability 124   International handbook on the economics of migration Hence, a decrease in D makes it less likely for the intermediary to recover his investment, involving that less low-­skilled, wealth-­constrained migrants can be smuggled into the host country By the same token, a decrease in D also means that migrants who are not wealth constrained find it more interesting to migrate, because they take lower risks of being sent home when trying to transit to the legal sector The total quantitative effect on migration is ambiguous, but the skill composition of migrants increases when D decreases Increases in the strictness of border controls have more straightforward consequences: they reduce the net present value of all migrants, and thus decrease migration of both types of workers The model thus generates quite different testable predictions about the effects of border controls versus deportation policies In macromodels without wealth constraints and smuggling intermediaries, however, border controls and deportation policies have quite similar effects: both policies reduce the net present value of migration In Friebel and Guriev’s model, the prediction is that border controls decrease the size of migration flows, while deportation policies change the skill composition These predictions are testable, in a cross-­country perspective, because countries invest differently in border controls, and have different degrees of strictness with respect to deportation policies, and longitudinally, because spending on enforcement measures changes when a government with a different ideology comes into office The model also provides an economic rationale for a policy against migration that is tough with respect to border controls, but soft on deportation polices Not sending migrants home who default on the debt contracts by trying to move out from the illegal sector is not only appealing from a moral point of view, but would also reduce the inflow of migrants This setup also highlights an additional positive effect of amnesties and legalization of illegal migrants (besides the one highlighted in Chau, 2001, in which large amnesties can be used as a commitment device to enhance the credibility of employer sanctions) In Chau’s model, however, there are no intermediaries In Friebel and Guriev’s model, expected amnesties increase the risk of migrant’s default on the debt to intermediary Therefore, the smugglers are less interested ex ante in entering into the relationship with wealth-­constrained migrants; this, in turn, may reduce the overall migration flow The Friebel and Guriev model is highly stylized and does not take into account important elements of the market for illegal migration In particular, in the model, the risk of ex post opportunistic behavior is only present on the side of the migrant In Tamura (2010), migrants have no financial constraints Rather, the migrant can commit to pay the smuggler upon successfully arriving in the host country The source of moral hazard is thus the smuggler, and not the migrant Some smugglers may be able to force the migrants to work without pay, once the migrant has entered a relationship with the smuggler Tamura (2010) hence assumes rationality and symmetric information in the transaction just as Friebel and Guriev (2006): migrants choose whether or not engage with an intermediary under full awareness that they may have to work without pay if they contract with ‘exploitative’ smugglers Thus, exploitation is quite similar to ‘indentured servitude’ in Friebel and Guriev, but the source of such bonded labor is not contractual consent but ex post opportunistic behavior Ex ante coercion, however, is not considered The policy setup is very rich The government may raise the intensity of enforcement as measured by the probability of apprehension at the border, or within the host country Human smuggling  ­125 It may also vary the penalty for border apprehension and for illegal employment (in both cases the penalty is imposed both on the migrant and on the smuggler) Smugglers vary in terms of their capacity to exploit and choose whether to exploit the migrants depending on the government policies The migrants observe the smugglers’ capacity to exploit and therefore infer their risk to be exploited In this setting – as in Friebel and Guriev – migration policies affect not only the total flow of migration but also its composition and characteristics In particular, through tightening the policies along any of the six dimensions (probabilities of detention within the country and at the border; fines in case of detention at the border and within, for migrant and smuggler, respectively), the government reduces the level of exploitation as certain exploitative smugglers exit the market or become non-­exploitative This result is based on the important assumption that the government imposes penalties for exploitative employment but not for a nonexploitative employment of illegal immigrants Hence, the effect of stricter immigration policy is different for exploitative and nonexploitative smugglers Such an assumption seems realistic and potentially related to the political difficulty of penalizing nonexploitative employers of illegal immigrants Tamura’s paper also provides scope for empirical testing, however, mainly along the lines of the amount of exploitation Tamura’s theory highlights the need for analyzing the business of intermediaries and smugglers, and exploitative and non­exploitative employers of illegal immigrants separately Different policies affect different links in the ‘illegal immigration value chain’ For example, border apprehensions affect smugglers, while employer sanctions affect employers Given that contracting between smugglers and employers (as well as migrants) takes place outside the legal system, these markets are highly imperfect and therefore it is difficult for the parties to pass the burden of potential penalties and sanctions up and down the ‘value chain’ The results are somewhat different once the symmetric information assumption is dropped As in the ‘lemons market’ of Akerlof (1970), Tamura’s (2011) follow-­up model shows that if the capacity to exploit is the smuggler’s private information, in equilibrium there is no way for ‘nonexploitative’ smugglers to signal their type Therefore, such smugglers are driven out of the market and only ‘exploitative’ smugglers prevail Here the policy implications become more nuanced Stricter enforcement reduces exploitation but increases the overall flow of illegal immigrants (as the nonexploitive smugglers re-­enter the market) Hence, Tamura (2011) allows testing along the dimensions of ­migration flows and thus complements Friebel and Guriev Auriol and Mesnard (2012) dig deeper into the industrial organization of the market for human smuggling They abstract from financing constraints; the intermediary is only providing the smuggling services required for migration The smuggling market is assumed to be concentrated, because of the legal restrictions to entry (other reasons of concentration could be limited access to transport technologies or to networks in the host country) The assumption about market concentration is not only reasonable, it also gives rise to interesting, novel effects In particular, if the smuggler has some market power, policies that raise the costs of the smuggling tend to increase the exploitation of migrants through the price that migrants have to pay A government can try to eliminate the smuggling market by providing visas to sell to migrants While this undermines the market power of the smuggling network, it increases the flow of migrants and reduces the 126   International handbook on the economics of migration skill composition The latter effect stems from the fact that higher-­skilled migrants have higher willingness to pay Hence, a decrease in prices raises the flows and attracts fewer skilled migrants The prediction concerning skill composition is similar to that in Friebel and Guriev, in which higher deportation policies have the same effects Auriol and Mesnard, however, have a richer policy space Hence, in the presence of the policy tradeoff ‘exploitation of migrants versus migration control (in terms of quantities and qualities)’ they can treat the question whether a combination of instruments can improve the situation Auriol and Mesnard show that it is indeed possible to improve on both dimensions of the objective function by using the funds raised by the sales of visa for investment into border controls and employers sanctions, thus, raiding the profits of smuggling networks Auriol and Mesnard also provide calibrations to show that the prices for visa sales needed to drive smugglers of Chinese migrants out of business would be up to 47 000 dollars Auriol and Mesnard’s (2012) paper does not only go a step further to better understand the microstructure of the market for smuggling, it also raises an interesting debate about the welfare implications of different policies, a topic we will get back to in the next section Djajic and Vinogradova (2011b) provide another complementary analysis They consider a situation of wealth-­constrained migration; their interest, however, is not so much the strategic interaction between intermediaries and migrants, but endogenizing the saving decision of migrants Migrants decide when to leave and whether or not to use a loan by the intermediary, maximizing the intertemporal utility stream of their lifetime If a migrant saves for a longer period of time, he or she can migrate without entering a debt/ labor contract, but if migration is desired to be earlier, bonded labor is the only option A number of noteworthy results originate from the model First, it is optimal for migrants to enter debt/labor contracts when migration costs are low and foreign wages are high In the case of higher migration costs or lower foreign wages, it becomes preferable to save longer and finance migration by proper funds Second, in terms of the policy prediction, tougher border controls reduce the incidence of debt/labor contracts, which is in contrast to Friebel and Guriev (2006) in which tougher border controls, by increasing the costs of migration, push more people into entering debt/labor contracts Third, employer sanctions are likely to reduce the incidence of debt/labor contracts Djajic and Vinogradova (2011b) and their companion paper (Djajic and Vinogradova, 2011a) in which the duration of the servitude period during which migrants pay back their debt is endogenized, have interesting normative implications: when saving and the payback period are endogenized, bonded labor is not necessarily a destiny of wealth-­ constrained migrants, but may be a choice variable similar to the one of many of the migrants in the first wave of indentured servitude during the seventeenth century This posits intricate welfare issues that we discuss in the next section 3  CHALLENGES The theoretical literature has looked at integrated (Friebel and Guriev, 2006) and nonintegrated intermediaries (Tamura, 2010) who may or may not have market power (Auriol and Mesnard, 2012) Moral hazard may be present on the side of the migrant (Friebel and Guriev, 2006) or the intermediary (Tamura, 2010, 2011) Also, the migrant’s capac- Human smuggling  ­127 ity to pay the smuggling services upfront can be endogenized (Djajic and Vinogradova, 2011a, 2011b) The literature thus captures a broad range of institutional settings and decision problems of the relevant parties The theories also look at a broad range of policy instruments, spanning from border controls, to deportation and legalization policies, employer sanctions and the sales of visa Many of the issues of human smuggling still remain open We believe that the applicability of human smuggling theories for policy advice could be largely increased by tackling the following challenges: (1) building richer theories that take into account the upstream and downstream factors that are relevant for human smuggling, and make it possible to adapt theories to the specificities of the respective home and host countries; (2) enhancing the theoretical and empirical knowledge about the parameters relevant for human smuggling activities; (3) creating better distinctions between smuggling and trafficking and generating more reliable data, and (4) clarifying welfare issues and policy coordination Richer Theories on Illegal Migration and Human Smuggling Current theories not model the labor market for illegal migrants in the host countries and, in this respect, fall back behind Ethier’s (1986) analysis that does incorporate institutional features of the host country’s labor market A more recent exception is Epstein and Heizler (2008) who investigate the effects of minimum wages in the host country on stocks of illegal migrants They show that there is a positive relationship between the two In the case of human smuggling, it is evident that the liquidity of the labor market, the productivity of the migrants and the distribution of bargaining power during and after the bondage period is crucial for wage determination In turn, the wages earned during and after bondage determine the incentives of migrants to use intermediaries, the contract terms and the composition of migrants Theories of the labor market for illegal immigrants and of entrepreneurship among migrants are an important building block for a better understanding of illicit migration with the help of intermediaries In particular, if membership in a social network is important for employment after bondage, the incentives to default on the intermediary are reduced This may explain why long-­haul migration from China has been maintained for quite a while: networks help enforcing contracts between migrants and intermediaries as the migrants seem to fear exclusion from the networks Ethnicity may, however, also increase the possibility of default on implicit agreements, when links are weak, the ethnic minority is in itself heterogeneous or too small to maintain an economic identity Related topics are further developed in Chapter Another theory gap consists in the household reactions to policies undertaken by home countries Particularly important are information campaigns, which so far have not been considered by the existing theories of human smuggling This is somewhat logical, because these theories usually build on symmetric information between migrant and smuggler A notable theoretical exception is Tamura (2011) Mahmoud and Trebesch (2010), in their empirical analysis, show that information asymmetries exist, and argue for information campaigns to reduce trafficking Furthermore, Dula et al (2006) investigate to what extent it is optimal to shift some of the burden of policies against illegal immigration to home countries 128   International handbook on the economics of migration Understanding the nature of the labor market for illegal immigrants is also important for estimating properly the long-­term costs of the illegal status of the migrants If illicit jobs not provide opportunities for accumulating human capital, migrants would fall further behind their counterparts in the legal sector of the host country, but possibly also in the home country; this issue is discussed by Kossoudji and Cobb-­Clark (2002) Both broader and deeper is the question of how the demand for illegal migration and smuggling services is affected by economic development and the resulting change in wage distribution, risk aversion and expectations In particular, whether or not workers are willing to subject themselves to the immense hardship and risks associated with illicit migration will depend on the standard of living in the home country and on norms Parameters Relevant for the Human Smuggling Market In the absence of large databases allowing econometric tests of the theories and their implications, calibrations of the theories seem the right approach Indeed, some of the theories have been complemented by calibrations to check for the internal consistency of the theory (as in the case of Friebel and Guriev, 2006) or to generate implications (as in the case of Auriol and Mesnard, 2012) Auriol and Mesnard gather a host of available data to determine the critical visa process that would eliminate the market for smuggling They also highlight the sensitivity of these prices to other policy variables, for instance, deportation policies But even for the relatively limited requirements involved for calibrations, the data situation is far from satisfactory, the need for better data is evident for trafficking (Laczko, 2002) and smuggling alike While information about short-­haul migration and human smuggling, in particular, between Mexico and the US is of quite high quality (for example, Gathmann, 2008; Hanson et al., 2002), data from long-­haul migration is much harder to come by Exceptions include some sociological work about Chinese migrants that is based on hundreds of interviews and provides quite detailed information (Chin, 1999) Other interesting sources are the New Immigrant Survey (Jasso et al., 2000)4 and the survey on irregular migration in Italy by Coniglio et al (2006) There are other sources that, while far from perfect, could provide useful information for calibrations Examples from reports of governmental and nongovernmental organizations (NGOs) include United Nations Office on Drugs and Crime (UNODC), Clandestino, case studies and reports by the International Organization for Migration (IOM), and even media reports, the latter when used with care These figures stem from quite different sources, and they are generated by different techniques Thus it would be very useful to review them systematically in order to generate a matching of different pairs of home and host countries with the respective information about wealth constraints and financing sources of migrants, risk aversion (for a recent theoretical contribution, see Woodland and Yoshida, 2006), information available for migrants, prices for smuggling service, physical risks involved with travel­ ing and during the (potential) periods of bonded labor, deportation and legalization probabilities in the host countries, work place inspections, penalties for employers, and so on Similarly, it would be useful to know more about the different employment characteristics in host countries of migrants coming from different countries.5 Human smuggling  ­129 Distinctions between Smuggling and Trafficking Throughout the chapter we have used a clear distinction between smuggling and trafficking However, the terms have been used interchangeably by some researchers and practitioners A lack of consensus on the use of the terms complicates the analysis of these activities (see Salt and Hogarth, in Laczko and Thompson, 2000, pp. 18–23) However, recent efforts to create legal instruments to fight against human smuggling and trafficking have helped in providing a much clearer distinction between these activities The widespread definition of smuggling versus trafficking that is based on whether or not migrants are coerced into a relationship or not, is useful for building theoretical frameworks However, Tamura’s (2010) analysis shows that the frontiers between the concepts are not that clear cut Even when migrants are not coerced to enter the relationship, they may nonetheless be exploited ex post Moreover, they may be aware of this ex ante and may still enter the relationship Mahmoud and Trebesch (2010, p. 174) use as a definition of trafficking the presence of coercion ex ante or ex post and argue that ‘high emigration areas are often disadvantaged  .  so that migrants  .  may be more willing to take risks, and possibly even consent to hazardous working conditions abroad’, which, again, shows that the distinction between the concepts is quite fuzzy The problem is far-­reaching not only because of the lack of conceptual clarity, but also because the data used often cannot distinguish perfectly whether a person is trafficked or smuggled We see two sources of the problem Traditionally, any migrant who was smuggled or trafficked was seen as a victim Only recently, we have got used to make the distinction necessary to investigate illicit migration with intermediaries Policy makers and NGOs now increasingly see the difference, and it can be expected that more care will be given interviewing migrants to identify clearer smuggled from trafficked migrants A second, potentially harder to correct problem lies in the incentives of the migrants interviewed As most migrants would like to stay in a host country, they have no incentives to reveal that they entered the relationship willingly, in particular, as victim protection programs and leniency programs are designed for victims of trafficking, not for the clients of smuggler intermediaries Sampling techniques should be developed that allow more detailed investigations about the precise circumstances of a migrant’s coming to a host country In particular, retrospective questions, to legalized migrants who need not fear deportation any more, could help to find out more about whether or not migration decisions involved ex post and/or ex ante coercion, while any migrant expecting to be deported if no coercion was involved would have strong incentives not to reveal the truth Besides retrospective questions, researchers could ask migrants whether or not they had access to people who had migrated or, even easier, to the Internet It seems doubtful that anyone who has access to the Internet would not have known about some of the risks involved in migration Notice that we are not advocating a different treatment of smuggled or trafficked migrants, but argue that identifying different types of migrants is crucial for research, and policy advice There is no easy solution for the problem of distinguishing smuggled from trafficked migrants, but it has serious implications While the technology of trafficking and smuggling is quite similar, the political and moral implications are very different Illicit immigration has features that distinguish it from the traditional cost-­benefits considerations that determine migration decisions without intermediaries, for instance in the Harris 130   International handbook on the economics of migration and Todaro (1970) framework This entails kinship and friendship networks, and political dimensions, like the vulnerability to conflicts and internal displacements Taking these dimensions into account introduce a new set of challenges to empirical work It requires more care in designing questionnaires and in qualifying people into smuggled versus trafficked migrants, understanding their motives, expectations, strategies and constraints Intricate Welfare Issues, and Policy Coordination Most of the papers we have discussed are purely positive They investigate how increases or decreases in policies such as border enforcement, employer sanctions, deportation polices or amnesties affect the optimizing behavior of the parties involved and thus the outcomes of migration The absence of normative analyses may be owing to the fact that positing a welfare function of a government creates three types of problems First, the objectives of ‘planners’ may be manifold and, quite often, contradicting For a broader view on the topic, see Chapter 22 on the political economy of migration in this chapter In respect of our topic of human smuggling, it seems that contradictions in objectives and policies are not always clear to policy makers The analysis of Auriol and Mesnard (2012) for instance shows that the goal of reducing migrant streams and eradicating the smuggling intermediaries contradict each other, which only comes to the fore when the market for smuggling is modeled explicitly Similarly, Friebel and Guriev (2006) point to tradeoffs between the quality and the quantity of illegal migrants that are far from evident Political economy considerations may also come into play Conservative voters tend to like a tough stance on illegal migrants, but entrepreneurs, often also constituents of conservative parties, tend to like the idea of cheap labor and labor market competition; further examples can be found in Chapter Hence, governments may take contradictory policy measures, for instance, tough border controls but lax employer sanctions, to please both types of constituents at the same time The second obstacle to the formulation of normative analyses stems from the fact that policy goals and measures are often formulated on different levels This has horizontal aspects: different federal agencies in a given country may be unable or unwilling to coordinate In the context of the European Union (EU), Mayr et al (2012) have investigated the problems of coordinating on a joint immigration policy While the border states like Italy must bear much of the costs of border enforcement, they are usually not the final destination for many smuggled migrants, who may migrate further to the more attractive labor markets in, say, Germany or the UK Consequently, the core countries of the EU may not agree on the desirability of legalization campaign of the border countries, and the willingness to invest in border enforcements may be lower in the border countries than at the core At the EU level, there are initiatives to take care of this conflict of interest, for instance, through FRONTEX.6 So far, the results of this type of coordination effort seem to be less than satisfactory The third point that makes a welfare analysis quite intricate are the moral tradeoffs involved Unless one takes the arguably quite extreme position that only the welfare of citizens should be considered, moral tradeoffs emerge necessarily when a country tries to protect its borders Unless perfect enforcement of borders is available, such that illegal Human smuggling  ­131 migration ceases to exist altogether, any policy involves tradeoffs with respect to smuggled and trafficked individuals These individuals are not only forced into humiliating conditions through the need to stay in the shadows, but many of them are deported, often into quite dangerous condition in the home country (see Chapter 24 on refugee migration) It could seem morally more acceptable to deport back smuggled people who knew about the risks and took a deliberate decision, rather than trafficked people But whether or not these can be discerned is a topic discussed above This short discussion is meant to suggest that a consensus as to what is morally sound behavior with respect to illegal migrants is very much needed, but we are far from it Thus, government policies against illegal migration are usually of a short-­term and partial nature, reacting to pressure by various interest groups, beliefs or the media While this justifies the positive perspective in many papers, and in-­depth political economy analyses, it makes policy advice difficult because of theories with limited scope, lacking knowledge about important parameters and fuzzy definitions of the type of migrants considered 4  CONCLUDING REMARKS To date, the research on human smuggling consists mainly of a small number of theoretical papers While there is some scope for improvement of the theories, the main challenges seem to lie in improving the empirical knowledge, in order to be able to adjust theories to the respectively relevant condition of home and host country pairs A host of empirical information is available and awaits systematic use, but additional efforts should be undertaken, by economists, legal scholars, sociologists and anthropologists alike to generate information that can be used for calibrations of the existing and new theories In order to have reliable information as a basis of policy advice, new sampling techniques should be developed and care should be given to distinguish smuggling from trafficking While smuggling and trafficking may look quite similar at first sight, both the mechanisms and the welfare implications are quite different We also argue that a broad discussion about welfare definitions is needed to develop a normative analysis of the problem NOTES * We thank the editors, Amelie F Constant and Klaus F Zimmermann, an anonymous referee, Nancy Chau and Alice Mesnard for useful comments and inputs All errors are ours For an early attempt to sketch the economics of human smuggling and trafficking, see Schloenhardt (1999) In the paper by Tamura (2010), discussed below, coercion and exploitation occurs, but only ex post Notice that the model assumes perfect correlation between skills and wealth, an assumption that could be relaxed The New Immigrant Survey is a survey of new legal immigrants But the data set also contains migrants retrospective answers about the mode of entry in the 1990s In the context of trafficking, Akee et al (2011) have matched the existing data on trafficked migrants between countries and have tested a gravity model FRONTEX is the EU’s agency that coordinates membership states’ efforts to protect external borders 132   International handbook on the economics of migration REFERENCES Akee, Randall, Arjun S Bedi, Arnab K Basu and Nancy H Chau (2011), ‘Transnational trafficking, law enforcement and victim protection: a middleman trafficker’s perspective’, IZA Discussion Paper No 6226, Institute for the Study of Labor (IZA), Bonn Akerlof, G (1970), ‘The market for “Lemons”: quality uncertainty and the market mechanism’, Quarterly Journal of Economics, 84 (3), 488–500 Auriol, Emmanuelle and Alice Mesnard (2012), ‘Sale of visas: a smuggler’s final song?’, NORFACE MIGRATION Discussion Paper No 2012-­07, NORFACE Research Programme on Migration, University College London Chau, Nancy H (2001), ‘Strategic amnesty and credible immigration reform’ Journal of Labor Economics, 19 (3), 604–34 Chin, Ko-­lin (1999), Smuggled Chinese: Clandestine Immigration to the United States, Philadelphia, PA: Temple University Press Coniglio, Nicola D., Giuseppe De Arcangelis and Laura Serlenga (2006), ‘Intentions to return of undocumented migrants: illegality as a cause of skill waste’, IZA Discussion Paper No 2356, Institute for the Study of Labor (IZA), Bonn Djajic, S (1987), ‘Illegal aliens, unemployment and immigration policy’, Journal of Development Economics, 25 (1), 235–49 Djajic, Slobodan and Alexandra Vinogradova (2011a), ‘Migrants in debt’, unpublished paper, The Graduate Institute, Geneva Djajic, Slobodan and Alexandra Vinogradova (2011b), ‘Immigration policy and debt-­bonded migration’, unpublished paper, The Graduate Institute, Geneva Dula, G., N Kahana and T Lecker (2006), ‘How to partly bounce back: the struggle against illegal immigration to the source countries’, Journal of Population Economics, 19 (2), 315–25 Epstein, G and O Heizler (2008), ‘Illegal migration, enforcement and minimum wage’, Research in Labor Economics, 28, 197–224 Ethier, W.J (1986), ‘Illegal immigration: The host-­country problem’, American Economic Review, 76 (1), 56–71 Friebel, G and S Guriev (2006), ‘Smuggling humans: a theory of debt-­financed migration’, Journal of the European Economic Association, (6), 1085–111 Gathmann, C (2008), ‘Effects of enforcement on illegal markets: evidence from migrant smuggling at the southwestern border’, Journal of Public Economics, 92 (10–11), 1926–41 Hanson, G.H., R Robertson and A Spilimbergo (2002), ‘Does border enforcement protect US workers from illegal immigration?’, Review of Economics and Statistics, 84 (1), 73–92 Harris, J and M Todaro (1970), ‘Migration, unemployment and development: a two-­sector analysis’, American Economic Review, 60 (1), 126–42 Jasso, G., D.S Massey, M.R Rosenzweig and J.P Smith (2000), ‘The New Immigrant Survey Pilot (NISP): overview and new findings about U.S legal immigrants at admission’, Demography, 37 (1), 127–38 Kossoudji, S and D Cobb-­Clark (2002), ‘Coming out of the shadows: learning about legal status and wages from the legalized population’, Journal of Labor Economics, 20 (3), 598–628 Laczko, F (2002), ‘Human trafficking: the need for better data’, International Organization for Migration, available at: www.migrationinformation.org/Feature/display.cfm?ID566 (accessed 22 January 2012) Laczko, Frank and David Thompson (eds) (2000), Migrant Trafficking and Human Smuggling in Europe: A Review of the Evidence with Case Studies from Hungary, Poland and Ukraine, Geneva: International Organization for Migration Mahmoud, O.T and C Trebesch (2010), ‘The economics of human trafficking and labour migration: micro-­ evidence from Eastern Europe’, Journal of Comparative Economics, 38 (2), 173–88 Mayr, K., S Minter and T Krieger (2012), ‘Policies on illegal immigration in a federation’, Regional Science and Urban Economics, 42 (1–2), 153–65 Salt, J (2000), ‘Trafficking and human smuggling: a European perspective’, International Migration, 38 (3), 31–56 Salt, J and Stein, J (1997), ‘Migration as a business: the case of trafficking’, International Migration, 35 (4), 467–94 Schloenhardt, A (1999), ‘Organised crime and the business of migrant trafficking: an economic analysis’, Crime, Law and Social Change, 32 (3), 203–33 Tamura, Y (2010), ‘Migrant smuggling’, Journal of Public Economics, 94 (7–8), 540–48 Tamura, Yuji (2011), ‘Illegal migration, people smuggling, and migrant exploitation’, Warwick Economics Research Paper 791(revised version), University of Warwick, Department of Economics Tirole, J (2006), The Theory of Corporate Finance, Princeton NJ: Princeton University Press Human smuggling  ­133 United Nations Office on Drugs and Crime (2004), United Nations Convention against Transnational Organized Crime and the Protocols Thereto, Vienna: United Nations Office on Drugs and Crime Woodland, A and C Yoshida (2006), ‘Risk preferences, immigration policy, and illegal immigration’, Journal of Development Economics, 81 (2), 500–513 PART III PERFORMANCE AND THE LABOR MARKET 7  Labor mobility in an enlarged European Union* Martin Kahanec Daughter: ‘What’s over there, Mom?’ Mother: ‘There is nothing there – there is the East Bloc.’ (Conversation between a mother and her daughter pointing in the direction of Slovakia on a hilltop in Austria near the Czechoslovak border sometime in 1987) 1  INTRODUCTION The freedom of movement of workers is one of the four fundamental pillars of economic integration in the European Union (EU), which also includes the free mobility of capital, goods and services A central objective of free mobility is to enable EU citizens to seek employment, and any social benefits attached with it, in any of the EU member states From the economic perspective free labor mobility improves the allocative efficiency of EU labor markets, thus buttressing the EU’s economy and alleviating some of its demographic challenges (Kahanec and Zimmermann, 2010; Zimmermann, 2005) Yet, with the process of EU enlargement expanding this freedom to new member states, free labor mobility constitutes one of the most sensitive, and often challenged, freedoms in the EU The controversies surrounding the freedom of movement of labor culminated when Cyprus, the Czech Republic, Estonia, Latvia, Lithuania, Malta, Hungary, Poland, Slovakia, and Slovenia, referred to as the EU10, joined the EU in 2004 and carried on in 2007 when Bulgaria and Romania, the EU2, followed suit.1 These controversies were probably rooted in the history of deep political, economic and social separation during the Cold War This separation had severely limited mobility and contact across the East– West limits and resulted in a fissure in the European identity along the Iron Curtain Presumably the economic disparities between the new and old member states, combined with the large scale of these enlargements, created grounds for a widespread perception in the EU15 of EU812 migrants as a threat to their labor markets and welfare systems, and explain the magnitude of such controversies at least partly.2 As a consequence, a policy instrument – transitional arrangements – was adopted allowing member states to keep their labor markets closed for citizens from new member states for up to seven years, with revisions required after two and five years, following their accession.3 In the EU812 free mobility was seen as a way out of the difficulties stemming from labor market mismatches and excess labor supply inherited from the process of their difficult post-­socialist transformation In spite of some fears of brain drain, overall, the expectations of faster convergence to the living standards of the old member states following their accession framed the expectations in the acceding countries quite favorably towards this process This chapter reviews what we know about labor mobility in the EU following the two waves of enlargement in 2004 and 2007 We in particular evaluate the experience with 137 138   International handbook on the economics of migration post-­enlargement migration in an enlarged EU in view of the fears and hopes attached to it in the sending and receiving countries, and by the migrants themselves The next section provides a theoretical account of possible effects of free mobility in sending and receiving labor markets We then describe how enlargement affected labor mobility in the EU, and what measurable effects can be documented empirically In the subsequent section we shed some light on what migration flows can be expected in the near future Finally, we discuss a number of lessons that can be learned, and conclude 2  A THEORETICAL ACCOUNT Does economic theory predict any significant migration flows in an enlarged EU that could justify the controversies pervading EU enlargement? Harris and Todaro (1970) point to the significance of (expected) regional disparities in the standard of living for the migration decision More generally, international disparities in the levels (and distribution) of earnings and income, net of migration costs, chances to pursue a rewarding career and avoid unemployment, the cost of living, or the availability and quality of public goods and amenities are proposed in the literature as key drivers of migration (Bonin et al., 2008; Borjas, 1999; Massey, 1990).4 Others, such as the generosity of the welfare system are more debatable (De Giorgi and Pellizzari, 2009; Giulietti et al., 2013) Kureková (2011) stresses the importance of skill-­mismatches and their interaction with the welfare state Stark (1991) advanced the view that for the household as a decision-­ making unit it may be worthwhile to have one or more of its members abroad as a strategy of risk sharing.5 Factors such as those listed above may affect various subpopulations differently The costs of migration and adjustment in the host economy, pecuniary and nonpecuniary, depend on the geographical, linguistic and cultural distances between (subpopulations in) the sending and receiving countries (Chiswick and Miller, 2012) The human capital  theory predicts that the migration decision also depends on age and skills of potential migrants, as these determine their capacity to adjust in the host country and thus benefit from migration (Becker, 1957; Sjaastad, 1962) As a result, people who decide to migrate and stay in the receiving country may be positively or negatively self-­selected based on their observable or unobservable characteristics (Borjas, 1987; Chiswick, 1999) Based on these arguments and given the initial disparities in many socio-­economic variables, the expectations of nonnegligible, and possibly uneven, migration rates between new and old member states were probably justified A key question then is whether such flows could negatively affect sending or receiving countries The impact of migration can be studied considering a sending and a receiving country, each with separate labor markets for high- and low-­skilled labor, and drawing on the idea that the redistributive effects of migration depend on the degree of substitutability or complementarity of migrant and nonmigrant (native or staying) labor (Chiswick, 1980, 1998; Chiswick et al., 1992) Then, in case of high-­skilled post-­enlargement migration, the winners of enlargement would be low-­skilled workers in the receiving countries, benefiting from increased demand for their labor as a consequence of the complementarity of low- and high-­skilled workers In the sending countries, the staying high-­skilled workers could Labor mobility in an enlarged European Union  ­139 also benefit from their increased scarcity in the labor market High-­skilled workers in the receiving countries could be among the losers of enlargement, but not if the increased demand for low-­skilled labor resulted in their higher employment and thus an increased productivity of high-­skilled workers in spite of their increased absolute abundance Low-­ skilled workers in the sending countries would lose under this scenario owing to lowered demand for their labor One can similarly track the redistributive effects of low-­skilled migration in such models.6 While this framework elucidates some potential redistributive effects of post-­ enlargement migration, many other important factors may considerably change or even reverse some of its predictions For example, economic migration can be expected to improve the allocation of labor and human capital Moreover, migration proliferates cross-­regional and cross-­border social ties, thus acting as a vehicle for international flows of goods and services, capital, as well as ideas and knowledge (Bonin et al., 2008) The resulting improved productivity may benefit all types of labor in sending as well as receiving countries Further economic benefits may result from increased ethnic diversity in receiving countries (Ottaviano and Peri, 2006) However, a range of psychological, linguistic, institutional or legislative barriers, as well as discrimination, may impede immigrant adjustment in the host society, thus hindering some of the positive effects migration may entail (Constant et al., 2009; Kahanec et al., 2013) Such barriers may, for example, result in weaker labor market outcomes and, as a consequence, an increase in migrants’ demand for welfare (Borjas, 1999; Brücker et al., 2002; Kahanec et al., 2013; Zimmermann et al., 2012) Another discrepancy may arise if skills are not perfectly transferable from sending to receiving countries and workers work in jobs below their level of qualification (downskilling).7 Ethnic identity is another factor that may positively or negatively affect adjustment in host labor markets (Chapter 14 in this volume) This theoretical account of migration illustrates that the scale and (properly measured) skill composition of post-­enlargement migration are particularly important for the evaluation of its effects in sending and receiving economies For the receiving countries the degree and speed of adjustment of immigrants is another important variable It also shows that without strong assumptions the effects of enlargement are hard to evaluate unequivocally based on theoretical grounds Similarly, any a priori fears of enlargement are hard to justify theoretically 3 THE SCALE AND COMPOSITION OF POST-­ENLARGEMENT MIGRATION Given the scarcity of migration data, to evaluate the scale and composition of post-­ enlargement migration is a formidable task We therefore look at various data sources and the available literature to triangulate some of the most important trends According to the data provided in Holland et al (2011), in 2004 there were about a million citizens from the EU8, and almost another million EU2 nationals, residing in the EU15.8 By 2009, just five years later, the total number of EU8 and EU2 citizens residing in the EU15 increased by about 150 percent and reached almost five million (Table 7.1) In effect, the combined populations of citizens from EU8 and EU2 countries residing in the 140   International handbook on the economics of migration Table 7.1  Citizens from new EU member states residing in the EU15 Migrants from the EU8 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Migrants from the EU2 Total Percentage of EU8 population Percentage of EU15 population Total Percentage of EU2 population Percentage of EU15 population 673 324 674 972 717 976 753 056 800 534 851 250 942 321 1 006 851 1 235 429 1 627 625 2 027 651 2 252 681 2 288 600 0.91 0.91 0.97 1.02 1.09 1.16 1.29 1.38 1.69 2.23 2.78 3.09 3.13 0.18 0.18 0.19 0.20 0.21 0.22 0.25 0.26 0.32 0.42 0.52 0.57 0.58 249 781 234 743 271 657 315 699 391 045 509 160 711 930 916 298 1 109 570 1 376 956 1 971 968 2 348 523 2 564 008 0.81 0.76 0.88 1.03 1.28 1.71 2.40 3.10 3.77 4.69 6.74 8.05 8.81 0.07 0.06 0.07 0.08 0.10 0.13 0.19 0.24 0.29 0.35 0.50 0.60 0.65 Source:  Based on data provided in Holland et al (2011), Eurostat population statistics, and own calculations EU15 constituted 1.2 percent of the total EU15 population and 4.8 percent of combined populations of EU8 and EU2 countries.9 Whereas over the five-­year period preceding 2004 the average annual inflow to the EU15 from the EU8 was about 58 000, in the five years after 2004 this has risen to 256 000 annually The corresponding figures for the EU2 were 129 000 and 330 000, respectively We observe an increasing dynamic of inflows from new to old member states until 2007, followed by a significant slow down during the financial crisis in 2008 and 2009 EU8 citizens reacted to enlargement with some delay, with peak migration attained only in 2006 and 2007, two years after their accession The response of EU2 citizens was considerably swifter and more pronounced, reaching peak migration flows already in the year of enlargement.10 The slow down of 2008 and 2009 indicates that the worsened economic prospects in some of the receiving countries may have discouraged potential migrants The most important sending countries are Romania and Poland, which in 2009 together accounted for three quarters of all migrants from the EU8 and EU2 in the EU15 Romania, Lithuania, and Bulgaria sent the highest numbers of their nationals relative to their populations The Czech Republic, Slovenia, and Hungary exhibit the lowest shares of their population residing in the EU15 As for the receiving countries, in 2009 the most significant of the EU15 host countries for EU8 citizens were Germany and the UK, jointly hosting 62 percent of them For EU2 citizens the two most significant destinations were Italy and Spain, in 2009 each hosting more than 40 percent of all EU2 citizens residing in the EU15.11 The growth in population of EU8 citizens increased significantly, although to a different degree, in each of the countries that liberalized access to their labor market as of Labor mobility in an enlarged European Union  ­141 May 2004 Remarkably, many of the countries that opened up their labor markets later – including Denmark, the Netherlands, Luxembourg, Finland, and Austria – have similarly experienced an increase in the rate of growth of their EU8 populations following the 2004 enlargement A possible explanation is that EU accession removed some bureaucratic and psychological barriers to moving to old member states or that EU8 citizens circumvented labor market barriers mainly by coming as self-­employers.12 Concerning citizens of EU2 countries, their access to most EU15 labor markets continued to be restricted throughout the studied period Nevertheless, EU2 populations increased significantly in southern Europe, most notably in Spain and Italy, but also in Greece In Scandinavia, Bulgarian and Romanian populations continued to be rather small, although since 2007 there appear to be significant growth rates in Denmark and Sweden Among the other EU15 countries Austria, Ireland, Luxembourg and Belgium hosted the most dynamic EU2 populations The trends discussed above point at an important phenomenon that characterizes post-­enlargement migration, namely, the geographic diversion of migration flows For EU8 citizens the relative importance of the UK, Ireland but also Spain as host countries increased substantially, while the traditional host countries, Germany and Austria, lost their share quite dramatically For EU2 citizens the shares of Spain and Italy increased steeply, at the expense of mainly Germany but also of Austria and France The effects of this diversion may be long lasting due to the power of immigrant networks for the ­migration decision (Delbecq and Waldorf, 2010) As concerns the skill composition of citizens from new member states residing in the EU15 in the early post-­enlargement period, a number of early studies indicate that the majority of EU8 immigrants had medium educational attainment, and almost a quarter of them attained higher education (Brücker and Damelang, 2009; Brücker et al., 2009; European Commission, 2008b) Brücker and Damelang (2009) report that in 2006 among EU8 migrants in the EU15 17 percent had low and 22 percent had high educational attainment The corresponding figures for EU2 migrants were 29 percent and 18 percent Among the natives in the EU15 27 percent of them had low and the same percentage had high educational attainment Holland et al (2011) find that Luxembourg, Denmark, Sweden and Ireland are most popular among high-­skilled workers while low-­skilled workers are more likely to go to Greece, Portugal, Spain, Belgium, the Netherlands and Finland Furthermore, this study finds that for most of the EU812 countries’ migrants heading to the EU15 over-­represent the high – (except for Estonia, Slovenia and Lithuania) as well as low-­skilled (excepting Hungary and Latvia) domestic populations, but under-­represent the medium-­skilled population A book edited by Kahanec and Zimmermann (2010) systematically summarizes the available evidence on the scale, composition and effects of free labor mobility in the early post-­enlargement period Kahanec et al (2010) provide a broad account of post-­ enlargement migration in the EU, documenting the cross-­country differences in the scale and composition of these flows and their effects They in particular argue that EU enlargement has had different effects in countries that opened up their labor market early on, such as the UK, and those that strictly applied transitional arrangements, such as Germany For example, they report that while the skill composition of EU8 immigrants improved after enlargement in the UK, it worsened in Germany Blanchflower and Lawton (2010) document that in the UK EU8 migrants had a high incidence of 142   International handbook on the economics of migration self-­employment and high employment rates, and were well skilled Barrett (2010) finds that the EU10 migrants in Ireland had very high employment rates and levels of education comparable to the natives He also finds evidence for downskilling accompanied by relatively lower wages Brenke et al (2010) document that post-­enlargement migrants from the EU8 in Germany were predominantly male and young but were less educated and older than EU8 migrants had been previously The authors also report higher self-­employment rates but lower earnings and lower-­quality jobs for these immigrants Self-­employment rates as high as 38 percent for post-­enlargement migrants from the EU10 in Germany and 51 percent for those coming from EU2 in the UK in 2007 reported by European Commission (2008b) may signify inefficient spurious self-­employment as a way to circumvent transitional arrangements imposed in these cases A study by de la Rica (2010) reports that EU812 immigrants in Spain were predominantly young and had secondary education, allowing them to achieve high employment rates, but they also struggled with relatively high unemployment Importantly, she also reports lack of adjustment as concerns job quality Gerdes and Wadensjö (2010) find that in Sweden post-­enlargement migrants were relatively young and highly educated, but their earnings and employment rates were not as high as those of the natives While before enlargement immigration to Sweden from the EU10 was dominated by females, in the post-­enlargement its gender composition became much more even Kaczmarczyk and Okólski (2008) report bimodal out-­migration of high-­skilled migrants and low-­skilled ones mainly from peripheries in Poland Hazans and Philips (2010) and Hazans (2012) find that in the Baltic states’ post-­enlargement migrants were significantly less educated than stayers, with medium-­skilled workers being most likely to move after accession They not find evidence for brain drain but report significant brain waste in the form of downskilling Galgóczi et al (2012) shed light on skill-­mismatches in an enlarged EU and the role of trade unions in bridging these mismatches Using an innovative web-­based survey WageIndicator, Tijdens and van Klaveren (2012) document that among EU15 residents born in the EU10 only 65 percent report a correct job-­education match compared to 74 percent for the whole sample and 72 percent for all migrants Using the EU Labour Force Survey, Kahanec (2012) reports that in 2009 among EU10 nationals in the EU15 the share of high educated was 26.1 percent and low educated 22.5 percent; that is, they were considerably more educated than EU2 nationals in the EU15 of whom 12.2 percent were high and 37.5 percent low educated.13 They were in fact more educated than the total population in the EU15 with 18.9 percent high and 45.7 percent low educated residents EU10 as well as EU2 nationals in the EU15 were each positively selected compared to their source populations (this does not hold for EU2 nationals if only prime working age populations are looked at), with 14.4 percent high educated and 27.4 percent low educated residents in EU10 and 10.3 percent high educated and 40.9 percent low educated residents in EU2 The author also retrospectively constructs immigrant cohorts by year of arrival and finds that with enlargement the share of EU10 migrants with high educational attainment residing in the EU15 increased substantially, but so did the share of their low-­educated counterparts after 2007 For EU2 migrants it is reported that during the initial period 2007–08 there was a steady share of highly educated and an increasing share of low-educated migrants among them, followed by a Labor mobility in an enlarged European Union  ­143 steep increase in the share of high-­educated migrants and a similarly sharp decrease in the share of low-­educated ones in 2009 4 THE EFFECTS OF POST-­ENLARGEMENT MIGRATION IN RECEIVING AND SENDING COUNTRIES To evaluate the effects of post-­enlargement migration in an enlarged EU we consider the welfare of three key stakeholders to this process: the sending countries, the receiving countries and the migrants themselves Migrants from the new member states in the EU15 appear to be over-­represented in low- and medium-­skilled occupations and sectors, such as construction, manufacturing, hotels and restaurants, and agriculture (Kahanec et al., 2010) Given their relatively favorable skill composition discussed above, this discrepancy signifies a degree of downskilling and possibly brain waste Accompanied with the separation from their families and relatives in their countries of origin, it is not too surprising that this leads to lack of satisfaction with their migration experience (Anderson et al., 2006; Blanchflower and Lawton, 2010) In spite of their possible dissatisfaction with some aspects of their experience as migrants, post-­enlargement migrants can hardly be considered elsewhere but among the winners of free labor mobility in the EU Given the wage and unemployment gaps between sending and receiving countries, post-­enlargement migrants have benefited in terms of higher salaries, improved career prospects and a generally higher standard of living in the EU15 Improved human capital and language skills in particular add to the benefits of their migration experience in the EU15 Kureková (2011) finds that potential employers value migrant’s work experience acquired abroad upon their return, especially if they are young By the revealed preferences argument, the sum of these benefits should exceed the pecuniary, but also psychological and social, costs migration typically entails As concerns the effects on receiving countries, the available empirical evidence paints a rather positive picture Very small if any effects of post-­enlargement migration on the unemployment rate or wages are found in the UK (Blanchflower et al., 2007; Gilpin et al., 2006; Lemos and Portes, 2008) Blanchflower and Lawton (2010) detect small effects in the least skilled sectors Blanchflower and Shadforth (2009) and Blanchflower et al (2007) point at the importance of immigration and the resulting fear of unemployment for suppressing inflationary pressures Doyle et al (2006) and Hughes (2007) report a similar picture for Ireland, where post-­enlargement immigration might have caused some substitution and a temporary slow-­down of wage growth in some sectors, but any displacement at the microlevel was not affecting aggregate unemployment and the effects on wage growth reversed soon Brenke et al (2010) find that EU8 migrants compete with immigrants from outside of Europe for low-­skilled jobs rather than with the natives.14 Barrett  (2010) argues that post-­enlargement immigration helped Ireland to moderate the rather high wage growth during the pre-­2008 boom, which helped the country in terms of gross national product (GNP) growth Kaminska and Kahancová (2011) report that in Slovakia postenlargement out-­migration enabled trade unions to obtain wage increases Kahanec and Zimmermann (2009) show that high-­skilled immigration can be expected to decrease inequality, which highlights the importance of adjustment of high-­skilled migrants into 144   International handbook on the economics of migration corresponding jobs As concerns the feared effects on the receiving countries’ welfare systems, they have been shown to be unjustified (Doyle, 2007; Gerdes and Wadensjö, 2010; Hughes, 2007) Giulietti et al (2013) reject the welfare magnet hypothesis for migration within and into the EU The massive outflow of workers from some of the EU10 and EU2 countries has sparked some fears that the risks of EU enlargement may actually be borne by the new member states Kadziauskas (2007) warns that on the background of adverse demographic trends, the Lithuanian social security system may collapse due to post-­enlargement out-­ migration Kaczmarczyk and Okólski (2008) and Kadziauskas (2007) report growing shortages in some segments of the labor market soon after Poland’s and Lithuania’s EU accession Kureková (2011) reports significant skill shortages in Slovakia in the post-­ enlargement period A new trend in the sending countries has emerged, whereby such skill shortages are filled in by immigrants from outside the EU, mainly from Ukraine, Belarus, Russia and some Balkan countries (Frelak and Kazmierkiewicz, 2007; Iglicka, 2005; Kureková, 2011) An important consideration for the sending countries is to what extent post-­ enlargement out-­migration represents a lasting loss of labor and human capital, and to what extent it might signify the beginning of an era of, ‘brain gain and circulation’ Early studies suggest that there were no signs of significant ‘brain drain’, although some skilled sectors, such as medical doctors, lost nonnegligible proportions of their workforce (Brücker et al., 2009; European Commission, 2008b; Frelak and Kazmierkiewicz, 2007; Hazans, 2012) Kaczmarczyk et al (2010) argue that the economic effects of relatively large out-­migration are moderate in Poland They propose that post-­enlargement migration may foster the process of modernization in Poland, to the extent that ‘brain circulation’, facilitates restructuring and a higher allocative efficiency The negative selection into return migration observed for migrants from the Baltic states more recently (Hazans, 2012) may pose additional risks for the growth potential and sustainability of social security in the sending countries Also important is to what extent the gains from migration are transmitted to the left-­ behinds in the form of remittances Kahanec et al (2010) report an increasing importance of remittances in a number of sending countries, most significantly in Bulgaria and Romania, but also the Baltic states In Romania and Bulgaria remittances constituted about percent of their GDP in 2007 (Dietz, 2009) Comini and Faes-­Cannito (2010) report that the overall volume of remittances to the EU8 and EU2 declined in 2009 after years of growth, probably owing to the worsened economic situation in the host economies caused by the financial crisis Kaczmarczyk and Okólski (2008) document that remittances were primarily used for consumption and durable goods during the early post-­enlargement period, but also report that more recently they have been invested in human capital as well In a general equilibrium model Baas et al (2010) argue that the aggregate GDP of an enlarged EU can be expected to increase by about 0.2 percent, that is, 24 billion euros in total or 28 571 euros per post-­enlargement migrant, from 2004 to 2007 as a consequence of post-­enlargement migration from the EU8 alone The authors predict no lasting effects on wages or unemployment in the sending and receiving countries Similarly small effects on wages and unemployment are predicted by Holland et al (2011), although they predict some of these effects to last Labor mobility in an enlarged European Union  ­145 5 THE POTENTIAL FOR FURTHER POST-­ENLARGEMENT MIGRATION To shed light on what migration flows can be expected in the foreseeable future we consider current migration intentions Eurostat (2010) reveals that the most migration prone appear to be Scandinavians, with more than half of the Danes reporting positive intentions to work abroad sometime in the future Next and very close come the Baltic states, in each of which more than a third of the population intend to work abroad Perhaps somewhat surprisingly, at least in view of their relatively low out-­migration rates following their EU accession, Hungary and Slovenia exhibit higher shares of people who envisage working outside their country than Poland or Slovakia, and all exhibit greater shares than seen in Bulgaria and Romania As concerns which EU15 destination countries are preferred by EU8 workers, according to Eurostat (2010) it is mainly Germany (25.4 percent) and the UK (25.3 percent), followed by Austria (13.5 percent) Workers from EU2 countries mainly prefer Italy (17.0 percent), Spain (14.5 percent) and Germany (14.5 percent), but also the UK (11.5 percent) A key question is how imminent, or concrete, these intentions are This can be measured by the share of the respondents who see themselves working in a country outside their own country within the next six or 12 months The results from Eurostat (2010) indicate that migration intentions are most imminent in the Baltic states as well as Romania and Bulgaria The remaining new member states, Slovakia and Poland not differ very much from the EU27 average, whereas Slovenia, the Czech Republic and Hungary exhibit the lowest imminence of migration intentions To answer the question from which new member states one can expect the highest migration outflows in the foreseeable future, we construct a simple analytical migration imminence matrix using the data on migration intentions from Eurostat (2010) We plot in Figure 7.1 the share of population envisaging work abroad against the share of those of them who indicate that they expect to work there during the next six months (panel (a)) and, as a robustness check, during the next year (panel (b)).15 We then interpret the relative position in the matrix as a measure of a country’s imminent migration potential In particular, countries that fall into the southwest quadrant of the migration imminence matrix can be interpreted to have low imminent out-­migration potential This includes the Czech Republic but also, somewhat surprisingly, Poland and Slovakia This may indicate that the migration potential of these countries had been already partly exhausted by the end of 2009 The countries that fall into the northwest quadrant, Romania and Bulgaria, exhibit relatively low shares of people planning to work abroad However, for a relatively large share of those planning to work abroad the indicated plans seem to be rather imminent Slovenia and Hungary fall into the southeast quadrant with a relatively high share of people envisaging work abroad, but only a relatively small share of them indicating this that will to happen during the next six months With Hungary and Slovenia sharing a weak economic prospect in late 2009 and up until then relatively low out-­migration rates, a possible interpretation is that in these countries larger numbers of people were starting to consider the option of finding a job abroad, but their plans were relatively recent and not yet concrete The highest imminent migration potential is observed in Lithuania and Latvia in the 146   International handbook on the economics of migration Normalized share envisaging to work abroad within the next months (a) Work abroad within six months RO LV BG LT EU8+2 0.5 SK CZ 0 EE PL EU27 SI HU 0.5 Normalized share envisaging to work abroad Normalized share envisaging to work abroad within the next 12 months (b) Work abroad within a year RO LV BG LT EU8+2 0.5 EE PL EU27 CZ SK SI HU 0 0.5 Normalized share envisaging to work abroad Notes:  Based on the answers to the question QC10: ‘Do you envisage to work in a country outside (our country) at some time in the future?’ and QC11: ‘How soon are you likely to work there?’ The share of population answering ‘Yes’ to the first question is on the x-­axis, and, of those, the share answering ‘During the next months’ is on the y-­axis Values normalized with representing the highest, and the lowest, value observed in the EU EU812 calculated as a simple average of EU812 countries Source:  Eurostat (2010), data collection November–December 2009 Figure 7.1  The migration imminence matrix, EU812 northeast quadrant, which in 2009 exhibited a relatively high share of people expecting to work abroad and for this to happen during the next six months Estonia is the borderline case with the largest share of people expecting to work abroad in the future among the EU812 countries, although the share of people expecting this to happen during the next six months is considerably lower than in Latvia and Lithuania and is close to the EU812 simple average The high degree of similarity between panels (a) and (b) indicates that these findings are robust within the studied horizon of migration intentions To fully grasp the prospects of future migration between the new and old member states, it is necessary to understand the prospect of return migration as well Migration intentions of EU10 migrants are known to be rather transitory For example, of workers registered in the Worker Registration Scheme in the UK in 2008 62 percent envisaged staying in the UK for less than three months, up from 59 percent in 2007 and 55 percent Labor mobility in an enlarged European Union  ­147 in 2006 (Kahanec et al., 2010) The long-­run trends in return migration are yet to be evaluated First evidence by Hazans (2012) for the Baltic countries indicates that significant shares of migrants are indeed returning Whereas they used to be positively selected from migrant populations in the period immediately following the 2004 enlargement, after 2006 the share of high-­skilled workers among returnees is lower than among emigrant cohorts they come from (ibid.) Such developments could undermine the prospects for gainful ‘brain circulation’ from the perspective of sending countries Hazans (2012) further reports that compared with the pre-­crisis period, out-­migration intensified in Estonia and even more so in Latvia during the crisis The worsened economic conditions disproportionally pushed the less skilled as well as ethnic minorities to migrate, mainly Russian-­speakers Latvian migration became more long-­term oriented during the crisis Perhaps the most detrimental effect of the crisis is that high-­skilled workers became under-­represented among returnees, undermining the prospects of gainful ‘brain circulation’ for these sending countries 6  CONCLUSIONS Aging, diminishing young cohorts and a lack of innovation potential, and structural mismatches in the labor market resulting in unemployment and skill shortages at the same time, are some of the most important labor market challenges in the EU These challenges have contributed to, and are themselves aggravated by, the current debt crisis in the Eurozone With this backdrop, embracing the freedom of movement of workers in an enlarged EU as a powerful tool to improve allocation of human capital and thus combat some of these challenges would seem rational Yet, fear and controversies entangled the implementation of free labor mobility vis-­à-­vis the countries that joined the EU in 2004 and 2007 Painstaking empirical analyses based on theoretical underpinnings and hard data surveyed in this study tell the true story, however The free movement of labor in an enlarged EU can, with little doubt, be considered a success story of EU integration and enlargement It resulted in substantial relocation of labor that has improved the allocation of human capital in the EU These new hands and brains appear to have been absorbed by the receiving labor markets rather seamlessly In particular, except for some downskilling, we not observe any significant negative effects on (un)employment or wages in the EU15 Similarly, the hypothesis of welfare tourism has not been substantiated.16 We thus conclude that the pre-­enlargement fears of labor market disruptions to be caused by immigrants from the new member states were unjustified.17 The sending countries appear to have been relieved of some currently redundant labor resulting from skill mismatches in their labor markets, as well being relieved of the related fiscal burden Some new skill shortages have emerged, however Additionally, the loss of young and skilled labor may be rather worrying in view of the dismal demographic trends in most of the new member states, as well as for the sustainability of their public finances In view of these potential risks, of key importance for the sending countries is their ability to benefit from ‘brain gain’ resulting from ‘brain circulation’ in an enlarged EU This includes having a proper policy approach to the issues of return and circular migration and inefficient downskilling.18 Remittances partly compensate for the 148   International handbook on the economics of migration loss of human capital possibly characterizing the early stages of post-­enlargement migration Migrants themselves, as well as their families, appear to have traded the benefits of migration against some pecuniary and nonpecuniary costs to their benefit Based on the migration imminence matrix, we conclude that whereas the Baltic states will continue to send relatively large numbers of workers abroad in the near future, migration fatigue has emerged in Romania and Bulgaria and even more evidently in Slovakia and Poland There are no signs that the low migration potential of the Czech Republic will change soon, but Hungary and Slovenia may be the future sources of migrants Transitional arrangements seem to have affected not only the direction, but also the composition, of post-­enlargement migration flows A full evaluation of their effects is yet to come, but the evidence so far is that the countries that delayed liberalizing access to their labor market for citizens from the new member states disproportionally lost skilled and young migrants, who chose more welcoming countries such as Ireland and the UK Another negative effect is that transitional arrangements apparently led to spurious self-­ employment as a strategy to circumvent them This leads us to conclude that transitional arrangements backfired The current debt crisis in the EU is a challenge on its own Although the effects of free labor mobility in the EU are yet to be fully evaluated, based on the available literature we propose that the freedom of movement in an enlarged EU not only contributes to the European Project by strengthening the social fabric and improving cohesion in the EU, but that it does so also by directly contributing to its economic viability Namely, it provides for an improved allocative efficiency of European labor markets, a higher innovation potential, increased utilization of resources and their higher productivity, and the resulting fiscal relief, all enabling the EU to thrive economically, socially and politically in a globalized world NOTES   * The author thanks the anonymous referee as well as the editors of this volume, Amelie F Constant and Klaus F Zimmermann, for providing a number of suggestions that helped to improve the chapter significantly I remain responsible for any mistakes still present   The respective abbreviations used in this chapter are: CY, CZ, EE, LV, LT, MT, HU, PL, SK, SI, BG and RO EU8 denotes EU10 minus Cyprus and Malta; EU812 includes EU8 and EU2 EU15 includes Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain, Sweden and the United Kingdom   Some of the early forecasts added to the fears in the EU15 by predicting rather high east–west migration flows (Sinn et al., 2000), possibly even undermining the welfare state in the receiving countries (Sinn and Ochel, 2003) More moderate migration rates in the vicinity of actual post-­enlargement migration flows were predicted by, for example, Layard et al (1992), Bauer and Zimmermann (1999), Dustmann et al (2003), IOM (1998); see also Zaiceva and Zimmermann, (2008) and Brücker et al (2009) See Canoy et al (2010) for a thorough account of the links between public perception, migrants’ labor market outcomes and migration policies   Cyprus and Malta were exempt from such restrictions Ireland, the UK and Sweden opened up their labor markets immediately following the 2004 accession, while Germany and Austria imposed restrictions up until the end of the seven-­year period, albeit simplifying some of the procedures The other old member states had gradually opened up by May 2009 As for the 2007 enlargement, 10 member states opened up their labor markets during the first two-­year phase: the Czech Republic, Estonia, Cyprus, Latvia, Lithuania, Poland, Slovenia, Slovakia, Finland and Sweden By the end of the second phase on January 2012, Denmark, Greece, Hungary, Portugal and Spain opened up as well, with Austria, Germany, Labor mobility in an enlarged European Union  ­149           10 11 12 13 14 15 16 17 18 Belgium, France, Ireland, Luxembourg, Malta, the Netherlands and the UK still applying transitional arrangements as of January 2012 In July 2011 the EC authorized Spain to reinstate restrictions for Romanian workers until the end of 2012 Besides these economic factors, family, ethnic or social ties; natural catastrophes; social and political crises; as well as discrimination or persecution may result in significant movement of people (Mincer, 1978; Massey, 1990) For further reference see Chapter in this volume See Kahanec (2012) for a detailed analysis of low- and high-­skilled migration in a model of this type and Kahanec and Zimmermann (2009) for more on redistributive effects of migration Whereas the formal recognition of qualification obtained within the EU in another EU member state has been significantly simplified by EU legislation, informational asymmetries, linguistic and other barriers still obstruct the adjustment of within-­EU migrants The statistics about post-­enlargement migration may due to various measurement issues some of which are discussed in Kahanec (2012) over-­or under-­represent true migration flows and need to be interpreted with such caveats in mind For 2007 these figures are slightly higher than those reported by Brücker and Damelang (2009) or Brücker et al (2009), and in the range of those provided by European Commission (2008a, 2008b) That the 2004 accession took place on May, whereas in 2007 it was January, can at best only partly explain this difference in response The size of these populations needs also to be interpreted in the context of total immigrant populations, as people originating from EU10 or EU2 constitute only a smaller fraction of all immigrants in EU15 (Kahanec et al., 2010) Even in countries applying transitional arrangements restrictions for EU812 migrants were relaxed upon their countries’ EU accession This includes preferential treatment in access to work permits vis-­à-­vis third country nationals, the freedom of establishment of a business for self-­employed, and the freedom to provide services and thus post workers in the EU15 (excepting Austria and Germany) High level of education includes International Standard Classification of Education (ISCED) and levels; medium level of education comprises ISCED and levels; and low level of education takes in ISCED 0, and levels For further details about this classification see UNESCO (1997) This may have been one of the causes behind the 50 percent drop in immigration from other important source countries, including Russia, Ukraine and Turkey, from 2004 to 2006 reported by these authors These shares are normalized on the interval [0,1] to range between the respective minimums and maximums observed in the EU For further reference see Chapter 26 See also Constant (2012) For further reference see Chapter REFERENCES Anderson, B., M Ruhs, B Rogaly and S Spencer (2006), ‘Fair enough? Central and East European migrants in low-­wage employment in the UK’, report written for the Joseph Rowntree Foundation, COMPAS, Oxford Baas, T., H Brücker and A Hauptmann (2010), ‘Labor mobility in the enlarged EU: who wins, who loses?’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post – Enlargement Migration, Berlin and Heidelberg: Springer, pp. 47–70 Barrett, A (2010), ‘EU enlargement and Ireland’s labor market’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 143–61 Bauer, T and K.F Zimmermann (1999), ‘Assessment of possible migration pressure and its labor market impact following EU enlargement to Central and Eastern Europe’, a study for the Department of Education and Employment, UK, IZA Research Report No 3, Institute for the Study of Labor (IZA), Bonn Becker, G.S (1957), The Economics of Discrimination, Chicago, IL: University of Chicago Press Blanchflower, D.G and H Lawton (2010), ‘The impact of the recent expansion of the EU on the UK labour market’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 181–215 Blanchflower, D.G and C Shadforth (2009), ‘Fear, unemployment and migration’, The Economic Journal, 119 (535), F136–F182 Blanchflower, D.G., J Saleheen and C Shadforth (2007), ‘The impact of the recent migration from Eastern Europe on the UK economy’, IZA Discussion Paper No 2615, Institute for the Study of Labor (IZA), Bonn Bonin, H., W Eichhorst, C.Florman, M.O Hansen, L Skiöld, J Stuhler, K Tatsiramos, H Thomasen and 150   International handbook on the economics of migration K.F Zimmermann (2008), ‘Geographic mobility in the European Union: optimising its economic and social benefits’, IZA Research Report No 19, Institute for the Study of Labor (IZA), Bonn Borjas, G.J (1987), ‘Self-­selection and the earnings of immigrants’, American Economic Review, 77 (4), 531–53 Borjas, G.J (1999), ‘Immigration and welfare magnets’, Journal of Labor Economics, 17 (4), 607–37 Brenke, K., M Yuksel and K.F Zimmermann (2010), ‘EU enlargement under continued mobility restrictions: consequences for the German labor market’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 111–29 Brücker, H and A Damelang (2009), ‘Labour mobility within the EU in the context of enlargement and the functioning of the transitional arrangements: analysis of the scale, direction and structure of labour mobility’, background report, IAB, Nürnberg Brücker, H., T Baas, I Beleva, S Bertoli, T Boeri, A Damelang, L Duval, A Hauptmann, A Fihel, P.  Huber, A Iara, A Ivlevs, E.J Jahn, P Kaczmarczyk, M.E Landesmann, J Mackiewicz-­Lyziak, M. Makovec, P Monti, K Nowotny, M Okólski, S Richter, R Upward, H Vidovic, K Wolf, N Wolfeil, P Wright, K Zaiga and A Żylicz (2009), ‘Labour mobility within the EU in the context of enlargement and the functioning of the transitional arrangements’, final report (European Integration Consortium: IAB, CMR, fRDB, GEP, WIFO, wiiw), Nürnberg Brücker, H., G.S Epstein, B McCormick, G Saint-­Paul, A Venturini and K.F Zimmermann (2002), ‘Managing migration in the European welfare state’, in T Boeri, G.H.Hanson and B McCormick (eds), Immigration Policy and the Welfare System, Oxford: Oxford University Press, pp. 1–168 Canoy, M., A Horvath, A Hubert, F Lerais and M Sochacki (2010), ‘Post-­enlargement migration and public perception in the European Union’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 71–107 Chiswick, B.R (1980), ‘An analysis of the economic progress and impact of immigrants’, prepared for the Employment and Training Administration, U.S Department of Labor, National Technical Information Service, PB80-­200454 Chiswick, B.R (1998), ‘The economic consequences of immigration: application to the United States and Japan’, in M Weiner and T Hanami (eds), Temporary Workers or Future Citizens? Japanese and U.S Migration Policies, New York: New York University Press, pp. 177–208 Chiswick, B.R (1999), ‘Are immigrants favorably self-­selected?’, American Economic Review, 89 (2), 181–5 Chiswick, B.R and P.W Miller (2012), ‘Negative and positive assimilation, skill transferability, and linguistic distance’, Journal of Human Capital, (1), 35–55 Chiswick, C.U., B.R Chiswick and G Karras (1992), ‘The impact of immigrants on the macroeconomy’, Carnegie-­Rochester Conference Series on Public Policy, 37 (1), 279–316 Comini, D and F Faes-­Cannito (2010), ‘Remittances from the EU down for the first time in 2009, flows to non-­EU countries more resilient’, Statistics in Focus, 40/2010, available at: http://epp.eurostsat.ec.europa eu/portal/page/portal/product_details/publication?p_product_code5KS-­SF-­10-­040 (accessed 11 January 2013) Constant, A.F (2012), ‘Sizing it up: labor migration lessons of the EU expansion to 27’, Scribani International Conference Proceedings, Bruylant: Belgium, pp. 49–77 Constant, A.F., M Kahanec and K.F Zimmermann (2009), ‘Attitudes towards immigrants, other integration barriers, and their veracity’, International Journal of Manpower, 30 (1/2), 5–14 De Giorgi, G and M Pellizzari (2009), ‘Welfare migration in Europe and the cost of a harmonised social assistance’, Labour Economics, 16 (4), 353–63 De la Rica, S (2010), ‘The experience of Spain with the inflows of new labor migrants’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 131–44 Delbecq, B.A and B.S Waldorf (2010), ‘Going west in the European Union: migration and EU-­enlargement’, Working Paper 10-­4, Purdue University, Department of Agricultural Economics, West Lafayette Dietz, B (2009), ‘Migration, remittances and the current economic crisis: implications for Central and Eastern Europe’, Osteuropa Institut Regensburg: Kurzanalysen und Informationen, 42, available at: http://www oei-­dokumente.de/publikationen/info/info-­42.pdf (accessed February 2013) Doyle, N (2007), ‘The effects of Central European labor migration on Ireland’, in J Smith-­Bozek (ed.), Labor Mobility in the European Union: New Members, New Challenges, Washington, DC: Center for European Policy Analysis, pp. 35–59 Doyle, N., G Hughes and E Wadensjö (2006), ‘Freedom of movement for workers from Central and Eastern Europe – experiences in Ireland and Sweden’, SIEPS Report No 5, Swedish Institute for European Policy Studies, Stockholm Dustmann, C., M Casanova, M Fertig, I Preston and C.M Schmidt (2003), ‘The impact of EU enlargement on migration flows’, Home Office Online Report 25/03, Research Development and Statistics Directorate, London Labor mobility in an enlarged European Union  ­151 European Commission (2008a), The Impact of Free Movement of Workers in the Context of EU Enlargement, communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions Brussels, COM(2008) 765 final, 18.11.2008 European Commission (2008b), ‘Geographical labour mobility in the context of EU enlargement’, in European Commission, Employment in Europe 2008, Luxembourg: Office for Official Publications of the European Communities, pp. 111–45 Eurostat (2010), ‘Geographical and labour market mobility’, Special Eurobarometer 337, Eurostat, European Commission, Luxembourg Frelak, J and P Kazmierkiewicz (2007), ‘Labor mobility: the case of Poland’, in J Smith-­Bozek (ed.), Labor Mobility in the European Union: New Members, New Challenges, Washington, DC: Center for European Policy Analysis, pp. 60–79 Galgóczi, B., J Leschke and A Watt (eds) (2012), EU Labour Migration in Troubled Times: Skills Mismatch, Return and Policy Responses, Aldershot: Ashgate Gerdes, C and E Wadensjö (2010), ‘Post-­enlargement migration and labor market impact in Sweden’, in M. Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 163–79 Gilpin, N., M Henty, S Lemos, J Portes and C Bullen (2006), ‘The impact of free movement of workers from Central and Eastern Europe on the UK labour market’, Working Paper No 29, Department of Work and Pensions, London Giulietti, C., M Guzi, M Kahanec and K.F Zimmermann (2013), ‘Unemployment benefits and immigration: evidence from the EU’, International Journal of Manpower, 34 (1), 24–38 Harris, J.R and M.P Todaro (1970), ‘Migration, unemployment and development: a two-­sector analysis’, American Economic Review, 60 (1), 126–42 Hazans, M (2012), ‘Selectivity of migrants from Baltic countries before and after enlargement and responses to the crisis’, in B Galgóczi, J Leschke and A Watt (eds), EU Labour Migration in Troubled Times: Skills Mismatch, Return and Policy Responses, Aldershot: Ashgate Hazans, M and K Philips (2010), ‘The post-­enlargement migration experience in the Baltic labor markets’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 255–304 Holland, D., T Fic, P Paluchowski, A Rincon-­Aznar and L Stokes (2011), ‘Labour mobility within the EU: the impact of enlargement and transitional arrangements’, NIESR Discussion Paper No 379, National Institute of Economic and Social Research, London Hughes, G (2007), ‘EU Enlargement and Labour Market Effects of Migration to Ireland from Southern, Central and Eastern Europe’, paper presented at Second IZA Migration Workshop: EU Enlargement and the Labour Markets, Bonn, 7–8 September Iglicka, K (2005), ‘The impact of the EU enlargement on migratory movements in Poland’, Reports and Analyses 12/05, Center for International Relations, Warsaw International Organization for Migration (IOM) (1998), Migration Potential in Central and Eastern Europe, Geneva: International Organization for Migration Kaczmarczyk, P and M Okólski (2008), ‘Economic impacts of migration on Poland and the Baltic states’, Fafo-­paper 2008:1, Fafo, Oslo Kaczmarczyk, P., M Mioduszewska and A Żylicz (2010), ‘Impact of the post-­accession migration on the Polish labor market’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­ Enlargement Migration, Berlin and Heidelberg: Springer, pp. 219–53 Kadziauskas, G (2007), ‘Lithuanian migration: causes, impacts and policy guidelines’, in J Smith-­Bozek (ed.), Labor Mobility in the European Union: New Members, New Challenges, Washington, DC: Center for European Policy Analysis, pp. 80–100 Kahanec, M (2012), ‘Labor mobility in an enlarged European Union’, IZA Discussion Paper No 6485, Institute for the Study of Labor (IZA), Bonn Kahanec, M and K.F Zimmermann (2009), ‘International migration, ethnicity and economic inequality’, in W Salverda, B Nolan and T.M Smeeding (eds), The Oxford Handbook of Economic Inequality, Oxford: Oxford University Press, pp. 455–90 Kahanec, M and K.F Zimmermann (eds) (2010), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer Kahanec, M., A Myung-Hee Kim and K.F Zimmermann (2013), ‘Pitfalls of immigrant inclusion into the European welfare state’, International Journal of Manpower, 34 (1), 39–55 Kahanec, M., A Zaiceva and K.F Zimmermann (2010), ‘Lessons from migration after EU enlargement’, in M Kahanec and K.F Zimmermann (eds), EU Labor Markets after Post-­Enlargement Migration, Berlin and Heidelberg: Springer, pp. 3–45 Kaminska, M.E and M Kahancová (2011), ‘Emigration and labour shortages: an opportunity for trade unions in the New Member States’, European Journal of Industrial Relations, 17 (2), 189–203 152   International handbook on the economics of migration Kureková, L (2011), ‘From job search to skill search Political economy of labor migration in Central and Eastern Europe’, PhD dissertation, Central European University (CEU), Budapest Layard, R., O Blanchard, R Dornbusch and P Krugman (1992), East–West Migration The Alternatives, Cambridge, MA: MIT Press Lemos, S and J Portes (2008), ‘The impact of migration from the new European Union member states on native workers’, Working Paper No 52, Department for Work and Pensions, Leeds Massey, D.S (1990), ‘Social structure, household strategies, and the cumulative causation of migration’, Population Index, 56 (1), 3–26 Mincer, J (1978), ‘Family migration decisions’, Journal of Political Economy, 86 (5), 749–73 Ottaviano, G.I.P and G Peri (2006), ‘The economic value of economic diversity: evidence from US cities’, Journal of Economic Geography, (1), 9–44 Sinn, H.-­W and W Ochel (2003), ‘Social union, convergence and migration’, Journal of Common Market Studies, 41 (5), 869–96 Sinn, H.-­W., G Flaig, M Werding, S Munz, N Düll and H Hofmann (2000), ‘EU-­Erweiterung und Arbeitskräftemigration Wege zu einer schrittweisen Annäherung der Arbeitsmärkte’, Studie für das Bundesministeriums für Arbeit und Sozialordnung, Ifo Institut, München Sjaastad, L.A (1962), ‘The costs and returns of human migration’, Journal of Political Economy, 70 (5), 80–93 Stark, O (1991), The Migration of Labor, Cambridge: Blackwell Tijdens, K and M van Klaveren (2012), ‘A skill mismatch for migrant workers? Evidence from WageIndicator survey data’, in B Galgóczi, J Leschke and A Watt (eds), EU Labour Migration in Troubled Times: Skills Mismatch, Return and Policy Responses, Aldershot: Ashgate United Nations Educational, Scientific, and Cultural Organization (UNESCO) (1997), International Standard Classification of Education, Paris: UNESCO Zaiceva, A and K.F Zimmermann (2008), ‘Scale, diversity, and determinants of labour migration in Europe’, Oxford Review of Economic Policy, 24 (3), 427–51 Zimmermann, K.F (2005), European Migration: What Do We Know? Oxford and New York: Oxford University Press Zimmermann, K.F., M Kahanec, A Barrett, C Giulietti, B Mtre and M Guzi (2012), ‘Study on active inclusion of immigrants’, IZA Research Report 43, Institute for the Study of Labor (IZA), Bonn 8  Minority and immigrant entrepreneurs: access to financial capital* Robert W Fairlie 1  Introduction A large literature examines the impact of financial capital on small business formation and performance The findings from this literature indicate that access to financial capital is one of the most important determinants of small business creation and success Many previous studies also examine the barriers that disadvantaged minorities face in obtaining access to capital for their businesses These studies find that access to capital, wealth inequality and lending discrimination create substantial barriers for minority business success Wealth inequality represents a troubling and persistent root of this problem because the owner’s wealth can be invested directly in the business or used as collateral to obtain business loans, and the disparities are extremely large For example, in the United States, African-­Americans and Latinos have median wealth levels that are one-­ninth to one-­thirteenth white levels (US Census Bureau, 2011) Much less research, however, has focused on access to and use of financial capital among immigrant entrepreneurs.1 Anecdotal evidence suggests that immigrant entrepreneurs rely heavily on informal sources to finance their businesses instead of banks or other institutions, but there is little direct evidence from nationally representative datasets carefully documenting these patterns The lack of research on the advantages and disadvantages that immigrant entrepreneurs face in obtaining capital for creating and maintaining successful businesses may represent an important omission from the literature Some of the barriers to access to capital might be similar between immigrants and overall minorities, but may also differ because immigrants lack familiarity with the host country’s language, institutions and culture A better understanding of the constraints faced by immigrant entrepreneurs may shed light on whether there is untapped potential for this group and whether their contributions to host country economies can be even greater.2 The entrepreneurial success of immigrants is well known For example, business ownership is higher among the foreign-­born than the native-­born in many developed countries such as the United States, UK, Canada and Australia (Borjas, 1986; Clark and Drinkwater, 2000, 2010; Fairlie et al., 2010; Lofstrom, 2002; Schuetze and Antecol, 2006).3 Businesses owned by some immigrant groups are also very successful with higher incomes and employment than native-­owned businesses Immigrants are found to contribute even more to specific sectors and regions of host countries’ economies (Fairlie, 2008) In particular, much recent attention has been drawn to the contributions of immigrant entrepreneurs to the technology and engineering sectors of the economy (Saxenian, 1999, 2000; Wadhwa et al., 2007) Immigration is also found to increase innovation 153 154   International handbook on the economics of migration measured as patents and even have positive spillovers in innovation for others (Hunt and Gauthier-­Loiselle, 2010; Kerr and Lincoln, 2010 In an attempt to attract immigrant entrepreneurs, many developed countries have created special visas and entry requirements for immigrant entrepreneurs (Schuetze and Antecol, 2006; see also Chapter 23 in this volume) This chapter reviews the literature on access to financial capital among ethnic minority entrepreneurs and businesses Access to capital among immigrant entrepreneurs is also examined New estimates from a nationally representative government source from the United States are presented because of the lack of previous research on access to capital among immigrant-­owned businesses owing to data limitations The descriptive analysis of Census data allows for an examination of a few questions regarding access to financial capital among small businesses owned by immigrants These questions have not been previously addressed in the literature First, immigrant entrepreneurs have access to less or more startup capital than non­immigrant entrepreneurs? Second, what are the sources of financial capital used by immigrant business owners? Do these sources differ from those used by nonimmigrant business owners, especially from informal sources compared with bank loans and other more formal sources? Finally, the single largest asset held by most households is equity in their home which can be invested directly into business startups or used as collateral to obtain business loans Do immigrants and natives differ in rates of home ownership and these differences have any impact on differences in rates of business formation? 2 Previous Research on Financial Capital and Minority Businesses One of the most important barriers preventing would-­be entrepreneurs from starting businesses and small businesses from growing is inadequate access to financial capital Starting with entry, the importance of personal wealth as a determinant of entrepreneurship has been the focus of an extensive body of literature Numerous studies, using various methodologies, measures of wealth and country microdata, explore the relationship between wealth and entrepreneurship.4 Previous research has examined the relationship between wealth and entrepreneurship using data from the United States (for example, Bates and Lofstrom, 2013; Demiralp and Francis, 2008; Evans and Jovanovic, 1989; Evans and Leighton, 1989; Fairlie, 1999; Holtz-­Eakin and Rosen, 1999; Holtz-­ Eakin et al., 1994; Hurst and Lusardi, 2004; Meyer, 1990; Zissimopoulos and Karoly, 2007; Zissimopoulos et al., 2009), Europe (for example, Blanchflower and Oswald, 1998; Giannetti and Simonov, 2004; Holtz-­Eakin and Rosen, 1999; Johansson, 2000; Lindh and Ohlsson, 1996, 1998; Nykvist, 2008; Schäfer et al., 2011; Taylor, 2001) and developing countries (for example, Paulson and Townsend, 2004; Yu, 2010) Most studies find that asset levels (for example, net worth) measured in one year increase the probability of starting a business by the following year This finding has generally been interpreted as providing evidence that entrepreneurs face liquidity constraints, although there is some debate over the interpretation (Fairlie and Krashinsky, 2012; Hurst and Lusardi, 2004) Although a large body of previous research provides evidence that is consistent with Minority and immigrant entrepreneurs  ­155 low levels of personal wealth resulting in lower rates of business creation, less research has focused on the related question of whether low levels of personal wealth and liquidity constraints also limit the ability of entrepreneurs to raise startup capital resulting in undercapitalized businesses The consequence is that these undercapitalized businesses will likely have lower sales, profits and employment, and will be more likely to fail than businesses receiving the optimal amount of capital at startup Evidence on the link between startup capital and owner’s wealth is provided by examining the relationship between business loans and personal commitments, such as using personal assets for collateral for business liabilities and guarantees that make owners personally liable for business debts Avery et al (1998) find that the majority of all small business loans have personal commitments The common use of personal commitments to obtain business loans suggests that wealthier entrepreneurs may be able to negotiate better credit terms and obtain larger loans for their new businesses possibly leading to more successful firms.5 Cavalluzzo and Wolken (2005) find that personal wealth, primarily through home ownership, decreases the probability of loan denials among existing business owners If personal wealth is important for existing business owners in acquiring business loans, then it may be even more important for entrepreneurs in acquiring startup loans Examining the relationship between startup capital and business performance directly, previous research indicates a strong positive correlation Firms with higher levels of startup capital are less likely to close, have higher profits and sales, and are more likely to hire employees (Fairlie and Robb, 2008) The estimates are large and consistent across outcomes This positive relationship is consistent with the inability of some entrepreneurs to obtain the optimal level of startup capital because of borrowing constraints Because these entrepreneurs are constrained in the amount of startup capital that could be used to purchase buildings, equipment and other investments, their businesses are less successful than if they could have invested the optimal amount of capital the positive correlation, however, may alternatively be partly due to potentially successful business ventures being more likely to generate startup capital than business ventures that are viewed as being potentially less successful Financial constraints are one of the most significant issues affecting minority business creation and performance To get an idea of the potential importance of access to financial capital in contributing to racial disparities in business ownership, one only has to look at the alarming levels of wealth inequality that exist Estimates from the US Census Bureau (2011) indicate that half of all black families have less than $8650 in wealth, and half of all Hispanic families have less than $13 375 Wealth levels among whites are $113 822, which is nine to 13 times higher Low levels of wealth and ­liquidity constraints may create a substantial barrier to entry for minority entrepreneurs because the owner’s wealth can be invested directly in the business or used as collateral to obtain business loans Investors frequently require a substantial level of owner’s investment of his/her own capital as an incentive (that is, ‘skin in the game’) Racial differences in home equity may be especially important in providing access to startup capital Less than half of blacks in the United States own their own home compared with three-­quarters of whites, and the median equity in their homes is $35 000 compared with $59 000 for whites (Fairlie and Robb, 2008) Homes provide collateral 156   International handbook on the economics of migration and home equity loans provide relatively low-­cost financing Without being able to tap into this equity many minorities will not be able to start businesses Previous studies find that relatively low levels of wealth among blacks and Hispanics contribute to why these groups have low business creation rates in the United States Indeed, recent research using statistical decomposition techniques provides evidence supporting this hypothesis Fairlie (2006) finds that the largest single factor explaining racial disparities in business creation rates are differences in asset levels Lower levels of assets among blacks account for 15.5 percent of the difference between the rates of business creation among whites and blacks This finding is consistent with the presence of liquidity constraints and low levels of assets limiting opportunities for blacks to start businesses The finding is very similar to estimates reported in Fairlie (1999) for men using earlier data from the Panel Study of Income Dynamics (PSID) Estimates from the PSID indicate that 13.9 to 15.2 percent of the black/white gap in business start rates can be explained by differences in assets Fairlie and Woodruff (2010) examine the causes of low rates of business formation among Mexican-­Americans One of the most important factors in explaining the gaps between Mexican-­Americans and non-­Hispanic whites in rates of business creation is assets Relatively low levels of assets explain roughly one-­quarter of the business entry rate gap for Mexican-­Americans Lofstrom and Wang (2009) using SIPP data also find that low levels of wealth for Mexican-­Americans and other Latinos work to lower self-­ employment entry rates Apparently, low levels of personal wealth limit opportunities for Mexican-­Americans and other Latinos to start businesses Although previous research provides evidence that is consistent with low levels of personal wealth resulting in lower rates of business creation among minorities, very little research has focused on the related question of whether low levels of personal wealth, and liquidity constraints also limit the ability of minority entrepreneurs to raise startup capital resulting in undercapitalized businesses The consequence is that these undercapitalized businesses will likely have lower sales, profits and employment and will be more likely to fail than businesses receiving the optimal amount of startup capital Estimates from US Census microdata indicate that black and Hispanic-­owned businesses have very low levels of startup capital relative to non-­Hispanic white-­owned businesses (US Census Bureau, 1997; Fairlie and Robb, 2008) For example, less than percent of black firms start with $100 000 or more of capital and 6.5 percent have between $25 000 and $100 000 in startup capital Black-­owned firms are also found to have lower levels of startup capital across all major industries (US Census Bureau, 1997) These low levels of startup capital are found to be a major cause of worse outcomes among black-­owned businesses Using earlier CBO data, Bates (1997) finds evidence that racial differences in business outcomes are associated with disparities in startup capital More recent estimates indicate that lower levels of startup capital among black firms are the most important explanation for why black-­owned businesses have lower survivor rates, profits, employment and sales than nonminority-owned businesses (Fairlie and Robb, 2008) Asian firms are found to have higher startup capital levels and resulting business outcomes Minority and nonminority entrepreneurs differ in the types of financing they use for their businesses Previous research indicates, for example, that black entrepreneurs rely less on banks than whites for startup capital (US Census Bureau, 1997) Blacks are also Minority and immigrant entrepreneurs  ­157 less likely to use a home equity line for startup capital than are whites, which may be partly due to the lower rates of home ownership reported above On the other hand, black business owners are more likely to rely on credit cards for startup funds than are white business owners Using earlier data, Bates (1997, 2005) finds large differences between black and white firms in their use of startup capital Black firms are found to be more likely to start with no capital, less likely to borrow startup capital and more likely to rely solely on equity capital than white firms Bates (2005) also finds that loans received by black firms borrowing startup capital are significantly smaller than those received by white firms even after controlling for equity capital and owner and business characteristics, such as education and industry Previous research also indicates that minority-­owned businesses are more likely to use credit cards and less likely to use bank loans to start their businesses than nonminority-­owned businesses (Minority Business DevelopmentAgency, 2008) Additional evidence on racial differences in access to financial capital is provided by self-­reports by owners with unsuccessful businesses on why their businesses were unsuccessful Black business owners were two to three times more likely than all business owners to report ‘lack of access to business loans/credit’ or ‘lack of access to personal loans/credit’ as a reason for closure (US Census Bureau, 1997) Hispanic business owners were also more likely to report these two reasons Minority firms also have trouble securing funds from venture capitalists and angel investors Private equity funds targeting minority markets are very small relative to the total, which is problematic because these funds appear to be important for success (Yago and Pankrat, 2000) The disparity in access does not appear to be driven by performance differences Bates and Bradford (2009) examine the performance of investments made by venture capital funds specializing in minority firms and find that these funds produce large returns Venture capital funds focusing on investing in minority firms provide returns that are comparable to mainstream venture capital firms Funds investing in minority businesses may provide attractive returns because the market is underserved Evidence of Lending Discrimination A factor that may pose a barrier to obtaining financial capital for minority-­owned businesses is lending discrimination Much of the recent research on the issue of discrimination in business lending uses data from various years of the Survey of Small Business Finances (SSBF) conducted by the US Board of Governors of the Federal Reserve The main finding from this literature is that minority businesses experience higher loan denial probabilities and pay higher interest rates than white-­owned businesses even after controlling for differences in creditworthiness, and other factors.6 Cavalluzzo and Wolken (2005) found that while greater personal wealth is associated with a lower probability of denial, even after controlling for personal wealth, there remained a large difference in denial rates across demographic groups They also found that denial rates for blacks increased with lender market concentration, a finding consistent with Becker’s (1957 [1971]) classic theories of discrimination Using earlier data, Cavalluzzo et al (2002) found that all minority groups were more likely than whites to have unmet credit needs Blacks were more likely to have been denied credit, even after 158   International handbook on the economics of migration controlling for many factors related to creditworthiness In fact, denial rates and unmet credit needs for blacks widened with an increase in lender market concentration The fear of denial often prevented some individuals from applying for a loan, even when they had credit needs Blacks and Hispanics most notably had these fears Blanchflower et al (2003) conducted a similar analysis with similar results, but did not have access to some of the proprietary information available to researchers from the Federal Reserve However, they did find that black-­owned businesses were more likely to have a loan application denied, even after controlling for differences in creditworthiness, and that blacks paid a higher interest rate on loans obtained They also found that concerns over whether a loan application would be denied prevented some prospective borrowers from applying for a loan in the first place The disparities between the denial rates between whites and blacks grew when taking these individuals into consideration along with those who actually applied for a loan Bostic and Lampani (1999) include additional geographic controls and continue to find a statistically significant difference in approval rates between blacks and whites Immigrant-­Owned Businesses Much less research, however, has focused on access to and use of financial capital among immigrant entrepreneurs Anecdotal evidence suggests that immigrant entrepreneurs rely heavily on informal sources to finance their businesses instead of banks or other institutions, but there is little direct evidence from nationally representative datasets carefully documenting these patterns One exception is provided by US Census data suggesting that there may be significant leveraging of personal wealth by immigrant entrepreneurs Asian-­immigrant businesses have substantially higher levels of startup capital than non-­Latino white-­owned businesses, but comparisons of overall personal wealth indicate similar levels between non-­Latino whites and Asians (Fairlie and Robb, 2008) The use of rotating credit associations among some immigrant groups has been argued to be important in financing immigrant businesses, but perhaps an equal number of studies suggest that they play only a minor role (see, for example, Bates, 1997; Light et al., 1990; Yoon, 1991) 3 New Estimates on Capital Use among Immigrant-­ Owned Businesses To address the limited previous research on capital use among immigrant-­owned businesses this section presents new estimates from a nationally representative, US government data source The main reason for the lack of previous research on access to financial capital among immigrant entrepreneurs has been data availability Datasets with large enough sample sizes and information on immigrant status typically not provide information on financial capital An exception is provided by a newly available dataset that includes information on both immigrant status and on the sources and levels of financial capital use – the 2007 Survey of Business Owners (SBO) For the first time since 1992, the US Census Bureau collected information on whether business owners are immigrants and the amount of startup capital used by the business as part of its main Minority and immigrant entrepreneurs  ­159 Table 8.1 Business ownership rates by immigrant status, Current Population Survey (2010) Total Immigrant US-­born Percentage of workforce Sample size 9.5 10.5 9.3 636 401 90 086 546 315 Notes: The sample consists of individuals aged 20–64 who work 15 or more hours per usual week All estimates are calculated using sample weights provided by the CPS business owner data collection effort These data, as well as data from the 2010 Current Population Surveys (CPS), are used to present some patterns of access to financial capital among immigrant entrepreneurs in the United States Immigrant Business Ownership and Performance Before examining capital use among immigrant-­owned businesses it is useful to first examine patterns of immigrant business ownership and performance.7 Previous research indicates higher levels of business ownership among the foreign-­born than the native-­born in many developed countries including the United States, the UK, Canada and Australia (Borjas, 1986; Clark and Drinkwater, 2000, 2010; Fairlie et al., 2010; Lofstrom, 2002; Schuetze and Antecol, 2006) The latest estimates from the 2010 CPS microdata confirm this finding Table 8.1 displays estimates of self-­ employed business ownership rates in 2010 for immigrants and nonimmigrants The self-­employed business ownership rate is the ratio of the number of self-­employed business owners to the total number of workers Business ownership in the CPS captures ownership of all types of businesses including incorporated, unincorporated, employer and nonemployer businesses The estimates indicate that business ownership rates are higher for immigrants than nonimmigrants Indeed, 10.5 percent of the immigrant workforce owns a business, compared with 9.3 percent of the nonimmigrant workforce The difference in business ownership rates of 1.2 percentage points implies that immigrants are more than 10 percent more likely to own a business than are nonimmigrants The business ownership rate captures the stock of business owners in the economy at a given point in time, but does not capture the dynamics of business creation It is useful to examine business formation among immigrants because it captures the startup potential of this group New businesses are often associated with economic growth, innovation and the creation of jobs To investigate, I estimate the rate of business formation for immigrants and compare it to nonimmigrants Table 8.2 displays estimates of business formation rates for 2010 Immigrants are found to create businesses at a faster rate than nonimmigrants The business formation rate per month among immigrants is 0.62 percent; that is, of 100 000 nonbusiness-­owning immigrants, 620 start a business each month This rate of business formation is much higher than the US-­born rate of 0.28 percent, or 280 of 100 000 US-­born nonbusiness owners per month Although higher 160   International handbook on the economics of migration Table 8.2 Business formation rates by immigrant status, Current Population Survey (2010) Percentage of nonbusiness owners Sample size 0.34 0.62 0.28 593 271 82 640 510 631 Total Immigrant US-­born Notes: The sample consists of individuals aged 20–64 who not own a business in the first survey month The business formation rate is the percentage of nonbusiness owners that start a business in the following month with 15 or more hours worked All observations with allocated labor force status, class of worker and hours worked variables are excluded All estimates are calculated using sample weights provided by the CPS Table 8.3 Average sales, employment and payroll for immigrant-­and nonimmigrant-­ owned firms, special tabulations from Survey of Business Owners (2007) Ownership All firms Immigrant (majority foreign-­born) Nonimmigrant (majority   native-­born) Hispanic immigrant Asian immigrant Average sales ($) Percentage hiring employees 1 108 464 433 592 608 703 257 416 465 296 Employer firms Average number of employees Average payroll ($) 21.2 27.9 26.3 20.5 8.0 11.9 840 862 252 758 428 546 19.8 36.0 7.2 7.0 198 404 200 530 Note:  All firms includes publicly held firms rates of business ownership have been documented extensively in the previous literature, the finding of substantially higher immigrant-­owned business formation rates is a relatively new and important finding Combined with the previous finding of slightly higher business ownership rates among immigrants relative to nonimmigrants, it indicates that immigrants move into and out of business ownership at a much higher rate than nonimmigrants.8 The performance of businesses started by immigrants is examined next using data from the newly released 2007 SBO The SBO is considered the most up-­to-­date, comprehensive dataset on minority businesses in the United States For the first time since 1992, the US Census Bureau collected information on the immigrant status of business owners in its main database of information on the ownership characteristics of US businesses Table 8.3 reports estimates from specially commissioned tabulations from the 2007 SBO for the average sales and employment of immigrant-­ and nonimmigrant-­owned businesses.9 Immigrant-­owned businesses represent 13.2 percent of all businesses in Minority and immigrant entrepreneurs  ­161 which foreign-­born status of the owners can be determined Immigrant-­owned firms have $434 000 in average annual sales and receipts The average level of sales is roughly 70 percent of the level of nonimmigrant-­owned firms at $609 000 Immigrant-­owned businesses are slightly more likely to hire any employees than are nonimmigrant-­owned businesses, however, they tend to hire fewer employees on average immigrant-­owned businesses that hire employees hire an average of 8.0 employees with an average payroll of $253 000 Nonimmigrant-­owned businesses that hire employees hire an average of 11.9 employees with an average payroll of $429 000 There are interesting differences by race and ethnicity for immigrant-­owned businesses Hispanic-­immigrant owned businesses have an average sales level of $257 000 compared with $465 000 for Asian-­immigrant owned businesses Asian-­immigrant owned firms are more likely to hire employees than Hispanic-­immigrant owned firms (36 percent compared with 20 percent), but have roughly similar levels of employment and payroll conditioning on being an employer firm On average, businesses owned by Hispanic immigrants are smaller than businesses owned by Asian immigrants Average sales levels can be influenced heavily by a few outliers of very successful firms and may be misleading of the more common performance levels of immigrant-­owned firms An examination of the entire sales distribution for immigrant and nonimmigrant firms reveals some interesting differences between the firms Immigrant firms are less likely to have very low levels of sales and are more likely to be in the middle of the sales distributions Immigrant firms are slightly less likely to have sales at the very high end of the distribution defined as $1 000 000 or more, but are slightly more likely to have sales in the $500 000 to $999 999 range than nonimmigrant firms Overall, 11.4 percent of immigrant firms have sales of $500 000 or more, which is similar to the percentage of nonimmigrant firms at this level Asian-­immigrant owned businesses tend to have higher levels of sales than immigrant-­ owned businesses overall They are more likely to be represented in the highest sales categories with 13.7 percent having sales of $500 000 or more, which is higher than for nonimmigrant owners at 11.8 percent Hispanic-­immigrant businesses have lower sales than the immigrant total and nonimmigrants Among Hispanic-­immigrant firms, 7.4 percent have sales of $500 000 or more Overall, immigrant businesses have lower average sales and hire fewer employees than nonimmigrant businesses They are less likely to have very high levels of sales of $1 000 000 or more Hispanic-­immigrant owned firms tend to have lower sales and employment, and Asian-­immigrant owned firms have higher sales and employment Although immigrant-­owned businesses are not substantially underperforming nonimmigrant-­owned businesses, there might be some untapped potential among this group of business owners One potential barrier is access to financial capital Limited access to financial capital may restrict immigrant business success, partly explaining why performance is lower than for businesses owned by nonimmigrants Capital Use Among Immigrant-­Owned Businesses In addition to providing new information on the immigrant status of the business, the 2007 SBO is the first US Census survey since 1992 to include information on levels of startup capital Given the importance of access to startup capital for business 162   International handbook on the economics of migration Table 8.4 Startup capital distributions for immigrant-­and nonimmigrant-­owned firms, special tabulations from the Survey of Business Owners (2007) Amount of startup capital No startup capital Less than $5000 $5000 to $9999 $10 000 to $24 999 $25 000 to $49 999 $50 000 to $99 999 $100 000 to $249 999 $250 000 to $999 999 $1 000 000 or more All firms Immigrant Nonimmigrant Hispanic immigrant Asian immigrant 23.5% 33.7% 9.2% 9.8% 6.4% 6.1% 5.8% 3.9% 1.6% 22.0% 29.7% 10.1% 11.1% 7.6% 7.4% 6.9% 4.1% 1.2% 23.8% 35.3% 9.2% 9.6% 6.2% 5.7% 5.4% 3.6% 1.2% 25.2% 34.8% 12.0% 10.9% 6.7% 4.8% 3.4% 1.8% 0.4% 17.3% 22.8% 8.8% 12.5% 9.6% 10.5% 10.7% 6.0% 1.7% Notes: All firms includes publicly held firms Excludes nonresponding firms and owners reporting ‘don’t know’ for level of startup capital ­ erformance found in the previous literature, this information is extremely valuable p for identifying potential barriers to business success Table 8.4 reports estimates for the amount of startup capital used by immigrant-­and nonimmigrant-­owned businesses from specially commissioned tabulations from the US Census Bureau Distributions for startup capital levels are reported because categorical responses were included on the questionnaire instead of write-­in values Immigrant firms are less likely to use low levels of startup capital than are nonimmigrant firms Combining the bottom two categories, 51.7 percent of immigrant firms start with less than $5000 in startup capital compared with 59.1 percent of nonimmigrant firms At the other end of the distribution, immigrant-­ owned firms are more likely to be represented in the highest startup capital levels Nearly 20 percent of immigrant-­owned firms started with $50 000 or more in startup capital compared with 15.9 percent of nonimmigrant-­owned firms The distributional estimates make it clear that immigrant-­owned businesses start with higher levels of startup capital than nonimmigrant-­owned businesses Similar to the patterns found for sales and employment, Hispanic immigrant firms have lower levels of startup capital than the immigrant total and Asian-­immigrant firms have higher levels of startup capital Among Hispanic-­immigrant firms, only 10.3 percent have startup capitals of $50 000 or more Among Asian-­immigrant firms, 29.0 percent have startup capitals of $50 000 or more The finding of relatively high levels of startup capital among Asian-­owned firms supports previous results indicating high startup capital levels (Fairlie and Robb, 2008) Types of Financing Do immigrant businesses differ in the types of financing they use from nonimmigrant businesses? The 2007 SBO includes information on sources of capital used to start or acquire the business and sources of capital to finance expansion or capital improve- Minority and immigrant entrepreneurs  ­163 Table 8.5 Sources of startup capital for immigrant-­and nonimmigrant-­owned firms, special tabulations from the Survey of Business Owners (2007) Amount of startup capital None needed Personal/family savings   of owner(s) Personal/family assets  other than savings of owner(s) Personal/family home  equity loan Personal/business credit  card(s) Business loan from  federal, state or local government Government-­guaranteed  business loan from a bank or financial institution Business loan from  a bank or financial institution Business loan/investment  from family/friends Investment by venture  capitalist(s) Grants Other source(s) of capital All firms Immigrant Nonimmigrant Hispanic immigrant Asian immigrant 21.7% 62.8% 20.5% 65.9% 22.1% 63.3% 23.6% 61.9% 16.2% 68.8% 8.1% 6.4% 8.3% 5.4% 7.8% 5.8% 6.9% 5.6% 6.6% 8.8% 10.8% 11.1% 11.0% 10.8% 11.5% 0.7% 0.6% 0.7% 0.4% 0.8% 0.7% 0.6% 0.7% 0.3% 0.9% 11.1% 8.3% 11.2% 5.7% 11.1% 2.7% 3.2% 2.6% 2.2% 4.7% 0.4% 0.3% 0.3% 0.3% 0.3% 0.4% 2.4% 0.2% 2.0% 0.2% 1.9% S 1.7% 0.2% 2.2% Notes: All firms includes publicly held firms Excludes nonresponding firms and owners reporting ‘don’t know’ for source of startup capital ments for the business Table 8.5 reports sources of startup capital for immigrant-­ and nonimmigrant-­owned businesses from specially commissioned tabulations from the 2007 SBO The reported totals are not restricted to sum to 100 percent because business owners were instructed to ‘mark at all that apply’ among a list of potential sources of startup capital The most common source of startup capital for both immigrant firms and nonimmigrant firms is from personal or family savings Roughly two-­thirds of both immigrant-­ and nonimmigrant-­owned firms report this source of startup capital The second most common source of startup capital used by immigrant businesses is personal or business credit cards (11.1 percent) Another common source of startup capital is a business loan from a bank or financial institution, but immigrant businesses are slightly less likely to use this source than are nonimmigrant businesses (8.3 percent compared with 11.2 percent) Immigrant business owners also commonly use personal 164   International handbook on the economics of migration and family assets (other than savings) and home equity loans to finance business starts Interestingly, only a small share of immigrant-­owned businesses report receiving business loans or investments from family and friends Consistent with higher levels of startup capital, Asian-­immigrant owned firms tend to use all sources of startup capital more often than the immigrant total ­Hispanic-­immigrant firms, in contrast, tend to use less of the reported sources of startup capital Asian-­immigrant firms are similarly likely to rely on business loans from banks or financial institutions for financing startups as are nonimmigrant firms The main finding from these results is that immigrant and nonimmigrant business owners rely on similar sources of startup capital to start their businesses Immigrant-­ owned firms rely heavily on personal and family savings to fund startup activities They also rely heavily on credit cards, bank loans, personal and family assets, and home equity loans These are also the most common sources of financing by nonimmigrant-­owned businesses in the United States In terms of sources of capital used to finance expansions reported in Table 8.6, immigrant-­owned businesses report personal and family savings as the most common source (36.1 percent) Immigrant-­owned businesses also commonly rely on personal and business credit cards and business profits and assets to finance expansions of their businesses The reported totals for sources of capital used for expansion not differ substantially between immigrant-­and nonimmigrant-­owned businesses Home Ownership The single largest asset held by most households is their home More than two-­thirds of families are home owners with a median home equity of $59 000 (US Census Bureau, 2008) Home equity represents 60 percent of all wealth Home equity as well as other forms of personal wealth is important for starting businesses because it can be invested directly in the business or used as collateral to obtain business loans Indeed, previous research indicates that home ownership and equity are found to be associated with entrepreneurship and obtaining business loans using Finnish data (Johansson, 2000), UK data (Black et al., 1996), and US data (Cavalluzzo and Wolken, 2005; Fairlie, 2013; Fairlie and Krashinshy, 2012) Although the SBO does not collect information on the use of home equity as collateral for loans, it does indicate that home equity loans are one of the most common sources of startup capital (see Table 8.5) Unfortunately, the previous literature does not examine whether immigrants and natives differ in rates of home ownership and whether these differences have any impact on differences in rates of business formation To remedy this omission I use data from the CPS to examine home ownership patterns and their relationship with business formation Table 8.7 reports home ownership rates for 2010 for immigrants and the US-­born from the CPS These are the latest data available on immigrant home ownership rates Immigrant rates of home ownership are much lower than the US-­born home-­ownership rate Among immigrants 52.1 percent own a home compared with 70.8 percent of US-­born What impact these differences have on business formation patterns? To investigate this question I first examine the relationship between home ownership and entrepreneurship Using matched CPS data on business formation for 2010, I examine the determi- Minority and immigrant entrepreneurs  ­165 Table 8.6 Sources of expansion capital for immigrant-­and nonimmigrant-­owned firms, special tabulations from the Survey of Business Owners (2007) Sources of expansion capital Personal/family savings   of owner(s) Personal/family assets  other than savings of owner(s) Personal/family home  equity loan Personal/business credit  card(s) Business loan from  federal, state or local government Government-­guaranteed  business loan from a bank or financial institution Business loan from  a bank or financial institution Business loan/investment  from family/friends Investment by venture  capitalist(s) Business profits and/or  assets Grants Other source(s) of capital Did not have access to  capital Did not expand or make  capital improvement(s) All firms Immigrant Nonimmigrant Hispanic immigrant Asian immigrant 31.7% 36.1% 31.6% 33.7% 39.4% 4.5% 4.2% 4.6% 3.7% 4.9% 5.0% 6.0% 4.9% 5.7% 7.1% 13.3% 13.1% 13.6% 12.6% 13.1% 0.5% 0.5% 0.5% 0.5% 0.7% 0.4% 0.4% 0.3% 0.3% 0.5% 9.5% 7.1% 9.7% 5.6% 8.3% 1.1% 1.7% 1.0% 1.2% 2.5% 0.2% 0.2% 0.1% 0.2% 0.2% 11.2% 8.5% 11.4% 7.0% 8.9% 0.4% 1.1% 1.6% 0.2% 1.0% 2.2% 0.2% 0.8% 1.5% 0.2% 0.8% 2.8% 0.3% 1.0% 2.1% 49.0% 45.9% 49.8% 47.8% 41.8% Notes: All firms includes publicly held firms Excludes nonresponding firms and owners reporting ‘don’t know’ for source of expansion capital nants of business formation as defined for Table 8.2 The following logit regression for the probability of entrepreneurship is estimated: Prof(yit51) F(a g1Hit b9Xit lt) (8.1) where yit equals if the individual starts a business by the second or subsequent survey month in the two-­month pair and otherwise, Hit is whether the individual owns his or her home, Xit includes individual characteristics, lt are month fixed effects to control 166   International handbook on the economics of migration Table 8.7 Home ownership rates by immigrant status, Current Population Survey (2010) Total Immigrant US-­born Percentage of home owners Sample size 67.8 52.1 70.8 967 917 137 635 830 282 Notes: The sample consists of individuals aged 20–64 All estimates are calculated using sample weights provided by the CPS for seasonal variation, and F is the cumulative distribution function for the logistic distribution The individual characteristics include gender, race/ethnicity, nativity, age, education, family income, marital status, region, urban status and initial employment status The parameter of interest is g1, which captures the relationship between whether an individual owns a home and entrepreneurship All specifications are estimated with logit regressions using CPS sample weights Marginal effects and their standard errors are reported.10 Marginal effects estimates are similar from probit and linear probability models, and are thus not reported Table 8.8 reports estimates of (8.1) The base specification includes controls for individual characteristics The estimates indicate that women are less likely to become entrepreneurs African-­Americans, Latinos and Asians are also less likely to start businesses, all else being equal Entrepreneurship increases with age and married people are more likely to start businesses The relationship between entrepreneurship and education is not linear Entrepreneurship rates are lower for high school graduates than for high school dropouts (the left out category), but entrepreneurship rates are similar between those with some college and high school graduates College graduates have higher rates of entrepreneurship and those with graduate degrees have the highest rates of entrepreneurship Thus, there is a U-­shaped relationship between entrepreneurship and education.11 Business formation rates tend to decline with total family income, and entrepreneurship rates are higher among the unemployed and those not in the labor force Turning to results for the two main variables of interest to this study, the logit­ estimates indicate that home owners are more likely to start businesses than nonhome owners The coefficient is large, positive and statistically significant Home owners have a 0.034 percentage point higher rate of entrepreneurship than nonhome owners, which is more than 10 percent of the mean rate of entrepreneurship In other words, home owners are roughly 10 percent more likely to start businesses than are nonhome owners, all else being equal In the presence of liquidity constraints, the ability of owners to borrow against the value of their homes may make it easier to finance new business ventures It is unlikely that the home ownership variable is simply picking up current or permanent income effects because the regressions control for family income, education, and unemployment The logit regressions also indicate that immigrants have higher entrepreneurship rates than the native-­born even after controlling for education, family income, region, initial Minority and immigrant entrepreneurs  ­167 Table 8.8 Regressions for probability of entrepreneurship, Current Population Survey (2010) Explanatory variables Female Black Latino Native American Asian Immigrant Age (00s) Age squared Married Previously married High school graduate Some college College graduate Graduate school Family income: $25 000 to $50 000 Family income: $50 000 to $75 000 Family income: $75 000 or more Unemployed Not in the labor force Noncentral city Non-­MSA Central city status not identified South (1) −0.00224 (0.00015) −0.00149 (0.00027) 0.00032 (0.00024) −0.00185 (0.00093) −0.00118 (0.00036) 0.00216 (0.00022) 0.03707 (0.00468) −0.04035 (0.00537) 0.00072 (0.00021) 0.00001 (0.00027) −0.00049 (0.00023) −0.00010 (0.00024) 0.00053 (0.00027) 0.00030 (0.00033) −0.00069 (0.00020) −0.00084 (0.00024) −0.00050 (0.00023) 0.00659 (0.00019) 0.00392 (0.00018) −0.00016 (0.00018) 0.00029 (0.00024) −0.00028 (0.00025) −0.00022 (0.00026) (2) −0.00192 (0.00016) −0.00131 (0.00027) 0.00032 (0.00024) −0.00187 (0.00093) −0.00096 (0.00036) 0.00213 (0.00022) 0.03530 (0.00472) −0.03783 (0.00542) 0.00066 (0.00021) −0.00005 (0.00027) −0.00043 (0.00023) 0.00004 (0.00024) 0.00070 (0.00027) 0.00059 (0.00034) −0.00070 (0.00020) −0.00084 (0.00024) −0.00052 (0.00023) 0.00624 (0.00020) 0.00606 (0.00123) −0.00015 (0.00018) 0.00048 (0.00025) −0.00019 (0.00025) −0.00013 (0.00026) 168   International handbook on the economics of migration Table 8.8 (continued) Explanatory variables Midwest West Home owner Industry controls Mean of dependent variable Sample size (1) 0.00077 (0.00022) 0.00061 (0.00023) 0.00034 (0.00018) No 0.00324 593 271 (2) 0.00072 (0.00022) 0.00058 (0.00023) 0.00038 (0.00018) Yes 0.00324 593 271 Notes: The sample consists of individuals (aged 20–64) who not own a business in the initial survey month of the two-­month pair Marginal effects and their standard errors are reported Additional controls include month dummies employment status, home ownership and other characteristics Immigrants have entrepreneurship rates that are 0.22 percentage points higher than US-­born rates The raw difference in entrepreneurship rates was 0.28 percentage points as reported in Table 8.2 The difference in these findings suggests that controlling for demographic factors and home ownership explains part, but only part, of why immigrants have higher entrepreneurship rates than nonimmigrants Specification of Table 8.8 reports logit regression estimates that include industry controls Industries differ in their propensity for individuals to start businesses and the industrial composition may be related to education, home ownership, immigrant status and other characteristics Construction has the highest rate of business creation followed by professional services The lowest rate of entrepreneurship is found in manufacturing The addition of industry controls, however, has little effect on the results for the immigrant variable It declines slightly from 0.00216 to 0.00213 The home ownership coefficient increases slightly to 0.38 percentage points Industry controls are not included in the main specification because of endogeneity concerns The main issue is that the choice of industry and the choice of starting a business may be simultaneously determined Workers are not constrained to starting businesses in their current industry and may choose different industries depending on the goals of their businesses But, these results provide a useful robustness check of the main results and indicate that the results are not sensitive to industry differences Given that home ownership has a positive effect on entrepreneurship rates it is useful to conduct a simple ‘back-­of-­the-­envelope’ calculation of how much low rates of home ownership restrict the business formation rate of immigrants Home owners have a 0.00034 higher rate of entrepreneurship than nonhome owners and the home ownership rate is 0.19 lower among immigrants than the native-­born (see Table 8.7) The product of these two estimates indicates that immigrant entrepreneurship rates would be 0.00006 higher if immigrants had home ownership rates that were at the same level as the US-­ born In other words, the high rate of business formation among immigrants could be Minority and immigrant entrepreneurs  ­169 slightly higher if they had higher rates of home ownership, which might provide better access to financial capital 4  Conclusions A review of the previous literature indicates that limited access to financial capital is one of the most important determinants of disparities in business creation and performance between minority businesses and nonminority businesses Inadequate access to financial capital is found to be a constraint limiting the creation and growth of minority-­owned businesses Minorities are found to have wealth levels that are an order of magnitude lower than nonminority levels Minority firms are also found to invest substantially less capital at startup Previous research indicates that they pay higher interest rates on loans, are more likely to be denied credit, and are less likely to apply for loans because they fear their applications will be denied There is less evidence in the previous literature on access to capital among immigrant-­owned businesses Some previous evidence indicates a heavier reliance on informal sources of capital among immigrant business owners, but there is some disagreement over the importance of sources such as rotating credit associations To fill this void in the literature, the estimates presented here from the 2007 SBO and 1996–2010 CPS data provide a new picture of capital use among immigrant-­owned businesses in the United States Immigrant-­owned businesses start with higher levels of startup capital than nonimmigrant-­owned businesses Nearly 20 percent of immigrant-­ owned firms started with $50 000 or more in startup capital compared with 15.9 percent of nonimmigrant-­owned firms The most common source of startup capital for immigrant firms is from personal or family savings with roughly two-­thirds of businesses reporting this source of startup capital Other commonly reported sources of startup capital used by immigrant businesses are credit cards, bank loans, personal or family assets, and home equity loans The most commonly reported source of capital used to finance expansions among immigrant-­owned businesses is personal and family savings, followed by credit cards and business profits and assets The sources of startup capital used by immigrant firms not differ substantially from those used by nonimmigrant firms Immigrants, however, are found to have substantially lower rates of home ownership than nonimmigrants These differences in home ownership have implications for business formation rates because regression estimates indicate that home owners are roughly 10 percent more likely to start businesses than are nonhome owners Given low rates of home ownership among immigrants, business formation could be slightly higher if they had rates of home ownership more similar to nonimmigrants The findings presented in this chapter contribute to our understanding of the use of financial capital among minority-­ and immigrant-­owned businesses, but more research is needed Uncovering additional barriers to capital access is especially important because of the potential contribution of minority-­ and immigrant-­owned businesses to their economies Although minority-­ and immigrant-­owned businesses already contribute greatly to their economies, there remains a lot of untapped potential for creating jobs and fostering innovation Barriers to entry and expansion are potentially costly to 170   International handbook on the economics of migration ­ roductivity, especially because minorities and immigrants represent a growing share of p the ­population in many developed countries NOTES   *                   10 11 I thank the editors, Amelie F Constant and Klaus F Zimmermann, and an anonymous referee for providing comments and suggestions to improve the chapter Immigrants are often, but not always, represented by minority racial and ethnic groups Minority groups are sometimes immigrants, but not always (for example, second generation born in host country) and in some cases rarely (for example, African-­Americans) Constraints to successful business ownership also may limit assimilation and integration of immigrants in their host countries (see Chapter in this volume) Interestingly, immigrants are found to have lower business ownership rates than non­immigrants in Germany (Constant and Zimmermann, 2006; Constant, et al., 2007) For example, see Evans and Jovanovic (1989), Evans and Leighton (1989), Meyer (1990), Holtz-­Eakin et al (1994), Lindh and Ohlsson (1996, 1998), Black et al (1996), Blanchflower and Oswald (1998), Dunn and Holtz-­Eakin (2000), Fairlie (1999), Earle and Sakova (2000), Johansson (2000), Taylor (2001), Holtz-­Eakin and Rosen (2005), Hurst and Lusardi (2004), Fairlie and Krashinsky (2012), Zissimopoulos and Karoly (2007), Zissimopoulos et al (2009), Giannetti and Simonov (2004), Nykvist (2005), Bates and Lofstrom (2013), Schäfer et al (2011), Yu (2010), and Paulson and Townsend (2004) Also, see Parker (2009), Kerr and Nanda (2011), and Fairlie and Krashinsky (2012) for recent discussions of the literature Astebro and Berhardt (2003) find a positive relationship between business survival and having a bank loan at startup after controlling for owner and business characteristics See Blanchard et al (2004), Blanchflower et al (2003), Cavalluzzo et al (2002), Cavalluzzo and Wolken (2005), Coleman (2002, 2003) and Mitchell and Pearce (2004) Estimates of business ownership and outcomes from these sources that are presented here are generated from special tabulations of confidential data, public-­release microdata and published sources from the US Census Bureau and Bureau of Labor Statistics More details about the data sources are provided in the Data Appendix Conditional on two groups having similar business ownership rates, the only way that one group can have a higher business entry rate is if it also has a higher business exit rate (see Fairlie, 2006, and Fairlie and Robb, 2008, for more discussion) The 2007 SBO microdata are not publicly available (see Data Appendix) The reported marginal effect provides an estimate of the effect of a 1-­unit increase in the independent varb (1 eXiˆb ) iable on the self-­employment entry probability It equals the sample average of eXiˆ/ See van der Sluis et al (2005) and Van Praag (2005) for reviews of the evidence on the relationship between education and entrepreneurship References Astebro, T and I Bernhardt (2003), ‘Start-­up financing, owner characteristics and survival’, Journal of Economics and Business, 55 (4), 303–20 Avery, R.B., R.W Bostic and K.A Samolyk (1998), ‘The role of personal wealth in small business finance’, Journal of Banking and Finance, 22 (6), 1019–61 Bates, Timothy (1997), Race, Self-­Employment & Upward Mobility: An Illusive American Dream, Washington, DC: Woodrow Wilson Center Press, and Baltimore, MD: Johns Hopkins University Press Bates, Timothy (2005), ‘Financing disadvantaged firms’, in Patrick Bolton and Howard Rosenthal (eds), Credit Markets for the Poor, New York: Russell Sage Foundation, pp. 149–78 Bates, T and W.D Bradford (2009), ‘Venture-­capital investment in minority business’, Journal of Money Credit and Banking, 40 (2–3), 489–504 Bates, Timothy and Magnus Lofstrom (2013), ‘African American pursuit of self-­employment’, Small Business Economics, 40 (1), 73–86 Becker, Gary (1957), The Economics of Discrimination, republished in 1971, Chicago, IL: University of Chicago Press Minority and immigrant entrepreneurs  ­171 Black, J., D de Meza and D Jeffreys (1996), ‘House prices, the supply of collateral and the enterprise economy’, Economic Journal, 106 (434), 60–75 Blanchard, Lloyd, John Yinger and Bo Zhao (2004), ‘Do credit market barriers exist for minority and women entrepreneurs?’, Working Paper No 74, Syracuse University Center for Policy Research, New York Blanchflower, D.G and A.J Oswald (1998), ‘What makes an entrepreneur?’, Journal of Labor Economics, 16 (1), 26–60 Blanchflower, D.G., P.L Levine and D Zimmerman (2003), ‘Discrimination in the small business credit market’, Review of Economics and Statistics, 85 (4), 930–43 Borjas, G (1986), ‘The self-­employment experience of immigrants’, Journal of Human Resources, 21 (4), 487–506 Bostic, R and K.P Lampani (1999), ‘Racial differences in patterns of small business finance: the importance of local geography’, in Jackson L Blanton, Alicia Williams and Sherrie L.W Rhine (eds), Business Access to Capital and Credit: A Federal Reserve System Research Conference: Proceedings of a Conference Held in Arlington, VA, March 8–9, 1999, Chicago, Il: Federal Reserve Bank of Chicago, pp. 149–79 Cavalluzzo, Ken and John Wolken (2005), ‘Small business loan turndowns, personal wealth and discrimination’, Journal of Business, 78 (6), 2153–77 Cavalluzzo, K., L Cavalluzzo and J Wolken (2002), ‘Competition, small business financing, and discrimination: evidence from a new survey’, Journal of Business, 75, 641–79 Clark, K and S Drinkwater (2000), ‘Pushed out or pulled in? Self-­employment among ethnic minorities in England and Wales’, Labour Economics, (5), 603–28 Clark, K and S Drinkwater (2010), ‘Patterns of ethnic self-­employment in time and space: evidence from British census microdata’, Small Business Economics, 34 (3), 323–38 Coleman, S (2002), ‘The borrowing experience of black and Hispanic-­owned small firms: evidence from the 1998 Survey of Small Business Finances’, The Academy of Entrepreneurship Journal, (1), 1–20 Coleman, Susan (2003), ‘Borrowing patterns for small firms: a comparison by race and ethnicity’, The Journal of Entrepreneurial Finance & Business Ventures, (3), 87–108 Constant, A.F and K.F Zimmermann (2006), ‘The making of entrepreneurs in Germany: are native men and immigrants alike?’, Small Business Economics, 26 (3), 279–300 Constant, A.F., Y Shachmurove and K.F Zimmermann (2007), ‘What makes an entrepreneur and does it pay? Native men, Turks, and other migrants in Germany’, International Migration, 45 (4), 69–98 Demiralp, Berna and Johanna Francis (2008), ‘Wealth, human capital, and the transition to entrepreneurship’, Department of Economics Discussion Paper No 2008-­09, Fordham University, New York Dunn, T.A and D.J Holtz-­Eakin (2000), ‘Financial capital, human capital, and the transition to self-­ employment: evidence from intergenerational links’, Journal of Labor Economics, 18 (2), 282–305 Earle, J.S and Z Sakova (2000), ‘Business start-­ups or disguised unemployment? Evidence on the character of self-­employment from transition economies’, Labour Economics, (5), 57–601 Evans, D and B Jovanovic (1989), ‘An estimated model of entrepreneurial choice under liquidity constraints’, Journal of Political Economy, 97 (4), 808–27 Evans, D and Linda Leighton (1989), ‘Some empirical aspects of entrepreneurship’, American Economic Review, 79 (3), 519–35 Fairlie, Robert W (1999), ‘The absence of the African-­American owned business: an analysis of the dynamics of self-­employment’, Journal of Labor Economics, 17 (1), 80–108 Fairlie, Robert W (2006), ‘Entrepreneurship among disadvantaged groups: women, minorities and the less educated’, in Simon C Parker (ed.), The Life Cycle of Entrepreneurial Ventures, International Handbook Series on Entrepreneurship, vol 3, New York: Springer, pp. 437–75 Fairlie, Robert W (2008), Estimating the Contribution of Immigrant Business Owners to the U.S Economy, Washington, DC: US Small Business Administration, Office of Advocacy Fairlie, R.W (2013), ‘Entrepreneurship, economic conditions, and the great recession’, Journal of Economics and Management Strategy, 22 (2), 207–31 Fairlie, R.W and H.A Krashinsky (2012), ‘Liquidity constraints, household wealth, and entrepreneurship revisited’, Review of Income and Wealth, 58 (2), 279–306 Fairlie, Robert W and Alicia M Robb (2008), Race and Entrepreneurial Success: Black-­, Asian-­, and White-­ Owned Businesses in the United States, Cambridge, MA: MIT Press Fairlie, R.W and C Woodruff (2010), ‘Mexican-­American entrepreneurship’, The Berkeley Electronic Journal of Economic Analysis & Policy: Contributions, 10 (1), art 10 Fairlie R.W., J Zissimopoulos and H.A Krashinsky (2010), ‘The international Asian business success story? A comparison of Chinese, Indian and other Asian businesses in the United States, Canada and United  Kingdom’, in Josh Lerner and Antoinette Schoar (eds), International Differences in Entrepreneurship, Chicago, IL: University of Chicago Press and National Bureau of Economic Research, pp. 179–208 172   International handbook on the economics of migration Giannetti, M and A Simonov (2004), ‘On the determinants of entrepreneurial activity: social norms, economic environment and individual characteristics’, Swedish Economic Policy Review, 11 (2), 271–313 Holtz-­Eakin, Douglas and Harvey Rosen (1999), ‘Cash constraints and business start-­ups: Deutschmarks versus dollars’, Center for Policy Research Working Paper No 11, Syracuse University, New York Holtz-­Eakin, Douglas and Harvey S Rosen (2005), ‘Cash constraints and business start-­ups: Deutschmarks versus dollars’, Berkeley Electronic Journals, Contributions to Economic Analysis & Policy, (1), art Holtz-­Eakin, D., D Joulfaian and H Rosen (1994), ‘Entrepreneurial decisions and liquidity constraints’, RAND Journal of Economics, 25 (2), 334–47 Hunt, J., and M Gauthier-­Loiselle (2010), ‘How much does immigration boost innovation?’, American Economic Journal: Macroeconomics, (2), 31–56 Hurst, E and A Lusardi (2004), ‘Liquidity constraints, household wealth, and entrepreneurship’, Journal of Political Economy, 112 (2), 319–47 Johansson, E (2000), ‘Self-­employment and liquidity constraints: evidence from Finland’, Scandinavian Journal of Economics, 102 (1), 123–34 Kerr, W.R and W.F Lincoln (2010), ‘The supply side of innovation: H-­1B visa reforms and U.S ethnic invention’, Journal of Labor Economics, 28 (3), 473–508 Kerr, William R and Ramana Nanda (2011), ‘Financing constraints and entrepreneurship’, in David B Audretsch, Oliver Falck, Stephan Heblich and Adam Lederer (eds), Handbook on Research on Innovation and Entrepreneurship, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing, pp. 88–103 Light, I., I Jung Kwuon and D Zhong (1990), ‘Korean rotating credit associations in Los Angeles’, Amerasia Journal, 16 (2), 35–54 Lindh T and H Ohlsson (1996), ‘Self-­employment and windfall gains: evidence from the Swedish lottery’, Economic Journal, 106 (439), 1515–26 Lindh, T and H Ohlsson (1998), ‘Self-­employment and wealth inequality’, Review of Income and Wealth, 44 (1), 25–41 Lofstrom, M (2002), ‘Labor market assimilation and the self-­employment decision of immigrant entrepreneurs’, Journal of Population Economics, 15 (1), 83–114 Lofstrom, M and C Wang (2009), ‘Hispanic self-­employment: a dynamic analysis of business ownership’, Research in Labor Economics, 29, 197–227 Meyer, Bruce (1990), ‘Why are there so few black entrepreneurs?’, NBER Working Paper No 3537, National Bureau of Economic Research (NBER), Cambridge, MA Minority Business Development Agency (2008), Characteristics of Minority Businesses and Entrepreneurs: An Analysis of the 2002 Survey of Business Owners, Washington, DC: US Department of Commerce, Minority Business Development Agency Mitchell, K and D.K Pearce (2004), Availability of Financing to Small Firms using the Survey of Small Business Finances, Washington, DC: United States Small Business Administration, Office of Advocacy Nykvist, Jenny (2005), ‘Entrepreneurship and liquidity constraints: evidence from Sweden’, Working Paper 2005:21, Department of Economics, Uppsala University Nykvist, J (2008), ‘Entrepreneurship and liquidity constraints: evidence from Sweden’, The Scandinavian Journal of Economics, 110 (1), 23–43 Parker, Simon C (2009), The Economics of Entrepreneurship, Cambridge: Cambridge University Press Paulson, A.L and R Townsend (2004), ‘Entrepreneurship and financial constraints in Thailand’, Journal of Corporate Finance, 10 (2), 229–62 Saxenian, AnnaLee (1999), Silicon Valley’s New Immigrant Entrepreneurs, San Francisco, CA: Public Policy Institute of California Saxenian, AnnaLee (2000), ‘Networks of immigrant entrepreneurs’, in Chong-­Moon Lee, William F Miller, Henry S Rowen and Marguerite Gong Hancock (eds), The Silicon Valley Edge: A Habitat for Innovation and Entrepreneurship, Stanford, CA: Stanford University Press, pp. 248–75 Schäfer, D., O Talavera and C Weir (2011), ‘Entrepreneurship, windfall gains and financial constraints: evidence from Germany’, Economic Modelling, 28 (5), 2174–80 Schuetze, Herbert J and Heather Antecol (2006), ‘Immigration, entrepreneurship and the venture start-­up process’, in Simon C Parter (ed.), The Life Cycle of Entrepreneurial Ventures, International Handbook Series on Entrepreneurship, vol 3, New York: Springer, pp. 107–35 Taylor, M (2001), ‘Self-­employment and windfall gains in Britain: evidence from panel data’, Economica, 68 (272), 539–65 US Census Bureau (1997), 1992 Economic Census: Characteristics of Business Owners Washington, DC: US Government Printing Office US Census Bureau (2008), Home Ownership Statistics, Washington, DC: US Department of Commerce, US Census Bureau US Census Bureau (2011), Median Value of Assets for Households, by Type of Asset Owned and Selected Minority and immigrant entrepreneurs  ­173 Characteristics: 2004, Washington, DC: US Department of Commerce, US Census Bureau, Housing and Household Economic Statistics Division van der Sluis, J., M van Praag and W Vijverberg (2005), ‘Education and entrepreneurship in industrialized countries: a meta-­analysis’, World Bank Economic Review, 19 (2), 225–61 Van Praag, Mirjam (2005), Successful Entrepreneurship: Confronting Economic Theory with Empirical Evidence, Cambridge, UK and Northampton, MA, USA: Edward Elgar Publishing Wadhwa, Vivek, AnnaLee Saxenian, Ben Rissing and Gary Gereffi (2007), America’s New Immigrant Entrepreneurs, Durham, NC: Master of Engineering Management Program, Duke University, and Berkeley, CA: School of Information, University of California Berkeley Yago, Glenn and Aaron Pankrat (2000), The Minority Business Challenge: Democratizing Capital for Emerging Domestic Markets, research report, Santa Monica, CA: Milken Institute, and Washington, DC: US Department of Commerce, Minority Business Development Agency Yoon, I.-­J (1991), ‘The changing significance of ethnic and class resources in immigrant business’, International Migration Review, 25 (2), 303–31 Yu, Chen (2010), ‘Entrepreneurship and credit constraints: evidence from rural households in China’, Chinese Academy of Science Working Paper Issue 2008.4.20-­24, Chinese Academy of Science, Institute of Policy and Management, Bejing Zissimopoulos, J and L Karoly (2007), ‘Transitions to self-­employment at older ages: the role of wealth, health, health insurance, and other factors’, Labour Economics, 14 (2), 269–95 Zissimopoulos, Julie, Lynn Karoly and Qian Gu (2009), ‘Liquidity constraints, household wealth, and self-­ employment: the case of older workers’, RAND Working Paper WR-­725, Santa Monica, CA 174   International handbook on the economics of migration Data Appendix The two main sources of data used in the study are the 2007 Survey of Business Owners (SBO) and the Current Population Survey (CPS) The SBO is conducted by the US Census Bureau every five years to collect statistics that describe the composition of US businesses by gender, race, and ethnicity This survey was previously conducted as the Survey of Minority-­ and Women-­Owned Business Enterprises (SMOBE/SWOBE) and Characteristics of Business Owners (CBO) The universe for the most recent survey is all firms operating during 2007 with receipts of $1000 or more that filed tax forms as individual proprietorships, partnerships, or any type of corporation The SMOBE, CBO and SBO data have undergone several major changes over time including the addition of C corporations and the removal of firms with annual receipts between $500 and $1000 starting in 1997 (see Fairlie and Robb, 2008, for more details) The most important change for 2007, however, was the reintroduction of questions on the immigrant status of the owner and the level of startup capital used by the business These two variables had not been asked since the 1992 CBO The 2007 SBO also includes new information on export levels that was not available in the 1992 CBO It provides information for roughly 2.5 million firms compared with 75 000 firms in the 1992 CBO The 2007 SBO provides the most comprehensive data available on businesses by detailed owner characteristics including nativity, race, ethnicity and gender In addition to immigrant status, information is available on whether the owners are male, female, white, Hispanic origin, African-­American, Asian or Native American The public-­use tables from the SBO/SMOBE are the most widely used source for tracking the number, performance, size, and industry composition of minority-­owned businesses in the United States Although microdata from the SBO are not publically available and require an extensive application and disclosure process prohibiting their use for this study, a few publications reporting the data are available on the SBO website at http://www.census.gov/econ/ sbo/ Most of the reported tables, however, were acquired through specially commissioned tabulations purchased from the US Census Bureau These special runs from the US Census Bureau allow very detailed analyses of the data by immigrant status The second dataset used for the study is the 1996 to 2010 CPS The CPS, conducted monthly by the US Bureau of the Census and the US Bureau of Labor Statistics, is representative of the entire US population and contains observations for more than 130 000 people for each monthly survey Although the CPS is most commonly used as a cross-­sectional dataset, longitudinal data can be created by linking CPS files over time, which allows for the examination of business creations The process of creating longitudinal or panel data takes advantage of the fact that households in the CPS are interviewed each month over a four-­month period Eight months later they are re-­interviewed in each month of a second four-­month period Thus, individuals who are interviewed in March of one year are interviewed again in March of the following year The CPS provides detailed information on the immigrant status, race, ethnicity, gender, age, education level and home ownership of the owner The CPS provides the only dataset in which business formation can be examined for immigrant groups because of the need for very large sample sizes and panel data Another useful feature of the CPS data is its timely release The basic monthly CPS Minority and immigrant entrepreneurs  ­175 data for all of 2010 are available No other large-­scale, nationally representative microdata set is released as quickly as the CPS The timeliness of the data is extremely useful for examining the impact of the recent recession and financial crisis on access to financial capital among immigrant entrepreneurs 9  Migrant educational mismatch and the labor market* Matloob Piracha and Florin Vadean 1  INTRODUCTION Labor market mismatch, particularly ‘over-­education’,1 has a long and controversial history in the labor economics literature Freeman (1976), who argued that an oversupply of university-educated individuals in the US since the start of 1970s had resulted in the fall in return to education, set the scene for further research on the topic Even though Freeman’s claims were challenged in a number of papers in subsequent years and the issue seemed to have been resolved with Smith and Welch (1978) declaring that ‘at best Freeman exaggerates the case for an oversupply of college-­educated manpower and that he may in fact be dead wrong’, the revival came in a paper by Duncan and Hoffman (1981) Unlike the previous literature which used aggregate data, Duncan and Hoffman used individual level data and compared those who were properly matched, that is, had the required level of education, with those who had either less or more education than their job required They found that there is indeed some ‘misallocation of education resources’ With this paper a subfield of economics of over-­education was born.2 In the past two decades a number of papers have extended the labor economics literature to analyze the incidence and effects of education mismatch for immigrants The emphasis in the migration literature has been to compare the extent of over-­education between natives and immigrants This research has typically focused on the formal education qualifications of migrants and has compared their possible labor market mismatch, and in some cases the labor market mismatch of ethnic minorities, with that of natives There is an almost universal consensus in the literature that immigrants are often more over-­educated than their native counterparts and researchers have forwarded different explanations for this disparity.3 These range from imperfect transferability of human capital across borders – owing to language as well as cultural and economic dissimilarities between home and host countries – to innate ability of immigrants to ­discrimination in the labor market This chapter presents the analysis and findings from this literature and is structured as follows Section addresses the measurement issues of educational mismatch and discusses the relevant theoretical explanations of the phenomenon from labor as well as migration economics literature Section then presents some stylized facts on the incidence of immigrants’ over-­ and under-­education and discusses empirical findings on its determinants and its impact on wages Section explores the state of current research and offers some suggestions for further research in this area The last section concludes the chapter 176 Migrant educational mismatch and the labor market  ­177 2  THE ECONOMICS OF EDUCATIONAL MISMATCH One of the key aspects of studying educational mismatch in labor markets is the way mismatch is defined/measured We therefore first explain the various ways it has been measured in the literature and then discuss some of the theoretical explanations for education–­occupation mismatch to occur Definition and Measurement Issues Educational mismatch occurs when the required level of education for a particular job diverges from the employee’s attained level of education.4 The level of attained education could be higher than needed for the job, in which case the worker is over-­educated, or lower than required, in which case the worker is under-­educated.5 There are three different ways in which the divergence from the required level of education has been measured in the literature: the first method is based on information included in job descriptions, the second method relies on workers assessments about the schooling requirements for the job they perform, while the third is a statistical method that uses data on realized matches The job analysis method relies on the information contained in the occupational classification documents like the Dictionary of Occupational Titles (DOT), which is based on a scale of to This scale is then translated into number of years of schooling from to 18 Even though this is an ‘objective’ way to measure a mismatch and has been used by a number of authors,6 there are certain disadvantages as well Some of the criticisms raised are the inability of the method to capture the dynamic nature of job structure as very little new information is added to the DOT on a regular basis as well as the problems resulting in measurement errors when translating job requirements into a single schooling variable (see Hartog, 2000) However, the advantage is that the education level is linked to a particular classification of occupations and is therefore relatively less ­subjective than the other methods Worker self-­assessment is the most subjective of the three measures of educational mismatch and has the advantage of drawing on the current available knowledge Workers are asked about the required education level of their job.7 However, the relevant questions included in various survey questionnaires differ significantly from each other and, as demonstrated by Green et al (1999), may lead to different answers The answers obtained may be biased if, for example, employees associate higher status with jobs for which more education is required In addition, even though the information may be the most up to date when asked of a recent employee, the standards/requirement are likely to have changed over time which will not have affected someone hired prior to, say, an increase in the education levels required for the same job (Hartog, 2000) This bias could distort the extent of over-­education (see Sicherman, 1991) Finally, realized matches method has been utilized using two similar approaches One, proposed by Verdugo and Verdugo (1989), is based on the mean level of schooling obtained from those who are working in the same occupations: any workers whose educational level is at least one standard deviation above the mean are deemed over-­ educated, whereas those with one standard deviation below the mean are considered under-­educated The other approach, proposed by Kiker et al (1997), is a variant of 178   International handbook on the economics of migration Verdugo and Verdugo in that it uses the mode, instead of the mean, of the acquired schooling for workers in the same occupation and does not use the two standard deviations interval around the centralized measure Workers with education level more or less than the modal value are considered over-­ or under-­educated, respectively The main criticism of this method is that the realized match does not reflect only requirements, but is the result of labor market supply and demand This method has been used by, among others, Chiswick and Miller (2008, 2009, 2010a, 2010b, 2010c) and, with slight variations, Quinn and Rubb (2006) and Battu and Sloane (2002, 2004) Reasons for Educational Mismatches There is no obvious reason for an educational mismatch to occur, especially if the labor markets are assumed to function efficiently However, the recent data shows that a substantial number (up to 50 percent) of employed are mismatched (see Leuven and Oosterbeek, 2011) So what are the possible reasons for this divergence from the ‘norm’? Education–occupation mismatch is a dynamic process that is theorized to be affected by the individual’s experience in the labor market Imperfect information, for instance, is one reason why the resultant mismatch might occur as the lack of information from both employer and potential employee perspectives cause frictions within the search-­ and-­match context Workers at the beginning of their career, therefore, might settle for relatively low-skilled jobs in the hope that they can engage in on-­the-­job search in pursuit to climb the occupational ladder (see Dolado et al., 2009; Gautier, 2002; Groot and Maassen van den Brink 2000) This ‘information-­adjustment’ model is equally relevant for immigrants as the initial search-­and-­match cost is relatively higher for those looking for jobs from either outside the country or soon after immigration, as they learn the new labor market structure (see Chiswick and Miller, 2009) Hence, in the initial stages of settling down, immigrants are likely to take up employment in jobs that not match their attained education but then search for a better match while employed With residence length and the accumulation of information about the host country labor market, the incidence of educational mismatch is likely to fall Sicherman and Galor (1990) have used the career mobility/human capital argument to explain the education–occupation mismatch Their model builds on the notion that workers with a given i­nnate ability may prefer to take a job that requires less education than what they have obtained in the understanding that the cost incurred in accepting a low-­pay job in the initial stages will be compensated by the much higher probability of rapid promotion in the future This theory has not been supported by empirical analysis (see Sicherman, 1991) and hence has not been used extensively in the labor economics literature; and to our knowledge has never been used in the migration literature However, there is some scope that a variant of the human capital argument could be developed to analyze educational mismatch for immigrants The above reasons for a possible educational mismatch apply well to both natives and immigrants However, there are some further reasons specific to immigrants only For instance, ethnic minorities (for example, both first and later generation immigrants) might be discriminated against in the labor market.8 Consequently, they would need to obtain more education for the same job to ‘counter’ the discrimination effect Following Chiswick (1978), it is generally argued that immigrants are positively self-­ Migrant educational mismatch and the labor market  ­179 selected and, therefore, the average educational/ability level of those who arrive from abroad is likely to be higher than that of natives.9 However, the jobs they are employed in often require lower levels of education compared to their actual qualifications More importantly, this over-­education incidence is more ‘severe’ than that of natives, that is, immigrants are more over-­educated than natives for similar kind of jobs Chiswick and Miller (2009) have argued that one of the main reasons for immigrant over-­education is the imperfect transferability of human capital across borders This could range from lack of language skills to not being familiar with host country labor regulations to requirement of licensing for some professions Given that there are likely to be significant differences in labor market structure between origin and host countries (for example, employment opportunities and hiring mechanisms), especially if migration is from a less developed country to an industrialized one, immigrants would be more likely to be over-­educated shortly after arrival Then, once they had worked in the host country for some time, they would eventually move into jobs that better correspond to their education and skills Chiswick and Miller also found that the immigrant mismatch in the destination country is likely to be exacerbated if the individual accumulates a greater amount of work experience in the home country.10 An implicit assumption in this argument is that there was no education mismatch in the immigrants’ country of origin In other words, it is assumed that the professional experience gained prior to immigration was in jobs requiring exactly the education level obtained from formal schooling Piracha et al (2012) diverge from the existing literature and argue that it is not only the education signal that determines the incidence of mismatch, but the signal from previous work experience is equally, if sometimes not more, important as an explanation for a mismatch in the host country In other words a mismatch experience in the country of origin might significantly determine the immigrants’ education mismatch probability in the host country, especially since the education signal attenuates with work experience (see Belman and Heywood, 1997) Those who were working in a job that required less education than they acquired are likely to be assessed by the host country employer to be of a lower ability than their education might show and, hence, are hired accordingly A simple, and therefore perhaps crude, example of this is that if an individual with an engineering degree drove a taxi in his home country before migration, then he is not likely to be hired as an engineer after migrating to another country The mismatch in this case is less likely due to discrimination and/or imperfect transferability of human capital and more likely due to the lower on-­the-­job skills accumulated and/or some other unobservable factors (for example, ability, motivation, ambition and/or energy) If, however, the individual was properly matched in the home labor market but is over-­educated in the host country, then perhaps the existing explanations of imperfect skill transferability and/or discrimination could be put forward for such an outcome 3  WHAT DO WE KNOW? The Incidence of Over-­and Under-­Education Among Immigrants Most studies indicate that the incidence of education–occupation mismatch is higher among immigrants compared with natives: the percentage of correctly matched 180   International handbook on the economics of migration i­mmigrant employees is, for example, about 5.0 percentage points lower compared with native employees in Denmark and reaches up to 15.6 percentage points in the USA (exceptions are Finland and Italy, where the mismatch incidence seems to be higher for natives; see Table 9.1) The percentage of mismatched immigrants differs from country to country with respect to the measurement method employed, the immigrant group, gender as well as the residence length The incidence of over-­education, thus, ranges from 13.2 percent in the case of ethnic Bangladeshi in the UK to 58.1 percent in the case of female immigrants resident in New Zealand for less than five years On the other hand, under-­education is less frequent and ranges from 5.4 percent in the case of ethnic Indians in the UK to 44.7 percent in the case of immigrants in the USA Different measurement methods often lead to significantly different estimates of incidence rates According to Leuven and Oosterbeek (2011), studies based on the self-­ assessment and job analysis methods not lead to large differences in the estimated incidence of mismatch However, the realized matches procedure based on the mean level of schooling has generally led to lower estimated levels of over-­education.11 On the other hand, Sanroma et al (2008) found that mismatch is more frequent when self-­reported rather than when objective measures are used He argues that this is so because employees are often biased in reporting that they perform tasks above their education level Compared with the statistically computed match rate for employees of Spanish companies – 34.3 percent for immigrants and 39.1 percent for natives – only 13.1 and 18.2 percent respectively perceived themselves to be employed in a job that matched their education level (see Table 9.1) The level of education mismatch in an immigrant group is often strongly linked to the definition of over-­/under-­education as well For example, the low over-­education rate (13.2 percent) for ethnic Bangladeshi in the UK in 1993/94 was primarily due to the low education level of the group (about 53 percent had no qualifications; see Battu and Sloane, 2002) Individuals with a low education level can, by definition, not work below their education level Contrarily, migrant groups with high average education are more likely to be over-­educated and less likely to be under-­educated, as were, for example, the ethnic Indian and Chinese The education mismatch incidence is also dependent on immigrants’ characteristics and background Poot and Stillman (2010) show that recent female migrants in New Zealand had about 6.4 percentage points higher over-­education rate and a 6.1 percentage points lower under-­education rate compared with their male counterparts Immigrants originating from countries with a similar language to that of the host country seem to have a significantly better matching rate As illustrated by Green et al (2007), immigrants in Australia coming from an English-­speaking background had about 10 percentage points lower over-­education rates compared with Asian immigrants Mismatch Dynamics The dynamics of education mismatch have received little attention compared with its determinants and wage effects that we review in the next two sections The few studies that touched the issue showed that there is important persistence in education mismatch (see Mavromaras et al., 2009) Simultaneously, however, there are significant labor market integration dynamics at work In New Zealand, for example, the incidence of Migrant educational mismatch and the labor market  ­181 Table 9.1  The incidence of educational mismatch among immigrants and natives Over­educated Correctly matched Under-­ educated Measurement method; year Natives Immigrants Natives Immigrants 11.0 16.3 32.1 27.5 71.1 66.1 43.4 27.8 17.9 15.6 24.5 44.7 RM3 (mode);  1995–2002 RM3 (mode);  2000 Natives Immigrants Natives Immigrants Whites Non-whites Caribbean Indian African-­Asian Pakistani Bangladeshi Chinese Native males Recent male migrants1 Earlier male migrants Native females Recent female migrants1 Earlier female migrants Natives ESB2 immigrants  (5 months) ESB2 immigrants  (17 months) ESB2 immigrants  (41 months) Asian immigrants  (5 months) Asian immigrants  (17 months) Natives Immigrants 31.8 30.3 29.7 39.4 19.7 24.0 16.3 33.0 33.2 16.8 13.2 30.8 36.3 51.7 41.0 34.1 58.1 42.9 7.4 21.3 18.2 13.1 39.1 34.3 70.3 66.9 77.6 61.6 59.8 74.0 50.9 56.8 44.2 33.6 40.3 48.0 33.2 43.4 50.0 56.6 31.2 26.3 9.9 9.1 6.1 5.4 7.1 9.2 35.9 12.3 19.5 14.7 18.7 17.9 8.6 13.7 WA;4 2000 21.4 24.3 70.5 71.5 8.0 4.2 Natives Immigrants Natives Immigrants Natives Immigrants 15.6 23.6 18.9 44.6 26.3 26.9 72.4 67.5 77.8 54.3 68.6 69.7 12.0 8.9 3.3 1.0 5.1 3.4 Country (reference) Population group Denmark   (Nielsen, 2011) United States  (Chiswick and Miller, 2009) Spain  (Sanroma et al., 2008) United Kingdom  (Battu and Sloane, 2002) New Zealand  (Poot and Stillman, 2010) Australia6  (Green et al., 2007) EU-­15  (Tijdens and van Klaveren, 2011) Belgium Denmark Finland RM3 (mode);  2000 RM3 (mode);  1993/94 RM3 (mode);  1996, 2001, and 2006 JA;5 1993–96 27.4 18.9 32.0 37.9 WA;4 2005–10 182   International handbook on the economics of migration Table 9.1  (continued) Country (reference) Population group France Natives Immigrants Natives Immigrants Natives Immigrants Natives Immigrants Italy Netherlands Sweden Over­educated Correctly matched Under-­ educated 13.2 22.3 19.5 17.1 18.1 26.1 20.7 23.3 80.8 75.6 68.1 77.5 68.9 63.6 76.2 72.8 6.0 2.1 12.3 5.4 13.1 10.3 3.1 3.9 Measurement method; year Notes: Recent migrants resided in New Zealand for less than five years ESB – English-­speaking background RM – realized matches procedure WA – worker self-­assessment procedure JA – job analysis procedure Green et al estimated the incidence for over-­educated only over-­education decreased after five years of residence for both male (from 51.7 to 41.0 percent) and female (from 58.1 to 42.9 percent) immigrants, while the rates of correctly matched and under-­educated immigrant employees rose accordingly (see Table 9.1; Poot and Stillman, 2010) Newly arrived immigrants are usually employed in jobs that are below their education level since host country employers are generally not properly able to assess foreign qualifications However, once employed for a period of time, skilled immigrants have the opportunity to prove their abilities on the job and to climb the occupation ladder Two dynamics are of particular interest in the case of immigrants: (1) education mismatch transitions between home and host country and (2) education mismatch transitions in the host labor market To our knowledge the only paper that captures both these dynamics is Piracha et al (2012) Using Australian data and a job analysis method to measure education mismatch, Piracha et al illustrate the persistence in education–occupation mismatch The biggest share of immigrants who were over-­ educated in their last job in the home country (41.1 percent) remained over-­educated in their first job in Australia, while only 19.4 percent of them found a job to match their qualification within five months of arrival (see first part of Table 9.2) However, from those who were correctly matched in their last job in the home country, almost 60 percent found jobs at their education level immediately after arrival, while over 63 percent of those who worked above their education level (that is, were under-­educated) before migration found jobs at or above their education level within five months of arrival in Australia The second part of Table 9.2 illustrates that despite educational mismatch persistence, about 30 percent of those initially over-­educated five months after immigration managed to find a job to match their qualification within one more year of residence The labor market integration dynamics have been also confirmed by Huber et al Migrant educational mismatch and the labor market  ­183 Table 9.2  Transition matrix of education mismatch Education mismatch in home country Over-­educated Correctly matched Under-­educated Education mismatch in Australia – months after arrival Over-­educated Correctly matched Under-­educated Education mismatch in Australia – months after arrival Unemployed Over-­educated Correctly matched Under-­educated Total 39.1 26.5 33.9 41.1 12.0   2.8 19.4 59.2 17.0   0.4   2.3 46.3 100 100 100 Education mismatch in Australia – 17 months after arrival Unemployed Over-­educated Correctly matched Under-­educated Total   6.1   3.0   4.5 61.5   6.5   1.2 30.1 86.1 15.7   2.3   4.4 78.6 100 100 100 Source:  Piracha et al (2012); data are from the Longitudinal Survey of Immigrants to Australia (LSIA); pooled Cohort and 2; male immigrants only Table 9.3 Share of over-­educated immigrants aged 151 by skill level, country of residence and duration of stay Receiving country EU-­15 Austria Belgium Denmark Spain France Greece Italy Luxembourg Netherlands Portugal Sweden UK Less than 10 years of residence More than 10 years of residence Medium skilled High skilled Medium skilled High skilled 27.8 22.9 19.3 (17.4) 36.6 22.9 33.8 27.3 (11.9) 23.1 25.9 17.4 21.5 49.6 29.5 31.9 36.7 72.9 41.5 72.9 60.9 (4.6) 29.1 52.9 39 29.2 13.5 18.8 12.2 (10.4) 17.4 12.9 25.7 15.7 (5.4) 11.6 7.8 8.9 11.7 23.1 29.1 25 16.8 29.7 21.3 53.2 27.5 (4.4) 17.5 13 26.1 21.5 Notes:  Medium skilled ISCED 3, 4; high ­skilled ISCED 5, 6; values in brackets have a low reliability Source:  Huber et al (2010); EU-­LFS (2010) in the case of immigrants in EU-­15 countries (see Table 9.3) They showed that for both medium-­and high-­skilled migrants in the EU-­15 the average over-­education rate of those with more than 10 years of residence (13.5 and 23.1 percent respectively) was only half of that for medium-­ and high-­skilled migrants with less than 10 years of residence (27.8 and 49.6 percent respectively) 184   International handbook on the economics of migration What Causes Education Mismatch? The determinants of over-­ and under-­education are usually estimated using binary outcome or multinomial models The specifications vary widely, making the comparison of results quite difficult A further complication is that several studies not take into account the eventual important selection bias into employment Education mismatch is observed only for employed individuals, and immigrant unemployment rates are sometimes considerable Piracha et al (2012) indicate that over 30 percent of the immigrants are unemployed at five months after arrival in Australia Moreover, some studies have separate estimations for males and females (for example, McGoldrick and Robst, 1996), education levels (for example, Poot and Stillman, 2010) or residence periods (for example, Piracha et al., 2012), while others run pooled estimations over population groups and/or time periods (for example, Chiswick and Miller, 2009; Green et al., 2007; Sanroma et al., 2008) Nevertheless, more or less coherent findings are that more recent immigrants, with less work experience, less fluent in the host country language, as well as originating from countries that are economically and culturally different are more likely to be over-­educated The country or region of origin is one of the most frequently analyzed determinants for immigrants’ education mismatch The reason for that is that it may eventually hide two important aspects related to easier labor market integration in the host country: (1) the quality of schooling and/or labor market experience in the home country and (2) the cultural and language similarities between home and host country By analysing the intensity of mismatch (that is, more than five years of over-­education versus less then five years of over-­education), Sanroma et al (2008) found that immigrants in Spain originating from a developed country had a considerably smaller likelihood of being severely mismatched (only 3.6 percent) compared with those originating from Eastern Europe, the Maghreb countries or sub-­Saharan Africa (over 20 percent) The most plausible explanation for that is that local employers put a low value on qualifications from education systems in developing countries as well as on the work experience gained in those countries Sanroma et al (2008) further found that cultural and language proximity between the immigrants’ home country and Spain reduces the intensity of over-­education as well, with immigrants from Latin America being less over-­educated compared with East European, Asian and African migrants These results were confirmed by studies on other host countries Using a self-­assessment measurement method of over-­education, Mavromaras et al (2009) estimated that the likelihood of immigrants from non-­English speaking countries to be moderately over-­educated in the Australian labor market is 12.9 percent, and that of being severely over-­educated is 14.2 percent higher compared with natives At the same time, they found no significant effect for immigrants from an English-­speaking background Similarly, Green et al (2007) and Poot and Stillman (2010) confirm that originating from an English-­speaking country (for example, the British Isles or North America) decreases the likelihood and/or intensity of over-­education in Australia and New Zealand respectively All these finding give support to the hypothesis of imperfect transferability of human capital There are several possible strategies to improve the transferability of home country Migrant educational mismatch and the labor market  ­185 human capital One is to have the home country qualifications assessed Green et al (2007) show, however, that the qualification assessment policy introduced in Australia in the 1990s had contradictory effects Having a home country qualification assessed in Australia decreases the probability of over-­education by 19 percent for migrants with an English-­speaking background and by 14 percent for Asian migrants, but increased it by percent for other migrants with a non-­English speaking background Another strategy is to obtain a host country qualification Nielsen (2011) shows that immigrants in Denmark that have Danish education have three times lower probability of being over-­educated compared with those having a foreign diploma Battu and Sloane (2002) argue that foreign employers are more likely to recognize foreign qualifications They found that working for a non-­white employer decreases over-­education of non-­whites in the UK by 16.6 percent and increases under-­education by 2.8 percent Moreover, working in an urban area – that eventually draws in highly skilled labor and generates positive spillover benefits of such labor pooling – has been found to decreases the intensity of over-­education by 1.3 percent for immigrants with no qualification, and up to 28.5 percent for immigrants with bachelor degrees (Poot and Stillman, 2010) The hypothesis of imperfect transferability of human capital predicts also that with the accumulated host country labor market experience the immigrants’ labor market situation should improve The pace of assimilation, however, can be slow Sanroma et al (2008) estimated that it would need about 15 years for immigrants living in Spain to close the educational mismatch gap between themselves and natives A quicker pace was found for Eastern Europeans, and immigrants from the Maghreb However, for Asians and sub-­Saharan Africans the gap seems not to narrow at all Immigrants from these regions are at high risk of remaining permanently trapped in the Spanish labor market in jobs with lower wages, regardless of their level of education By estimating a multinomial model of mismatch, Chiswick and Miller (2009) found that residence/experience in the USA helps migrants to climb up the occupational ladder The predicted over-­education rate decreased after 30 years of residence from 34.3 to 25.0 percent, while the predicted under-­education rate increased from 36.5 to 46.9 percent Nevertheless, they found that the experience in the home country had an even stronger effect: 30 years of home country labor market experience decreasing the predicted over-­ education rate from 36.2 to 23.2 percent and increasing the under-­education rate from 21.4 to 53.7 percent They explain this result by the probably favorable selection of migrants with respect to skills valued in the US labor market Among other countries, Australia has long experience of migrant selection on the basis of point system that assesses the potential immigrants’ labor market skills The merits of such a selection is confirmed by the findings that migrants who were not selected on the basis of skills (for example, preferential family and concessional family visa holders) had highest likelihood of being over-­educated in the Australian labor market A quite important but often neglected cause of immigrants’ education mismatch is previous education mismatch experience in the home country As shown by Piracha et al (2012) the home country mismatch significantly adds to the explanation of the variation in the immigrants’ mismatch in Australia Compared with a model that has as covariates only socio-­economic controls used in other studies (that is, age, age squared, a dummy for having the qualification assessed in Australia, dummies for the former region of residence, dummies for the entry visa type, a dummy for school age children 186   International handbook on the economics of migration present, a dummy for having financial funds at time of entry and regional dummies), the inclusion of covariates controlling for the mismatch in the home country almost doubles the explanatory power of the probit estimation for over-­education five months after arrival in Australia (that is, adjusted R-­squared increase from 0.14 to 0.26) and almost quadruples the explanatory power of probit estimation for under-­education five months after arrival (that is, adjusted R-­squared increase from 0.14 to 0.51 percent) The findings show that at time of arrival in the host country, employers use the available information about the immigrants’ labor market experience in the home country to derive ability signals from it How Does Mismatch Affect Wages? The main model used in the literature to estimate the returns to over-­ and under-­ education is the so-­called ORU (over-­, required-­, under-­education) model It is an extended Mincerian wage equation introduced by Duncan and Hoffman (1981) and has the form: o r ln (wi) ao Ei ar Ei au Eiu xri b 1ei (9.1) where wi is individual’s i wage, Eto are the number of years of surplus or over-­education, Eir are the number of years of required education for the job, Eiu are the number of years of deficit or under-­education, and xi9 are a vector of control variables, including experience and experience squared The actual years of education for the individual i are either Eir, Eir + Eio or Eir − Eiu; Eio and Eiu cannot be simultaneously positive (that is, the individual cannot be both over-­and under-­educated at the same time) The estimations of equation (9.1) face two often ignored problems: (1) omitted variable bias (that is, individuals are likely to be non-­randomly assigned with respect to both completed and required education), and (2) measurement error with respect to required education (as alluded to in section 2; also see section for further discussion on this issue) The solutions are far from trivial and some scholars have tried to address them using instrumental variables and individual fixed effects models (for a more detailed discussion see Leuven and Oosterbeek, 2011) Duncan and Hoffman’s (1981) model has been replicated extensively using different data and, sometimes, different measures of required education.12 The general findings are that: (1) in a job requiring a given amount of education, the earnings of over-­educated employees are higher, while the earnings of under-­educated employees are lower compared with employees who have the required education level; (2) the returns to years of education above the level required as well as the ‘penalty’ for the years of education below the level required are both lower than the returns to the years of required education; and (3) under-­education is less severely punished than over-­education is rewarded (see Hartog, 2000; Leuven and Oosterbeek, 2011) The empirical findings focusing on immigrants summarized in Table 9.4 are in accordance with the general estimation results mentioned above Moreover, immigrants’ returns to required schooling seem to be quite similar to those of natives and range from about percent in Denmark up to about 15 percent in the USA and Australia Nevertheless, immigrants seem to get significantly less return from over-­education, but Migrant educational mismatch and the labor market  ­187 Table 9.4  Wage effect of educational mismatch among immigrants and natives Country (reference) Population group USA  (Chiswick and Miller, 2008) USA  (Chiswick and Miller, 2010b) USA  (Chiswick and Miller, 2010b) Australia  (Chiswick and Miller, 2010c) Australia  (Chiswick and Miller, 2010c) Spain  (Sanroma et al., 2008) Denmark  (Nielsen, 2011) Over-­ education Required education Under-­ education Measurement method Natives Immigrants 0.056*** 0.045*** 0.153*** 0.153*** −0.066*** −0.022*** RM5 (mode)  procedure Natives  (Bachelor1)1 Immigrants  (Bachelor1)1 Natives  (Master1)2 Immigrants  (Master1)2 Natives ESB3  immigrants Non-­ESB3  immigrants Natives ESB3  immigrants Non-­ESB3  immigrants Natives Immigrants  (EU-­15 & NA4) Immigrants  (Rest Europe) Immigrants  (Rest America) Immigrants  (Africa) Immigrants  (Asia) Natives Immigrants  (Danish educ.) Immigrants  (foreign educ.) 0.020*** 0.122*** 0.019*** 0.140*** 0.027*** 0.132*** −0.018*** 0.091*** 0.060*** 0.056*** 0.152*** 0.152*** −0.037*** −0.027*** 0.032*** 0.152*** −0.014*** 0.053*** 0.053*** 0.112*** 0.127*** −0.083*** −0.060*** 0.035*** 0.096*** −0.052*** 0.044*** 0.044*** 0.090*** 0.098*** −0.043*** −0.051*** 0.056*** −0.054*** 0.017*** 0.065*** −0.042*** 0.001 0.031*** −0.017*** 0.018 0.100*** −0.046*** 0.054*** 0.037*** 0.079*** 0.079*** −0.047*** −0.032*** 0.011*** 0.072*** 0.019* −0.003 Notes: Bachelor1 – individuals with bachelor or higher education Master1 – individuals with master or higher education ESB – English-speaking background NA – North America RM – realized matches procedure JA – job analysis procedure RM5 (mode)  procedure RM5 (mode)  procedure RM5 (mode)  procedure JA6 procedure RM5 (mode)  procedure RM5 (mode)  procedure 188   International handbook on the economics of migration also a lower penalty for under-­education Immigrants in the United States, for example, have 4.5 percent higher wages for each year of surplus education (compared with 5.4 percent for natives) and a −2.2 percent penalty for each year of education deficit (compared with −6.6 percent for natives) There are important variances with respect to migrants’ origin Immigrants in Spain originating from the EU-­15 and North America had similar returns for education compared with natives At the same time, immigrants from Eastern Europe and Africa had 3.4 and 5.9 percentage points respectively lower returns for required schooling and no returns for over-­education (see Sanroma et al., 2008) Fluency in the host country language certainly improves the international transferability of human capital Originating from a country with more similar culture and language seems to increase not only the probability of a matched employment (as discussed above), but the returns for education as well Chiswick and Miller (2010c) found that immigrants in Australia originating from an English-­speaking country have 2.4 percentage points higher returns for each year of over-­education Similarly, Latin American immigrants in Spain had higher returns for both required and surplus education compared with East Europeans and Africans (see Sanroma et al., 2008) Another aspect is the familiarity of employers with credentials held Nielsen (2011) differentiated between immigrants in Denmark with foreign versus host country education and found that the latter had 2.6 percentage points higher returns per year of over-­ education and 0.7 percentage points higher returns for each year of required education Chiswick and Miller (2010a) tested whether returns to schooling vary with the quality of the foreign education obtained They estimated a two-­step model that has the ORU equation as its first step In the second-­step equation, the payoff coefficients from the first are used as dependent variable and the Programme for International Student Assessment (PISA) scores as well as the gross domestic product (GDP) per capita for the countries of origin as explanatory variables They found that a better education acquired abroad in terms of PISA test scores (a 100 points increase) is associated with greater returns for required education (1.0 to 1.4 percent) on the US labor market, modest returns for surplus education (0.3 to 0.4 percent) and a greater penalty for years of under-­education (−1.3 to −1.6 percent) This suggests that employers assess, at least partly, objectively the skills and abilities of immigrants with different backgrounds 4  POSSIBLE EXTENSIONS Research work on the topic may be extended in two different directions of equal relevance The first relates to the more general problems of measurement error and omitted variable bias already mentioned While there have been first attempts to tackle the problems using individual fixed effects and instrumental variables estimations (see Dolton and Silles, 2008; Korpi and Tahlin, 2009; Tsai, 2010), the use of these econometric models is limited by the availability of suitable data Individual fixed effects can be applied only for panel data (which is quite scarce with respect to immigrants), while the second demands good instruments for required and completed education (for example, information on individual ability and motivation, number of siblings, economic problems and/or disrup- Migrant educational mismatch and the labor market  ­189 tions in the family of origin) are seldom available in immigrants’ surveys Perhaps better theoretical foundations and more appropriate data collection could disentangle some of the elements that affect the results The second direction is related to specific research questions in the field of international migration While so far immigrants’ educational mismatch has been analyzed only with respect to its determinants and effects on wages, there is still little or nothing known of the mismatch effect on migrants’ self-­employment, integration, return migration and remittance decisions as well as the consequent effects for migrant-­sending countries If migrants working in jobs below their education level decide to settle abroad, could over-­education in the labor market cause immigrants to become self-­employed? And/or, if foreign employers in the destination country are more likely to recognize foreign qualifications than are native employees (as found by Battu and Sloane, 2002, 2004), would that cause migrants to become more encapsulated in their own ethnic group in terms of both labor activity and social life, even if that is more likely to undermine their integration process in the host society? Stark and Taylor (1991) argued that immigrants might decide to return to their home countries if they rank higher in the income distribution of their home reference group compared with that of their reference group in the host country (that is, relative deprivation) Could education mismatch make immigrants feel ‘relatively deprived’ and contribute to return decisions? If over-­education is an important aspect of the return decision of highly educated migrants, does that not undermine the expectations of ‘brain drain’ compensations through transfer of know-­how and technology (as they probably gained no professional experience in high-skilled jobs)? Another important impact of possible labor market mismatch in the destination country is likely to be on immigrants’ remittance behavior If there is, for instance, a high incidence of over-­education then, as the existing literature shows, that will result in relatively lower wages compared with correctly matched levels, which can then have a negative impact on remittance flows.13 5  CONCLUSIONS This chapter has reviewed the possible causes and consequences of educational mismatch of immigrants in the labor market of their countries of destination Within the labor economics literature, the issue of over-­ and under-­education is controversial and has been under discussion for a number of years However, research on educational mismatch of immigrants is still in its infancy and there is limited research in this area The extant literature has shown that immigration, at least in the initial stages, has not been successful in allocating skills of the new entrants efficiently This result has been shown when comparing with natives for a number of high immigrant-­receiving countries (for example, Australia, the USA and Canada) A number of reasons have been forwarded for this mismatch including imperfect transferability of human capital, the lack of host country language skills, the lack of innate ability, and discrimination This survey has therefore not only explored the theoretical and empirical literature on immigrant mismatch, but also has presented some possible directions for further research We see scope for a lot of interesting work to be done on the effects of 190   International handbook on the economics of migration i­mmigrants’ education mismatch on self-­employment and integration in the host country as well as on return migration and remittances Moreover, the measurement error and omitted variable bias problems related to required and completed education leave the estimation of mismatch effects a challenging econometric exercise NOTES   *                   10 11 12 13 We would like to thank two anonymous referees and the editors of this volume, Amelie F Constant and Klaus F Zimmermann, for helpful suggestions and comments We are, of course, responsible for any remaining errors Over-­education (under-­education) is defined as an employee having more (less) education than their job requires For a comprehensive survey of this literature see Hartog (2000), McGuinness (2006) and Leuven and Oosterbeek (2011) There is a general debate in the literature about the overall impact of migrants on the host country’s economic and social structure The literature has looked at migrants’ level of assimilation in terms of how similar, in economic terms in general and in labor market terms in particular, they are to natives For more discussion on this, see Chapter in this volume This type of mismatch is often referred in the literature as ‘vertical mismatch’ On the other hand, ‘horizontal mismatch’ refers to employees that have acquired education in another field than their job requires (see CEDEFOP, 2010) In this paper only the ‘vertical’ aspects of mismatch are discussed The concept of measuring over-­education is controversial For instance, one argument against this measurement is that education is typically general and is not acquired for a particular job This then raises one of the main challenges for the measurement issue, that is, of the definition of over-­education See for example, Hartog (1980), Rumberger (1987), Kiker and Santos (1991) and Piracha et al (2012) This method has been used by a number of authors, for example, Hartog and Oosterbeek (1988), Duncan and Hoffman (1981) and Galasi (2008) See Chapter 10 in this volume for further discussion on this Selection is also related to immigrants’ visa category on which they enter the host country See Chapter 23 in this volume It is worth pointing out here that it is perhaps not easy to compare the needs of a particular occupation a person has, since the characteristics of a job are not usually observed in the data That was most likely owing to the fact that the two standard deviations interval around the mean – applied in the mean measurement procedure – increases the estimated number of realized matches A quite special case is Battu and Sloane (2002, 2004) They have used dummies for over-­ and under-­ education instead of years, making the results incomparable with those from other studies To date, however, only the study of McDonald and Valenzuela (2009) considers the effect of over-­ education on remittance behavior Using the data on Filipino migrants, they found that over-­educated women tend to remit less but the mismatch has no impact on men in the same category Men tend to work longer hours to compensate for the lower wage REFERENCES Battu, H and P.J Sloane (2002), ‘To what extent are ethnic minorities in Britain over-­educated?’, International Journal of Manpower, 23 (3), 192–208 Battu, H and P.J Sloane (2004), ‘Over-­education and ethnic minorities in Britain’, The Manchester School, 72 (4), 535–59 Belman, D and J.S Heywood (1997), ‘Sheep skin effects by cohort: implications of job matching in a signalling model’, Oxford Economic Papers, 49 (4), 623–37 CEDEFOP [European Centre for the Development of Vocational Training] (2010), ‘The skill matching challenge Analysing skill mismatch and policy implications’, Publications Office of the European Union, Luxembourg Chiswick, B.R (1978), ‘The effect of Americanization on the earnings of foreign-­born men’, Journal of Political Economy, 86 (5), 897–921 Migrant educational mismatch and the labor market  ­191 Chiswick, B.R and P.W Miller (2008), ‘Why is the payoff to schooling smaller for immigrants?’, Labour Economics, 15 (6), 1317–40 Chiswick, B.R and P.W Miller (2009), ‘The international transferability of immigrants’ human capital skills’, Economics of Education Review, 28 (2), 162–9 Chiswick, Barry R and Paul W Miller (2010a), ‘The effects of school quality in the origin on the payoff to schooling for immigrants’, in Gil Epstein and Ira Gang (eds), Migration and Culture, Frontiers of Economics and Globalization, Vol 8, Bingley: Emerald Publishing, pp. 67–103 Chiswick, Barry R and Paul W Miller (2010b), ‘Education mismatch: are high-­skilled immigrants really working at high-­skilled jobs and the price they pay if they aren’t?’, in Barry R Chiswick (ed.), High Skilled Immigration in a Global Labor Market, Washington, DC: American Enterprise Institute Press, pp. 111–54 Chiswick, B.R and P.W Miller (2010c), ‘The effects of educational-­occupational mismatch on immigrant earnings in Australia, with international comparisons’, International Migration Review, 44 (4), 869–98 Dolado, J.J., M Jansen, and J.F Jimeno (2009), ‘On the job-­search in a matching model with heterogeneous jobs and workers’, Economic Journal, 119 (534), 200–28 Dolton, P.J and M.A Silles, (2008), ‘The effects of over-­education on earnings in the graduate labour market’, Economics of Education Review, 27 (2), 125–39 Duncan, G and S Hoffman (1981), ‘The incidence and wage effects of overeducation’, Economics of Education Review, (1), 75–86 Freeman, Richard (1976), The Overeducated American, New York: Academic Press Galasi, Peter (2008), ‘The effect of educational mismatch on wages for 25 countries’, Budapest Working Papers on the Labour Market BWP – 2008/8, Institute of Economics, Hungarian Academy of Sciences, Department of Human Resources, Corvinus University of Budapest Gautier, P (2002), ‘Unemployment and search externalities in a model with heterogeneous jobs and workers’, Economica, 69 (273), 21–40 Green, C., P Kler and G Leeves (2007), ‘Immigrant overeducation: evidence from recent arrivals to Australia’, Economics of Education Review, 26 (4), 420–32 Green, Francis, Steven McIntosh and Anna Vignoles (1999), ‘Overeducation and skills – clarifying the concepts’, CEP Discussion Papers DP No 0435, Centre for Economic Performance, London School of Economics and Political Science (CEP) Groot, W and H Maassen van den Brink (2000), ‘Overeducation in the labor market: a meta-­analysis’, Economics of Education Review, 19 (2), 149–58 Hartog, J (1980), ‘Earnings and capability requirements’, Review of Economics and Statistics, 62 (2), 230–40 Hartog, J (2000), ‘Over-­education and earnings: where are we, where should we go?’, Economics of Education Review, 19 (2), 131–47 Hartog, J and H, Oosterbeek (1988), ‘Education, allocation and earnings in the Netherlands: overschooling?’, Economics of Education Review, (2), 185–94 Huber, P., M Landesmann, C Robinson and R Stehrer (2010), ‘Migration, skills and productivity: a European perspective’, National Institute Economic Review, 213 (1), R20–R34 Kiker, B.F and M.C Santos (1991), ‘Human capital and earnings in Portugal’, Economics of Education Review, 10 (3), 187–203 Kiker, B.F., M.C Santos and M de Oliveira (1997), ‘Overeducation and undereducation: evidence for Portugal’, Economics of Education Review, 16 (2), 111–25 Korpi, T and M Tahlin (2009), ‘Educational mismatch, wages, and wage growth: overeducation in Sweden, 1974–2000’, Labour Economics, 16 (2), 183–93 Leuven, Edwin and Hessel Oosterbeek (2011), ‘Overeducation and mismatch in the labor market’, IZA Discussion Paper No 5523, Institute for the Study of Labor (IZA), Bonn Mavromaras, Kostas, Seamus McGuinness and Yin King Fok (2009), ‘Overskilling dynamics and education pathways’, IZA Discussion Paper No 4321, Institute for the Study of Labor (IZA), Bonn McDonald, James T and M Rebecca Valenzuela (2009), ‘The impact of skill mismatch among migrants on remittance behaviour’, SEDAP Research Paper No 242, McMaster University, Hamilton, Ontario McGoldrick, K and J Robst (1996), ‘Gender differences in overeducation: a test of the theory of differential overqualification’, American Economic Review, 86 (2), 280–84 McGuinness, S (2006), ‘Overeducation in the labour market’, Journal of Economic Surveys, 20 (3), 387–418 Nielsen, C.P (2011), ‘Immigrant overeducation: evidence from Denmark’, Journal of Population Economics, 24 (2), 499–520 Piracha, M., M Tani and F Vadean (2012), ‘Immigrant over-­and under-­education: the role of home country labour market experience’, IZA Journal of Migration, (3), 1–21 Poot, Jacques and Steven Stillman (2010), ‘The importance of heterogeneity when examining immigrant education-­occupation mismatch: evidence from New Zealand’, IZA Discussion Paper No 5211, Institute for the Study of Labor (IZA), Bonn 192   International handbook on the economics of migration Quinn, M and S Rubb (2006), ‘Mexico’s labor market: the importance of education–occupation matching on wages and productivity in developing countries’, Economics of Education Review, 25 (2), 147–56 Rumberger, R.W (1987), ‘The impact of surplus schooling on productivity and earnings’, Journal of Human Resources, 22 (1), 24–50 Sanroma, Esteban, Raul Ramos and Hipolito Simon (2008), ‘The portability of human capital and immigration assimilation: evidence from Spain’, IZA Discussion Paper No 3649, Institute for the Study of Labor (IZA), Bonn Sicherman, N (1991), ‘“Overeducation” in the labor market’, Journal of Labor Economics, (2), 101–22 Sicherman, N and O Galor (1990), ‘A theory of career mobility’, Journal of Political Economy, 98 (1), 169–92 Smith, James and Finis Welch (1978), ‘The overeducated American? A review article’, UCLA Economics Working Papers 147, University of California (UCLA), Los Angeles Stark, O and J.E Taylor (1991), ‘Migration incentives, migration types: the role of relative deprivation’, Economic Journal, 101 (408), 1163–78 Tijdens, Kea and Maarten van Klaveren (2011), ‘Over-­ and underqualification of migrant workers Evidence from WageIndicator survey data’, AIAS Working Paper No 110, Amsterdam Institute for Advanced Labour Studies Tsai, Y (2010), ‘Returns to overeducation: a longitudinal analysis of the U.S labor market’, Economics of Education Review, 29 (4), 606–17 Verdugo, R and N Verdugo (1989), ‘The impact of surplus schooling on earnings: some additional findings’, Economics of Education Review, 22 (4), 690–95 10  Ethnic hiring* David Neumark 1  INTRODUCTION Economic migration brings with it the challenge of racial, ethnic or national minorities assimilating into their adopted labor markets Barriers to the employment of these minorities will clearly inhibit their successful assimilation.1 This chapter focuses on the hiring side of the equation: are there barriers to the hiring of racial and ethnic minorities, what is the nature of these barriers and how workers overcome these barriers? The chapter focuses on three key influences on the hiring of racial, ethnic or national minorities: discrimination, spatial mismatch and networks The barriers posed by discrimination and spatial mismatch are obvious Networks can also pose barriers to the extent that ethnic minorities have fewer network connections than majority groups, but networks can also be a way for ethnic minorities to overcome barriers to employment The chapter integrates recent research my co-­authors and I have done on these topics It is not an exhaustive survey of all of the research on these specific topics, nor does it cover other topics that could bear on ethnic hiring The focus is not on economic migrants per se, but on ethnic and racial minorities, with much evidence coming from research on Hispanics and blacks in the United States Although much of the research I discuss is limited to the US setting, I touch on related recent work on ethnic hiring in European countries The first part of the chapter focuses on methodological issues regarding the estimation of discrimination in what is widely regarded as the best way to test for ­discrimination – audit and correspondence studies The remainder of the chapter discusses evidence on other factors that could impede good economic outcomes for minorities, including spatial mismatch and labor market networks that, to some extent at least, may less to connect minorities to jobs 2 DISCRIMINATION: EVIDENCE FROM FIELD EXPERIMENTS, AND PROBLEMS WITH THAT EVIDENCE2 Research on discrimination in labor markets has a long history The ‘workhorse’ of early economics research on labor market discrimination is the ‘residual discrimination’ approach of Oaxaca (1973), in which wage regressions are estimated on individual-­ level data, and discrimination is estimated from the ethnic differential that remains unexplained after including many proxies for productivity.3 This approach suffers from numerous criticisms First, the proxies may not adequately capture group differences in productivity, in which case the ‘unexplained’ differences cannot be interpreted as discrimination Second, mean differences between groups are treated as nondiscriminatory, and differences in the equation coefficients as discriminatory But differences in 193 194   International handbook on the economics of migration c­ oefficients can arise for other reasons – for example, a productivity-­enhancing effect of marriage on men but not women (Korenman and Neumark, 1991, 1992) – and differences in means can reflect discrimination (Neumark and McLennan, 1995) Finally, arbitrary assumptions have to be made about the counterfactual – that is, what the wage structure would look like in the absence of discrimination (Neumark, 1988) Audit or correspondence studies are a response to these criticisms.4 In these studies, fictitious individuals who are identical except for race, sex or ethnicity apply for jobs Evidence of group differences in outcomes, such as fewer job offers for blacks, is generally viewed as compelling evidence of discrimination, because there is no reason to expect important differences between, for example, black and white job applicants As a consequence, this strategy has come to be widely used in testing for discrimination in labor markets (as well as in housing markets).5 An advantage of audit or correspondence studies with regard to the present inquiry is that they directly address discrimination in hiring, rather than pay discrimination At the same time, these studies not provide all the evidence we might like; in particular, they provide us with evidence on ethnic differences in hiring for the specific artificial samples they create, but not descriptive evidence on differences in hiring experienced by representative populations Across numerous countries and minority groups, audit or correspondence studies find evidence consistent with discrimination, including, for example, discrimination against blacks, Hispanics and women in the United States (Bertrand and Mullainathan, 2004; Mincy, 1993; Neumark, 1996), Moroccans in Belgium and the Netherlands (Bovenkerk et al., 1995; Smeeters and Nayer, 1998), Turks in Germany (Kaas and Manger, 2012) and lower castes in India (Banerjee et al., 2008).6 Researchers have, over time, shifted from audit to correspondence studies, in response to critiques of the audit study method (for example, Bertrand and Mullainathan, 2004) For example, Heckman and Siegelman (1993) noted that in the well-­known Urban Institute audit studies (Mincy, 1993), white and minority testers were told, during their training, about ‘the pervasive problem of discrimination in the United States’, possibly introducing experimenter effects Correspondence studies address this problem by using applications on paper, cutting out the potential influence of live job applicants However, a fundamental critique that applies equally well to correspondence studies has not been addressed by researchers In particular, Heckman and Siegelman (1993) and Heckman (1998) consider what most researchers view as the ideal conditions for an audit or correspondence study – when the observable average differences between groups are eliminated, and the applications are sufficiently detailed that it is safe to assume that employers believe there are no average differences in unobservable characteristics between groups Heckman and Siegelman show that, even in this case, these studies can generate evidence of discrimination (in either direction) when there is none, and can also mask evidence of discrimination when it in fact exists To see this in a simple setting, suppose that productivity depends on two individual characteristics, X9 (X I, X II) Two testers (or applications) are sent to firms to apply for jobs; R is an indicator equal to for ethnic minority applicants (E) and for nonminorities (NM) Denote by XEj and XNjM the values of X I and X II for the two groups, j I, II The study controls only X I in the resumes or interviews, standardizing XI across applicants so XEI XNIM X I* PE* and PN*M denote expected productivity of the two groups Ethnic hiring  ­195 Assume that productivity is P(X I, X II) b I9X I X II, and that the parameter g9 captures possible discrimination against the ethnic minority, resulting in the ‘discounting’ of the productivity of the ethnic minority, as in Becker (1971) If the outcome T(P, R) depends on productivity and ethnicity additively (such as the wage in the Becker employer discrimination model), then each individual test (two applicants to a firm) provides an observation equal to: T(PE*,1) −­T(PN*M,0) PE* g9 − PN*M bI9XEI E(XEII) g9 −­(b I9XNIM E(XNIIM)) g9 E(XEII) −­E(XNIIM).7 (10.1) These observations identify g9 under the assumption that E(XEII) E(XNIIM) We cannot, in practice, rule out employers holding different expected means for unobserved productivity, although by including a rich set of resume characteristics this can be mitigated Moreover, one can always reinterpret {g9 (E(XEII) −­ E(XNIIM))} as capturing illegal discrimination, at least in countries like the United States where statistical discrimination is illegal As Heckman and Siegelman emphasize, a more troublesome problem arises because in the hiring process firms likely evaluate a job applicant’s productivity relative to a standard, and offer the applicant a job (or an interview) if the standard is met In this case, even when E(XEII) E(XNIIM) a correspondence study can generate spurious evidence of discrimination in either direction, or of its absence; in other words, discrimination is unidentified.8 In this ‘best-­case’ scenario, suppose the correspondence study standardizes on a low value of X I Because an employer offers a job interview only if it perceives the sum b I9X I X II to be sufficiently high, the employer has to believe that X II is high (or that the probability that it is high is large) to offer an interview For example, although the employer does not observe X II, if the employer knows that the variance of X II is higher for nonminorities, the employer correctly concludes that nonminorities are more likely than minorities to have a sufficiently high sum of b I9X I X II, and will therefore be less likely to offer jobs to minorities The opposite holds if the standardization is at a high value of X I, in which case the employer only needs to avoid very low values of X II, which will be more common for nonminorities And the results flip if the unobserved variance is higher for minorities.9 To formalize this, suppose that a job offer or interview is given if a worker’s perceived productivity exceeds the threshold c9 The hiring rules are: Hire E if b I9X I* XEII g9 c9 (10.2) Hire NM if b I9X I* XNIIM c9. (10.3) Assume that the unobservables XEII and XNIIM are normally distributed, with equal means (set to zero, without loss of generality), and standard deviations sEII and sNIIM Then the hiring probabilities are: Pr[Hire E] −­F[(c9 − b I9XI* − g9)/sEII] F[(b I9X I* g9 – c9) /sEII] (10.4) 196   International handbook on the economics of migration Pr[Hire NM] −­F[(c9 −­b I9X I*)/sNIIM] F[(b I9X I* − c9)/sNIIM] (10.5) where F denotes the standard normal distribution function.10 The difference between equations (10.4) and (10.5) is supposed to be informative about discrimination But even if g9 0, so there is no discrimination, these two expressions need not be equal because sEII and sNIIM can be unequal Thus, even when the means of the unobserved productivity-­related variables are the same, and firms use the same hiring standard (g9 0), correspondence studies can generate evidence consistent with discrimination against ethnic workers, or in their favor So are correspondence (or audit) studies rendered useless by this criticism? In a recent paper, I show that with the right data from a correspondence study, and using the same framework as in Heckman and Siegelman (1993), it is possible to recover an unbiased estimate of discrimination, conditional on an identifying assumption The intuition is as follows The Heckman and Siegelman critique rests on differences between groups in the variances of unobserved productivity The fundamental problem, as equations (10.4) and (10.5) show, is that we cannot separately identify the effect of ethnicity (g9) and a difference in the relative variance of the unobservables (sEII/sNIIM) However, a higher variance for one group implies a smaller effect of observed characteristics on the probability that an applicant from that group meets the standard for hiring Consequently, information on how variation in observable qualifications is related to employment outcomes can be informative about the relative variance of the unobservables, and this, in turn, can identify the effect of discrimination Based on this idea, the identification problem is solved by invoking an identifying assumption – specifically, that there is variation in some applicant characteristics in the study that affect perceived productivity and have effects that are homogeneous across groups This is an assumption, but it implies overidentifying restrictions that can be tested in the data, because if the effects on hiring of multiple productivity controls differ between two groups only because of the difference in the variance of the unobservables, the ratios of the estimated probit coefficients for the two groups, for each variable, should be equal To see why this works, the difference in hiring probabilities is: F[(b I9X I* g9 – c9)/sEII] − F[(b I9X I* − c9)/sNIIM]. (10.6) Impose the normalization that sNIIM The parameter sEII is then the variance of the unobservable for the ethnic group relative to nonminorities Denoting that parameter II sREL , and dropping the prime subscripts to indicate that the coefficients in equation (10.6) are now ratios relative to sNIIM, equation (10.6) becomes: II F[(bIX I* g − c)/sREL ] − F[bIX I* − c]. (10.7) If there is variation in the level of qualifications used as controls (XI*), and these II qualifications affect hiring outcomes, then we can identify c, bI/sREL , and bI The ratio II II of the latter two parameters identifies sREL, and identification of sREL then implies identification of g Note that variation in X I* that affects hiring is essential, since otherwise II we cannot separately identify sREL , c, and g The parameters can be estimated using Ethnic hiring  ­197 a heteroskedastic probit model, with the variance of the unobservable varying across groups As an application, Bertrand and Mullainathan’s well-­known correspondence study of race discrimination is unusual in that, for different reasons, it used data on applicants with, roughly speaking, two different levels of qualifications Bertrand and Mullainathan (2004) report probit models estimated for whites and blacks separately (their table 5) These estimates reveal higher callbacks for whites, and substantially stronger effects of measured qualifications for whites than for blacks Viewed in light of the preceding discussion, and assuming that the true effects of qualifications are the same for blacks and whites, the smaller estimated probit coefficients or marginal effects for blacks implies that blacks have a larger variance of the unobservable If Bertrand and Mullainathan standardized applicants at low levels of qualifications, then the Heckman and Siegelman analysis would imply that there is a bias towards finding discrimination in favor of blacks, as the high-­variance group would be preferred; that is, the evidence of discrimination against blacks would be stronger without the bias from differences in the variances of the unobservables But there is no way to assess whether the characteristics of applicants were low, since we not know the population of applicants Hence, implementation of the estimation procedure outlined above is likely the only way even to sign the bias, let alone to recover an unbiased estimate of discrimination Because Bertrand and Mullainathan’s data include applicants with different levels of qualifications, and the qualifications predict callbacks, their data can be used to implement the methods described above Panel A of Table 10.1 shows their baseline results Marginal effects are reported for two different specifications Estimates are shown for males and females combined, and for females only; as the sample sizes indicate, the male sample is considerably smaller Callback rates are much lower (about 33 percent) for blacks than for whites, and the race difference is robust It turns out that a number of the resume characteristics have statistically significant effects on the callback probability (in the expected direction), as needed Panel B in Table 10.1 begins by reporting the estimated overall marginal effects of race from the heteroskedastic probit model These estimates are slightly smaller than the estimates from the simple probits, but trivially so They remain statistically significant and indicate callback rates that are lower for blacks by about 2.4−2.5 percentage points (or about 25 percent) However, the marginal calculation is more complicated in the heteroskedastic probit model, because if the variances of the unobservable differ by race, then when race ‘changes’ both the variance and the level of the latent variable that determines hiring can shift As long as we use the continuous version of the partial derivative to compute marginal effects from the heteroskedastic probit model, there is a natural decomposition of the effect of race into two pieces: the partial derivative with respect to race affecting the level of the latent variable, corresponding to the counterfactual of race changing the valuation of the worker without changing the variance of the unobservable; and the partial derivative with respect to changes via the variance of the unobservable In Table 10.1, these two separate effects are reported The effect of race via shifts in the latent variable, or how race affects the employer’s valuation of worker productivity, is of greatest interest The point of the Heckman and Siegelman critique is that differential treatment of blacks and whites based only on differences in variances of the unobservable should not be interpreted as discrimination And the effect of race via the latent 198   International handbook on the economics of migration Table 10.1  Heteroskedastic probit estimates for callbacks: full specifications Males and females A Estimates from basic probit Black B Heteroskedastic probit model Black (unbiased estimates) Marginal effect of race through level Marginal effect of race through variance Standard deviation of unobservables, black/white Wald test statistic, null hypothesis that ratio of   standard deviations (p-­value) Other controls: Individual resume characteristics Neighborhood characteristics N Females (1) (2) (3) (4) −.030 (.006) −.030 (.006) −.030 (.007) −.030 (.007) −.024 (.007) −.086 (.038) 062 (.042) 1.37 22 −.026 (.007) −.070 (.040) 045 (.043) 1.26 37 −.026 (.008) −.072 (.040) 046 (.045) 1.26 37 −.027 (.008) −.054 (.040) 028 (.044) 1.15 56 X X X 4784 X X X 3670 4784 3670 Notes: Standard probit marginal effects are reported in Panel A Panel B uses equations (16) − (160) from Neumark (2012) Marginal effects are evaluated at sample means Standard errors are computed clustering on the ad to which the applicants responded, and are reported in parentheses; the delta method is used to compute standard errors for the marginal effects Individual resume characteristics include bachelor’s degree, experience and its square, volunteer activities, military service, having an email address, gaps in employment history, work during school, academic honors, computer skills and other special skills Neighborhood characteristics include the fraction high school dropout, college graduate, black, and white, as well as log median household income in the applicant’s zip code Source:  Neumark (2012) variable captures discrimination likely to be manifested in the real economy, whereas its effect through the variance is more of an artifact of the study (Neumark, 2012) The marginal effect via the level of the latent variable is larger than the marginal effect from the probit estimation, ranging from −0.054 to −0.086 The effect of race via the variance of the unobservable, in contrast, is positive, ranging from 0.028 to 0.062 (not statistically significant) The implication of the first marginal effect is that race discrimination is more severe than indicated by the analysis that ignores the role of differences in the variances of the unobservables.11 Thus, in the context of the Bertrand and Mullainathan study, implementing a method that addresses the Heckman and Siegelman unobservables critique and recovers an unbiased estimate of discrimination leads to stronger evidence of discrimination More generally, the method proposed here can be easily implemented in any future correspondence (or audit) study All that is needed is for the resumes or applicants to include some ­variation in characteristics that affect the probability of being hired Ethnic hiring  ­199 As it turns out, some additional past audit and correspondence studies that have helped to establish the consensus that ethnic minorities (or other disadvantaged groups) face discrimination in hiring also use applicants of different quality, along a number of dimensions.12 Results from these studies are described briefly in Table 10.2 In every case, these studies indicate discrimination against an ethnic group But given the Heckman and Siegelman critique, it is possible that in some cases these studies overstate discrimination We cannot determine this without fully implementing the methods described above However, if a low level of standardization was used, then we would expect to find stronger evidence of discrimination in cases where qualifications more to boost hiring among the nonminorities than among the minorities, and vice versa The evidence appears to be mixed on this point For example, for Pager et al (2009), the ratio of callback rates for the more-­ versus less-­qualified applicants is higher for nonminorities than for African-­Americans or Hispanics – or the higher level of qualifications does more to boost nonminority employment The same is true for Middle Eastern or Chinese versus white Australians in Booth et al (2010) In contrast, Pager (2003) reports the opposite: the noncriminal/criminal ratio is 2.0 for whites but 2.8 for African-­Americans The same is true for Ravaud et al (1992), and indigenous versus white Australians in Booth et al (2010) We not actually know whether the level of standardization was high or low The implication is that the conclusion we might draw from this body of literature, once we apply methods that can identify discrimination, could give us a different impression from the current one that nearly all the evidence points to discrimination It is possible that the case for ethnic discrimination is not as overwhelming as it appears.13 3 SPATIAL MISMATCH: IS THE PROBLEM WHERE MINORITIES LIVE?14 Another potential source of hiring barriers for ethnic minorities is a lack of jobs near where they live, or ‘spatial mismatch’, driven by exogenous residential segregation and other frictions As a result of the segregation of minorities in areas with fewer jobs, the net wage (defined as the wage minus commuting costs) is more likely to be below their reservation wage, and fewer will choose to work This will be truer of lower-­skilled minorities for whom commuting costs represent a larger share of earnings Spatial mismatch requires frictions that prevent labor markets from reaching an equilibrium in which employment rates are largely equalized across neighborhoods Spatial mismatch has been widely invoked to explain black employment problems in the United States In that context, the disequilibrium is attributed to numerous factors, including the movement of jobs out of central city areas, discrimination in housing that prevents mobility of blacks to where jobs are located, customer discrimination against blacks that can reduce black employment prospects in white areas, employer discrimination that deters employers from moving to urban black areas where wages are lower, and poor information about jobs in other areas (Ihlanfeldt and Sjoquist, 1998) The role of spatial mismatch in Hispanic employment problems in the United States has been much less studied This is unfortunate because, in the international context, there appear to be more parallels between the Hispanic immigrant population in the United States and major immigrant populations in Western Europe than there are 200 394 454 388 432 447 451 381 460 413 390 559 567 556 556 200 150 200 150 171 171 171 169 169 169 Highly qualified Highly qualified Modestly qualified Modestly qualified No criminal record No criminal record Criminal record Criminal record No criminal record No criminal record No criminal record Criminal record Criminal record Criminal record Number of resumes sent High quality High quality High quality High quality High quality Low quality Low quality Low quality Low quality Low quality Type of resume Source:  Neumark and Rich (in progress) Pager et al (2009) United States Pager (2003) France Ravaud et al (1992) Australia Booth et al (2010) Study/year of test African-­American White African-­American White African-­American Latino White African-­American Latino White Disabled Able-­bodied Disabled Able-­bodied Chinese Italian Indigenous Middle Eastern White Australian Chinese Italian Indigenous Middle Eastern White Australian Ethnic or other group Table 10.2  Results for callback rates for labor market studies of discrimination with different qualifications 14.0 34.0 5.0 17.0 15.2 25.2 31.0 13.0 15.4 17.2 2.3 5.1 1.1 4.5 24.0 38.0 33.0 22.0 42.0 18.0 24.0 21.0 22.0 28.0 Callback rate % Ethnic hiring  ­201 between the situations of blacks in the United States and immigrants in Europe, including: language differences in some cases, such as Turks in Germany (Hillman, 2002) and Asians in Sweden (Zenou et al., 2010); residence in ethnic enclaves (Drever and Clark, 2006; Schönwälder, 2007); continuing economic and political ties with the origin countries of the immigrants; and of course the absence of a history of slavery Like for blacks and Hispanics in the United States, there is considerable residential segregation of minorities in Europe (Musterd, 2005) Newer research testing spatial mismatch tries to incorporate direct information on access to jobs that is related to either travel time or the extent of nearby jobs (for example, Ellwood, 1986; Ihlanfeldt and Sjoquist, 1990; Weinberg, 2000) These studies tend to show that blacks live in places with fewer jobs per person, and that this lower job access implies that blacks face longer commute times to jobs – although the differences may not be large and could conceivably be overcome relatively easily (Ellwood, 1986) However, if blacks with jobs and therefore higher incomes choose to live in areas with less job access (for example, consuming suburban amenities), this generates a downward bias in the estimated relationship between job access and employment (Ihlanfeldt, 1992) Evidence of longer commute times for blacks also does not necessarily point to spatial mismatch, as simple employment discrimination against blacks can imply fewer job offers and hence on average longer commute times for blacks even if they live in the same places as whites Overall, two comprehensive reviews argue that there is a good deal of evidence consistent with the spatial mismatch hypothesis (Holzer, 1991; Ihlanfeldt and Sjoquist, 1998), although Jencks and Mayer (1990) provide a more negative assessment of the hypothesis In recent work, Hellerstein et al (2008) ask whether other sources of barriers to minority employment could erroneously be interpreted as spatial mismatch The pure spatial mismatch hypothesis implies that it is only the location of jobs, irrespective of whether they are held by blacks or whites (but perhaps conditional on skill), which affects employment prospects But if discrimination, or labor market networks in which race matters, play important roles, then the distribution of jobs held by members of one’s own race may be a more important determinant of employment status Given that urban areas with large concentrations of black residents may also be areas into which whites tend to commute to work, it is possible that the employment problems of low-­skilled inner-­city blacks may not reflect simply an absence of jobs where they live, even at appropriate skill levels, but rather that the jobs that exist are more likely to be available to whites Hellerstein et al (2008) therefore study whether the relationship between job access and employment of blacks is driven solely by the spatial distribution of jobs, or whether the racial composition of those jobs is also important in explaining black employment They construct measures of job access at a disaggregated level, using confidential Census information on place of work In particular, they define local labor markets as the zip code in which a person resides, plus all contiguous zip codes, based on evidence that about one-­third of people work in these areas These job access measures are also constructed by skill (jobs at a skill level per resident at that same skill level) The research departs from the spatial mismatch literature by introducing the idea of racial mismatch, constructing measures of job density not only by location and skill, but also by race The regression models then estimate whether black employment is more sensitive to the spatial distribution of jobs held by blacks than to job density measured 202   International handbook on the economics of migration Table 10.3 Employment regressions for black men, alternative race-­specific density measures (1) (2) (3) (4) (5) (6) Job density measure: Nonblack jobs or black jobs/ black resident Density defined for: All LTHS1HSD LTHS All LTHS1HSD LTHS Non-­black or white   job density Black job density 001 (.0001) 008 (.002) 140 001 (.0001) 013 (.002) 140 0005 (.0001) 016 (.003) 140 0009 (.0001) 010 (.002) 140 0006 (.0001) 014 (.002) 140 0005 (.0002) 018 (.002) 140 R2 Male white jobs or male black jobs/ black male resident Notes: There are 533 198 observations on black men, and 4 030 425 on white men ‘LTHS’ refers to those without a high school diploma and ‘HSD’ represents high school graduates Regression estimates are from linear probability models, with standard errors (robust to nonindependence of observations within zip code areas) in parentheses All specifications include controls for age (linear and quadratic terms), marital status (a dummy variable for currently married), highest education (six categories including less than high school, high school degree, some college, Associate’s degree, Bachelor’s degree and advanced degree), residence in the central city, noncentral city, or suburbs, and MSA fixed effects Source:  Hellerstein et al (2008) without regard to race Note that if racial mismatch is important but one simply estimates models of the effects of overall (or skill-­specific) job density on black employment, one can still find evidence suggesting that job density matters, consistent with the spatial mismatch hypothesis The evidence is far more consistent with racial mismatch than with simple spatial mismatch Black job density (the ratio of local jobs held by blacks to black residents) strongly affects black employment, whereas white job density (the ratio of local jobs held by whites to black residents) does not.15 And the own-­race relationship is stronger at low skill levels For blacks, then, the spatial distribution of jobs alone is not an important determinant of black urban employment, rather it is the interaction of the spatial distribution of jobs combined with a racial dimension in hiring, or ‘racial mismatch’, that matters Some of the main evidence is reported in Table 10.3 Columns (1) and (4) report linear probability estimates for employment using a measure of overall job density broken down by race (for men and women combined and for black and white men only) Only job density for blacks is substantively related to the employment of blacks In each case, the estimated coefficient on the black job density measure is larger than that of the nonblack or white job density measure by a factor of about 10 When job density is measured based on lower educational levels – at most a high school degree (columns (2) and (5)), and high school dropouts (columns (3) and (6)) – the main difference is that the estimated effects of black job density are higher.16 The sharp differences in the estimated coefficients of the black versus the nonblack or white job density measures Ethnic hiring  ­203 Table 10.4 Calculation of effects of spatial distribution of jobs on black–white employment differential, black male high school dropouts A Mean employment rates Black male employment White male employment B Regression estimates of job density coefficients White male jobs/black male resident Black male jobs/black male resident C Mean job densities for black males White male jobs/black male resident Black male jobs/black male resident D Mean job densities for representative white males in same MSA as black males White male jobs/black male resident Black male jobs/black male resident E Predicted black male employment rate if black males faced job densities  of average white male in MSA (substituting job densities from Panel D into employment model) 0.459 0.690 002 (.0004) 028 (.004) 1.985 0.432 7.868 0.886 0.484 Notes: The estimates are from the same specification as in Table 10.3, column (6), including only high school dropouts in the sample The sample size is 129 348 The estimates in Panel D come from computing the average job densities (on a per black resident basis) for white male high school dropouts, taking the mean across whites in the MSA, assigning these to each black based on their MSA of residence, and then averaging across blacks Source:  Hellerstein et al (2008) (differences that are strongly statistically significant) still indicate that black job density is a much more important determinant of black employment than is nonblack or white job density.17 Together, this evidence is consistent with the notion that the spatial distribution of jobs matters for the employment of less-­educated blacks, but it is only the spatial distribution of jobs held by blacks that matters Thus, the racial mismatch hypothesis is a better characterization of how the spatial distribution of jobs affects black employment Even if blacks reside in areas that are dense in jobs at their skill level, if these jobs tend to be held by whites, the employment of black residents can be quite low Moreover, descriptive statistics reported in Hellerstein et al (2008) show that the density of jobs where blacks live is in fact quite high, even at blacks’ skill levels, suggesting that what is more important is which group is more likely to get hired Table 10.4 presents another way of making the point that the spatial distribution of jobs, per se, is not very important The estimated coefficients from an employment model are used to calculate the employment probability that would be implied if blacks lived where the representative white lived Attention is restricted to high school dropouts, for whom spatial mismatch (whether race-­specific or not) is most important.18 Panel A shows an employment rate gap of 0.231 Panel B reports estimates from the simplest 204   International handbook on the economics of migration model with race-­specific job densities The estimates reflect the same finding as above; the estimated effect of black job density is more than 10 times that of white job density Panel C reports the means of the job density measures for blacks There are considerably more white male jobs per black resident than black male jobs per black resident, averaged across blacks Panel D instead reports the means of the job density measures that blacks would face if they lived where the representative white in their Metropolitan Statistical Area (MSA) lived Whites on average live in areas where there are more jobs per black resident, whether held by whites or by blacks, although the difference is far greater for jobs held by whites Finally, employment probabilities are predicted using the estimated employment model in Panel B, but substituting the job density measures in Panel D for those in Panel C Because both job density estimates in Panel D are higher, the predicted employment rate for blacks is higher However, because the effect on black employment of white job density – which is what would increase most sharply if blacks lived where whites lived – is so small (0.002), the simulated change in residential location has very little effect on the predicted probability of employment for blacks The new predicted black probability is only higher than the actual mean by 0.025, which is a small share (10.8 percent) of the race difference in employment rates for these groups Thus, the evidence indicates that, in contrast to the spatial mismatch hypothesis, changing the spatial distribution of black residents would very little to increase black employment In contrast, the racial mismatch model predicts exactly that – simply shifting black residents to areas with high job density (even at the appropriate skill level) is unlikely to much to increase black employment More recent research establishes that the results are very similar for Hispanics in the US labor market (Hellerstein et al., 2010) Columns (1) and (4) of the top panel of Table 10.5 Table 10.5 Employment regressions for Hispanic men, alternative ethnicity-­specific density measures (1) (2) (3) (4) (5) (6) Job density measure: Non-­Hispanic jobs or Hispanic jobs/Hispanic resident Density defined for: All Poor English Immigrant All Poor English Immigrant Non-­Hispanic or   white job density Hispanic job density 001 (.0003) 022 (.006) 058 0003 (.0007) 016 (.003) 058 −.0001 (.0013) 028 (.005) 058 001 (.0003) 018 (.005) 058 003 (.002) 007 (.001) 058 001 (.001) 017 (.004) 058 R2 Male white jobs or male Hispanic jobs/Hispanic male resident Notes: There are 625 523 observations on Hispanics See notes to Table 10.3 Additional controls include dummy variables for the four Census categories of English proficiency, and for immigrant status ‘Poor English’ refers to the bottom two (of four) Census categories Source:  Hellerstein et al (2010) Ethnic hiring  ­205 report estimates of the similar equation to that used for blacks, but using a measure of overall job density broken down by Hispanic ethnicity The estimates indicate that only job density for Hispanics is substantively related to the employment of Hispanics In columns (2) and (5) job density is based on poor-­English speakers (that is, jobs held by poor-­English speakers, divided by poor-­English-­speaking residents, computed for Hispanics and non-­Hispanics or whites) The estimated effects of non-­Hispanic job density are very small, and insignificant in one case, while the estimated effects of Hispanic job density are much larger Columns (3) and (6) present similar evidence, but distinguishing workers by immigrant status instead Again, it is principally Hispanic job density that matters for Hispanic employment Labor market discrimination at a local level could give rise to a finding that blacks (Hispanics) are much more likely to be employed when they live in areas where many other blacks (Hispanics) hold jobs, but not when they live in areas where many nonblacks (non-­Hispanics) are employed For example, if the distribution of discriminatory employers or employees varies across areas, this kind of variation in minority employment could arise Yet when similar ‘racial mismatch’ specifications are estimated for white males, only white job density is associated with increases in white employment (Hellerstein et al., 2008) This casts doubt on discrimination as the principal explanation of the evidence 4 NETWORKS: IS THE PROBLEM WHO MINORITIES KNOW OR DON’T KNOW? The similar ‘racial mismatch’ findings for whites as well as Hispanics or blacks suggest that racially or ethnically stratified networks rather than discrimination may explain the results Stronger evidence of racial mismatch for low-­skilled workers and low-­skilled job density measures (Hellerstein et al., 2008, 2010) also suggests that networks may be important; networks may operate along many dimensions, but geographical links between workers are likely to be more important for lower-­skill jobs and workers, for which labor markets are more local A large body of evidence is consistent with labor market networks, much of it simply survey evidence indicating widespread reliance on friends, relatives, and acquaintances to find jobs (Ioannides and Datcher Loury, 2004).19 The evidence points to little difference between blacks and whites in the use of informal contacts in job search, higher rates of use of informal contacts among low-­educated workers, and substantially higher rates of use of informal contacts among Hispanics Subsequent work has noted that labor market networks may be race (or ethnic) based so that, for example, reliance on informal referrals in a predominantly white labor market benefits whites at the expense of other groups (Kmec, 2007) Focusing on the geographic or spatial dimension of networks, Bayer et al (2008) look for evidence of network effects among neighbors, using confidential Census data on Boston-­area workers They find that two individuals living on the same Census block are more likely to work on the same Census block than are two individuals living in the same block group but not on the same block As long as informal networks are stronger within blocks than within block groups, but unobserved differences are similar within blocks 206   International handbook on the economics of migration and block groups, this evidence suggests that residence-­based labor market ­networks affect hiring In recent work, Hellerstein et al (2011) assess evidence on the importance of labor market networks among neighbors, improving on Bayer et al (2008) by looking explicitly at who works at which establishment The approach is also used to ask whether networks are racially stratified, which can help explain the evidence of racial mismatch The study tests for the importance of residence-­based labor market networks in determining the establishments at which people work, using matched employer–employee data at the establishment level, based on a large-­scale dataset covering most of the United States (the 2000 DEED, described in Hellerstein and Neumark, 2003) The measure of labor market networks captures the extent to which employees of a business establishment come disproportionately from the same sets of residential neighborhoods (defined as Census tracts), relative to the residential locations of other employees working in the same Census tract but in different establishments.20 The method first computes the share of an individual’s co-­workers who are his or her residential neighbors, relative to the share that would result if the establishment hired workers randomly from the geographic areas where all individuals who work in the Census tract reside Residence-­based networks would predict that the share of neighbors among a worker’s co-­workers would be higher than would result from the random hiring process While random hiring provides a lower bound for the sorting of workers by neighborhoods across establishments, it is also important to construct an upper bound, because if establishments are larger than networks, perfect sorting by residence-­based networks across establishments cannot occur The measure of the importance of residential labor market networks is then the fraction of the difference between the lower and upper bounds of the extent to which a worker can work with neighbors that is actually observed in the data; this is called the ‘effective network isolation index’ In some of the analyses this is computed conditional on skill measures Overall, the evidence indicates that residence-­based labor market networks play an important role in hiring For blacks and whites, about 10 percent of the maximum amount to which residential networks could contribute to the sorting of workers by establishment is actually reflected in the sorting of workers into establishments However, blacks and whites work in very differently-­sized establishments, and when looking at much more homogeneous samples by race with respect to establishment size, the effective network isolation index for blacks is nearly double that for whites In addition, networks appear more important for less-­skilled workers, as would be expected for network connections among neighbors because of the more local nature of low-­skill markets For Hispanics, residence-­based networks are considerably more important; the grouping of workers from the same neighborhoods in the same business establishments is about 22 percent of the maximum, and as much as twice as high for Hispanic immigrants and those with poor English skills These results suggest that informal labor market networks may be particularly important for migrant workers who are not as well-­integrated into the labor market, and for whom employers may have less reliable information Hellerstein et al (2010) present a different kind of analysis, for Hispanics, intended to ask whether network effects likely underlie the racial (or in this case ‘ethnic’) mismatch evidence described earlier Traditional receiving areas for Hispanic immigrants have Ethnic hiring  ­207 been metropolitan Los Angeles, South Texas and South Florida The persistent spatial distribution of immigration suggests the importance of immigrant enclaves in helping immigrants integrate into the labor market Strikingly, however, between 1990 and 2000, when the Hispanic US population doubled, Hispanics established sizable communities in cities that traditionally had small Hispanic populations, with the growth of Hispanic communities in these cities driven primarily by changes in the destinations of new migrants to the United States For example, 1990 and 2000 Census data indicate that the Greensboro–Winston Salem–Highpoint MSA had fewer than 1000 non-­US-­born Hispanic adult males in 1990, but a decade later had over 20 000 (Hellerstein et al., 2010) Given the high transaction costs of migration, net migration of over 2000 percent in a decade suggests that these new migrants had information that the returns to moving to the Greensboro area were high, or more specific information that would make the returns high for them – exactly the kind of information that labor market networks might provide Moreover, network contacts in these new communities may have been especially important in securing employment for new migrants, given that the local economies did not have long histories of Hispanic employment and employers in these areas did not have much experience with Hispanic workers, especially poor-­English speakers As a consequence, if the relationship between density of jobs for Hispanics and employment of poor-­ English-­speaking Hispanic residents is particularly strong in the cities that experienced rapid recent growth of Hispanic immigrants, it is likely that this relationship is driven by network effects Regressions similar to those reported earlier, but for the top 50, 30 and 10 metropolitan areas in terms of the rate of growth between 1990 and 2000 of the nonnative Hispanic working-­age male population, are consistent with this prediction Table 10.6 reports estimates of the specification from column (1) of Table 10.5 The estimation uses the aggregate job density measures defined for either all Hispanic or all non-­Hispanic workers rather than the measure defined for poor-­English speakers alone because networks may well cross skill boundaries when workers are recruited or induced to move Table 10.6 Employment regressions for Hispanic men, ethnicity-­specific job density measures, cities with high growth rates of non-­US-born Hispanics (1990–2000) Growth rate of non-­US-­born Hispanics in MSA/PMSA Non-­Hispanic job density Hispanic job density R2 (1) All (2) Top 50 (3) Top 30 (4) Top 10 001 (.0003) 022 (.006) 058 −.0001 (.0003) 040 (.009) 045 −.0002 (.0002) 037 (.012) 044 −.001 (.0008) 088 (.028) 033 Note:  The specification corresponds to column (1) of Table 10.5 MSA: Metropolitan Statistical Area; PMSA: Primary Metropolitan Statistical Area Source:  Hellerstein et al (2010) 208   International handbook on the economics of migration to new locales to find employment Nonetheless, the expectation is still that the effect of Hispanic job density would be particularly pronounced for the poor-­English speakers for whom networks are likely to be most important Relative to the baseline estimates in column (1), which repeat the earlier estimates for the full sample, the effects of Hispanic job density are quite a bit larger for the metropolitan areas with the highest Hispanic immigrant growth, especially the narrowest set of such MSAs (the top 10) There is also evidence (not in Table 10.6) that the effects of Hispanic job density for those who speak poor English are much stronger in the MSAs with high Hispanic immigrant growth, again most markedly for the top 10 cities Perhaps more relevant with regard to hiring challenges faced by minorities is the question of whether labor market networks are racially or ethnically stratified The simple fact that networks based on neighborhood of residence are important points to ethnically or racially stratified networks After all, given pervasive ethnic and racial residential segregation in the United States, networks among neighbors have to be partially ethnicity or race based However, the methods used in Hellerstein et al (2011) can also be used to see whether there is ethnic or racial stratification of networks even within neighborhoods, with labor market information less likely to flow between, for example, black and white co-­residents than between co-­residents of the same race To examine this, the analysis of blacks is modified, treating the relevant set of a black worker’s neighbors and co-­workers to include either blacks or whites, and hence measuring the extent to which black workers are clustered in establishments with black or white co-­workers who are their neighbors – not just with black co-­workers who are their neighbors.21 If networks among co-­residents are racially stratified, then the likelihood that a black works with a neighbor regardless of race should be smaller than the likelihood that a black works with a black neighbor The evidence points to weaker network connections between black and white neighbors than between black neighbors Specifically, the empirical importance of networks disregarding the race of neighbors and co-­workers falls by more than 40 percent.22 The two findings from this research – that labor market networks are important and that these networks are racially stratified – can potentially explain the evidence of racial mismatch, that is, that higher local job density for one’s own race or ethnic group affects employment probabilities, but higher job density for other race or ethnic groups does not An area rich in jobs held by members of a group that is not networked strongly with residents may little to boost employment among that group Moreover, the existence of labor market networks that are stratified along racial or ethnic lines is consistent with evidence of establishment-­level segregation by race and ethnicity, documented in Hellerstein and Neumark (2008) for the United States, and Åslund and Skans (2010) for Sweden 5  CONCLUSIONS AND DISCUSSION This chapter has discussed research on three aspects of potential barriers to ethnic hiring: discrimination, spatial mismatch and networks This research presents challenges to, or at least questions about, the importance of discrimination and spatial mismatch With regard to discrimination, the essentially unanimous conclusion from audit or Ethnic hiring  ­209 correspondence studies is called into question by the possibility that these studies not actually identify discrimination A new method of dealing with this identification problem, applied to Bertrand and Mullainathan’s (2004) study of race discrimination, ends up reinforcing the finding of discrimination But this method, or other approaches to the identification problem that may be developed in the future, needs to be applied to data from both new and existing studies to see how robust the evidence of ethnic discrimination actually is Spatial mismatch is widely viewed as a partial contributor to the employment problems of minorities in the United States, especially blacks for whom the topic has been studied most extensively, and is often invoked in Europe as well (for example, Gobillon and Selod, 2007; Patacchini and Zenou, 2005) But new evidence for the United States suggests that the spatial distribution of jobs and workers disadvantages minorities for reasons more to with the hiring side, with the density of jobs held by one’s own race or ethnic group in an area being the only type of local job market density that matters Networks provide a potential explanation of this evidence Areas rich in jobs for unskilled Hispanics, for example, may provide more network connections to the labor market for other Hispanics, and hence increase their employment In contrast, given some evidence that networks are ethnically or racially stratified, areas rich in jobs for groups other than these minority groups, even at the same skill level, may not afford many opportunities for minorities At the same time, networks may be a two-­edged sword, as ethnic and racial minority groups may take advantage of labor market networks to find jobs when connections to the broader labor market are weak or unavailable This can give rise to the well-­known phenomenon of ethnic enclaves, which offers short-­term labor market gains, although there is debate about whether in the longer run these enclaves inhibit or encourage economic assimilation (Edin et al., 2003).23 Questions about the role of discrimination, spatial mismatch and networks have a long history in economics and sociology (for example, Becker, 1971; Granovetter, 1974; Kain, 1968).With respect to all three of these, the economics literature is still – in some cases newly – fertile and active, especially regarding the reinvigoration of research on discrimination spurred by the increasing popularity of field experiments and the new hypothesis of implicit discrimination (for example, Bertrand et al., 2005; Rooth, 2010), as well as the seeming flood of recent labor economics research on labor market networks The rising flows of economic migration into the developed economics (OECD, 2010) emphasize the continuing importance of understanding barriers to ethnic hiring, with the eventual goal of shaping policies to help eliminate these barriers Future research testing for spatial mismatch versus ‘racial mismatch’, and trying to study the role of labor networks and how they function, as well as additional direct evidence on discrimination, should give policy makers a better understanding of the tools needed to reduce barriers to ethnic hiring NOTES   * I am grateful to the editors, Amelie F Constant and Klaus F Zimmermann, and an anonymous referee for helpful comments   For a detailed discussion of the evidence on assimilation of migrants, see Chapter in this volume 210   International handbook on the economics of migration   This section draws heavily on Neumark (2012)   The approach is sometimes applied to employment (for example, Fairlie, 2005)   Another approach is to use worker and firm data to compare productivity and wage differentials between groups (Hellerstein et al., 1999)   Thorough reviews are contained in Fix and Struyk (1993), Riach and Rich (2002) and Pager (2007)   Chapter in this volume discusses attitudes toward immigrants that can underlie ethnic discrimination   The expectations Pk* and E(XkII), k E, NM, can be interpreted as conditional on XjI   Since this problem arises even in correspondence studies, I now refer exclusively to correspondence studies   Differences in the variances of unobservables across groups were introduced in early models of statistical discrimination (Aigner and Cain, 1977) 10 In an audit study the randomness that generates a statistical model arises naturally, as variables unobserved by the econometrician but observed by the firm can generate variation in hiring In a correspondI (assumed equal), the employer hires the higher variance group if the level ence study, given XEI and X NM of standardization is low, and vice versa One way to introduce unobservables that generate random variation, with different variances across groups, is to assume that there are random productivity differences across firms that are multiplicative in the unobserved productivity of a worker Alternatively, employers may make expectational errors and rather than assigning a zero expectation to the unobservable assign random draws based on the distributions of the unobservables 11 The implication of the second is that, as the preceding argument implies, with low standardization a high variance increases the probability of a callback 12 This was originally pointed out to me by Judith Rich 13 Analysis of these data using the methods described here is underway (Neumark and Rich, in progress) 14 Some of the discussion in this section and the next is taken from Hellerstein and Neumark (2012) 15 The finding that black employment tends to be higher when black job density is higher is not tautological The job density measure captures jobs located in an area divided by residents of that area; it is not simply the employment rate of residents 16 In these specifications, the density measure is low-­skilled jobs divided by low-­skill residents 17 Hellerstein et al (2008) also estimate specifications including interactions of the race-­ and education-­ specific job density measures with dummy variables for individuals’ education levels Regardless of the education level for which density is measured, the effect of black job density on black employment is much stronger than the effect of the corresponding nonblack job density, but the difference is much larger for less-­educated workers In addition, the effect of black job density for less-­educated blacks is stronger when this job density is defined based on less-­educated workers and residents, and the relationship is strongest when looking at less-­educated blacks, using the density measures defined for lower education levels 18 This simulation ignores any general equilibrium effects of people moving, and is therefore best thought of as calculating the change in predicted employment if a small number of black males moved to areas in which they faced the job densities of the representative white male in their MSA Results are similar for the broader low-­skill group with at most a high school education, and for black and nonblack males 19 Pellizzari (2010) provides recent, similar evidence for many European countries 20 Place of residence is treated as predetermined, potentially influencing place of work This appears to be a reasonable assumption, because the results reported below are similar when the sample is restricted to people who have lived at the same location for five or more years but have worked at their current employer for fewer than five years 21 This analysis was only done for blacks and whites 22 There is some other evidence consistent with racially or ethnically stratified networks in both the United States and Europe (Kasinitz and Rosenberg, 1996; Semyonov and Glikman, 2009) 23 This echoes a more general question as to whether jobs found through network contacts are better or worse than jobs found in other ways (Pellizzari, 2010) For evidence of positive effects once one accounts for firm-­level heterogeneity, see Dustmann et al (2011) REFERENCES Aigner, D.J and G Cain (1977), ‘Statistical theories of discrimination in labor markets’, Industrial and Labor Relations Review, 30 (2), 175–87 Åslund, O and O Nordstrom Skans (2010), ‘Will I see you at work? Ethnic workplace segregation in Sweden, 1985–2002’, Industrial and Labor Relations Review, 63 (3), 471–93 Ethnic hiring  ­211 Banerjee, A., M Bertrand, S Datta and S Mullainathan (2008), ‘Labor market discrimination in Delhi: evidence from a field experiment’, Journal of Comparative Economics, 38 (1), 14–27 Bayer, P., S Ross and G Topa (2008), ‘Place of work and place of residence: informal hiring networks and labor market outcomes’, Journal of Political Economy, 116 (6), 1150–96 Becker, Gary S (1971), The Economics of Discrimination, 2nd edn, Chicago, IL: University of Chicago Press Bertrand, M and S Mullainathan (2004), ‘Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination’, American Economic Review, 94 (4), 991–1013 Bertrand, M., D Chugh and S Mullainathan (2005), ‘New approaches to discrimination: implicit discrimination’, American Economic Review, 95 (2), 94–8 Booth, A.L., A Leigh and E Varganova (2010), ‘Does racial and ethnic discrimination vary across minority groups? Evidence from a field experiment’, IZA Discussion Paper No 4947, Institute for the Study of Labor (IZA), Bonn Bovenkerk, F., M Gras and D Ramsoedh (1995), ‘Discrimination against migrant workers and ethnic minorities in access to employment in the Netherlands’, International Migration Papers, Drever, A.I and W.A.V Clark (2006), ‘Mixed neighborhoods, parallel lives? Residential proximity and inter-­ ethnic group contact in German neighborhoods’, unpublished paper, University of Tennessee, Knoxville Dustmann, Christian, Albrecht Glitz and Uta Schönberg (2011), ‘Referral-­based job search networks’, IZA Discussion Paper No 5777, Institute for the Study of Labor (IZA), Bonn Edin, P.-­A., P Fredriksson and O Åslund (2003), ‘Ethnic enclaves and the economic success of immigrants – evidence from a natural experiment’, Quarterly Journal of Economics, 118 (1), 329–57 Ellwood, David (1986), ‘The spatial mismatch hypothesis: are there jobs missing in the ghetto?’, in Richard Freeman and Harry Holzer (eds), The Black Youth Employment Crisis, Chicago, IL: University of Chicago Press, pp. 147–85 Fairlie, R.W (2005), ‘An extension of the Blinder-­Oaxaca decomposition technique to logit and probit models’, Journal of Economic and Social Measurement, 30 (4), 305–16 Fix, Michael and Raymond Struyk (eds) (1993), Clear and Convincing Evidence: Measurement of Discrimination in America, Washington, DC: Urban Institute Press Gobillon, Laurent and Harris Selod (2007), ‘The effect of segregation and spatial mismatch on unemployment: evidence from France’, CEPR Discussion Paper No DP6198, Centre for Economic Policy Research (CEPR), London Granovetter, Mark S (1974), Getting a Job: A Study of Contacts and Careers, Cambridge, MA: Harvard University Press Heckman, J.J (1998), ‘Detecting discrimination’, Journal of Economic Perspectives, 12 (2), 101–16 Heckman, James and Peter Siegelman (1993), ‘The Urban Institute audit studies: their methods and findings’, in Michael Fix and Raymond Struyk (eds), Clear and Convincing Evidence: Measurement of Discrimination in America, Washington, DC: Urban Institute Press, pp. 187–258 Hellerstein, J.K and D Neumark (2003), ‘Ethnicity, language, and workplace segregation: evidence from a new matched employer–employee data set’, Annales d’Economie et de Statistique, 71–72, 19–78 Hellerstein, J.K and D Neumark (2008), ‘Workplace segregation in the United States: race, ethnicity, and skill’, Review of Economics and Statistics, 90 (3), 459–77 Hellerstein, J.K and D Neumark (2012), ‘Employment in black urban labor markets: problems and solutions’, in P.N Jefferson (ed.), The Oxford Handbook of the Economics of Poverty, Oxford: Oxford University Press, pp 164–202 Hellerstein, J.K., M McInerney and D Neumark (2010), ‘Spatial mismatch, immigrant networks, and Hispanic employment in the United States’, Annales d’Economie et de Statistique, 99/100, 141–67 Hellerstein, J.K., M McInerney and D Neumark (2011), ‘Neighbors and co-­workers: the importance of residential labor market networks’, Journal of Labor Economics, 29 (4), 659–95 Hellerstein, J.K., D Neumark and M McInerney (2008), ‘Spatial mismatch vs racial mismatch?’, Journal of Urban Economics, 64 (2), 467–79 Hellerstein, J.K., D Neumark and K Troske (1999), ‘Wages, productivity, and worker characteristics: evidence from plant-­level production functions and wage equations’, Journal of Labor Economics, 17 (3), 409–46 Hillman, F (2002), ‘A look at the “hidden side”: Turkish women in Berlin’s ethnic labour market’, International Journal of Urban and Regional Research, 23 (2), 267–82 Holzer, H.J (1991), ‘The spatial mismatch hypothesis: what has the evidence shown?’, Urban Studies, 28 (1), 105–22 Ihlanfeldt, Keith R (1992), Job Accessibility and the Employment and School Enrollment of Teenagers, Kalamazoo, MI: W.E Upjohn Institute for Employment Research Ihlanfeldt, K.R and D Sjoquist (1990), ‘Job accessibility and racial differences in youth employment rates’, American Economic Review, 80 (1), 267–76 212   International handbook on the economics of migration Ihlanfeldt, K.R and D Sjoquist (1998), ‘The spatial mismatch hypothesis: a review of recent studies and their implications for welfare reform’, Housing Policy Debate, (4), 849–92 Ioannides, Y.M and L Datcher Loury (2004), ‘Job information, networks, neighborhood effects, and inequality’, Journal of Economic Literature, 42 (4), 1056–93 Jencks, Christopher and Susan E Mayer (1990), ‘Residential segregation, job proximity and black job opportunities’, in Laurence Lynn and Michael McGeary (eds), Inner-­City Poverty in the United States, Washington, DC: National Academy Press, pp. 187–222 Kaas, L and C Manger (2012), ‘Ethnic discrimination in Germany’s labour market: a field experiment’, German Economic Review, 13 (1), 1–20 Kain, J (1968), ‘Housing segregation, negro employment, and metropolitan decentralization’, Quarterly Journal of Economics, 82 (2), 175–97 Kasinitz, P and J Rosenberg (1996), ‘Missing the connection: social isolation and employment on the Brooklyn waterfront’, Social Forces, 43 (2), 180–96 Kmec, J.A (2007), ‘Ties that bind? Race and networks in job turnover’, Social Problems, 54 (4), 483–503 Korenman, S and D Neumark (1991), ‘Does marriage really make men more productive?’, Journal of Human Resources, 26 (2), 282–307 Korenman, S and D Neumark (1992), ‘Marriage, motherhood, and wages’, Journal of Human Resources, 27 (2), 233–55 Mincy, R (1993), ‘The Urban Institute audit studies: their research and policy context’, in Michael Fix and Raymond Struyk (eds), Clear and Convincing Evidence: Measurement of Discrimination in America, Washington, DC: Urban Institute Press, pp. 165–86 Musterd, S (2005), ‘Social and ethnic segregation in Europe: levels, causes, and effects’, Journal of Urban Affairs, 27 (3), 331–48 Neumark, D (1988), ‘Employers’ discriminatory behavior and the estimation of wage discrimination’, Journal of Human Resources, 23 (3), 279–95 Neumark, D (1996), ‘Sex discrimination in restaurant hiring: an audit study’, Quarterly Journal of Economics, 111 (3), 915–41 Neumark, D (2012), ‘Detecting discrimination in audit and correspondence studies’, Journal of Human Resources, 47 (4), 1128–57 Neumark, D and M McLennan (1995), ‘Sex discrimination and women’s labor market outcomes’ Journal of Human Resources, 30 (4), 713–40 Neumark, D and Judith Rich (in progress), ‘Do field experiments of markets overestimate discrimination?’ Oaxaca, R (1973), ‘Male–female wage differentials in urban labor markets’, International Economic Review, 14 (3), 693–709 Organisation for Economic Co-­operation and Development (OECD) (2010), ‘International migration database’, OECD International Migration Statistics, available at: http://stats.oecd.org/BrandedView.aspx?oecd_ bv_id5mig-­data-­en&doi5data-­00342-­en (accessed December 2011) Pager, D (2003), ‘The mark of a criminal record’, American Journal of Sociology, 108 (5), 937–75 Pager, D (2007), ‘The use of field experiments for studies of employment discrimination: contributions, critiques, and directions for the future’, The Annals of the American Academy of Political and Social Science, 609 (1), 104–33 Pager, D., B Western and B Bonikowski (2009), ‘Discrimination in a low-­wage labor market: A field experiment’, American Sociological Review, 74 (5), 777–99 Patacchini, E and Y Zenou (2005), ‘Spatial mismatch, transport mode and search decisions in England’, Journal of Urban Economics, 58 (1), 62–90 Pellizzari, M (2010), ‘Do friends and relatives really help in getting a good job?’, Industrial and Labor Relations Review, 63 (3), 494–510 Ravaud, J.-­F., B Madiot and I Ville (1992), ‘Discrimination towards disabled people seeking employment’, Social Science & Medicine, 35 (8), 951–8 Riach, P.A and J Rich (2002), ‘Field experiments of discrimination in the market place’, The Economic Journal, 112 (483), F480–F518 Rooth, D.-­O (2010), ‘Automatic associations and discrimination in hiring: real world evidence’, Labour Economics, 17 (3), 523–34 Schönwälder, Karen (2007), ‘Residential segregation and the integration of immigrants: Britain, the Netherlands and Sweden’, WZB Discussion Paper No SP IV 2007-­602, WZB Wissenschaftszentrum Berlin für Sozialforschung, Berlin Semyonov, M and A Glikman (2009), ‘Ethnic residential segregation, social contacts, and anti-­minority attitudes in European societies’, European Sociological Review, 25 (6), 693–708 Smeeters, B and A Nayer (1998), ‘La discrimination l’accès l’emploi en raison de l’origine étrangère: le cas de la Belgique’, International Migration Papers, 23 Ethnic hiring  ­213 Weinberg, B (2000), ‘Black residential centralization and the spatial mismatch hypothesis’, Journal of Urban Economics, 48 (1), 110–34 Zenou, Y., O Åslund and J Östh (2010), ‘How important is access to jobs? Old question – improved answer’, Journal of Economic Geography, 10 (3), 389–422 11  Immigrants in risky occupations* Pia M Orrenius and Madeline Zavodny 1  INTRODUCTION Although immigrants can be found in virtually every occupation across the globe, many immigrants hold ‘three D’ jobs: jobs that are dirty, dangerous and difficult Risky jobs may be attractive to immigrants who have low skills, little education and limited fluency in the host country language These jobs also may pay more than other jobs immigrants would hold in the host country A growing literature examines whether immigrants are disproportionately employed in risky jobs and, if so, why This chapter surveys the literature on immigrant–native differences in occupational risk After a brief explanation of the economic theory of occupational risk and compensating differentials, the chapter surveys the literature on whether immigrants are disproportionately employed in risky jobs and whether they are more likely than natives to experience work-­related injuries or fatalities It then discusses the limited literature on immigrant–native differences in risk premiums It closes with a discussion of areas for future research 2  ECONOMIC THEORY The standard model of efficient labor markets predicts that risky occupations will pay higher wages to compensate workers for incurring more risk Profit-­maximizing employers trade off higher compensation costs and the cost of reducing occupational risk The result is a concave wage offer curve in which wages increase with occupational risk at a declining rate, as shown in Figure 11.1 Workers view risk as a disamenity and are willing to hold risky occupations only in exchange for higher wages, called ‘compensating differentials’ or risk premiums Workers’ convex indifference curves show the tradeoff between wages and occupational risk If workers differ in their willingness to bear risk, those who are the most willing to trade off increased risk for higher wages will sort into the riskier occupations In terms of Figure 11.1, workers with steeper indifference curves, like UN, will sort into relatively safe occupations while workers with flatter indifference curves, like UI, will sort into relatively risky occupations Workers in riskier jobs will earn a compensating differential or risk premium, w(RI) − w(RN) The standard model posits that all workers face the same wage offer curve However, there are likely to be different wage offer curves for different types of workers For example, if employers can assess workers’ productivity, less-­productive workers will face a lower wage offer curve At every level of job risk, the employer will offer a lower wage to less-­productive workers, which is equivalent to a downward shift of the wage offer curve In addition, the wage offer curve may be flatter for some workers This is the case 214 Immigrants in risky occupations  ­215 Wage UI UN Wage offer curve w(RI) w(RN) RI RN Occupational risk Figure 11.1 Wage offer curve and indifference curves for workers with different willingness to bear risk Wage Wage offer curve N UN wN(RN) UI wN(0) Wage offer curve I wI(RI) wI(0) RN RI Occupational risk Figure 11.2  Different wage offer curves for workers with different cost of reducing risk if some workers impose higher safety costs on the employer at greater occupational risk levels The wage curve is flatter to offset the higher costs to the employer of reducing occupational risk for such workers Workers who both are less productive and impose higher safety costs on the employer at greater occupational risk levels will face a lower and flatter wage offer curve This situation is illustrated in Figure 11.2 The flatter the wage offer curve, the smaller the compensating differential for occupational risk If the wage offer curve is both low and 216   International handbook on the economics of migration flat enough for some workers, those workers may not earn any compensating differential for occupational risk This situation is illustrated in Figure 11.2 Workers in riskier occupations earn less than workers in safer occupations, or wI(RI) , wN(0) as in Figure 11.2 Hersch and Viscusi (2010) refer to this situation as segmented labor markets The model thus far assumed that markets are efficient In efficient markets, workers have full information about occupational risk Dávila et al (2011) discuss another possibility: some workers may underestimate occupational risk If some workers underestimate occupational risk, and increasingly so for riskier occupations, they think they are on a different wage curve than they actually are Their perceived wage curve is like the higher, steeper wage offer curve in Figure 11.2, while their actual wage offer curve is like the lower, flatter curve Such misperceptions can also result in some workers not earning compensating differentials The difference from the earlier scenario is that such misperceptions can create rents for employers at the expense of the workers who underestimate occupational risk The empirical literature on compensating differentials in the general population is mixed Reviews by Smith (1979) and Viscusi (1993) conclude that workers earn compensating differentials for the risk of death but that there is little clear evidence that workers earn compensating differentials for less extreme job hazards, such as non-­fatal illnesses or injuries Implications for Immigrants Immigrants may differ from natives in several ways related to the compensating differentials model First, immigrants may have different tradeoffs between wages and risk than natives, on average Immigrants typically have less human capital and less wealth, which may cause them to be more willing to accept occupational risk in exchange for higher wages In terms of Figure 11.1, immigrants may have flatter indifference curves than natives This is particularly likely to be the case for target earners who migrate temporarily to work It also may be true of immigrants who have relatively limited labor market opportunities, such as unauthorized immigrants or those with small networks Second, immigrants may face different wage offer curves than natives because of higher safety-­related costs Hersch and Viscusi (2010) posit that this is particularly likely to be the case for immigrants with limited fluency in the host country language It is more expensive for employers to provide safety training to such workers Safety-­related costs also may be higher for immigrants from countries with lower standards of job safety than the host country because those workers may need more training or take too many risks by host country standards Third, immigrants may face different wage offer curves than natives as a result of imperfect markets As noted by Dávila et al (2011), immigrants may underestimate occupational risk in their host country This could occur because occupational risk is very high in their home country and so they mistakenly believe that risk is even lower in the host country than it actually is Alternatively, it could arise from employers deliberately misinforming immigrants about occupational risk Employers may be more able to mislead immigrants who are not proficient in the host country language, are relatively recent arrivals, are unauthorized or have smaller networks Immigrants are thus theoretically more likely to work in risky occupations than Immigrants in risky occupations  ­217 natives for a variety of reasons They may earn smaller compensating differentials for doing so or even no compensating differentials at all This chapter next turns to the empirical evidence on these issues 3  EMPIRICAL EVIDENCE Recent research in several advanced economies indicates that immigrants are more likely to work in risky occupations than natives (See also Chapter 12 in this volume.) Consistent with this, immigrants experience higher rates of occupational injuries and fatalities than natives in most advanced economies There are exceptions, however This section first reviews the evidence in these areas It then discusses the evidence on the causes of these immigrant–native differences and on whether immigrants and natives earn similar compensating differentials for working in risky occupations Immigrant–Native Differences in Occupational Risk Studies of immigrant–native differences in occupational risk typically examine whether immigrants work in occupations or industries with higher injury and fatality rates, on average, than natives This is an indirect measure of occupational risk that reflects the occupational distribution of immigrants and natives In the United States, several recent studies show that immigrants work in riskier jobs than natives Using data from 2003–05, Orrenius and Zavodny (2009) conclude that immigrants worked in occupations and industries with higher fatality and injury rates than natives The average immigrant worked in an industry with a 38 percent higher fatality rate than the average native Using data from the same period, Hersch and Viscusi (2010) similarly conclude that immigrants worked in industries with higher fatality rates than natives, on average, using either overall or native-­specific fatality rates The difference is concentrated among Mexican immigrants, who worked in sectors with fatality rates 36 to 46 percent higher than non-­Mexican immigrants Consistent with this finding, Dávila et al (2011), using data from 1999–2000, show that Hispanic immigrant men worked in occupations with higher fatality and injury rates than native-­born Hispanic, non-­Hispanic white and non-­Hispanic black men These findings mark a change from older research on immigrant–­native differences in occupational risk in the United States Using data from 1979–80, Berger and Gabriel (1991) show that the average immigrant worked in an industry with a 21 percent lower fatality rate than the average native Using data from 1991, Hamermesh (1998) shows that immigrants did not work in industries with higher injury rates than white natives This change in findings coincides with substantial changes in the origin countries of immigrants to the United States and in their average characteristics The 1965 Immigration and Nationality Act caused immigration to shift from Europe to Latin America and Asia over ensuing decades The relative position of immigrants in the wage structure worsened (Borjas, 1995) The fraction of US immigrants able to speak English well declined, and average educational attainment fell relative to US natives The unauthorized population increased, swelling from about to million in 1980 (Warren and Passel, 1987) to 8.5 to 10 million in 2000 (Costanzo et al., 2002; Hoefer et al., 2006) 218   International handbook on the economics of migration Research indicates that immigrants work in riskier jobs in Canada Premji et al (2010) find that the proportion of immigrants working in a job is positively related to the job’s risk of injuries and illnesses in Montreal Smith et al (2009) conclude that Canadian immigrants are more likely to be employed in physically demanding occupations, putting them at greater risk of work-­related injuries Research also indicates that immigrants work in riskier jobs than natives in Spain Solé et al (2010) report that 36 percent of immigrants hold jobs that expose them to risks, compared with 26 percent of natives Immigrants from Africa, Latin America and the European periphery (non-­EU 15 countries) were more likely than Spanish natives to work in risky jobs, while Asian immigrants were less likely Díaz-­Serrano (2010) concludes that African immigrants work in riskier jobs than Spanish natives in Catalonia, Spain Immigrant–Native Differences in Occupational Injuries and Fatalities Studies of immigrant–native differences in occupational injuries and fatalities examine whether immigrants experience more such adverse events than natives Whereas the studies of occupational risk discussed above apply industry-­ or occupation-­specific injury or fatality rates to workers, these studies examine directly whether immigrants are more likely than natives to be injured or killed because of workplace incidents Some studies analyze official reports of workplace injuries and fatalities, while others use data from surveys that ask about work-­related injuries In the United States, immigrants have higher work-­related fatality rates than natives Loh and Richardson (2004) report that the work-­related fatality rate was 33 percent higher among immigrants than the overall rate during 1996–2001 The workplace fatality rate rose among immigrants during the second half of that period while falling among natives Fatalities were particularly high among immigrants from Mexico, accounting for 40 percent of all fatalities to foreign-­born workers; Mexico accounts for about 30 percent of US immigrants Richardson et al (2003) note that foreign-­born Hispanic workers had higher fatality rates during 1995–2000 than both Hispanic and non-­ Hispanic native-­born workers Death rates due to workplace homicides were also higher among the foreign-­born, particularly among Asians, than among natives (Sincavage, 2005) Some evidence also indicates that immigrants are more likely to experience work-­ related injuries than natives in the United States Hao (2008) finds that immigrants had a 32 percent greater risk of experiencing a nonfatal workplace injury than natives, based on surveys from 1996 to 2004 Marvasti (2010) reports that official workplace injury rates are higher in US states with higher fractions of workers who are immigrants, particularly Hispanic immigrants Sinclair et al (2006) conclude that immigrants were more likely than natives to have experienced a work-­related injury that required medical attention, based on data from 2000 to 2003 However, Zhang et al (2009) conclude that immigrants were less likely to have experienced a work-­related injury that required medical attention during 1997–2005 than US natives The source of the difference is unclear since both studies use data from the National Health Interview Surveys In Canada, Smith and Mustard (2009) conclude that immigrants were not significantly more likely to report having experienced a work-­related injury than natives However, Immigrants in risky occupations  ­219 controlling for education and other observable characteristics, recent male immigrants were more likely to report having experienced a work-­related injury requiring medical attention than natives In Spain, Ahonen and Benavides (2006) conclude that official reports indicate higher rates of fatal and non­fatal occupational injuries among immigrants than among natives in 2003 In 2005, López-­Jacob et al (2008) estimate that immigrants had a 34 percent higher workplace fatality rate and a 13 percent higher non­fatal workplace injury rate than Spanish natives overall; immigrants’ occupational risk was lower than natives’ in construction, commerce, restaurants and hotels, however In Australia, the overall rate of work-­related fatalities was similar among immigrants and natives, but fatality rates were higher among immigrants than natives in rural (farming, fishing, hunting, timber and related workers) and mining occupations (Corvalan et al., 1994) Research indicates higher workplace injuries and fatalities in several other advanced economies, although there are exceptions According to Bollini and Siem (1995), reports indicate higher occupational accident rates among immigrant workers than native workers in France, Germany, the Netherlands and Switzerland The European Agency for Safety and Health at Work (2008) reports that immigrant workers have higher workplace accident rates in France, Germany and Spain; in Sweden and Finland, in contrast, studies of specific industries indicate no significant difference between immigrants and natives in the risk of work-­related accidents Wu et al (1997) conclude that the overall rate of workplace injuries was lower among legal migrant workers in Taiwan than among Taiwanese natives working in the same industries, but higher for female migrant workers than for female natives Interestingly, immigrants might be expected to have lower rates of work-­related injuries and fatalities than natives because of the ‘healthy immigrant’ effect Immigrants tend to be positively selected in terms of health attributes when they migrate, and they experience negative assimilation towards natives’ health in some countries (Antecol and Bedard, 2006) Despite greater exposure to risks at work, immigrants are less likely to become disabled than natives in Spain (Solé et al., 2010) The authors note this could be due to the healthy immigrant effect Reasons for Immigrant–Native Differences There are several reasons why immigrants tend to have higher occupational injury and fatality rates than natives The simplest explanation is that immigrants are overrepresented in risky occupations, as discussed above But why, in turn, are immigrants more likely to work in risky jobs? Research has focused on the role of immigrants’ lower levels of human capital and greater willingness to incur risk Workers with relatively low levels of human capital are more likely to work in manual labor jobs that involve more risk Less-­educated workers typically have fewer job choices, lower incomes and less wealth The compensating differentials model predicts that these factors make workers more willing to trade off higher wages for increased job risk It is therefore not surprising that immigrants are more likely to work in risky jobs and have higher injury and fatality rates in countries in which immigrants have lower average educational attainment than natives An interesting, relatively unexplored 220   International handbook on the economics of migration ­ uestion is whether immigrants and natives with similar levels of educational attainment q (and other characteristics) have similar levels of occupational risk.1 Limited proficiency in the host ­country language appears to play an important role in immigrants’ higher occupational risk In the United States, limited-­English-­proficient Hispanic male immigrants work in occupations with higher fatality and injury rates than their English-­proficient counterparts (Dávila et al., 2011) In Australia, immigrant workers from non-­English-­speaking countries have higher work-­related mortality rates than immigrants from English-­speaking countries or natives, particularly in the first few years after arrival (Corvalan et al., 1994) Other research also reports a negative relationship between occupational risk and years since migration in Taiwan (Wu et al., 1997) and in the United States (Hao, 2008), which could be due in part to increased proficiency in the host country language over time Immigrants also may have different risk preferences than natives, on average, for reasons unrelated to human capital Immigrants may be less risk averse than natives After all, the fact that immigrants have chosen to move to another country is consistent with them being less risk averse than their countrymen who stayed behind However, Bonin et al (2009, 2012) find that first-­generation immigrants in Germany are more risk averse than German natives Immigrants who migrate in search of higher earnings may also be more willing to accept greater risk in exchange for higher wages Interviews with recent migrant workers in England and Wales suggest that they work in risky sectors and occupations because of the importance they place on earning as much as possible as quickly as possible (McKay et al., 2006) In Spain, the occupation-­industry injury and fatality rate is negatively related with job satisfaction among natives but not immigrants (Díaz-­Serrano, 2010) This is consistent with immigrants being less risk averse than natives Alternatively, immigrants and natives may have similar risk preferences but immigrants perceive risks to be lower than natives There is little direct evidence on whether immigrants underestimate risk compared with natives For example, interviews of migrant workers in England and Wales conclude that some interviewees underestimated job risks (McKay et al., 2006), but the study does not include a comparison with natives The negative relationship between years since migration and occupational risk reported in some studies (Hao, 2008; Orrenius and Zavodny, 2009; Wu et al., 1997) is consistent with immigrants’ estimate of occupational risk becoming more accurate over time as well as with increased fluency in the host country language Other market imperfections could also result in immigrants working in riskier jobs Ethnographic studies conclude that many immigrants, particularly the undocumented, are reluctant to complain about unsafe working conditions because they are concerned about losing their job or being deported or because they are not aware of their rights (for example, Ahonen et al., 2009; Brown et al., 2002; Walter et al., 2002) Underreporting Underreporting is a continual concern in studies of occupational risk Underreporting is a well-­documented phenomenon in the workers’ compensation system in the United States (for example, Biddle et al., 1998) and Canada (for example, Shannon and Lowe, 2002) Employers have an incentive to underreport payroll and injuries in order to reduce Immigrants in risky occupations  ­221 the workers’ compensation insurance premiums they have to pay Employers may give workers paid time off as an incentive to not claim workers’ compensation, or employers may threaten to dismiss workers if they file a claim Work-­related injuries, illnesses or fatalities also may not be correctly attributed to work, particularly if the injury, illness or fatality occurs later, such as asbestos exposure that results in cancer decades later Underreporting may be higher for immigrants than natives for several reasons Immigrants may be less likely than natives to report occupational injuries and illnesses, perhaps because they are less aware of their rights or less likely to correctly attribute them to their employment Alternatively, employers may be better able to discourage immigrants from filing a claim for a work-­related injury or illness In addition, surveys and official statistics will underestimate workplace injuries, illnesses and fatalities among immigrants if they leave the host country after incurring an injury or illness Lower rates of health insurance coverage among immigrants than natives – an issue in the United States and other countries without universal coverage – also may result in greater underreporting among immigrants if the uninsured are less likely to see a healthcare provider when injured or ill The evidence on whether immigrants underreport work-­related injuries and illnesses to their employers is inconclusive In the United States, between 63 and 71 percent of low-­wage, low-­skilled immigrant workers surveyed in California and New York who had experienced a work-­related injury or illness said they had reported it to a supervisor (Brown et al., 2002; Gany et al., 2011) Migrant workers interviewed in Britain said they often did not report workplace accidents for fear of being dismissed (McKay et al., 2006) A study of recent immigrants in Canada concluded that most, but not all, injured workers reported their injury to a healthcare provider or their employer; many said they were discouraged from filing a workers’ compensation claim or misled by their employer about their rights (Kosny et al., 2011) None of these studies include a comparison to native-­born workers, so it is unclear whether immigrants are less likely to report work-­ related injuries and illnesses than natives Some evidence does suggest that immigrants’ work-­related injuries and illnesses are underreported in official records relative to those of natives Recent immigrants to Canada were less likely than natives to report receiving benefits from workers’ compensation or other programs after experiencing work-­related injury (Smith et al., 2009) A survey conducted in Trentino, Italy, concluded that immigrants were less likely to officially report occupational injuries than natives, often because their employer did not want them to, because they were employed illegally, or because they feared losing their job (Martinelli, 2011) Higher rates of self-­employment may result in greater underreporting of work-­related injuries and illnesses among immigrants Immigrants are more likely than natives to be self-­employed in Canada (Hou and Wang, 2011) and the United States (Orrenius and Zavodny, 2011), among others Self-­employed workers may fall through the cracks of the workers’ compensation system and other official records, particularly if they are day laborers working off the books Finally, immigrant–native differences in occupational risk will be underestimated if immigrants misrepresent their nativity for fear of deportation This may be particularly true of unauthorized immigrants Employers who hire unauthorized immigrant workers also may misrepresent workers’ nationality because they not know it or because they are concerned about penalties for hiring such workers (See also Chapter in this volume.) 222   International handbook on the economics of migration Compensating Differentials Economic theory predicts that workers will earn smaller or even no compensating differentials, or risk premiums, if they impose greater safety costs on employers or work in imperfectly competitive labor markets Markets might not be perfectly competitive if workers not have complete information about occupational risk or if employers have some monopsony power over workers, among other reasons As discussed above, these situations may apply disproportionately to immigrants However, some studies of immigrant–native differences in compensating differentials find that immigrants earn larger, not smaller, compensating differentials than natives The literature to date has only examined the United States Berger and Gabriel (1991) report that risk premiums are 25 percent higher among immigrants than among natives; immigrants who work in industries with higher fatality rates not only earn more relative to those working in safer industries, but their average return for doing so is 25 percent larger than the corresponding return among natives Dávila et al (2011) conclude that limited-­English-­proficient Hispanic male immigrants earn larger risk premiums than other Hispanic immigrants or US natives for working in occupations with higher fatality rates Hersch and Viscusi (2010) find no significant immigrant–native difference in compensating differentials among most groups of immigrants There are cases where immigrants earn smaller compensating differentials than natives Dávila et al (2011) find that English-­proficient Hispanic male immigrants incur an earnings penalty – a negative risk premium – for working in jobs with higher fatality rates Hersch and Viscusi (2010) conclude that immigrants from Africa and Mexico not earn compensating differentials for working in jobs with higher fatality rates Their further investigation of Mexican immigrants reveals that those who understand English receive a higher wage premium for fatality risks than those who not This is the opposite of Dávila et al (2011)’s conclusion, warranting further research on this issue These results leave it unclear why some groups of immigrants earn larger risk premiums than natives while others – mainly from Mexico, according to Hersch and Viscusi (2010) – earn smaller or no risk premiums If smaller risk premiums are due to higher safety costs, workers who are not proficient in the host country language would be expected to earn smaller risk premiums If they are due to market imperfections, workers who are more likely to underestimate risk or be taken advantage of by employers would be expected to earn smaller risk premiums These circumstances may disproportionately apply to unauthorized immigrants There is no direct evidence this is the case, however Dávila et al (2011) find that Hispanic immigrant males who are not naturalized US citizens earn positive risk premiums Hersch and Viscusi (2010) report that foreign-­born workers who now have US permanent resident status but previously were unauthorized immigrants not earn significantly different risk premiums than other immigrants 4  DISCUSSION Although there is a growing literature on immigrants and occupational risk, there are several important areas for further research Researchers need a better understanding of why immigrants are more likely to work in risky jobs in most countries Do immi- Immigrants in risky occupations  ­223 grants choose to work in risky jobs because they are less risk averse and simply want to earn higher wages or limited choices push them into risky jobs? The finding that occupational risk declines with years since migration is consistent with either of these possibilities Mixed evidence on whether limited-­English-­proficient Hispanic immigrants earn compensating differentials in the United States also makes it difficult to distinguish between these two possibilities It is also important to determine whether immigrants earn large enough risk premiums to compensate them for greater occupational risk The limited evidence suggests that most groups of immigrants earn compensating differentials for fatality risk, but it is not clear whether these differentials fully compensate them for higher risks Future research could examine immigrant–native differences in workers’ compensation and insurance coverage, and how such differences relate to differences in compensating differentials There are several other relatively unexplored areas Do immigrants and natives have similar perceptions of occupational risks? What is the role of networks in immigrants’ occupational risk? Within occupations and industries, employers assign immigrants to riskier jobs? Answering the last question would require careful analysis of whether there are systematic differences between immigrants’ occupational risk as measured by overall industry and occupation injury and fatality rates versus immigrants’ actual ­workplace injury and fatality rates Research in this area has focused almost exclusively on the first generation, or people who immigrate How the second generation, the children of immigrants, compares with the first generation and to natives has received little attention Bonin et al (2009, 2012) show that in Germany, the second generation has risk attitudes similar to natives The extent of assimilation in occupational risk both within and across immigrant generations is worthy of additional research Another interesting question is whether immigration leads to changes in occupational risk Do employers decrease safety standards when immigration increases, or does immigration induce changes in the wage structure that make it possible for employers to reduce occupational risk? Research using German data from 1976 concludes that a higher share of foreign guestworkers in a firm is associated with fewer severe accidents among the firm’s native workers but has no effect on the number of accidents among the guestworkers within a firm (Bauer et al., 1998) Meanwhile, a higher share of skilled guestworkers is associated with more nonsevere accidents among the guestworkers within a firm Further research on the interaction between immigrant shares, skill levels and occupational risk among both natives and immigrants is needed Most research on immigrant–native differences in occupational risk has focused on low-­skilled workers Hersch and Viscusi (2010) restrict their sample to blue-­collar jobs, for example More research on high-­skilled immigrants and occupational risk is warranted Skilled immigrants may be at heightened risk for workplace accidents if they experience occupational downgrading after migrating and are not accustomed to manual labor Skilled immigrants also may take on risk in other ways, such as starting their own business Finally, this survey only includes studies on developed countries Given the high levels of South–South migration, research on immigrants and occupational risks in developing countries is a key area for future research Occupational risks may be higher in 224   International handbook on the economics of migration ­ eveloping countries, and workers may have less access to workers’ compensation bend efits or healthcare if they experience work-­related injuries or illnesses NOTES * We thank Amelie F Constant, Klaus F Zimmermann, and an anonymous referee for helpful comments We also thank Amy Chapman for her excellent research assistance The opinions expressed herein are those of the authors and not necessarily reflect the views of the Federal Reserve Bank of Dallas or the Federal Reserve System Orrenius and Zavodny (2009) report that the immigrant–native gap in occupation or industry injury and fatality risks is smaller when controlling for education, language ability and other observable characteristics, but it remains significant for some measures REFERENCES Ahonen, E.Q and F.G Benavides (2006), ‘Risk of fatal and non-­fatal occupational injury in foreign workers in Spain’, Journal of Epidemiology and Community Health, 60 (5), 424–6 Ahonen, E.Q., V Porthé, M.L Vázquez, A.M García, M.J López-­Jacob, C Ruiz-­Frutos, E Ronda-­Pérez, J Benach and F.G Benavides (2009), ‘A qualitative study about immigrant workers’ perceptions of their working conditions in Spain’, Journal of Epidemiology and Community Health, 63 (11), 936–42 Antecol, H and K Bedard (2006), ‘Unhealthy assimilation: why immigrants converge to American health status levels?’, Demography, 43 (2), 337–60 Bauer, Thomas K., Andreas Million, Ralph Rotte and Klaus F Zimmermann (1998), ‘Immigration labor and workplace safety’, IZA Discussion Paper No 16, Institute for the Study of Labor (IZA), Bonn Berger, M.C and P.E Gabriel (1991), ‘Risk aversion and the earnings of US immigrants and natives’, Applied Economics, 23 (2), 311–18 Biddle, J., K Roberts, K.D Rosenman and E.M Welch (1998), ‘What percentage of workers with work-­ related illnesses receive workers’ compensation benefits?’, Journal of Occupational and Environmental Medicine, 40 (4), 325–31 Bollini, P and H Siem (1995), ‘No real progress towards equity: health of migrants and ethnic minorities on the eve of the year 2000’, Social Science & Medicine, 41 (6), 819–28 Bonin, H., A.F Constant, K Tatsiramos and K.F Zimmermann (2009), ‘Native–migrant differences in risk attitudes’, Applied Economics Letters, 16 (15), 1581–6 Bonin, H., A.F Constant, K Tatsiramos and K.F Zimmermann (2012), ‘Ethnic persistence, assimilation and risk proclivity’, IZA Journal of Migration, 1, Article Borjas, G (1995), ‘Assimilation and changes in cohort quality revisited: what happened to immigrant earnings in the 1980s?’, Journal of Labor Economics, 13 (2), 201–45 Brown, Marianne P., Alejandra Domenzain and Nelliana Villoria-­Siegertt (2002), ‘Voices from the margins: immigrant workers’ perceptions of health and safety in the workplace’, UCLA Labor Occupational Safety and Health Program Report, University of California Los Angeles Corvalan, C.F., T.R Driscoll and J.E Harrison (1994), ‘Role of migrant factors in work-­related fatalities in Australia’, Scandinavian Journal of Work and Environmental Health, 20 (5), 364–70 Costanzo, Joseph M., Cynthia David, Caribert Irazi, Danidel Goodkind and Roberto Ramirez (2002), ‘Evaluating components of international migration: the residual foreign born’, Population Division Working Paper No 61, US Census Bureau, Suitland, MD Dávila, A., M.T Mora and R González (2011), ‘English-­language proficiency and occupational risk among Hispanic immigrant men in the United States’, Industrial Relations, 50 (2), 263–96 Díaz-­Serrano, Luis (2010), ‘Do legal immigrants and natives compete in the labour market? Evidence from Catalonia’, IZA Discussion Paper No 4693, Institute for the Study of Labor (IZA), Bonn European Agency for Safety and Health at Work (2008), ‘Literature study on migrant workers’, available at: http://osha.europa.eu/en/publications/literature_reviews/migrant_workers (accessed 10 September 2011) Gany, F., R Dobslaw, J Ramierz, J Tonda, I Lobach and J Leng (2011), ‘Mexican urban occupational health in the U.S.: a population at risk’, Journal of Community Health, 36 (2), 175–9 Hamermesh, Daniel S (1998), ‘Immigration and the quality of jobs’, in Daniel S Hamermesh and Frank D Immigrants in risky occupations  ­225 Bean (eds), Help or Hindrance? The Economic Implications of Immigration for African Americans, New York: Russell Sage Foundation, pp. 75–106 Hao, Lingxin (2008), ‘Workplace nonfatal injuries among immigrants to the U.S.’, mimeo, Johns Hopkins University Department of Sociology, Baltimore, MD, available at: http://paa2008.princeton.edu/download aspx?submissionId580386 (accessed 15 November 2011) Hersch, J and W.K Viscusi (2010), ‘Immigrant status and the value of statistical life’, Journal of Human Resources, 45 (3), 749–71 Hoefer, M., N Rytina and C Campbell (2006), Estimates of the Unauthorized Immigrant Population Residing in the United States: January 2005, Annual Population Estimate 2006, Washington, DC: US Department of Homeland Security, Office of Immigration Statistics Hou, Feng and Shunji Wang (2011), ‘Immigrants in self-­employment’, Perspectives on Labour and Income (Statistics Canada), available at: http://www.statcan.gc.ca/pub/75-­001-­x/2011003/article/11500-­eng.pdf (accessed December 2011) Kosny, Agnieszka, Marni Lifshen, Ellen MacEachen, Peter Smith, Gul Joya Jafri, Cynthia Neilson, Diana Pugliese and John Shields (2011), ‘Delicate dances: immigrant workers’ experiences on injury reporting and claim filing’, Institute for Work and Health, Toronto, available at: www.iwh.on.ca/system/files/at-­work/ at_work_65.pdf (accessed November 2011) Loh, K and S Richardson (2004), ‘Foreign-­born workers: trends in fatal occupational injuries, 1996–2001’, Monthly Labor Review, 127 (6), 42–53 López-­Jacob, M.J., E.Q Ahonen, A.M García, Á Gil and F.G Benavides (2008), ‘Comparación de las Lesiones por Accidente de Trabajo en Trabajadores Extranjeros y Espoles por Actividad Económica y Comunidad Autónoma’, Revista Espola de Salud Pública, 82 (2), 179–87 Martinelli, D (2011), ‘Victims of occupational injuries: a comparison between migrants and Italians’, Rivista di Criminologia, Vittimologia e Sicurrezza, (2), 101–21 Marvasti, A (2010), ‘Occupational safety and English language proficiency’, Journal of Labor Research, 31 (4), 332–47 McKay, Sonia, Marc Craw and Deepta Chopra (2006), ‘Migrant workers in England and Wales: an assessment of migrant worker health and safety risks’, Research Report Series RR502, London Metropolitan University, Working Lives Research Institute Orrenius, P.M and M Zavodny (2009), ‘Do immigrants work in riskier jobs?’, Demography, 46 (3), 535–51 Orrenius, Pia M and Madeline Zavodny (2011), ‘From brawn to brains: how immigration works for America’, Federal Reserve Bank of Dallas Annual Report, available at: http://dallasfed.org/fed/annual/2010/ar10.pdf (accessed 10 May 2011) Premji, S., P Duguay, K Messing and K Lippel (2010), ‘Are immigrants, ethnic and linguistic minorities over-­represented in jobs with a high level of compensated risk? Results from a Montréal, Canada study using Census and workers’ compensation data’, American Journal of Industrial Medicine, 53 (9), 875–85 Richardson, S., J Ruser and P Suárez (2003), ‘Hispanic workers in the United States: an analysis of employment distributions, fatal occupational injuries, and non-­fatal occupational injuries and illnesses’, in Board on Earth Sciences and Resources, Safety is Seguridad: A Workshop Summary, Washington, DC: National Academies Press, pp. 43–82 Shannon, H.S and G.S Lowe (2002), ‘How many injured workers not file claims for workers’ compensation benefits?’, American Journal of Industrial Medicine, 42 (6), 467–73 Sincavage, J.R (2005), ‘Fatal occupational injuries among Asian workers’, Monthly Labor Review, 128 (10) 49–55 Sinclair, S.A., G.A Smith and H Xiang (2006), ‘A comparison of nonfatal unintentional injuries in the United States among U.S.-­born and foreign-­born persons’, Journal of Community Health 31 (4), 303–25 Smith, P.M and C.A Mustard (2009), ‘Comparing the risk of work-­related injuries between immigrants to Canada and Canadian-­born labour market participants’, Occupational and Environmental Medicine, 66 (6), 361–7 Smith, P.M., A.A Kosny and C.A Mustard (2009), ‘Differences in access to wage replacement benefits for absences due to work-­related injury or illness in Canada’, American Journal of Industrial Medicine, 52 (4), 341–9 Smith, R.S (1979), ‘Compensating wage differentials and public policy: a review’, Industrial and Labor Relations Review, 32 (3), 339–52 Solé, Meritxell, Luis Díaz-­Serrano and Marisol Rodríguez (2010), ‘Work, risk and health: differences between immigrants and natives in Spain’, IZA Discussion Paper No 5338, Institut zur Zukunft der Arbeit (IZA), Bonn Viscusi, W.K (1993), ‘The value of risks to life and health’, Journal of Economic Literature, 31 (4), 1912–46 Walter, N., P Bourgois, H.M Loinaz and D Schillinger (2002), ‘Social context of work injury among ­undocumented day laborers in San Francisco’, Journal of General Internal Medicine, 17 (3), 221–9 226   International handbook on the economics of migration Warren, R., and J.S Passel (1987), ‘A count of the uncountable: estimates of undocumented aliens counted in the 1980 United States Census’, Demography, 24 (3), 375–93 Wu, T.-­N., S.-­H Liou, C.-­C Hsu, S.-­L Chao, S.-­F Liou, K.-­N Ko, W.-­Y Yeh and P.-­Y Chang (1997), ‘Epidemiologic study of occupational injuries among foreign and native workers in Taiwan’, American Journal of Industrial Medicine, 31 (5), 623–30 Zhang, X., S Yu, K Wheeler, K Kelleher, L Stallones and H Xiang (2009), ‘Work-­related non-­fatal injuries among foreign-­born and US-­born workers: findings from the U.S National Health Interview Survey, 1997–2005’, American Journal of Industrial Medicine, 52 (1), 25–36 12  Occupational sorting of ethnic groups* Krishna Patel, Yevgeniya Savchenko and Francis Vella 1  INTRODUCTION Immigrant occupational choice has an important impact on labor market outcomes of the natives and immigrants in the host country This impact manifests itself on the wage rates and employment opportunities of both groups Thus occupational sorting among immigrants is an important area of study particularly to inform policy ­questions focused on identifying occupations in which new waves of immigrants may be employed, understanding their sorting process, estimating the impact from ­externalities on natives from immigrants’ occupational choices, and total economy welfare Prior to evaluating the impact of the occupational sorting of immigrants, it is worth examining why immigrants choose to seek employment in the certain occupations Closely related to this question is the type of immigrant that is attracted to the United States For example, some studies have shown that immigrants sort into occupation based on their skill sets, which reflects both the initial immigration decision and the immigration policies in the host country Immigrants who decide to migrate have to frequently overcome legal and financial barriers, and take on a considerable risk to enter the host country.1 Only those with sufficient motivation and ability to overcome these obstacles decide to immigrate Such unobserved characteristics can result in positive selection of immigrants to the host country On the other hand, if the cost of migration is sufficiently low, one might observe a negative selection into migration based on ­educational level Skill-­based immigration policies attract immigrants with certain professional qualifications to enter the country based on labor demand Such immigrant skill composition affects the associated occupational outcomes Alternatively, when immigrants enter the host country under a policy that is not skill based, their occupational choices may be limited or unpredictable at the time of entry Moreover immigrants often rely on social networks for job referrals Job referrals through this channel tend to reflect the set of occupations in which the network members are employed This may increase the job offer arrival rate for such jobs for unemployed immigrants, a phenomenon potentially explaining why certain jobs are dominated by certain ethnic groups in a host country This chapter discusses recent studies on the topic of immigrant occupational sorting Section discusses studies on skill-­based occupational sorting, section discusses the role of social networks in occupational outcomes and section concludes 227 228   International handbook on the economics of migration 2  Occupational Sorting based on immigrant skill 2.1  Introduction Immigrant occupational choices depend on the supply of the skills that immigrants bring with them and demand for these skills in the host country On the supply side, potential migrants self-­select into migration Once in the host country, immigrant skills become one of the major determinants of the immigrant’s occupational outcomes Unobserved individual characteristics can lead to either positive or negative selection, based on skills, into migration These characteristics also influence occupational sorting in the host country One strand of literature discusses positive self-selection, as only the highly motivated, higher-­ability immigrants can overcome the costs of immigration (Chiquiar and Hanson, 2005; Docquier and Marfouk, 2006) Alternatively, Borjas (1987) shows that negative selection can occur if income inequality is relatively higher in the country of origin than in the host country, while the costs to immigrate are relatively low Once in the host country, the type of occupations immigrants hold depends on the skills they possess, including prior education, the ability to speak the host-­country language and compatibility of credentials 2.2  Immigration Policies From a demand perspective, immigration policies determine the skill composition of immigrants in the host country Host countries form their immigration policies to achieve a desired skill composition of the labor force and might be directed either towards skilled or unskilled migration For example, Australia has a skilled migration program A migrant must hold one of the occupations that are specified by the Australian government to qualify under this program Skilled migration is regulated by a similar program in Canada If a potential migrant does not have a job offer in Canada, in order to be eligible for the skilled migration program she must have at least one year of experience in an occupation from the list specified by the Canadian government In the United States skilled immigration is regulated by H1-­B visa policy under which employers can hire foreign workers in specialty occupations for the period of up to six years Sometimes the governments of countries issue special visas to temporarily attract unskilled labor For example, the H2-­B visa in the United States is designed to hire temporary or seasonal unskilled workers The UK also has a history of skill-­based immigration policies, including policies under which it recruits seasonal unskilled workers, as well as a points-­based policy under which it recruits high-­skilled workers.2 2.3  Skill-­Based Occupational Sorting A skill-­based immigration policy may result in suitable occupational outcomes for high-­ skilled immigrants This is noted in Chiswick (2010), which studies the impact of being over-­or-­under-­educated on earnings for immigrants in Australia, a country with a skill-­ based immigration policy Chiswick (2010) uses the over-­education/required education/ under-­education (ORU) methodology to measure the effect of over-­ or-­under educa- Occupational sorting of ethnic groups  ­229 tion on earnings separately for immigrants and natives using data from the Australian Standard Classification of Occupations.3 Comparing the return to education for immigrants with that for natives helps determine whether immigrants are more or less suitably matched than natives The ORU methodology focuses on three variables in a linear regression with wages as the outcome variable The first variable denotes the number of years an individual is over-­educated compared with the standard education requirement of his or her occupation The variable takes the value zero if the individual is not over-­educated The second variable measures the number of years of education that is required for the occupation The third variable is the number of years an individual is under-­educated The immigrants in the sample, on average, are more educated than natives and this reflects the skill-­based nature of the Australian immigration policy However, the estimates show that the return for immigrants to being over-­educated is lower than the return for native-­born workers These findings suggest that immigrant education earned abroad is less valuable than that earned in the host country The return to over-­education for immigrants from non-­English speaking countries is even lower On the other hand, the wage penalty for under-­education is lower for immigrants than it is for natives This may be due to unobservable characteristics of immigrants such as perseverance and high motivation that typically reflects the positive self-­selection of immigrants One issue that arises when determining the impact of over-­ or under-­education is the measurement of education Studies on this topic are limited by the datasets to hand Years of schooling is the typical measure for education The obvious limitation of this variable is that it does not capture years of relevant schooling Additionally, this measure does not distinguish between the quality and quantity of education Also, differences in the length of schooling for a specific occupation may vary by country For example, an immigrant may have spent many years training to be a nurse in the home country, when the schooling requirement in the destination country may be only a few years The immigrant would be classified as being over-­educated, when there a similar level of training Finally, the quantity of education may not have a direct relationship with job ability for certain occupations For example, many of the most successful workers in the computer sciences field have acquired relevant skills without formal training In addition to measuring education, studies on this topic need to measure required education for the occupation needs to be determined According to Chiswick and Miller (2010a), this could be done by either worker self-­reported assessment of the job requirement, the realized mean or median education level for each occupation within a sample, or some specified dataset that compiles information on job requirements, such as the Australian Classification of Occupations Host countries with immigration policies that are not based on skill can result in occupational mismatch for high-skilled immigrants For example, immigrants that have entered the country as refugees or from family reunification policies may not have suitable employment opportunities that match their skill sets The results from Chiswick (2010) show that the return with schooling can be influenced by such differences in immigration policy Chiswick (2010) compares the returns with schooling based on data on Australia with those based on US data The study finds that a larger part of the earnings gap between immigrants and natives is explained by being over-­or ­under-educated 230   International handbook on the economics of migration in Canada and Australia while it explains a lower part of the gap in the United States, where many immigrants have not entered under a skill-­based policy It explains an even smaller part for immigrants from less developed countries This suggests that characteristics other than years of schooling impact earnings differentials between immigrants and natives in the United States Chiswick and Miller (2010b) further investigate the mismatch of high-­skilled immigrants in the United States This paper uses the ORU approach to measure the impact of skill mismatch on wages among high-skilled adult male immigrants, comparing the wage impact on natives using the 2000 US Census data As in Chiswick (2010), Chiswick and Miller (2010b) find that the payoff for education is higher for natives For high-­skilled workers with graduate degrees, the payoff to education is 11 percent for natives and only 5.5 percent for immigrants; the payoff to skilled workers with a bachelor’s degree was about the same for both groups at 11 percent Overall, the return to over-­education is much lower than the return to having the required level of schooling This conclusion is robust across broad occupation categories Interestingly, there can even be a negative impact of over-­education on immigrant earnings For immigrants with a bachelor’s degree, the impact of over-­education is negative until after about nine years spent in the United States, while for those with a master’s degree it is negative for 20 years In addition to examining the impact of years of schooling, Chiswick and Miller (2010b) look at work experience as a determining factor of wages in the host country Comparing high-­skilled and low-­skilled immigrants, the payoff for labor market experience is much higher for high-­skilled immigrants than it is for lower-­skilled workers The payoff for pre-­immigration labor market experience is 1.62 percent for foreign-­born skilled workers while it is 0.86 percent for all foreign-­born workers The evidence in the literature suggests that immigrant skills acquired in the origin country are not directly transferable to the host country, particularly when an immigrant entry to the host country is not based on skill For example, Chiswick and Miller (2010b) compare the impact of a change in immigration policy, to one that is more skill based, on earnings within the United States They note that the 1990 Immigration Act in the United States increased the number of labor certification/employer sponsored visas The degree of over-­education for immigrants is lower for individuals who immigrated after 1990 according to data, as more of these immigrants are suitably matched If immigrants are not placed within the context of a skill-­based immigration policy, they may rely on social networks to find the jobs The challenge arises for skilled immigrants who may not know other immigrants skilled in their profession who may provide them with information about job vacancies Mismatch can arise due to the lack of professional contacts Patel (2010) provides theoretical motivation for immigrant mismatch in the host country using a search model based on a Mortensen and Pissarides framework (Mortensen and Pissarides, 1994; Pissarides, 2000) and Calvó-­Armengnol and Zenou (2005) to explain the impact of networks on earnings and employment of skilled natives and immigrants In the United States, average wages of skilled immigrants are lower than those of natives Additionally, fewer skilled immigrants are employed in professional occupations These labor market outcomes are modeled as being the result of differences in productivity as well as differences in the size of professional networks The model is calibrated using observed data on wages and underemployment rates Occupational sorting of ethnic groups  ­231 from the Current Population Survey Implied values of productivity and network size from the model show that natives have larger networks than immigrants Naturalized citizens – established immigrants – have higher productivity than natives but lower network sizes Immigrants that are not naturalized (newer immigrants) have both lower productivity and lower network sizes 2.4  Immigrant Task Specialization Another way that immigrants might sort themselves into different occupations is sorting based on comparative advantage Under comparative advantage immigrants will sort into occupations in which they can perform relatively better than other individuals in the economy Moreover, sorting based on comparative advantage has implications for the economy’s wage distribution (see Roy, 1951) This section starts by reviewing the literature on task specialization of low-­skilled workers and then discusses the literature on occupational sorting of high-­skilled workers The literature on immigrant task specialization is relatively new Ottaviano and Peri (2012) show that immigrants and natives are imperfect substitutes within education-­ experience groups This means that immigrants and natives have comparative advantage in different tasks or occupations and thus will be better off specializing in those tasks This finding spurred growth in the literature on immigrant and native task specialization (Amuedo-­Dorantes and De la Rica, 2011; Peri and Sparber, 2009, 2011) In general, the literature finds that immigrants are specializing in manual task-­intensive occupations and natives in communication task-­intensive occupations Peri and Sparber (2009) present a model in which immigrants have a comparative advantage in manual task-­intensive occupations while natives have a comparative advantage in communication task-­intensive occupations An increase in the number of low-­skilled immigrants will decrease wages in manual tasks and increase wages in communication tasks Since those tasks are complementary, natives will reallocate to communication task-­intensive occupations where wages will increase in response to immigrant inflow Moreover, the increase in natives’ wages paid for communication tasks will compensate for losses in natives’ wages paid for manual tasks The authors highlight that this would explain the almost nonexistent effect of immigrants on natives’ wages found by previous literature (Card, 2001, 2007; Card and Lewis, 2007; Lewis, 2005) Peri and Sparber (2009) provide empirical support for the theoretical model using the data from 1960–2000 decennial US Census and the Occupational Information Network (O*NET) database on occupational skills requirements They find that on average less-­ educated immigrants supplied more manual labor relative to communication tasks than did natives In states with large low-­skilled immigrant inflows native workers are more likely to shift to more communication-­intensive occupations than they are in states with smaller immigrant inflows Moreover, in states with a large influx of low-­skilled immigrants, wages paid in communication-­intensive occupations rose Finally, the authors find that immigration decreased the average real wages of native low-­skilled workers by 0.3 percent between 1990 and 2000 However, the simulations show that without such task specialization the loss to native wages would have been four times higher – 1.2 percent 232   International handbook on the economics of migration Amuedo-­Dorantes and De la Rica (2011) apply the Peri and Sparber (2009) model to Spanish data Spain presents an interesting empirical case since the share of immigrants in the total population rose from to 11 percent from 2000 to 2008 Moreover, the authors find that occupations in Spain are often segregated by gender They account for this additional labor market characteristic in their analysis of immigrants and natives Amuedo-­Dorantes and De la Rica (2011) using 1999–2007 Encuestas de Población Activa (EPA) consider the impact of lower-­skilled immigrants4 on task specialization and occupational distribution of natives in Spain First, they find that the impact of an increase in male (female) immigrants on the occupational distribution of female (male) natives is negligible Then, similar to Peri and Sparber (2009), they find that increase in share of male (female) immigrants leads to lower supply of manual tasks by male (female) natives Moreover, there are implications for the occupational distribution An increase in male immigrants leads to a decrease in male natives employed in both skilled blue-­collar jobs and nonskilled blue-­collar jobs (these are jobs that include craftsmen, operators and assemblers) and an increase in male natives employed in skilled white-­ collar jobs (these are jobs that include managers and other professionals) For females, an increase in female immigrants leads to a decline in native women employed in nonskilled blue-­collar jobs (such as domestic service) and an increase in their employment in skilled white-­collar jobs (such as nursing and teaching) Both Peri and Sparber (2009) and Amuedo-­Dorantes and De la Rica (2011) acknowledge the potential endogeneity problem in identifying the effect of the inflow of immigration on the task specialization of natives and use an instrumental variable methodology to correct for it Peri and Sparber (2009) use two instruments: (1) the imputed share of Mexican workers based on 1960 state demographics, and (2) the interaction of decade dummies with distance of a state’s center of gravity to the closest Mexico–US border and with a dummy variable for the states that border with Mexico Amuedo-­Dorantes and De la Rica (2011) use the share of low-­skilled recent immigrants as an instrumental ­variable for the share of low-­skilled, long-­term immigrants Chiswick and Taengnoi (2007) and Peri and Sparber (2011) consider occupational choice based on task specialization among high-­skilled immigrants in the United States Chiswick and Taengnoi (2007) provide an empirical study of occupational choice of highly educated immigrants (those with educational attainment of college and above) using 2000 US Census data Using a multinomial logit model, they find that immigrants whose mother tongue is distant5 from English tend to choose technical professions such as engineering and IT where English and communication skills are not crucial Peri and Sparber (2011) – using the Occupational Information Network (O*NET) database, the 2003–08 Current Population Survey, and the 1990 US Census and 2002–07 American Community Survey – find that immigrants and natives with graduate degrees are not perfect substitutes The authors conclude that immigrants with graduate degrees typically specialize in occupations that require quantitative and analytical skills, while natives go for occupations that require interactive and communication skills Moreover, they find that increases in the share of immigrant workers with a graduate degree in an occupation cause highly educated natives to switch to occupations with more interactive and communication-based t­ asks Occupational sorting of ethnic groups  ­233 3 IMMIGRANT OCCUPATION SORTING BASED ON SOCIAL NETWORKS Ethnic social networks play an important role in immigrant occupational sorting Networks may have a positive or negative impact on occupational choice and labor market outcomes of immigrants Social networks also may influence the immigration decision from the country of origin, which impacts selection of immigrants and their skill composition in the host country 3.1  Networks and Immigrant Selection Occupational choices of immigrants in the host country depend on the skills obtained in the country of origin Thus, self-­selection into migration has implications not only for immigrants’ labor market outcomes in the host countries, including wages and the probability of employment, but also for occupation choices There is ample literature on self-­selection into migration that explains that the two major components of the migration decision are the wage differential between the host country and origin country, and the immigration costs Some researchers find positive or intermediate selection of immigrants (Chiquiar and Hanson, 2005; Docquier and Marfouk, 2006) based on education while others find negative selection (Borjas, 1987) Most of these papers focus on wage differentials as the determining factor of migration while keeping the cost of migration fixed Another wave of literature studies the self-­selection into migration based on the costs (Beine et al., 2011; Bertoli, 2010; McKenzie and Rapoport, 2010) These papers highlight the fact that ethnic networks may reduce the cost of migration, including financial and psychological costs, and help immigrants to settle in destination countries In general, this literature finds that networks are associated with negative sorting into migration McKenzie and Rapoport (2010) suggest that, a priori, one should expect networks to cause negative self-­selection into migration based on skills This arises as potential migrants with low skills are more likely to face credit constraints and more likely to need more help in assimilating in the host country than are high-­skilled immigrants They are more likely to benefit from network ties, particularly if they not speak the language of the host country McKenzie and Rapoport (2010) study Mexican immigration into the United States Using the 1997 Ecuesta Nacional de la Dinamica Demografica, they find a slightly positive or neutral self-­selection of Mexicans into migration among Mexican communities with small networks of migrants On the other hand, there is a negative self-­selection into migration in communities with strong migration networks, as strong migration networks lower migration costs Bertoli (2010) studies migrants from Ecuador to the United States, Spain and other countries The study uses the Ecuador Labor Force Survey (2005), American Community Survey (2005–08 rounds) and the Encuesta Nacional de Inmigrantes (2007) for Spain, and comes to a similar conclusion that networks increase the probability of negative self-­selection into migration based on education Both McKenzie and Rapoport (2010) and Bertoli (2010) acknowledge the endogeneity problem of networks and selection into immigration McKenzie and Rapoport (2010) use historic state-­level migration rates as an instrumental variable for networks Bertoli 234   International handbook on the economics of migration (2010) argues that the variation associated with the severe financial crisis of 1998 in Ecuador can be considered as a natural experiment The crisis was initiated by a number of unexpected events: the historic low oil prices,6 halt in oil exports due to a damage in oil pipeline after the earthquake, and major disruptions in infrastructure and losses due to the hurricane El Niño Thus the major driving factor of post-­crisis (post 1998) migration was the state of the economy and not the network.7 These findings imply that immigrants who rely on networks when deciding to migrate are low skilled One can expect that those immigrants would also rely on networks when they arrive in the host country Thus, networks in the host country will shape the labor market outcomes and the occupational choices of recent immigrants 3.2  Theoretical Background on Social Networks Immigrant social networks in the host country play an important role in immigrant labor market and occupational outcomes This section discusses some of the papers in a large body of literature on the effects of social networks on labor market outcomes The literature provides theoretical insights into the process by which social networks influence wages and employment probabilities of network members Different models predict different impacts of networks on labor market outcomes of members Montgomery (1991) is the seminal paper in the literature on the impact of social networks on labor market outcomes It presents a two-­period model with firms and workers The model predicts that in the equilibrium workers that are connected through a network may have higher wages than those who not belong to a network Additionally, firms may increase profits by hiring through referrals from existing employees Moreover, the model predicts that changes in the socio-­economic structure alter the wage distribution by increasing the network density and fostering network inbreeding (that is, similarity between network members) which generates greater wage dispersion Calvó-­Armengol and Jackson (2004) develop a model which explains the role of social networks on individual employment The authors find that employment is positively correlated across time and agents (for individuals within a network) Moreover, the authors show that if staying in the labor market is costly, then the difference between the employment rate for a network that started with a low employment rate and a network that started with a high employment rate will persist over time Calvó-­Armengol and Jackson (2007) extend Calvó-­Armengol and Jackson (2004) to study the wage effects Similarly, they find that wages are positively correlated across time and agents (for individuals within a network) and that wage inequality may persist between the network that started with a low average wage and the network that started with a high average wage Also, the model shows that while networks generally have a positive effect on labor market outcomes, such as wages and the probability of being employed, under certain conditions they may have a negative effect on these outcomes Bentolila et al (2010) study the impact of social networks on occupational choice The authors develop a theoretical model based on Pissarides (2000) The model makes the following predictions: (1) the workers who find jobs through social networks have a lower unemployment duration than those who find jobs through formal channels, (2) jobs found through social networks pay lower wages than jobs found through formal channels, (3) an increase in the number of social contacts reduces a worker’s expected Occupational sorting of ethnic groups  ­235 productivity in equilibrium, and (4) under some parameters an increase in the share of workers with many contacts may decrease social welfare In other words, under some conditions the workers who find the jobs through networks may be mismatched in terms of their occupation In addition, the authors find empirical support for the model’s predictions based on an empirical analysis using US and European data They find that social contacts reduce the time of search, suppress wages and lead to lower aggregate productivity among individuals in the network 3.3  Empirical Studies on Social Networks and Immigrant Occupational Choice This section discusses some empirical studies that show that immigrant networks impact immigrant labor market outcomes and occupational choice Information about job opportunities within a social network reflects the occupation distribution of the employed network members Therefore, new immigrants who rely on networks for job opportunities are more likely to be employed in those occupations chosen by their peers If these are low-­skilled occupations – those with low barriers to entry – then network effects can result in a sizeable presence of certain immigrant groups in certain occupations Patel and Vella (forthcoming) study these effects using US Census data Specifically, the empirical question is whether a new immigrant is more likely to be employed in the popular occupation of his country-­group if there are more countrymen employed in that occupation The paper notes that immigrants tend to cluster in specific occupations across metropolitan areas Patel and Vella (forthcoming) show examples of these clusters, presenting some of the occupations with the highest growth rates of immigrant concentrations For example, immigrants from Jamaica represented 39 percent of nursing aids in Fort Lauderdale in 2000, up from percent in 1980, while Jamaicans represented only percent of the population It is interesting to note that these occupation concentrations are present in the metropolitan area level, while there is no substantial concentration present at the national level; Jamaicans represented only percent of total workers in the nursing aids profession nationwide Most of these occupations are low skilled, which suggests that immigrant sorting is not occurring due to comparative advantage reflecting a skill set predominant in the ethnic group Furthermore, certain immigrant groups who are dominant in one occupation in a particular metropolitan area may be dominant in another occupation elsewhere, while the occupation may be dominated by another ethnic group elsewhere Patel and Vella (2007) show the number of unique popular occupations across metropolitan areas for each country in the sample Since not every country-­group is represented in each metropolitan area, the table also includes a measure that is normalized by the number of metropolitan areas The higher the value, the higher is the occupation dispersion across metropolitan areas For most countries, this measure  is greater than 0.5 indicating substantial dispersion An interesting feature of these trends is that these trends are occurring based on three digit occupation definitions The empirical model is based on the assumption that new immigrants – defined as those who have immigrated within five years of the survey – are interacting with other countrymen in the metropolitan area in which they reside First, for each metropolitan area and a country-­group, the most popular occupation among the ‘old’ countrymen – immigrants who have been in the United States for at least five years – is determined The social network is then defined as the share of old countrymen in the popular occupation 236   International handbook on the economics of migration in the metropolitan area Estimates are based on individual level data of new immigrants, with the outcome variable being a binary variable that takes the value if the individual is employed in the popular occupation and otherwise Estimates show the presence of network effects on the occupation choice of new immigrants The probability of a new immigrant selecting a popular occupation increases between and percent as a result of a one percentage point increase in the share of countrymen in the popular occupation These network effects also have a positive impact on wages showing that new immigrants are better off working in the popular occupation than then they would be otherwise The wage premium for being in the most popular occupation is as high as 12 percent for male immigrants in the 2000 Census While Patel and Vella (forthcoming) estimate the wage effect of selecting the popular occupation, other studies (Beaman, 2012; Lalonde and Topel, 1991; Savchenko and Vella, 2012) have focused on estimating the impact of the size of previous immigrant cohorts on immigrant wages These papers emphasize that an influx of immigrants increases competition for existing immigrants, which has a negative wage impact Beaman (2012) shows that immigrant networks have a negative impact on wages for new immigrants if the new immigrants arrived after a large influx of immigrants from the same country Additionally, Lalonde and Topel (1991) show a negative wage impact for immigrants resulting in a rise of the immigrant population, also capturing a competition effect Savchenko and Vella (2012) examine the role of networks in the labor market of immigrants in Australia They find that on average an increase in the established network size is associated with an increase in the employment probability and income of the new immigrants However, the magnitude of effect and even the sign of the impact depend on the type of the network For example, the British network has a positive effect on the employment probability and income of new British immigrants while the Chinese network has a negative effect on labor market outcomes of its compatriots Moreover, the authors present evidence that the labor market outcomes of newly arrived immigrants improve with network quality Many studies on immigration in the United States have focused on Mexican immigrants, since Mexicans have a long history of immigration to the United States and represent a large share of US immigration flows The literature documents the impact of networks on occupational sorting for Mexican workers in the United States For example, Munshi (2003) shows that networks not only increase the probability of employment for Mexican workers, but also increase the probability of being employed in higher paid nonagricultural jobs The intuition is that networks not only provide job offers, but also provide other forms of support that enable members to wait for better job offers A novel feature of the identification strategy is using lagged rainfall in the community of origin as the instrumental variable for network size which is endogenous to employment outcome The intuition is that rainfall is correlated with size of the network in the host country as low rainfall adversely affects labor market outcomes in Mexico and thus induces migration and the establishment of networks Evidence of network effects on Hispanic (which includes Mexican) occupational choice in the United States is also found in Hellerstein et al (2010) This paper takes a slightly different approach in that networks are represented by a measure for job density, which is the number of jobs in the area relative to the population residing there Density is defined separately for the Hispanic population The paper finds that the local density of jobs held by Hispanics Occupational sorting of ethnic groups  ­237 matters for Hispanic employment, especially for low-­skilled workers The result of this paper provides evidence for network effects influencing occupation choices While the result may provide evidence of network effects, it also may reflect complementarities based on language These network effects may persist over time Patel and Vella (forthcoming) show that certain ethnic groups represent a growing share of certain occupations over time While persistence of network effects was not the focus of Patel and Vella (2007), Munshi and Wilson (2011) show that the impact networks have on occupational choice can last across centuries Their paper studies the impact of ethnic fractionalization in the United States – a measure for ethnic competition – in the nineteenth century on the probability of holding a professional job in the twenty-­first century The rationale is that professional jobs often require geographic mobility and therefore come at a cost to those who have strong ethnic ties and are more firmly rooted geographically These ethnic ties have persisted from the formation of networks during the early immigration period Certain geographic roots were formed during the immigration wave in the nineteenth century when groups of immigrants established concentrations in certain occupations Munshi and Wilson (2011) note that Polish workers were prevalent in a gas refinery in Detroit; Croatians largely held jobs at Indiana’s oil refineries as stillmen helpers, firemen and still cleaners Italians dominated Pittsburgh’s steel industry in carpentry, repair and rail shops Fractionalization helped foster these ethnic clusters as higher ethnic competition increases the reward for solidarity within the network These networks helped shape a cultural identity that persisted and has impacted occupational choice of workers in 1994 and 2000 Cultural institutions such as family and church help perpetuate cultural traits through time Munshi and Wilson (2011) show that individuals born in countries with higher ethnic fractionalization in the nineteenth century are less likely to hold professional jobs Additionally, higher fractionalization also has a negative impact on annual earnings Network effects are also found to exist among immigrants who enter the host country as refugees Unlike other types of immigrants that may have pre-­existing networks in the host country from family ties, refugee location decisions are typically made exogenously, often by a government agency Refugees may be assigned locations where refugees from the same country have been assigned in the past The fact that location decisions are not made endogenously enables Beaman (2012) to identify the impact of networks on their occupational outcomes Beaman (2012) finds that networks have a negative impact on employment outcomes if there was recently a larger population of refugees that had settled in the area that competes with new refugees within the network On the other hand, the larger the old refugee cohort, the more positive is the employment outcome for the new refugee Damm (2009) also shows a positive impact of ethnic enclave size on refugee wages seven years after immigration However, enclave size has a negative impact on employment probability for high-­skilled workers 3.4  Econometric Issues One key issue that arises in empirical social network studies is the reflection problem, as discussed in Manski (1993) This problem arises as individuals within a social network tend to behave similarly due to endogenous effects (where individual behavior depends 238   International handbook on the economics of migration on group behavior), exogenous effects, (where individual behavior is influenced by factors external to the group) or correlated effects (where individuals within a group behave similarly because they have similar characteristics) It is important to be able to separately identify the correlated and exogenous effects, as these effects not generate the social multipliers that these studies strive to capture An inability to identify these effects is the reflection problem It is called the reflection problem because it is similar to observing movements of a person in front of a mirror It is hard to identify whether the individual is causing the observed movements or if it is his reflection, or whether they simply move together There may be features in the dataset used to overcome this problem (see, for example, Laschever, 2009) Many studies avoid this problem as they study the outcome of a particular cohort of immigrants based on the behavior of another cohort of immigrants For example, Beaman (2012) uses a dynamic model of social networks using data on multiple cohorts of immigrants Patel and Vella (forthcoming) and Savchenko and Vella (2012) separate immigrant groups by ‘old’ and ‘new’ studying the impact of old immigrants on the behavior of new immigrants Additionally, studies that define networks based on population shares may be subject to attenuation bias This problem arises due to a sampling error that may be present for population share variables based on a small sample Aydemir and Borjas (2011) find a sizeable sampling error in the US and Canadian population data For the United States, estimates suggest that the attenuation bias reduces when the average cell size (which the population share variable is based on) is around 1000 observations It is virtually eliminated when the average cell size is closer to 10 000 observations The authors run simulations creating subsamples based on the full range of observations in the Canadian Census to compare the bias across different sample sizes Regression estimates based on different cell sizes show that estimates are lower for variables based on smaller sample sizes than those that are based on larger sample sizes Aydemir and Borjas (2011) suggest some methods that may help reduce the bias when it is necessary to use population share variables based on small cell sizes This clearly is limiting for the analysis done at the local level, which is particularly the case for studies on immigrant occupational sorting which occurs at the local level However, many interesting questions in the network literature are done at this level The paper discusses some techniques that could be used to reduce the bias, though they each have their limitations So far, studies that include population share variables based on small samples have not been able to suitably address this bias Networks are typically defined based on population shares or numbers of countrymen employed within a specified geographic area, typically a metropolitan area One issue that arises as unobservable or omitted is city and nationality group characteristics that are typically correlated with the network This bias can be eliminated under a panel setting, using fixed effects controls for nationality and city, or a pseudo panel setting Estimates on the impact of social networks on occupation sorting by various groups of immigrants can be biased if the model does not adequately account for ‘unobserved skill’ This arises if immigrants from certain countries specialize in different occupations because they are more skilled in those occupations Patel and Vella (forthcoming) use a pseudo panel method to eliminate this ‘unobserved skill’ component from the error term The pseudo panel represents a dataset of aggregated individual level data with the unit of observation being occupation, country and metropolitan area The outcome variable Occupational sorting of ethnic groups  ­239 is the share of new immigrants in the popular occupation and the explanatory variables are population means The network variable in this dataset is the share of established immigrants in the occupation Aggregating individual cross section datasets for two time periods enables the formation of the pseudo panel dataset Taking the difference of each variable across time eliminates the ‘unobserved skill’ component from the error term, which is time invariant, and enables identification of the network effect in the regression model 4  DISCUSSION AND CONCLUSIONS Occupational outcomes of immigrants in host countries depend on immigrant skill sets, the skills demanded in the host countries, and social network effects both in the origin and host countries The skill composition of immigrants partly reflects the immigration decision at the country of origin Various studies find that the higher the costs of immigration the more positive is the selection of immigrants based on skills Social networks may reduce costs of immigration which could lead to negative selection of immigrants These skills represent observable factors such as education and ability to speak the host-country language, as well as unobservable characteristics such as motivation and ability On the demand side, skill sets demanded by host countries are reflected in their immigration policies These policies affect the skill composition of immigrants in the host countries and their occupational outcomes Immigrants are more likely to be suitably matched to their skill sets in host countries that have skill-­based immigration policies In a sense, the occupational outcomes of these immigrants are predetermined Occupational outcomes of these immigrants differ from immigrants who enter the host country through family reunification policies or as refugees High-­skilled immigrants who enter the host country under an immigration policy other than one based on skill are more likely to be mismatched in their occupation, potentially reflecting assimilation barriers and incompatible credentials Social networks also influence occupational outcomes for immigrants in the host country New immigrants rely on social networks for financial and moral support once in the host country Social networks are also sources of information about jobs and often decrease the job search time Since job offers from social networks tend to reflect the occupation distribution of the social networks, new immigrants may disproportionately select occupations that are popular among their countrymen in a particular geographic area This can generate occupational footholds along ethnic communities Many of these footholds are established in low-­skilled occupations, suggesting that these patterns are not explained by comparative advantages of ethnic groups While skill and social network effects impact occupational outcomes of new immigrants, immigration costs may also be a determining factor if ethnic groups face different costs in the host country This could be the case if occupations in the host country have higher costs for certain ethnic groups compared to other ethnic groups This idea is presented in Oyelere and Belton (2009) which studies the probability of self-­employment across ethnic groups.8 The results indicate that immigrants from developed countries are more likely to be self-­employed in the United States than immigrants from other 240   International handbook on the economics of migration countries This is an interesting result since self-­employment rates are generally higher in the developing countries than in the United States The results potentially reflect the fact that immigrants from developed countries may have better access to credit, and they may be better informed about institutional arrangements in the United States NOTES * The authors would like to thank an anonymous referee and the editors, Amelie F Constant and Klaus F Zimmermann, for very useful comments and suggestions For further references see also Chapters 1, 2, and 23 in this volume More details on this topic can be found in Chapter 23 in this volume and Chiswick (2010) See in this context also Chapter in this volume Immigrants without university education Linguistic distance is measured by one divided by a linguistic score as measured in the US Census The linguistic score measures how hard for an English speaker it is to learn a given foreign language The larger the score is the easier it is for English speaker to learn a foreign language Revenue from oil is the largest revenue item in the Ecuadorian balance of payments For further details on this topic see Chapter 17 in this volume For further reference on immigrant self-­employment and entrepreneurship see Chapter and Chapter in this volume REFERENCES Amuedo-­Dorantes, C and S De la Rica (2011), ‘Complements or substitutes? Task specialization by gender and nativity in Spain’, Labor Economics, 18 (5), 697–707 Aydemir, A and G Borjas (2011), ‘Attenuation bias in measuring the wage impact of immigration’, Journal of Labor Economics, 29 (1), 69–112 Beaman, L (2012), ‘Social networks and the dynamics of labor market outcomes: evidence from refugees resettled in the US’, Review of Economic Studies, 79 (1), 128–61 Beine, M., F Docquier and Ç Ưzden (2011), ‘Diasporas’, Journal of Development Economics, 95 (1), 30–41 Bentolila, S., C Michelacci and J Suarez (2010), ‘Social contacts and occupational choice’, Economica, 77 (305), 20–45 Bertoli, Simone (2010), ‘Networks, sorting and self-­selection of Ecuadorian migrants’, Annales d’Économie et de Statistique, 97/98, 261–88 Borjas, G.J (1987), ‘Self-­selection and the earnings of immigrants’, American Economic Review, 77 (4), 531–53 Calvó-­Armengol, A and M Jackson (2004), ‘The effects of social networks on employment and inequality’, American Economic Review, 94 (3), 426–54 Calvó-­Armengol, A and M Jackson (2007), ‘Networks in labor markets: wage and employment dynamics and inequality’, Journal of Economic Theory, 132 (1), 507–17 Calvó-­Armengol, A and Zenou, Y (2005), ‘Job matching, social network and word-­of-­mouth communication’, Journal of Urban Economics, 57 (3), 500–22 Card, D (2001), ‘Immigrant inflows, native outflows, and the local labor market impacts of higher immigration’, Journal of Labor Economics, 19 (1), 22–64 Card, David (2007), ‘How immigration affects U.S cities’, CReAM Discussion Paper No 11/07, Centre for Research and Analysis of Migration (CReAM), Department of Economics, University College London Card, David and Ethan Lewis (2007), ‘The diffusion of Mexican immigrants during the 1990s: explanations and impacts’, in George J Borjas (ed.), Mexican Immigration to the United States, Chicago, IL and London: University of Chicago Press, pp. 193–228 Chiquiar, D and G.H Hanson (2005), ‘International migration, self-­selection, and the distribution of wages: evidence from Mexico and the United States’, Journal of Political Economy, 113 (2), 239–81 Chiswick, Barry R (ed.) (2010), High Skilled Immigration in a Global Labor Market, Washington, DC: American Enterprise Institute Press Chiswick, B and P.W Miller (2010a), ‘The effects of educational-­occupational mismatch on immigrant earnings in Australia, with international comparisons’, International Migration Review, 44 (4), 869–98 Chiswick, Barry and Paul W Miller (2010b), ‘Educational mismatch: are high-­skilled immigrants really Occupational sorting of ethnic groups  ­241 working at high-­skilled jobs and the price they pay if they aren’t?’, in Barry R Chiswick (ed.), High Skilled Immigration in a Global Labor Market, Washington, DC: American Enterprise Institute Press, pp. 111–54 Chiswick, B.R and S Taengnoi (2007), ‘Occupational choice of high skilled immigrants in the United States’, International Migration, 45 (5), 3–34 Damm, A.P (2009), ‘Ethnic enclaves and immigrant labor market outcomes: quasi-­experimental evidence’, Journal of Labor Economics, 27 (2), 281–314 Docquier, Frederic and Abdeslam Marfouk (2006), ‘International migration by educational attainment (1990–2000)’, in Caglar Ozden and Maurice Schiff (eds), International Migration, Remittances and the Brain Drain, London: Palgrave-­Macmillan, pp. 151–200 Hellerstein, J.K., M McInerney and D Neumark (2010), ‘Spatial mismatch, immigrant networks, and Hispanic employment in the United States’, Annales d’Economie et de Statistique, 141–67 LaLonde, Robert J and Robert H Topel (1991), ‘Labor market adjustments to increased immigration’, in John M Abowd and Richard B Freeman (eds.), Immigration, Trade, and the Labor Market, Chicago, IL: University of Chicago Press, pp. 167–99 Laschever, Ron (2009), ‘The doughboys network: social interactions and the employment of World War I veterans’, working paper, University of Illinois Lewis, Ethan (2005), ‘Immigration, skill mix, and the choice of technique’, Working Paper No 05-­08, Federal Reserve Bank of Philadelphia, Philadelphia, PA Manski, C (1993), ‘Identification of endogenous social effects: the reflection problem’, The Review of Economics Studies, 60 (3), 531–42 McKenzie, D and H Rapoport (2010), ‘Self-­selection patterns in Mexico–U.S migration: the role of migration networks’, Review of Economics and Statistics, 92 (4), 811–21 Montgomery, J.D (1991), ‘Social networks and labor-­market outcomes: toward an economic analysis‘, American Economic Review, 81 (5), 1408–18 Mortensen, D and C Pissarides (1994), ‘Job creation and job destruction in the theory of unemployment’, Review of Economic Studies, 61 (3), 397–415 Munshi, K (2003), ‘Networks in the modern economy: Mexican migrants in the US labor market’, Quarterly Journal of Economics, 118 (2), 549–97 Munshi, Kaivan, and Nicholas Wilson (2011), ‘Identity, occupational choice, and mobility: historical conditions and current decisions in the American Midwest’, revised paper, April, available at: http://www.econ brown.edu/fac/Kaivan_Munshi Ottaviano, G.I.P and G Peri (2012), ‘Immigration and national wages: clarifying the theory and the empirics’, Journal of the European Economic Association, 10 (1), 152–97 Oyelere, Ruth Uwaifo and Willie Belton (2009), ‘Coming to America: does immigrant’s home country economic status impact the probability of self-­employment in the U.S.?’, IZA Discussion Paper No 4178, Institute for the Study of Labor (IZA), Bonn Patel, Krishna (2010), ‘Network effects on labor markets for immigrants and natives: its what you know and whom you know’, SSRN Working Paper No 1373885 Patel, K and F Vella (forthcoming), ‘Immigrant networks and their implications for occupational choice and wages’, Review of Economics and Statistics Peri, G and C Sparber (2009), ‘Task specialization, immigration and wages’, American Economic Journal: Applied Economics, (3), 135–69 Peri, G and C Sparber (2011), ‘Highly-­educated immigrants and native occupational choice’, Industrial Relations, 50 (3), 385–411 Pissarides, Christopher (2000), Equilibrium Unemployment Theory, 2nd edn, Cambridge, MA: MIT Press Roy, A.D (1951), ‘Some thoughts on the distribution of earnings’, Oxford Economic Papers, (2), 135–46 Savchenko, Yevgeniya and Francis Vella (2012), ‘Immigrant networks in Australia: they help newly arrived immigrants to find jobs and get higher income?’, working paper (forthcoming online) 13  Immigrants, wages and obesity: the weight of the evidence* Susan L Averett, Laura M Argys and Jennifer L Kohn 1  INTRODUCTION The worldwide obesity epidemic has not spared even those in developing countries (World Health Organization, 2006).1 There is a large and growing literature on migration and health that supports the ‘healthy immigrant effect’, that is, that those who are healthier are more likely to migrate (for example, Antecol and Bedard, 2006; Park et al., 2009) There is also substantial literature on the link between obesity and labor market outcomes (for example, Averett and Korenman, 1996; Cawley, 2004) The purpose of this chapter is to explore the complex interaction among immigration, obesity and labor market outcomes We know of only two studies that specifically examine this combined effect (Cawley et al., 2009, and Averett et al., 2012) This chapter proceeds as follows The next section examines the healthy immigrant hypothesis This is followed by a summary of the literature on obesity and labor market outcomes Then, we review the two studies that bring these literatures together The chapter concludes with a discussion of some policy implications and directions for future research 2  IMMIGRANT HEALTH AND OBESITY 2.1  The Healthy Immigrant Hypothesis The global obesity epidemic and the important role that migration plays in the demographic makeup of a country’s population has prompted a number of recent studies that have examined patterns of overweight and obesity among immigrants These studies focus on comparing body mass index (BMI) calculated as weight in kilograms divided by the square of height in meters and the clinical indicators of overweight (BMI 25) and obesity (BMI $ 30) of migrants from different countries of origin and different destination countries.2 A number of regular patterns have emerged Recent research using data on US immigrants is consistent with the healthy immigrant effect (Antecol and Bedard, 2006; Choi, 2011; Hao and Kim, 2009; Park et al., 2009) suggesting that recent migrants tend to be healthier than the native-­born American population In addition to reporting better initial health across a number of more traditional dimensions, including health behaviors, the prevalence of chronic conditions and self-­reported health (Jasso et al., 2004), many studies examining this phenomenon also note that immigrants have lower BMIs and are less likely to be classified as overweight or obese at the time of migration (Bates et al., 2008; Hao and Kim, 2009; Kaplan et al., 2004; Park et al., 2009) Using US data from the 1998 National Health Interview Survey 242 Immigrants, wages and obesity  ­243 (NHIS), Kaplan et al (2004) report that newly arrived Hispanic immigrants are healthier and have lower BMIs than the native-­born population Lower BMIs among this group translate into lower rates of obesity: only 19.7 percent of Hispanics who migrated to the US were over the threshold for obesity compared with 28 percent of US-­born Hispanics Similar patterns of significantly higher obesity rates for native-­born whites, blacks, Hispanics and Asians than for immigrants of the same race are found using US data between 1995 and 2005 (Park et al., 2009) Another recent study, also using US data, that compared Latino and Asian immigrants with their US-­born counterparts found that the proportion of third-­generation Latinos and Asians who were obese was significantly higher than immigrants to the US from Latin and Asian countries (Bates et al., 2008) For example, 24 percent of third-­generation Asians were categorized as obese compared with only 6.4 percent of Asian immigrants to the US Second-­generation descendants also faced higher rates of obesity than immigrants, but these differences did not reach conventional levels of statistical significance The healthy immigrant effect may be rooted in the fact that individuals choosing to migrate face very different circumstances in their countries of origin and have different health-­related preferences and behaviors that are often unobservable to the researcher Recent immigrants will appear healthier if obesity rates are substantially lower in countries from which these migrants originate, if there is self-­selection such that individuals investing in human capital through migration are healthier and have chosen greater levels of health investment than nonmigrators, or if the host-­country immigrant-­ screening process favors the healthy (Kennedy et al., 2006) Patterns of return migration also reinforce the healthy migrant effect because those in ill health are more likely to return to their country of origin (Hao and Kim, 2009) Borjas (1985, 1995) finds evidence of substantial heterogeneity among immigrants and some evidence that individuals who invest in migration have greater earning potential Accounting for differences in cohort quality in their examination of immigrant health, Antecol and Bedard (2006) confirm that even after controlling for the fixed characteristics of various immigrant cohorts, recent immigrants to the US are healthier than comparable natives As suggested by the results of the studies discussed above, there is heterogeneity in the relative fitness of immigrants based on their country of origin Asian migrants in particular face a substantially lower rate of obesity than their native-­born counterparts but the relative advantage for Latinos is much smaller (Bates et al., 2008; Hao and Kim, 2009) The obesity differential favoring immigrants is large for white and Hispanic men and smaller for women, especially white and black women (Antecol and Bedard, 2006) Although immigrants to the US are found to be initially healthier, research in this area also examines the degree to which changes over time alter this advantage There are a number of reasons to expect changes in health as immigrants settle into their new lives Upon arrival, immigrants may begin the process of assimilation including the gradual adoption of diet and exercise behaviors prevalent among the native population (Goel et al., 2004; Hao and Kim, 2009; Sorlie et al., 1993) For immigrants to the US, the adoption of native habits typically results in increases in obesity and worsening of health Strong cultural ties and living in areas with high proportions of immigrants have been found to slow the unhealthy assimilation process (Hao and Kim, 2009) Other factors that could alter the health of immigrants over time include increased income and access to healthcare (Jasso et al., 2004), both of which are associated with improved health 244   International handbook on the economics of migration Empirical evidence to date points to a deterioration in immigrant health and an increase in BMI as time since migration increases Whether or not such increases lead to a convergence of obesity rates for immigrants and natives depends on the relative rates of increase in obesity for immigrant and native populations Such a comparison is the focus of a study by Park et al (2009) Using data from NHIS, they find that Hispanic immigrants begin with lower rates of obesity than those who are native born, but that these rates converge as time since migration increases In examining the assimilation of migrants to the US, Antecol and Bedard (2006) compare rates of overweight and obesity of recent migrants to the descendants of earlier migrants Black, white and Hispanic male and female immigrants are less likely to be overweight upon arrival, but after remaining in the US for at least 15 years they exhibit rates of overweight equal to or exceeding that of natives Similar patterns of convergence of obesity rates are evident in their data, with the exception of Hispanic and black male immigrants whose obesity rates remain substantially below those of Hispanic and black male natives even 15 years after migration Similar results are found in a number of other studies (Goel et al., 2004; Kaplan et al., 2004) Assimilation patterns also vary by education level Increases in body weight are significantly more rapid for immigrants with less than a college degree (Kaushal, 2009) The health characteristics and transitions of immigrants to Canada are similar to those experienced by migrants to the US (Cairney and Ostbye, 1999; McDonald and Kennedy, 2005; Perez, 2002; Tremblay et al., 2005) These studies make use of two nationally representative datasets, the National Public Health Survey and the Canadian Community Health Survey Immigrants to Canada in these studies are significantly less likely to be classified as obese at the time of migration, but for many ethnic groups obesity rates converge to those of native-­born Canadians Linking these data to information on the location of immigrants from the same country of origin, McDonald and Kennedy find that the larger the ethnic enclave, the slower the transition to native rates of obesity Findings from studies on immigrants to Australia are similar (Hauck and Hollingsworth, 2009; Renzaho et al., 2006) Studies that have examined initial obesity rates and assimilation for migrants to Europe find mixed evidence when comparing the health of immigrants to that of the native population This may be due, in part, to differences in the predominant countries of origin among immigrants to Western Europe Kirchengast and Schober (2006) find higher rates of overweight and obesity among adolescents recently migrating from Turkey and Yugoslavia to Austria Studies from other countries report increased obesity and obesity-­related health risks for immigrants compared with natives in the Netherlands (Brussaard et al., 2001) and Germany (Bongard et al., 2002) In contrast, Averett et al (2012) and Kennedy et al (2006) find evidence supporting the healthy immigrant hypothesis in their analysis of immigrants to the UK 3  OBESITY AND LABOR MARKET OUTCOMES 3.1  Theory Until recently, economists had little to say on the issue of bodyweight in contemporary industrialized societies, although the medical literature had for decades warned of the Immigrants, wages and obesity  ­245 medical costs of obesity Currently, there is a large and growing literature that is concerned with determining whether or not obesity causes adverse labor market outcomes An association between obesity and labor market outcomes does not imply causality Obesity may cause lower wages if employers discriminate against the obese or if obesity results in lower labor market productivity However, it is also possible that the obese simply possess less desirable personality traits that affect their productivity such as laziness or a lack of social skills Sobal (2004) notes that these latter two productivity traits are often associated with the obese by the general public It is also possible that the obese more heavily discount the future, making them more prone both to overeat and to invest less in wage-­enhancing human capital (Cawley, 2004; Zhang and Rashad, 2008) In addition, since the obese are likely to incur higher healthcare costs, their employers may pay higher premiums for employer-­sponsored health insurance and compensate the obese with lower wages (Bhattacharya and Bundorf, 2009) Reverse causality is possible such that those with lower wages become obese in part because they cannot afford healthy food and rely on calorie dense fast foods (Drewnowski, 2009) Conversely, the obese, believing their marriage market prospects are low, may invest more heavily in labor market-­oriented human capital and thus have higher wages (Averett and Korenman, 1996) Finally, cultural norms may play a role in whether or not there is a labor market penalty associated with obesity (Costa-­Font and Gil, 2004; Garcia and Quintana-­ Domeque, 2006) Although most of the economics research has focused on the link between wages and obesity, employment is another dimension along which the obese may suffer In fact, it is likely that the endogeneity here may be even more salient because obesity can be the cause of substantial work disabilities This in turn has spillover effects in other functional spheres, such as home production, indicating that obesity may have broader implications for both home and personal productivity (Klarenbach et al., 2006) 3.2  Econometric Strategies and Findings The characteristics of empirical studies of obesity and labor market outcomes including their data sources, empirical strategies and principal findings are summarized in Table 13.1.3 In this table, we report OLS results from a study if they are the only results reported by the authors If the study includes multiple methods, Table 13.1 reports the results from the methods that attempt to control for endogeneity The effects of obesity on labor market outcomes in the US have been assessed in a large number of studies, and these studies have generally been concerned with establishing the direction of causality from obesity to labor market outcomes One of their most robust findings is that obese women tend to earn less than their nonobese counterparts, and that there are differences in the obesity wage penalty associated with race and ethnicity (Averett, 2011) Studies using European data report more mixed evidence Ascertaining if there are wage penalties to obesity in European countries is more difficult owing to the different institutional features of labor markets there, which are generally characterized by more compact wage distributions, a considerable share of the labor force in the public sector, and more fixed wage structures (Garcia and Quintana-­Domeque, 2006; Greve, 2008) The studies that address the potential endogeneity between labor market outcomes 246 NLSY (USA) NLSY (USA) Minnesota Twins (USA) NLSY (USA) NLSY (USA) WES (USA) GSOEP (Germany) PSID (USA) ECHP (Europe) DHS (France) ECHP (Europe) PSID (USA) ECHP (Europe) CHS (Canada) HSE (England) QLFS (England) ECHP (Europe) Health Finland SHARE (Europe) Averett and Korenman (1996) Pagán and Dávila (1997) Behrman and Rosenzweig (2001) Baum and Ford (2004) Cawley (2004) Cawley and Danziger (2005) Cawley et al (2005) Paraponaris et al (2005) Sousa (2005) Conley and Glauber (2006) Garcia and Quintana-­Domeque (2006) Klarenbach et al (2006) Morris (2006) HSE (England) QLFS (England) NLSAH (USA) ECHP (Europe) Morris (2007) Norton and Han (2008) Atella et al (2008) Brunello and d’Hombres (2007) Johansson et al (2009) Lundborg et al (2007) Fahr (2006) Data (Country) Authors (date) IV (local area average BMI) IV (local area proportion obese) IV (BMI of family members) OLS OLS IV (had any sisters, was an only child) IV (local area average BMI) IV (local area proportion obese) Propensity score matching IV (genentic information) Quantile regression IV (BMI of family members) Lag BMI OLS OLS OLS Deviations from optimal or peer BMI Lag BMI Propensity score matching Lag BMI Sibling FE OLS Twin differences Individual FE Sibling FE IV (BMI of sibling) Individual FE IV (BMI of child) IV (BMI of family members) Methods Table 13.1 Summary of selected empirical literature on obesity and labor market outcomes No significant effects Wage penalty No significant effects on employment Wage and employment penalties Wage and employment penalities  for women Wage penalty for women Wage penalty Less likely to be employed More likely to be absent No significant effects on  occupation Wage penalty No significant effects No significant effects Wage penalty Wage and employment penalities Wage penalty Wage penalty for white women Wage penalty for women No significant effects Wage penalty Wage penalty for white women Findings* 247 RLMS (Russia) FTCG (Tiwan) NLSY, MEPS (USA) NCDS (UK) Audit Study (Sweden) RLMS (Russia) Schultz (2008) Tao (2008) Bhattacharya and Bundorf (2009) Lindeboom et al (2010) Rooth (2009) Huffman and Rizov (2010) IV (BMI of family members) Lag BMI, Individual FE Lag BMI Semi-­parametric regression IV (Sibling BMI) Lag BMI Difference in differences OLS IV (BMI of parents) Experiment Lag BMI Random and Individual FE IV (parent prescription of obesity  medicine) Individual FE Lag BMI Individual FE Lagged BMI IV (child BMI, market food prices) Semi-­parametric regression OLS Wage penalty Less likely to be employed Wage penalty for women Wage penalty for women Wage penalty for women Less likely to be employed Lower interview call-­back rates No wage penalty Higher wages for obese men Less likely to be employed No significant effects Wage penalty Wage penalty for white women Women less likely to be employed No significant effects Wage penalty Women less likely to be employed Notes: ECHP: European Community Household Panel NLSAH: National Longitudinal Study of Adolescent Health FTCG: Female Taiwanese College Graduates PSID: Panel Study of Income Dynamics HILDA: Household Income and Labor Dynamics in Australia QLFS: Quarterly Labour Force Survey HSE: Health Survey for England RLMS: Russian Longitudinal Monitoring Survey NCDS: National Child Development Study SHARE: Survey of Health, Aging and Retirement in Europe NLSY: National Longitudinal Survey of Youth WES: Women Employment Study MEPS: Medical Expenditure Panel Survey CHNS: China Health and Nutrition Survey DHS: Decennial Health Survey GSOEP: German Socioeconomic Panel Study CHS: Canadian Community Health Survey NHANES: National Health and Nutrition Examination Survey * When possible we report findings from models that use IV, lagged BMI or FE estimators HILDA (Australia) NLSY, NHANES PSID (USA) NLSY (USA) ECHP (Europe) CHNS (China) Han et al (2009) Sanz-­de-­Galdeano (2008) Shimokawa, S (2008) Kortt and Lee (2010) Wada and Tekin (2010) Gregory and Ruhm (2011) Danish Admin Greve (2008) 248   International handbook on the economics of migration and obesity employ a variety of strategies For example, many researchers use a lagged measure of BMI, arguing that this temporal ordering of events cannot reflect reverse causality This may not be sufficient if the error term captures some omitted variable, such as motivation or time preference, related to both past BMI and the contemporaneous labor market outcome of interest Sibling fixed-­effects models have also been used to examine the relationship between obesity and wages on the assumption that differences between siblings remove the variation in weight attributable to a shared family environment However, even after differencing across siblings there remains variation in weight ascribed to genetic makeup unshared by siblings and the variation in weight attributable to nongenetic factors To the extent that these factors are not captured by the observable factors, such as education, or if parents treat children in the same family differently as a response to early academic potential that is related to wages, the sibling fixed-­effects estimates will be biased Twin difference models are another way to remove unobservable genetic factors correlated with obesity In this case the researcher can control completely for genetic propensity to be obese since monozygotic twins share 100 percent of their genes, however, sample sizes of monozygotic twins tend to be small Individual fixed-­effects models have also been used to control for time-­invariant individual-­specific factors that might be correlated with both BMI and earnings Cawley (2004) argues that using an individual fixed-­effects model is superior to analyzing differences between siblings because it eliminates more variation due to unobserved nongenetic factors Cawley cites evidence from previous studies that have been unable to detect any effect of common household environment on bodyweight However, there is a body of literature indicating that the family environment may be ‘obesogenic’ suggesting that parents shape the eating environment of their children by making food available and by their own eating habits and food choices (see Birch, 1999, and citations therein for more discussion) To the extent that unobservable factors affecting both weight and wages are time-­varying, however, they will not be accounted for by individual fixed-­effects models The method of instrumental variables (IV) is preferred to individual fixed-­effects if the primary concerns are that important unobservable factors are time-varying and/or that there is reverse causality As shown in Table 13.1, using the BMI of a biological family member as an IV, as first proposed by Cawley (2004), is most common, although several studies have used other instruments Finally, two studies has made use of propensity score matching This method allows the researcher to mimic an experimental setting in that those individuals from the treatment group (for example, in this case obesity is the treatment) are matched with observationally equivalent members of the control group (those who are not obese) While this is intuitively appealing, it is only possible to match on observable covariates and it is possible that the distribution of unobservable factors associated with obesity are quite different Furthermore, this method requires large samples so that there is substantial overlap between the treatment and the control groups (see Pearl, 2009, for further discussion of matching methods) Many studies using OLS find a wage penalty for women, with mixed results for men However, when controlling for endogeneity using one or more of the methods above, most studies find that these effects disappear A careful reading of the literature that uses US data reveals that to the extent there is a wage penalty for obesity, it is suffered prima- Immigrants, wages and obesity  ­249 rily if not exclusively by white women In Europe the results are more mixed perhaps due to heterogeneity in culture and labor market features across countries Studies for both Europe and the US often report negative effects of obesity on employment, and this also seems to disproportionately affect women However, not all of these studies control for the endogeneity of weight in the employment regressions 4  WAGES AND OBESITY AMONG IMMIGRANTS As noted above, immigrants tend to gain weight after they migrate and there are often penalties in the labor market for obesity We now ask how immigrants fare in the labor market of their new destination country if they are obese.4 Thus far, we know of only two studies that directly examine the link between wages and obesity for immigrants: Cawley et al (2009) and Averett et al (2012), hereafter CHN and AAK respectively.5 These studies use data from two of the most popular destination countries for migrants, the US and the UK, as measured by inflows of foreign population (OECD, 2011).6 Perhaps one explanation for the paucity of empirical work in this area is that such analysis requires information on immigrant status, labor market outcomes and health including height and weight It is difficult to find a data source that includes sufficient information in all three areas and that surveys adequate numbers of immigrants Furthermore, the endogeneity issues described in the previous section are best addressed using either longitudinal data or an instrumental variable Neither of the datasets used in the two studies that directly examine immigration, obesity and labor market outcomes contain the necessary data As a result, these studies provide empirical evidence on correlation but not causation which is an important area for future research and future data collection efforts CHN use the New Immigrant Survey (NIS), the first nationally representative survey of legal immigrants to the US This is the only US dataset of which we are aware that contains information on labor market outcomes and obesity for immigrants They use the NIS to estimate equation (13.1) in their examination of four labor market outcomes: the probability an individual is employed, has a white-­collar job, suffers health-­related work limitations, and the natural logarithm of wages Li a bxi gOWi hOBi tDURi ei  (13.1) They estimate logit models for the dichotomous dependent variables and OLS log wage equations OW and OB are dichotomous indicators of being overweight and obese, DUR is the duration of time in the US since migration and X is a vector of covariates including age, education, marital status, smoking, drinking and English language proficiency They have a sample of 2321 women and 2171 men comprised solely of immigrants to the US from developing countries These immigrants originate from China, Cuba, Dominican Republic, El Salvador, Ethiopia, Guatemala, Haiti, India, Jamaica, Mexico, Nigeria, Philippines and Vietnam, or from unspecified countries in the following regions: Latin America/Caribbean, Middle East/North Africa, sub-­Saharan Africa, East Asia/ South Asia/Pacific or Oceania They exclude immigrants born in Canada, Korea, Poland, Russia, Ukraine and the UK, and unspecified countries in Europe and Central 250   International handbook on the economics of migration Asia, North America or the Arctic The top five countries of origin in their sample are Mexico, India, El Salvador, Philippines and China CHN focus on immigrants from developing countries to gain insight into the observed pattern that weight is negatively correlated with wages in developed countries, but positively correlated in developing countries The only significant association between labor market outcomes and obesity that CHN find is among immigrant women who have been in the US for a relatively short period of time Specifically, for women in the US for less than a year, a one unit increase in BMI is associated with a 1.5 percentage point lower probability of employment They find obesity and overweight are associated with 18.3 and 9.4 percentage point lower probabilities of employment respectively The marginal effects remain significant for women who are in the US for up to five years, though the effects for BMI and obesity decline while the effect for overweight increases slightly In addition, CHN find that immigrant women, particularly those in the US less than five years, have a statistically significant association between BMI and the probability of reporting work limitations However, the magnitude of this association is very small (0.05 percentage points for all immigrant women and 0.07 percentage points for those in the US less than five years) They find very small point estimates that are not statistically significant for the association between BMI and wages for men and women regardless of duration AAK extend the work of CHN by analyzing data from the UK The British Household Panel Survey (BHPS) has information on immigration status, labor market outcomes and height and weight reports in two waves, 2004 and 2006 In addition to extending the analysis to a popular destination for immigrants in Europe, use of the BHPS also allows for comparisons between immigrants and natives However, since height and weight were only asked in two years, the sample size is limited reducing the precision of the estimates and making individual fixed effects analysis implausible The BHPS sample consists of 21 292 person year observations (14 493 individuals) of which 584 of these individuals are immigrants Two hundred and fifty-­eight immigrants provided data in both waves resulting in an immigrant sample of 842 person-­year observations Immigrants comprise 4.6 percent of the female sample and 3.6 percent of the male sample Because they have a smaller sample, AAK are unable to limit their analysis to immigrants from developing countries or stratify them by duration in the UK The largest number of immigrants in the British sample comes from other European countries, including Ireland.7 Next are immigrants from Africa, India, Pakistan, Bangladesh, Sri Lanka, sub-­Saharan African and former British colonies The remaining immigrants are from the Far East, the Middle East, the Caribbean, and Central and South America As a reflection of the longitudinal design of the BHPS, the average duration in the UK is 32 years with only 69 observations for immigrants in the UK five years or less Thus, while AAK extend the work of CHN, they so with a dataset that has a markedly different profile AAK begin by replicating the CHN estimates from equation (13.1) on the immigrant-­ only BHPS sample and find a negative association between being overweight or obese and wages for immigrants in the UK who are employed at the time of the survey The magnitude of this wage effect is larger than that found by CHN, but owing to the small sample size it is only statistically significant for men, indicating that obese immigrant Immigrants, wages and obesity  ­251 men face a wage penalty in excess of 20 percent relative to the wages of immigrant men whose BMI is less than 25 They find no significant association between obesity or being overweight and the probability of being employed However, there is evidence of a strong link between obesity and work limitations for both male and female immigrants, and a statistically significant association between being overweight and work limitations for females Finally, in this immigrant-­only sample they find a negative association between both being overweight and obese and the probability of being in a white-­collar job, which is classified as managerial or professional in the BHPS data The associations they find are only marginally significant but of reasonably high magnitudes and similar for both men and women: overweight and obese immigrants are approximately 11 percentage points and 17 points less likely to be in white-­collar jobs respectively AAK then extend the CHN model to directly compare immigrants with the native population Consistent with the literature on immigrants to the UK (Dustmann et al., 2010), they find a wage premium and higher levels of education for immigrants relative to natives in unadjusted means in the BHPS sample However, after controlling for education and other observable productivity measures, but not for weight, they find that immigrant women suffer a wage penalty as immigrant men from non-­Organisation for Economic Co-­operation and Development (OECD) countries In order to examine the interrelationship between immigration and obesity they estimate equation (13.2) for the combined sample of immigrants and natives This specification augments equation (13.1) with IMM indicating immigrant status and includes interaction effects between immigrant status and weight classification as shown in ­equation (13.2): Lit 5a1bxit 1GIMMit 1gOWit 1hOBit 1␾ (IMMit *OWit)1l (IMMit*OBit) 1tDURit 1eit  (13.2) As before, L represents one of four labor market outcomes, the log of earnings for those working, and dichotomous indicators of employment, health-­related work limitations and a white-­collar job Consistent with other studies using OLS (see Brunello and d’Hombres, 2007; Morris, 2006), they find a wage premium for overweight and obese native-­born men However, for immigrant men, they find a strongly significant wage penalty for being overweight (25.1 percent) or obese (30.9 percent) These point estimates are nearly five times the wage premium for overweight and obese natives They find sizeable (11.8 percent and 16.7 percent) wage penalties for overweight and obese immigrant women, though given the small proportion of immigrants these are not statistically significant at conventional levels While CHN found that obese immigrant women were less likely to work than their healthy-­weight immigrant peers, AAK find that obese immigrant women are percentage points more likely to work than natives in the healthy weight range One potential explanation for this finding is, as noted earlier, BMI is not the best measure of adiposity, someone who is very muscular may be classified as overweight on the basis of BMI They also find a strongly significant positive association between overweight status for immigrant women and work limitations Finally, AAK find that the association between 252   International handbook on the economics of migration weight and social class (as measured by working in a white-­collar job) is significant for men only While immigrant men as a group are 18.7 percentage points more likely to be white-­collar workers, overweight immigrant men are 10.5 percentage points less likely to have managerial or professional jobs than immigrant men of recommended weight The point estimate for obese immigrant men is 12.1 percentage points, but again, this is not statistically significant owing to a small obese immigrant sample The point estimates for the association between weight and social class for women are all insignificant and in most cases much smaller Overall, the findings from the two studies that address the intersection of immigration and obesity and labor market outcomes offer mixed results The differences are likely attributable to the different institutional settings in the US and the UK including different countries of origin and different immigration policies as well as different sample sizes and sample composition Nonetheless, both studies suggest that obese immigrant women appear to have a higher probability of work limitations and are less likely to be in the labor force in both the US and the UK At least in the UK, overweight and obese immigrant men experience a substantial wage penalty and are less likely to be in white-­collar jobs Still, it should be noted that these associations not establish a causal relationship More complete data that can tease out causality by addressing substantial issues of endogeneity that have been demonstrated in the general literature on obesity and labor market outcomes are necessary 5  CONCLUSION This chapter examines the economic consequences of obesity for immigrants by combining two rather disparate strands of literature The first explores immigrants’ obesity at the time of immigration and how their weight changes as they assimilate The second examines how obesity affects labor market outcomes We know of only two studies that have examined the combined labor market consequences of immigration and obesity Researchers have come to refer to the phenomenon that new immigrants are typically more healthy and less likely to be obese than their native-­born counterparts as the ‘healthy immigrant effect’ (Hao and Kim, 2009; Park et al., 2009) As these immigrants assimilate to their new environment and culture, their health and propensity for obesity begin to converge to those of natives These patterns are relatively robust and have been found for immigrants from Asia, Latin America and Europe to the United States, Canada, Australia (Antecol and Bedard, 2006; Bates et al., 2008; Hauck and Hollingsworth, 2009; Kaplan et al., 2004; McDonald and Kennedy, 2005; Park et al., 2009; Tremblay et al., 2005) We know far less about the initial health and obesity trajectories for immigrants to Europe Analyses of data from samples of immigrants to Germany, Austria, the Netherlands and the UK find mixed evidence of a healthy immigrant effect (Averett et al., 2012; Bongard et al., 2002; Brussaard et al., 2001; Kennedy et al., 2006; Kirchengast and Schober, 2006) Even if immigrants are healthier on average than natives, obese immigrants may be more likely to suffer adverse labor market outcomes than healthy-­weight immigrants and obese natives The two papers that offer empirical evidence on the combined effect of immigration and obesity offer mixed results and highlight the need for future research Immigrants, wages and obesity  ­253 Despite different institutional settings in the US and the UK, different sample sizes and data limitations that prevent addressing endogeneity, these studies suggest that there may be a double effect of obesity and immigration particularly for women entering the labor force, and in the UK for male wages and white-­collar work There are several possible avenues for future research Previous literature has shown the importance of controlling for the endogeneity of weight in wage equations Unfortunately, few data sources include information on immigration status, labor market outcomes and height and weight either over time to use fixed-­effects estimation or with additional information that can be used as instruments Therefore, it is most pressing for researchers to push for better data collection in order to better address the causal effect rather than merely the association between immigration, obesity and labor market outcomes In addition, a better understanding of the process by which immigrants converge to native-born levels of obesity could inform policy Is it through diet alone or a more complex interaction between diet, exercise and the new environment? Similarly, any initial effects of obesity on the labor market outcomes of immigrants converge to the labor market outcomes of natives over time? Again, data that follows immigrants over time would be necessary to address questions of assimilation Given our increasingly global world and the rise in obesity spreading around the globe, it seems paramount to understand more about the complex interaction between health, immigrant status and labor market outcomes NOTES * We are grateful to the editors, Amelie F Constant and Klaus F Zimmermann, and two anonymous referees for helpful comments The data used in this research were made available through the ESRC Data Archive The data were collected by the ESRC Research Center on Micro-­social Change at the University of Essex Neither the original collectors of the data nor the Archive bear any responsibility for the analysis nor the interpretations presented here A recent issue of the Lancet contained several articles on the obesity epidemic further underscoring the importance of this issue See the Lancet, 27 August 2011 (vol 378, no 9793) The BMI is the most often used measure of adiposity since it can be readily calculated with self-­reported height and weight, measures that are often included in survey data However, it has been criticized because it does not distinguish between fat and fat-­free mass such as muscle and bone (Romero-­Corral et al., 2006) Most studies use data where the only measure of adiposity is BMI, ­reflecting the fact that many datasets with labor market information not collect detailed information on body fat Yet, BMI does not distinguish between fat and fat-­free mass such as muscle and bone (Romero-­Corral et al., 2006) Burkhauser and Cawley (2008) recommend using more accurate measures of fatness such as total body fat, percentage body fat, fat-­free mass and waist circumference, and they present a method for adjusting self-­reported height and weight using US data Furthermore, both men and women tend to systematically misreport their weight, but Lakdawalla and Philipson (2009) find that this misreporting is small enough that it does not affect the qualitative conclusions of their empirical work However, this classical measurement error in reported weight or height could lead to attenuation bias in coefficients on BMI/obesity There is an extensive literature on immigration and labor market outcomes that does not include obesity (Borjas, 1985, 1995) Some of the studies cited in Table 13.1 include a control for whether the respondent is born in the country from which the data are drawn (for example, Kortt and Leigh, 2010) but only Cawley et al (2009) and Averett et al (2012) specifically look at how obesity affects the earnings of immigrants In 2009, the latest year for which data are available, the countries receiving the largest inflows of permanent foreign population immigrants according to the OECD were, in order: US, Germany, UK, Spain, Canada and Australia After May 2004 the addition of new EU states combined with a strong UK economy increased immigration from other European countries to the UK (Dustmann et al., 2010) 254   International handbook on the economics of migration REFERENCES Antecol, H and K Bedard (2006), ‘Unhealthy assimilation: why immigrants converge to American health status levels?’, Demography, 43 (2), 337–60 Atella, V., N Pace and D Vuri (2008), ‘Are employers discriminating with respect to weight? European evidence using quantile regression’, Economics and Human Biology, (3), 305–29 Averett, S (2011), ‘Labor market consequences: employment, wages, disability, and absenteeism’, in John Cawley (ed.), The Oxford Handbook of the Social Science of Obesity, New York: Oxford University Press, pp. 531–52 Averett, S and S Korenman (1996), ‘The economic reality of the beauty myth’, Journal of Human Resources, 31 (2), 304–30 Averett, S., L Argys and J Kohn (2012), ‘Immigration, obesity and labor market outcomes in the UK’, IZA Journal of Migration, (1), 1–19 Bates, L.M., D Acevedo-­Garcia, M Alegria and N Krieger (2008), ‘Immigration and generational trends in body mass index and obesity in the United States: results of the national Latino and Asian American Survey, 2002–2003’, American Journal of Public Health, 98 (1), 70–77 Baum, C II and W Ford (2004), ‘The wage effects of obesity: a longitudinal study.’ Health Economics, 13 (9), 885–99 Behrman, Jere R and Mark R Rosenzweig (2001), ‘The returns to increasing body weight’, PIER Working Paper 01-­052, University of Pennsylvania, Department of Economics, Philadelphia Bhattacharya, J and K Bundorf (2009), ‘The incidence of the healthcare costs of obesity’, Journal of Health Economics, 28 (3), 649–58 Birch, L (1999), ‘Development of food preferences’, Annual Review of Nutrition, 19, 41–62 Bongard, S., S Pogge, H Arslaner, S Rohrmann and V Hodapp (2002), ‘Acculturation and cardiovascular reactivity of second-­generation Turkish migrants in Germany’, Journal of Psychosomatic Research, 53 (3), 795–803 Borjas, G.J (1985), ‘Assimilation, changes in cohort quality, and the earnings of immigrants’, Journal of Labor Economics, (4), 463–89 Borjas, G.J (1995), ‘Assimilation and changes in cohort quality revisited: what happened to immigrant earnings in the 1980s?’, Journal of Labor Economics, 13 (2), 201–45 Brunello, G and B d’Hombres (2007), ‘Does body weight affect wages? Evidence from Europe’, Economics and Human Biology, (1), 1–19 Brussaard, J.H., M.A van Erp-­Baart, H.A.M Brants, K.F.A.M Hulshof and M.R.H Lowik (2001), ‘Nutrition and health among migrants in the Netherlands’, Public Health Nutrition, (2B), 659–64 Burkhauser, R.V and J Cawley (2008), ‘Beyond BMI: the value of more accurate measures of fatness and obesity in social science research’, Journal of Health Economics, 27 (2), 519–29 Cairney, J and T Ostbye (1999), ‘Time since immigration and excess bodyweight’, Canadian Journal of Public Health, 90 (2), 120–24 Cawley, J (2004), ‘The impact of obesity on wages’, Journal of Human Resources, 39 (2), 451–74 Cawley, J and S Danziger (2005), ‘Morbid obesity and the transition from welfare to work’, Journal of Policy Analysis and Management, 24 (4), 727–­43 Cawley, J., M Grabka and D Lillard (2005), ‘A comparison of the relationship between obesity and earnings in the U.S and Germany’, Journal of Applied Social Science Studies (Schmollers Jahrbuch), 125 (1), 119–29 Cawley, J., E Han and E Norton (2009), ‘Obesity and labour market outcomes among legal immigrants to the United States from developing countries’, Economics and Human Biology, (2), 153–64 Choi, J (2011), ‘Prevalence of overweight and obesity among US immigrants: results of the 2003 New Immigrant Survey’, Online First in Journal of Immigrant Minority Health, available at doi: http://dx.doi org/10.1007/s10903-­011-­9560-­8 (accessed February 2012) Conley, D and R Glauber (2006), ‘Gender, body mass and socioeconomic status: new evidence from the PSID’, in Kristian Bolin and John Cawley (eds), The Economics of Obesity, Advances in Health Economics and Health Services Research, Vol 17, Bradford: Emerald Group, pp 253–75 Costa-­Font, J and J Gil (2004), ‘Social interactions and the contemporaneous determinants of individuals’ weight’, Applied Economics, 36 (20), 2253–63 Drewnowski, A (2009), ‘Obesity, diets, and social inequalities’, Nutrition Reviews, 67 (supplement 1), S36–39 Dustmann, C., G Albrecht and T Vogel (2010), ‘Employment, wages and the economic cycle: differences between immigrants and natives’, European Economic Review, 54 (1), 1–17 Fahr, René (2006), ‘The wage effects of social norms: evidence of deviations from peers’ body-­mass in Europe’, IZA Discussion Paper No 2323, Institute for the Study of Labor (IZA), Bonn Garcia, Jaume and Climent Quintana-­Domeque (2006), ‘Obesity, employment and wages in Europe’, in Immigrants, wages and obesity  ­255 Kristian Bolin and John Cawley (eds) The Economics of Obesity, Advances in Health Economics and Health Services Research, Vol 17, Bradford: Emerald Group, pp 187–217 Goel, M.S., E.P McCarthy, R.S Phillips and C.C Wee (2004), ‘Obesity among U.S immigrant subgroups by duration of residence’, Journal of the American Medical Association, 292 (23), 2860–67 Gregory, C and C Ruhm (2011), ‘Where does the wage penalty bite?’ NBER Working Paper No 14984, in Michael Grossman and Naci H Mocan (eds), Economic Aspects of Obesity, Chicago, IL: University of Chicago Press, pp 315–47 Greve, J (2008), ‘Obesity and labor market outcomes in Denmark’, Economics and Human Biology, (3), 350–62 Han, E., E Norton and S Stearns (2009), ‘Weight and wages: fat versus lean paychecks’, Health Economics, 18, 535–48 Hao, L and J Kim (2009), ‘Immigration and the American obesity epidemic’, International Migration Review, 43 (2), 237–62 Hauck, Katharina and Bruce Hollingsworth (2009), ‘The impact of immigration, income and marriage on obesity’, paper presented at the Seventh International Health Economics Association (iHEA) World Congress, Beijing, 12–15 July Huffman, Sonya and Marian Rizov (2010), ‘Obesity and labor market outcomes in post-­soviet Russia’, paper presented at the Agricultural and Applied Economics Association’s 2010 AAEA, CAES and WAEA Joint Annual Meeting, Denver, CO, 25–27 July Jasso, Guillermina, Douglas Massey, Mark Rosenzweig and James P Smith (2004), ‘Immigrant health – selectivity and acculturation’, in Norman B Anderson, Randy A Bulatao and Barney Cohen (eds), Critical Perspectives on Racial and Ethnic Differences in Health in Later Life, Washington, DC: National Academies Press, pp 227–66 Johansson, E., P Bokckerman, U Kiiskinen and M Heliovaarta (2009), ‘Obesity and labour market success in Finland: the difference between having a high BMI and being fat’, Economics and Human Biology, (1), 36–45 Kaplan, M., N Huguet, J Newsom and B McFarland (2004), ‘The association between length of residence and obesity among Hispanic immigrants’, American Journal of Preventive Medicine, 27 (4), 323–6 Kaushal, N (2009), ‘Adversities of acculturation prevalence of obesity among immigrants’, Health Economics, 18 (3), 291–303 Kennedy, Steven, James McDonald and Nicholas Biddle (2006), ‘The healthy immigrant effect and immigrant effect and selection: evidence from four countries’, SEDAP Research Paper No 164, McMaster University, Program for Research on Social and Economic Dimensions of an Aging Population (SEDAP), Hamilton, Ontario Kirchengast, S and E Schober (2006), ‘To be an immigrant: a risk factor for developing overweight and obesity during childhood and adolescence?’, Journal of Biosocial Science, 38 (5), 695–705 Klarenbach, S., R Padwal, A Chuck and P Jacobs (2006), ‘Population-­based analysis of obesity and workforce participation’, Obesity, 14 (5), 920–7 Kortt, M and A Leigh (2010), ‘Does size matter in Australia?’, Economic Record, 86 (272), 71–83 Lakdawalla, D and T Philipson (2009), ‘The growth of obesity and technological change’, Economics and Human Biology, (3), 283–93 Lindeboom, M., P Lundborg and B van der Klaauw (2010), ‘Assessing the impact of obesity on labor market outcome’, Economics and Human Biology, (3), 309–19 Lundborg, Petter, Kristian Bolin, Sören Höjgård and Björn Lindgren (2007), ‘Obesity and occupational attainment among the 501 of Europe’, in Kristian Bolin and John Cawley (eds), The Economics of Obesity, Advances in Health Economics and Health Services Research, Vol 17, Amsterdam: Elsevier, pp. 221–54 McDonald, J.T and S Kennedy (2005), ‘Is migration to Canada associated with unhealthy weight gain? Overweight and obesity among Canada’s immigrants’, Social Science and Medicine, 61 (12), 2469–81 Morris S (2006), ‘Body Mass Index and occupational attainment’, Journal of Health Economics, 25 (2), 347–64 Morris, S (2007), ‘The impact of obesity on employment’, Labour Economics, 14 (3), 413–33 Norton, E and E Han (2008), ‘Genetic information, obesity and labor market outcomes’, Health Economics, 17 (9), 1089–104 Organisation for Economic Co-­operation and Development (OECD) (2011), ‘Statistics from A to Z: migration: inflows of foreign population into OECD countries’, available at: http://www.oecd.org/document/0,37 46,en_2649_201185_46462759_1_1_1_1,00.html (accessed 31 January 2012) Pagán, J and A Dávila (1997), ‘Obesity, occupational attainment, and earnings’, Social Science Quarterly, 78 (3), 756–70 Paraponaris, A., B Saliba and B Ventelou (2005), ‘Obesity, weight status and employability: empirical evidence from a French national survey’, Economics & Human Biology, (2), 241–58 256   International handbook on the economics of migration Park, J., D Myers, D Kao, and S Min (2009), ‘Immigrant obesity and unhealthy assimilation: alternative estimates of convergence or divergence, 1995–2005’, Social Science and Medicine, 69 (11), 1625–33 Pearl, J (2009), ‘Understanding propensity scores’, in J Pearl, Causality: Models, Reasoning, and Inference, 2nd edn, Cambridge: Cambridge University Press Perez, C (2002), ‘Health status and health behaviour among immigrants’, Health Reports, 13 (supplement) Renzaho, A., C Gibbons, B Swinburn, D Jolley and C Burns (2006), ‘Obesity and undernutrition in sub-­ Saharan African immigrant and refugee children in Victoria, Australia’, Asia Pacific Journal of Clinical Nutrition, 15 (4), 482–90 Romero-­Corral, A., V Montori, V Somers, J Korinek, R Thomas, T Allison, F Mookadam and F Lopez-­ Jimenez (2006), ‘Association of bodyweight with total mortality and with cardiovascular events in coronary artery disease: a systematic review of cohort studies’, The Lancet, 368 (9536), 666–78 Rooth, D (2009), ‘Obesity, attractiveness, and differential treatment in hiring: a field experiment’, Journal of Human Resources, 44 (3), 710–35 Sanz-­de-­Galdeano, Anna (2008), ‘An economic analysis of obesity in Europe: health, medical care and absenteeism costs’, FEDEA Working Paper No 2007-­38, Fundación de Estudios de Economía Aplicada, Madrid Schultz, T Paul (2008), ‘Health disabilities and labor productivity in Russia in 2004’, in Cem Mete (ed.), Economic Implications of Chronic Illness and Disability in Eastern Europe and the Former Soviet Union, Washington, DC: World Bank, pp. 85–118 Shimokawa, S (2008), ‘The labor market impact of body weight in China: a semiparametric analysis’, Applied Economics, 40 (8), 949–68 Sobal, Jeffery (2004), ‘Sociological analysis of the stigmatisation of obesity’, in John Germov and Lauren Williams (eds), A Sociology of Food and Nutrition: The Social Appetite, 2nd edn, Oxford: Oxford University Press, pp. 383–402 Sorlie, P., E Backlund, N Johnson and E Rogot (1993), ‘Mortality by hispanic status in the United States’, Journal of the American Medical Association, 270 (20), 2464–8 Sousa, Silvia (2005), ‘Does size matter? A propensity score approach to the effect of BMI on labour market outcomes’, paper presented at the 19th ESPE Conference and General Assembly, Paris Tao, H (2008), ‘Attractive physical appearance vs good academic characteristics: which generates more earnings?’, KYKLOS, 61 (1), 114–33 Tremblay, M.S., C.E Perez, C.I Ardern, S.N Bryan and P.T Katzmarzyk (2005), ‘Obesity, overweight and ethnicity’, Health Reports, Statistics Canada, Catalogue 82-­003, 16 (4), 23–36 Wada, R and E Tekin (2010), ‘Body composition and wages’, Economics and Human Biology, (2), 242–54 World Health Organization (WHO) (2006), ‘Obesity and overweight’, Fact Sheet No 311, WHO, Geneva Zhang, L and I Rashad (2008), ‘Obesity and time preference: the health consequences of discounting the future’, Journal of Biosocial Science, 40 (1), 97–113 PART IV NEW LINES OF RESEARCH 14  Immigrants, ethnic identities and the nation-­state* Amelie F Constant and Klaus F Zimmermann 1  INTRODUCTION The concept of identity and its importance for many aspects of life as well as in the political, social and psychological realm have been studied by fellow social scientists for a long time Sociologists, social psychologists, political scientists, anthropologists and human geographers have developed theories about the identity of individuals and created surveys to test them empirically Indeed, they have found that identity is a significant characteristic and a distinguishing attribute that affects many facets and phases of the individual’s, the group’s and the society’s sphere Following the neoclassical economic theory, economists have been reluctant to delve into ‘exotic’ questions such as how does the identity of an individual affect his or her utility function, demand and supply of goods and services, and demand and supply of labor or even tackle the fundamental economic question of how limited resources are distributed among the different ethnicities or minorities in the host country.1 With the exception of Amartya Sen, Gary Becker and a few others, economists started seriously looking into the identity ‘variable’ only in the 1990s The notable work summarized in Akerlof and Kranton (2010) gave economists a unifying analytical framework to study the new economic man and woman who are more real than the neoclassical agent By inserting identity in the utility function, Akerlof and Kranton (2000, 2005) opened the possibilities of the individual agent who not only derives utility from consumption and reveals her tastes but who can also draw boundaries between herself and ‘others’ as well For example, norms, conformity, exclusion and social distance are now able to influence one’s labor force participation, wages and well-­being As the authors proclaim ‘identity widens the scope of choices that economists should study’ (Akerlof and Kranton, 2005, p. 15) and provides ‘a new window of inequality’ The identity of an individual affects the way she reasons and makes decisions However, the strong rise in interest to understand and model identity in economics, has been met with a firm increase in the complexity of identity formation In fact, there are many identities individuals may have or can cultivate For instance, with the fall of the Iron Curtain in 1989 and with migration becoming a pressing issue again in so many countries, the economics field witnessed renewed interest in migrants, their economic performance and assimilation in the host country Consequently, countries such as Greece, Italy and Cyprus have become more heterogeneous in this process The importance of ethnicity has become more visible also in other parts of the world By the end of the twentieth century it was hard to find ‘homogeneous’ countries in terms of culture, religion, ethnicity and race Multiculturalism became the way to view and form ‘nationhood’ in the Western world, as it was manifested in the US, Canada, and Britain.2 In the early 2000s, with ‘9/11’ (2001) in the US and the bombings in London (7 July 2005) and Madrid (11 March 2004), economists started paying attention to less 259 260   International handbook on the economics of migration ­ bservable but more salient characteristics of the immigrants’ ‘being’, such as ethnic o identity and national identity An individual’s identification with a certain group, ethnicity or religion can be paramount in how this individual behaves in the host country, and especially towards its labor market One can have multiple identities, such as being a mother, a nurse, a soccer fan, a punk, an immigrant, and a naturalized citizen These identities can be compatible and reinforcing, but they can also be conflicting Moreover, one can identify with different ethnicities and be a cosmopolitan However, this does not necessarily mean that one does not have a national identity or does not pledge allegiance to the host country In the wide gamut of ethnic and national identities, it is possible that individuals can be patriotic, nationalistic, indifferent, apathetic or subversive, and undermine the host country Identity formation and preservation is, however, a two-­way street Perceptions and attitudes are central to the identity process Permanent immigrants are particularly challenged; they face the pressure to replace the national identity of the country of origin by that of the country of immigration Their ethnic identities may be preserved or adapted to the native ethnic identities of the host countries The openness of the people in the new country, their embracing of new culture and their respect towards the newcomers can play a major role in how immigrants react and how close they remain with the country of origin The arrival of immigrants in a country is bound to bring about social changes As identities are malleable and evolve through time and space, it is very possible that the identities of the natives or autochthones will alter and normalize after they come in contact with immigrants Finally, the laws of the host country together with the ideals, self-­understanding and the foundation of the sovereign nation can also affect the identities of immigrants and natives at the individual level and at the nation-­building level All this can apply not only to the first-­generation immigrants, but also transcend to the second and even third generations The intensity and strength of allegiance to the ‘old country’ can vary from being purely symbolic to being substantial as, for example, when an active diaspora mobilizes support of the ancestor country.3 In this new era of economics in which identity directly affects the utility function of economic agents and where individuals can reach a suboptimal economic equilibrium, the role of ethnic and national identity can perhaps explicate these results, especially in the context of immigrant nations Understanding and measuring migrant and native ethnic identities and their relationships to national identities then becomes an important task In this chapter, we seek to investigate the nature, role of and relationships between ethnic and national identities by using migrants as the natural innovators In the next section we first review ethnic and national identities in four different countries that together provide excellent real-­life paradigms of identity perceptions, formations and issues in host countries Modeling the ethnic identity of individuals, providing definitions and reviewing the models of identity formation solidifies this section Discussing the ramifications of the divergence between ethnic and national identities as well as the possibilities of reconciliation concludes this section Section presents empirical results concerning ethnic and national identities We discuss the available data and methodology, reviewing surveys and experimental contributions, study identity formation and its consequences for economic behavior We wrap up this section with the debate of the endogeneity issue of identity Section concludes Immigrants, ethnic identities and the nation-­state  ­261 Table 14.1  Types of national identity and native ethnic identity National identity Weak/loose Strong/narrow Native ethnic identity National identity No Yes The UK The US Germany France 2 A CONCEPTUAL BASIS OF ETHNIC AND NATIONAL IDENTITY 2.1  Contrasting Four Immigration Countries Upon entry into the new country, immigrants carry a mixture of ethnic and national identities from their home country and face an assimilation challenge concerning local ethnic and national identities Each country typically has a majority ethnic identity and a general national identity; however, there also exist sub-­national identities like Bavarians and Prussians in Germany or the Scottish or Welsh in Great Britain For simplicity, we ignore sub-­national identities here Therefore, the way a migrant adjusts and perceives him or herself depends very much on his or her background home and the concrete cultural structure in the receiving country For instance, a Turkish migrant to Germany or the UK will face different challenges according to the identity culture in the receiving country and depending on whether he is from Anatolia or Istanbul France, Germany, Great Britain and the US can serve as paradigms of sovereign nations that are also immigrant countries Each country/nation has applied and pursued different treatments in what political scientists call ‘negotiations of identities’ (Kastoryano, 2002) In short, we justify the selection of the four nations through the taxonomy used in Table 14.1 Migrants who enter one of those countries are facing a specific culture combining the strength of the national identity with its closeness to the native ethnic identity We argue that these four country examples are all in a separate class of national identity cultures The US and France are both strong nation-states, while France attempts to keep the native ethnic French close to the national identity and the United States of America is a collection of many very different ethnicities Great Britain and Germany have a weak or loose national identity, both countries depart from regions that have been independent across centuries and Germany national identity also still suffers from its Nazi history However, the German ethnic identity is much closer to the national identity than in Great Britain, where the colonist history and the history of the island has created a larger difference between ethnic identities of the citizens (take an Indian who is part of the British Empire) and the British national identity This section provides some insights in the struggle of the nations to preserve their profound ideals and maintain their sovereignty and national integrity without forcible assimilation and with respect to other people’s cultures At the same time, the presence of immigrants in the host country can cause reactionary changes in policies, which can feed back in the identity of immigrants and generate a new round of changes In political science language this is called ‘negotiating identities’ How the state or government 262   International handbook on the economics of migration negotiates identities with its people has a bearing on immigrants’ ethnic and national identities with serious feedback loops and with indeterminate economic behavior Much like open economies benefit from international trade by trading different goods, exporting the goods in which they have a comparative advantage and importing goods they need at a lower price, countries can benefit from migration The benefits for the host country are highest when people with different skills than the natives move to the host country It is differences in skills, abilities, talents and experience that make immigrants valuable to the host country Below, we contrast the ideals and laws among these four countries and the labor market performance of their immigrants as well as the ­immigrants’ ethnic identities France and Germany in continental Europe represent two almost diametrically opposing nation-­building models In France, it is the political unity and the tripartite Republican ideals of ‘liberté, égalité, fraternité’ that have shaped the nation and defined citizenship.4 The ‘law of soil’ but also the ‘law of blood’ have determined citizenship Ethnic identity was always viewed as an obstacle to national solidarity and to immigrant integration Therefore, the country’s immigration policy has always been to integrate foreigners into the nation by putting into practice the Enlightenment and the Republican assimilation model that aspires to efface ethnic and national origins in the second generation, so that immigrant children can hardly be distinguished from French children This model was strengthened by the relatively relaxed citizenship laws and the integrating institutions, such as schools, the military, unions, the French communist party and the Catholic Church It aimed to inculcate both French and immigrant children with a common civic culture and the pride of French values (Hutchinson and Smith, 1996) This is manifested by the recent decision of the French supreme court to prohibit any ostensible ethnic or religious manifestations in public and in schools, such as wearing the chador or the Christian cross Most immigrants in France come from France’s ex-­colonies, speak the same language, have citizenship, can vote and, thus, can make a difference in choosing their representation After the Second World War, other immigrants arrived in France under guestworker schemes In official statistics, the population in France is either French by birth (and origin is omitted), French by naturalization (and origin is omitted) or foreign On average, however, immigrants in France, remain un-­integrated in society, occupying a ‘different’ social class and maintaining a social distance from the white native French Their labor market attachment is poor and their economic situation is far worse than that of native French (Constant, 2005) In the past decade, France has introduced the terms ‘integration’ and ‘insertion’ and consciously tries to avoid social exclusion In Germany, in contrast, the nation was built as a natural community, homogeneous in language and culture where ‘Einigkeit und Recht und Freiheit’ were the defining ideals The ‘Volk’ preceded the geopolitical formation of the state and citizenship ideas were formed later (Brubaker, 1992) This explains the emphasis of the country in the origins and bloodlines of its people Often, individuals unfamiliar with the German culture, language and ideals who could prove German ancestry have been granted German citizenship and access to all benefits as the natives They are the ethnic Germans, and their enumeration as immigrants is effaced from official statistics In summary, the idea of a nation came late and has also suffered from Nazi history Mass migration to Germany started after the Second World War and was, after the Immigrants, ethnic identities and the nation-­state  ­263 war-­adjustment period of the 1950s, mostly demand driven.5 While initially low-­skilled laborers were recruited in the 1960s, within a decade after recruitment stopped in 1973, spouses and children comprised the majority of immigrants Thereafter, mostly refugees came In the late 1990s and in the new century, Germany has re-­evaluated its definition, liberalized its naturalization rules, established the ‘law of soil’, acknowledged politically the status of an immigrant country and is actively seeking to integrate its immigrants While Germany opposes the universalist and assimilationist French model, it grapples with the accusation of having created ‘parallel societies’ Immigrants in Germany not fare well in the labor market (Bauer et al., 2005) With the exception of some immigrant groups in self-­employment, compared with native Germans, immigrants have higher unemployment rates, lower labor force participation rates and lower earnings, which are mostly owing to individual characteristics of immigrants which are different from those of the natives and largely caused by selection Immigrants in Germany not often have German citizenship and, thus, they not have political representation While the French and German models are completely different in their inception of the nation-­state, the situation of their immigrants is similar Ethnic enclaves and banlieus (suburban ghettos) exist in both countries, indicating areas of poverty and foreignness (Kastoryano, 2002) In addition, the national consciousness or identity of their immigrants is also similar and does not always align with the host country Great Britain is a nation that has allowed people from its ex-­colonies to immigrate and be British; it practices yet a different political and civic philosophy Citizenship and the right to vote is not only open to the British but also to those from the Commonwealth Immigrants from Jamaica, Trinidad, Nigeria, Ghana, Uganda, Kenya, India, Pakistan and Bangladesh, for example, when they arrive in Britain have the right to vote They can choose representatives who are more likely to be sympathetic to them and their situation (Sen, 2001) After the Second World War, immigrants started flooding the country, which received them with an open mind, under the ideals of cultural diversity, and with antidiscrimination laws safeguarding their religious practices Riots and clashes between immigrants and the police in the 1980s made the nation reconsider its nonchalant multicultural stance and enforce the learning of the English language, culture, history and values if immigrants want to stay in the country These upheavals also prompted the British politician Norman Tebbitt to propose ‘the cricket test’ for the immigrants’ national identity and allegiance.6 While immigrants’ labor market performance in Great Britain varies by ethnic origin, some immigrant ethnicities fare very well The underlying motif in ethnic clashes in all these countries is that the immigrant minority ethnic groups almost unite in identification to stand against the larger dominant society with which they share the same territorial space but not the same resources However, it also observed that immigrant groups not share the same identity and often oppose other immigrant groups This is another manifestation of multiple identities The US, a prototypical immigrant nation, was created in a different historical context It is heralded by the ‘e pluribus unum’ etched on its Great Seal It means that ‘many are uniting into one’, as it is the cultural diversity of the immigrants that makes up the nation and molds into one national identity: ‘we the people’ The US as a nation emphasized commitment to three principles: liberty, equality and self-­government.7 It was identification with these principles that created the nationality of its people In 264   International handbook on the economics of migration the newly created nation in the eighteenth century any person could become American as long as he or she was willing to commit to the political ideology centered on these abstract ideals (Gleason, 1980) The ‘melting pot’ symbolizes the blending of cultures, languages, religions, ideals, beliefs and ideologies It does not mean that one culture or ethnicity is assimilated by another nor that the blending always produces the same peoplehood What comes out of this blending is a new peoplehood with an underlying common national identity It is important to note that this blending symbolizes the ever evolving American peoplehood as new ‘blood’ comes in the country and blends with the rest In the early 1900s ‘Americanization’ was encouraged and only immigrants from Germany, Scandinavia and the British Isles were allowed to enter; they were expected to resemble the Anglo-­Protestant model In 1965 the US Immigration Law abolished discriminationary quotas and opened up to immigrants from other hemispheres By the end of the twentieth century, one finds more and more hyphenated Americans, such as Greek-­Americans, African-­Americans or Italian-­Americans Hence, the myth of the melting pot today does not refer to assimilation between the ethnicities On the contrary, the United States of America are a collection of ethnic communities with a strong and rising Spanish speaking sub-­population Membership of the civic and national consciousness does not require a single religious or cultural identity, but support of the democratic political values The national security challenge with ‘9/11’ precipitated border controls, renewed xenophobia and anti-­migration discourse, victimizing some minority groups The question of competing identities and how they threaten the nation was not just academic, but real This section shows that the institutional and structural dimensions of national identities can shape the performance of immigrants and interfere with their ethnic identities This suggests that one needs to take care of the potentials and challenges of multiethnic dimensions for the dynamics of identity formation The host country or state has power over its citizens or legal population that cannot be ignored This power can in turn form perceptions and attitudes – for both natives and immigrants – that are vital to the ethnic identity process and to the economic performance of immigrants 2.2  Modeling Identity Formation Identities emerge from social interactions and become most relevant when social, cultural and linguistic clashes are present In the context of immigrant nations, immigrants are most likely to reconsider and alter their ethnic identities after they arrive in the host country Multiethnic dimensions may appear, as discussed by Constant et al (2009); national identities are likely to adjust too As discussed in section 2.1, the ethnic and national identity culture should matter a great deal for resulting adjustments However, the available theoretical literature has made only limited suggestions as to how to deal with this complex challenge Hence, our knowledge is incomplete and we can only present some core approaches to the economics of identity formation which have found different forms of applications While based on a very similar framework, some economists present it as a model of conformity (Bernheim, 1994) Bernheim incorporated social factors such as status directly in the individual preferences of agents His model can explain widely obeyed and persistent Immigrants, ethnic identities and the nation-­state  ­265 norms as well as transitory norms such as fads Others present it as a model of social distance and social decisions (Akerlof, 1997), as a model of economics and identity (Akerlof and Kranton, 2000), and others as a model of social identity and behavior (Benjamin et al., 2010; Georgiadis and Manning, 2013) At the beginning of the economic literature on identity formation are also studies on racial identity norms Darity et al (2006), in a central paper, use evolutionary game theory to study the origin and persistence of racial identity norms In particular they study the relationship between wealth accumulation and racial identity They argue that the formation of identity norms imposes externalities (both positive and negative) on an individual’s identity actions The authors consider three types of agents: a racialist, an individualist and a mixed-­identity agent Both intra-­ and inter-­group interaction are responsible for the construction of a racial identity One equilibrium of their model is achieved when all persons follow an individualist identity strategy in which race is not important for market or nonmarket interactions The other equilibrium is the racialist social norm where race is very important for both market and nonmarket interactions The equilibrium of the mixed-­identity strategy has both the individualist and the racialist strategy persistent in social interactions The latter is more likely to exist under a laissez-­faire organization of social interactions Their model provides a framework for understanding acculturation and other social phenomena Drawing on Stewart (1997), Darity et al (2006) discuss also the use of Becker’s household production framework for modeling the demand racial-­cultural identity, an approach that can be easily generalized for ethnic and national identities Such identities are just treated as commodities that are included in the utility function of the individual in the usual fashion.8 Commodities are produced through the use of marked goods and household production time The production technology is affected by the size of the ethnic network and the cultural conditions towards national and native ethnic identities The private production of identity by members of some other group and the negative externalities associated with that production can lead to a negative impact on identity production of other groups Most important, ‘the identity formation of individuals is affected by the prevailing norms of one’s own-­and other-­group racial-­cultural identity’ (Darity et al., 2006, p. 292) Others have an effect on production and, hence, commodities and utility Akerlof and Kranton (2000, p. 719) place the identity variable directly into the utility function of the individual, which then depends on the individual’s actions, its assigned social categories and the actions of others They concentrate on general identity formation and call decreases or increases in utility that come from variations of the individual identity losses or gains in identity Others, including Benjamin et al (2010) and Georgiadis and Manning (2013), model the choice between identity formation and normal actions using quadratic loss functions with different degrees of complexity to derive informed conclusions about the identity choices To show the basic ideas, no differences are made between the many types of identities an individual may have We also choose a simple framework related to the models used in particular by Benjamin et al (2010) and Georgiadis and Manning (2013) Let xi be some action choice by individual i such as the choice to assimilate in the host country Let | xi be the individual’s preferred action or her ideal action Then, any devia| ) The individual can tion from the ideal action will be a loss to the individual, (xi –x i 266   International handbook on the economics of migration also belong to a social club such as be a member of her ethnic group, C This belonging relates to her ethnic identity The group may require from its members to exert action xc Any deviation from the ethnic ‘directive’ will be a loss to the individual or it could be thought as the penalty she will have to pay to be a member of this ethnic group, (xi – xc*) We express this loss in a quadratic form and a weight or a parameter to it The utility function of the individual then is:9 | ) 2 w (s) (x 2x *) Ui (xi, c) (12w (s)) (xi2x i i c The first term in this function is the utility from choosing to act and the second term is utility from choosing to be a member of a group Note that w is the weight with strength s and # w # To find the optimal action xi* that will maximize her utility, we take the first order conditions and set them to zero Then: xi* | xi (12 w (s)) w (s) xc* This is a weighted average of the preferred action without identity considerations and with identity constraints By inserting xi* into the utility function, the maximized utility becomes: Ui (xi, c) w (s) (1 w (s)) (| xi x*c) Behavioral responses can be modeled by varying s, | xi or xc* Simple calculations show that total utility is the sum of utility from acting and utility from being a member in the social club Ui (xi, c) Uxi Uc , and relative utility is: (Uc Uxi) 2w (Uc Uxi) which implies: Uc Uxi for w 0.5 Uc Uxi for w , 0.5 and Uc , Uxi for w 0.5 2.3  Balancing Ethnic and National Identity In homogeneous countries ethnic and national identities fully overlap one another Ethnic identities are important because they can give meaning to individuals’ existence They provide a purpose in life and a strong link with ancestors and descendants National identity presumes a devotion to a country.10 However, in reality there is bifurcation between national and ethnic identities The notions of citizenship, nationality and dual nationality offer additional complexity to the issue.11 Immigrants of our times not have a singular identity; they often oscillate between the old and the new world.12 The question that arises is, how can a nation preserve its integrity and sovereignty while Immigrants, ethnic identities and the nation-­state  ­267 allowing its citizens to freely choose their ethnic identity and be happy and productive members of the new society? In modern societies, it is efficiency and success that is sought and not conformity or filial piety Remarkably, as Park (1950) reflects, even though natives as a whole may appear conspicuously uniform – at least to immigrants – they are also individuals and can have very different opinions and beliefs from each other Yet they all unite when it comes to their national identity, allegiance and creed The idea of ‘nation building’, that is, applying policies that encourage and reinforce a national identity, has been advanced as a means for peaceful integration and conflict reduction in countries with high levels of ethnic diversity or fractionalization The United States of America is the country that is closest in achieving this reconciliation While Americans are from all corners of the world with different cultures, maternal tongues and ethnic self-­identification, they all say ‘I am an American first’ and pledge allegiance to the United States of America The following section provides examples and findings about the balancing of ethnic and national identities for the United States of America and other countries 3  EMPIRICAL FINDINGS 3.1  Methodology and Data The US has come a long way in balancing ethnic, racial and national identities This is best reflected by the ethno-­racial classification in the decennial censuses over the years (Lee, 2009) In 1860, for example, there was only an entry for race that had three options: White, Black and Mulato In 2000, the options for race were: White, Black, African American or Negro, American Indian or Alaska Native, Chinese, Japanese, Filipino, Korean, Asian Indian, Vietnamese, Native Hawaiian, Guamanian or Chamorro, some other Pacific Islander, some other Asian, some other Race In addition, there was an entry for ethnicity with the following options: Mexican, Mexican-­American, Chicano, Puerto Rican, Cuban, other Spanish/Hispanic/Latino Lee (2009) observes that the census questions changed ‘from enumerator observations to self-­identification’ with the 1960 census (p. 116), and that self-­identification with one race was replaced with a ­multiracial identification with the 2000 census Citrin and Sears (2009) describe the policy of the US to achieve unity – given its polyethnic population – as one that does not publicly state one particular ‘ethnic or religious preference as the nation’s defining identity but simply to demand as a price of nationhood that everyone endorses democratic political principles and tolerates everyone else’s customs’ (p. 146) Thus, although ethnic and national identities in the US not overlap, they need not be competing either In fact, in a successful ‘blending of identities’ scenario it is possible that ‘strong identifications with both nation and ethnic group are not merely compatible but may even be mutually reinforcing’ (Citrin and Sears, 2009, p. 152) A critical question that all social scientists face is how to rigorously define and successfully measure ethnic and national identities In spite of the academic literature in the social sciences, this is such a difficult task that it has led many scientists to egregiously advocate abandoning it (Brubaker and Cooper, 2000) Economists have ignored the 268   International handbook on the economics of migration issue for a long time as it involves arguments of endogeneity and sample biases Few words can mean such different things at the same time Identity is one of them As Citrin and Sears (2009) put it, identity means ‘both sameness and difference, both commonality and ­individuality’ (p. 146) To empirically measure ethnic identity scientists have used small-­scale identity tailored surveys with open-­ended questions, larger nationally representative data, and experiments In a report prepared for the UK Longitudinal Studies Centre (ULSC) about understanding society – a major longitudinal survey in the UK with 40 000 individuals – Nandi and Platt (2009) describe the process of development of a series of new ethnic identity questions and the rational for asking these questions based, inter alia, on a thorough review of the literature and focus groups meetings In Europe there are a few large datasets that include questions on ethnic identity such as the German Socio-­Economic Panel (GSOEP) but they ask this question only to immigrants or those who are foreign-­born There are hardly any surveys that ask natives about their ethnic and national identity Other datasets such as the International Social Survey Programme offer the advantage of cross-­country comparisons vis-­à-­vis ethnic identity Other surveys ask respondents how proud they are to have an identity as a particular nationality and can capture ethnic and national identity and feelings of patriotism Such a survey is the World Values Survey (WVS) that asks respondents how proud they are to have an identity as, for example, Greek Using experiments, on the other hand, is not an easy task; validity and reliability cannot be compromised, and the correct choice of the control group is critical Researchers have to spend a lot of time observing and learning the right way to run experiments Experiments can provide intuition and open new avenues of research Combining surveys and experiments is another way to produce useful data for analysis Behavioral and experimental economists have run several experiments in their quest to test identity, social identity, natural identity, preferences, religion, and other human traits such as altruism, fairness and reciprocity 3.2  Findings on Ethnic and National Identities13 Citrin and Sears (2009) provide an excellent overview of studies that attempt to measure identities and examine the competing hypothesis of the ethnic and national identities of Americans Acknowledging the important nuance between identify ‘as’ and identify ‘with’, these studies employed the following datasets: the 1992 and 2002 American National Election Studies (ANES), the 1994 and 1996 General Social Surveys (GSS) and the 1994 to 2000 series of Los Angeles County Social Surveys (LACSS).The LACSS survey respondents were asked ‘When it comes to political and social matters, you think of yourself mainly as just an American, mainly as a member of an ethnic group, or both?’ and responses determined their classification as a ‘national’, ‘ethnic’, or ‘hyphenated’ American category These three surveys also included variables that were used to create a measure of the strength of patriotic feeling From the LACSS an ethnicity measure was created to quantify the strength of connectedness with specific ethnic groups as well as the importance of ethnicity to one’s overall sense of identity To determine identity choice, the authors used the question: ‘When you think of social and political issues, you think of yourself mainly as a member of a particular Immigrants, ethnic identities and the nation-­state  ­269 ethnic, racial, or nationality group, or you think of yourself as just an American?’ Ninety percent of respondents stated ‘just American’ as their association About half (54 percent) of those respondents felt American on all issues and another 28 percent felt that way for most issues Results from the LACSS are similar and reinforce the conclusion that the majority of respondents see themselves as American over any particular ethnic association The majority of white respondents claimed ‘just American’, but interestingly a large majority of Hispanic and black respondents placed nationality over ethnicity in their responses When given a follow-­up choice in the LACSS survey minority respondents would often list themselves as having dual or hyphenated identity Differences can be seen among the minority groups, however A larger percentage of black respondents stuck with identifying as ‘just an American’; more Hispanics and Asians responded with ethnic identity only Using Hispanic respondents as an example, results show that native-­born Hispanics are more likely to identify as ‘just American’ while foreign-­born respondents are likely to identify more with ethnicity solely Overall there is a majority of respondents who chose to identify as ‘just American’ across all minorities If given the option, there is a majority of minority respondents who will identify using a hyphenated label Those who prefer a purely ethnic identity are foreign-­born immigrants In another survey, the Pilot National Asian American Political Survey, with 1218 adults older than 18 years of age respondents were asked which of the following they were identifying ‘as’ in general: American, Asian-­American, Asian, ethnic-­American, or their own national origin The overwhelming majority (61 percent) chose some form of American identity Citrin and Sears (2009) further created a model of identity choice with independent variables including age, income, education, race and ethnicity, and an index of immigration status Results of their multinomial logit model show that the more experiences and claims (birth, citizenship) that one has in America, the more likely they are to identify as ‘just an American’ Age also increases the response as solely American identification although the authors not have an explanation for this Whites still exhibit the strongest identification as American, but as immigrants move further away from their immigration time/experience their likelihood of identifying as ‘just American’ increases The same study also tested patriotism with questions about love and pride in America and the American flag using the 2002 American National Election Study Ninety-­ one percent of respondents across all ethnic groups expressed an extremely or very strong love for the United States and 85 percent expressed very high levels of pride in the American flag Like identification, feelings of patriotism increase as time-­since-­ migration increases Regarding the strength of ethnic attachment, the authors used the questions: ‘How strongly you identify with other (ethnicity) people?’, ‘How important is being (ethnicity) to your sense of identity?’ and ‘How often you think of yourself as an (ethnicity) person?’ Interestingly, all ethnic and racial minorities identified ethnically more than whites, however, Asians had a rate only marginally higher than that of whites As with other measures in this study, more recent immigrants exhibit a stronger ethnic attachment Citrin and Sears (2009) next examined if a strong ethnic identity compromises national unity Using the LACSS, they studied three ethnic groups: whites, blacks and Hispanics America has traditionally been dominated by white, European Christians so identification collisions for whites should not be a problem For minority groups, however, ethnic 270   International handbook on the economics of migration and national identification may collide There is also the possibility that no matter what background, Americans will endorse the idea of a melting pot, a unified country no matter what their individual ethnicity is As would be expected, those identifying as ‘just an American’ have the weakest ethnic connection Those who identify as hyphenated are more likely to respond as solely ethnic identifying respondents Those choosing to identify as solely ethnic are the only groups that show diminished levels of patriotism Overall, the authors find that ethnic identity ‘appears to compromise patriotism toward America only at the extremes’ (p. 171), and that the ‘strength of national identity is ­pervasive’ (p. 173) The United Nations Educational, Scientific, and Cultural Organization’s (UNESCO’s) International Journal on Multicultural Societies published a series of research papers in 2005 about national identity and attitudes towards migrants, based on findings from the International Social Survey Programme (ISSP) Within this series, Heath and Tilley (2005) focused their research on British national identity and attitudes towards immigration, exploring the distinction between ethnic and civic conceptions The authors find that Britain is not as ‘open arms’ to migrants who would like to immigrate The majority of the sample (75 percent) said that the number of immigrants to Britain should be reduced and four-­fifths said that illegal immigrants should be expelled On the other hand, the authors find that once immigrants are in the country the British are more forbearing For example, respondents did not have particular views about the assimilation of immigrants or about multiculturalism; they favored, however, laws that outlaw racial discrimination While civic aspects are very important for British identity, patriotism is not The authors conclude with the speculation that Britain may gradually move towards a ‘civic only’ conception of identity Manning and Roy (2010) study the extent and determinants of national identity in Britain using the UK Labour Force Survey Overall, the authors find evidence for a culture club, and in fact one connected with Muslims The vast majority of those born in Britain, of whatever ethnicity or religion, think of themselves as British While newly arrived immigrants not think of themselves as British, with additional years of living in the UK they Interestingly, this acculturation into the British national identity occurs sooner for immigrants from poorer and less democratic countries of origin Georgiadis and Manning (2013) investigate the correlates of national identity in Great Britain using the 2007 Citizenship Survey of England and Wales administered by the Department for Communities and Local Government Respondents reveal what they consider to be their national identities, identifying whether they consider themselves to be British and how strong they feel they are This is correlated with factors covering ethnicity, religion, English language proficiency, discrimination and the economic background, among others White and nonwhite natives exhibit a very high British national identity, while nonwhite immigrants report much lower values The main finding of the paper is that people feeling well treated, with their values and actions tolerated, are more likely to identify with the country and to feel that they belong to society People who are more surrounded by likes have a stronger feeling to belong to the country Religious minorities are more likely to experience conflicts between religion and national identity Clots-­Figueras and Masella (2008) study how governments can affect and alter individual identity through the education curricula After Catalan was taught in Spanish Immigrants, ethnic identities and the nation-­state  ­271 schools in 1983, pupils and their parents were more likely to have a Catalan identity and vote for Catalans in elections even if they had no Catalan origins Using data from the World Values Survey, Masella (2013) finds no evidence that individuals in more ethnically diverse countries have a less intense national identity Examining the issue separately for minorities and majorities, the author finds that the minority has stronger national sentiments than the majority when the country is less ethnically diverse, and less intense national sentiments than the majority when the country is more ethnically diverse The author finally finds weaker national sentiments among larger groups and a lower connection with the ethnic group when individuals have high incomes 3.3  Identities in Economics and the Labor Markets People’s identities, and especially the ethnic identities of migrants, are important in economics because ethnic identity defines who people think they are, how they perceive themselves with respect to others and how they make decisions in their everyday life These decisions, for example, may have to with the purchasing preferences of individuals as consumers, with their decisions to invest in human capital, to work or to pay taxes.14 Identity economics can now explain the sub-­optimal (from the point of view of the neoclassical economics) behavior of choosing a profession and accepting lower than competitive wages rates Chapter 1, on migration and ethnicity, in this volume provides more analysis of how the ideal self and ideal fitting of an individual into a group can lead to ‘not-­rational’ economic outcomes Alesina and La Ferrara (2005) provide an excellent survey of the economic costs and benefits of ethnic diversity in developed and developing countries They also reflect on the endogenous formation of political jurisdictions and call for more research on the endogenous formation of ethnic identity and the measurement of ethnic diversity Bisin et al (2011) examine the correlation between ethnic identity and labor market outcomes of non-­EU immigrants in Europe using the European Social Survey The authors find that immigrants who express a strong identity are penalized in the labor market Notably, there is a penalty of 17 percent for first-­generation immigrants While second-­generation immigrants have the same probability of being employed as natives, they have a lower probability of finding a job when they have a strong ethnic identity Interestingly, the authors find that labor market policies and types of integration significantly affect the relationship between ethnic identity and employment prospects While more elastic labor markets facilitate immigrants’ employment, no market can protect immigrants from unemployment if they have strong ethnic identities Using the German Socio-­Economic Panel, Casey and Dustmann (2010) study the self-­ identification of immigrants with the home and host countries, and investigate how identification with either country relates to immigrants’ and their children’s labor market outcomes The authors find that identifying with either country is only weakly related to labor market outcomes Lastly, they find a strong intergenerational transmission of identity from one generation to the next Benjamin et al (2010), in their social identity and preferences paper, use experimental data to infer causal effects between identity norms and economic decision-­making Specifically, they tested the effect of ethnic, racial and gender category norms on time 272   International handbook on the economics of migration preference and risk preference Using methodology from social psychology and taking the self-­categorization theory as given, they ran experiments on Asian-­Americans, native black Americans, immigrant black Americans, white Americans, and men and women By ‘priming’ the laboratory subjects, they were able to find the marginal effect of increasing the strength of affiliation with that category The authors find that Asian-­Americans are less patient and more risk averse than whites when not primed, that is, when the social identity and norms of the group are not considered However, when they are ethnically primed, Asian-­Americans are significantly more patient with time This result confirms the characteristic of Asian-­Americans being more patient Regarding risk preferences, priming did not change their responses Further testing intertemporal choices, the authors found that for Asian-­Americans to differ payment when their ethnic identity is made salient, the interest rate must decrease dramatically Interesting results were also found for blacks in the US Making race salient to native blacks decreases the interest rate in intertemporal choices This is also in line with representative national data that show that black Americans are risk averse However, race salience had no effect on the intertemporal choices of black immigrants or whites Regarding risk preferences, the authors found that black Americans are more risk averse when race is salient, a result that is consistent with norms that black Americans not participate in the stock market and are less likely to be entrepreneurs Pendakur and Pendakur (2005) study how ethnic identity affects labor market behavior in Canada Using a direct measure of ethnic identity from a survey question which asks ‘Is your ethnic origin very important to you, somewhat important, not very important or not at all important?’, they examine the impact of ethnic identity on the use of informal networks to find jobs and on the quality of jobs found The authors find that ethnic identity is associated with the tendency to use informal methods to find a job The correlation varies across ethnic groups and depends on the size of ethnic communities White European ethnic minorities, for example, who said that ethnic origin is ‘very important’, and who live in areas with large co-­ethnic populations are about 25 percent more likely to use informal networks to find a job compared with comparable Europeans in small co-­ethnic populations areas In addition, men with strong ethnic identity who are also phenotypical minorities, have much lower occupational prestige Schandevyl (2010) looks at the Belgian trade unions in relation to immigrant members and their ethnic identity In a country that is divided into two cultures and identities – French speaking and Dutch speaking – immigrants and their various cultures and ethnic identities have added a serious challenge to the national trade unions Diversity and even conflicts of identity among the union members jeopardize the sense of unity and solidarity that characterizes labor unions Based on transcripts of trade union conventions and meetings, and on interviews the author conducted with activists and officials within unions, Schandevyl shows both the evolving thinking on identity and the development of union structures in relation to migration 3.4  Endogeneity and Causality The way the models are set up, and because identity can alter and remain fluid, economists are always facing the potential of endogeneity; often, the causality interpretation Immigrants, ethnic identities and the nation-­state  ­273 is dubious Hence, one needs to admit that estimated relationships are often ‘just’ correlations and not reveal causal structures This does not mean, however, that such an analysis is useless given the few meaningful settings (exogenous variation, important research issue and interpretable regressors) we have in economics In studies of racial identities, for example Darity et al (2006), there is an exogenous and an endogenous part of identity Looking at blacks and mulattos in the deep South in the US in the mid-­nineteenth century, Bodenhorn and Ruebeck (2003) are able to show that mulattos could alter their racial identity to advance and enjoy labor market benefits While adopting a mulatto identity generates pecuniary gains, it also inflicts psychic costs to individuals In conclusion, the authors reflect that their results imply that race is contextual A way to circumvent the endogeneity issue is with experiments Benjamin et al (2010) follow social psychology methods to introduce an exogenous variation in identity In ‘self-­categorization theory’, ‘primes’ or environmental cues can render – temporarily – a person’s behavior to lean toward the norms associated with the salient category Accordingly, experimental scientists can identify the marginal effect of a particular social category by priming subjects Indeed, the authors prime identities with unknown norms so that they can infer what those norms are via the behavioral response to the prime The idea here is that priming a particular social category reveals the marginal effect of increasing the strength of affiliation with that category 4  CONCLUSION While there is a rising interest in economics in the determinants and consequences of ethnic and national identities among individuals, groups and nations, and the various cultural backgrounds, research in this area is just at the beginning There are some ­theories and studies on ethnic identity formation and its consequences, see also Chapter 1 in this volume With few notable exceptions there is only limited evidence for national identity Even more, the joint evolution of migrant ethnicity, native ethnicity and national identities are not well understood This is particularly true if one studies and incorporates the complex picture of national identity structures and policies across the nation-­states NOTES   * We thank the anonymous referee for many helpful comments on earlier drafts   With the exception of Jews and the State of Israel, migrants are typically members of the majority in their country of origin and become members of the minority in the new country   Putnam (2007) has stressed the strong challenges caused by the rising social and ethnic heterogeneity in most advanced countries   See Chapter 27 in this volume   See Constant (2005) for a deeper inspection of the French immigration policy   For a further analysis of the German situation see Zimmermann (1996) and Bauer et al (2005)   Interestingly, the cricket test does not involve racial or other observable characteristics that can have any additional discriminatory consequences   For instance, it is understood that the state should exert limited involvement in society and economics 274   International handbook on the economics of migration   See Stewart (1997) and Darity et al (2006, p. 291 and fn 3)   This formulation follows Benjamin et al (2010) with the weights adding up to one for simplicity, while Georgiadis and Manning (2013) use a set of independent parameters in the utility function to impose flexibility to derive more meaningful implications for their empirical investigation 10 Another strand of literature about national identity perceives it as a substitute for religion (Harttgen and Opfinger, 2012) They constructed an index of national identity and found that social heterogeneity affects national identity through religious diversity That is, when religious diversity is high, individuals cannot identify with religion; they find national identity as a substitute for common values and norms Note that it is also possible that religion is a substitute for national identity This is often the case in Arab nations 11 See for these issues Chapter 25 in this volume and Zimmermann et al (2009) 12 See Chapter in this volume for the analysis of circularity 13 Chapter in this volume contains more literature on the determination of ethnicity, including Constant et al (2009) which does not have to be surveyed here again 14 Li (2010) using the European and World Values Surveys investigates the impact of individuals’ social identities on their tax attitudes controlling for the ethnic fragmentation of the country In the analysis, the author considers majority and minority groups She finds that both ethnic and national identities play important roles shaping tax morale, namely, in countries with high ethnic fractionalization tax morale is poorer REFERENCES Akerlof, G.A (1997), ‘Social distance and social decisions’, Econometrica, 65, 1005–27 Akerlof, G.A and R.E Kranton (2000), ‘Economics and identity’, Quarterly Journal of Economics, 115 (3), 715–53 Akerlof, G.A and R.E Kranton (2005), ‘Identity and the economics of organizations’, Journal of Economic Perspectives, 19 (1), 9–32 Akerlof, G.A and Rachel E Kranton (2010), Identity Economics How our Identities Shape our Work, Wages, and Well-­being, Princeton, NJ, and Oxford: Princeton University Press Alesina, A and E La Ferrara (2005), ‘Ethnic diversity and economic performance’, Journal of Economic Literature, 43, 762–800 Bauer, T., B Dietz, E Zwintz and K.F Zimmermann (2005), ‘German migration: development, assimilation, and labour market effects’, in Klaus F Zimmermann, (ed.), European Migration: What Do We Know? Oxford and New York: Oxford University Press, pp. 197–261 Benjamin, D.J., J.J Choi and A.J Strickland (2010), ‘Social identity and preferences’, American Economic Review, 100, 1913–28 Bernheim, B.D (1994), ‘A theory of conformity’, Journal of Political Economy, 102 (5), 821–77 Bisin, A., E Patacchini, T Verdier and Y Zenou (2011), ‘Ethnic identity and labour market outcomes of immigrants in Europe’, Economic Policy, 26 (65), 57–92 Bodenhorn, H and C.H Ruebeck (2003), ‘The economics of identity and the endogeneity of race’, NBER Working Paper No 9962, National Bureau of Economic Research, Cambridge MA Brubaker, Rogers (1992), Citizenship and Nationhood in France and Germany, Cambridge MA: Harvard University Press Brubaker, R and F Cooper (2000), ‘Beyond “Identity”’, Theory and Society, 29 (1), 1–47 Casey, T and C Dustmann (2010), ‘Immigrants’ identity, economic outcomes and the transmission of identity across generations’, Economic Journal, 120 (542), F31–F51 Citrin, Jack and David O Sears (2009), ‘Balancing national and ethnic identities: the psychology of e pluribus unum’, in Rawi Abdelal, Yoshiko M Herrera and Alastair I Johnston (eds), Measuring Identity: A Guide for Social Scientists, Cambridge: Cambridge University Press Clots-­Figueras, I and P Masella (2008), Education, language, and identity, mimeo, available at: http://www econ.ed.ac.uk/papers/catalan.pdf (accessed September 2012) Constant, A (2005), ‘Immigrant adjustment in France and impacts on the natives’, in Klaus F Zimmermann (ed.), European Migration: What Do We Know? Oxford: Oxford University Press, pp. 263–302 Constant, A.F., L Gataullina and K.F Zimmermann (2009), ‘Ethnosizing immigrants’, Journal of Economic Behavior and Organization, 69 (3), 274–87 Darity, W.A Jr, P.L Mason and J.B Stewart (2006), ‘The economics of identity: the origin and persistence of racial identity norms’, Journal of Economic Behavior and Organization, 60, 283–305 Georgiadis, A and A Manning (2013), ‘One nation under a groove? Understanding national identity’, Journal of Economic Behavior and Organization, forthcoming Immigrants, ethnic identities and the nation-­state  ­275 Gleason, Philip (1980), ‘American identity and Americanization’, in Stephan A Thernstrom, Ann Orlov and Oscar Handlin (eds), Harvard Encyclopedia of American Ethnic Groups, Cambridge, MA: Belknap Press Harttgen, K and M Opfinger (2012), ‘In the nation we trust: national identity as a substitute for religion’, Discussion Paper No 491, University of Hannover Heath, A.F and J.R Tilley (2005), ‘British national identity and attitudes towards immigration’, International Journal on Multicultural Societies, (2), 119–32 Hutchinson, John and Anthony D Smith (eds) (1996), Ethnicity, Oxford and New York: Oxford University Press Kastoryano, Riva (2002), Negotiating Identities: States and Immigrants in France and Germany, Princeton, NJ: Princeton University Press Lee, Taeku (2009), ‘Between social theory and social science practice: toward a new approach to the survey measurement of “race”’, in Rawi Abdelal, Yoshiko M Herrera, Alastair I Johnston and Rose McDermott (eds), Measuring Identity A Guide for Social Scientists, Cambridge, MA: Cambridge University Press, pp. 113–44 Li, X.S (2010), ‘Social identities, ethnic diversity, and tax morale’, Public Finance Review, 38 (2), 146–77 Manning, A and S Roy (2010), ‘Culture clash or culture club? National identity in Britain’, The Economic Journal, 120, F72–F100 Masella, P (2013), ‘National identity and ethnic diversity’, Journal of Population Economics, 26 (2), 437–54 Nandi, A and L Platt (2009), ‘Developing ethnic identity questions for understanding society, the UK Household Longitudinal Study’, Understanding Society, Working Paper Series, No 2009 – 03, Swindon: Economic and Social Research Council Park, R.E (1950), Race and Culture, New York: Free Press Pendakur, K and R Pendakur (2005), ‘Ethnic identity and the labour market’, Research on Immigration and Integration in the Metropolis, Working Paper Series No 05-­10 Putnam, R.D (2007), ‘E pluribus unum: diversity and community in the twenty first century’, Scandinavian Political Studies, 30, 137–74 Schandevyl, E (2010), ‘Identity, migration and diversity in Belgian trade unions’, National Identities, 12 (4), 351–64 Sen, A (2001), ‘Other people’, Proceedings of the British Academy, 111 Stewart, J (1997), ‘NEA presidential address, 1994: toward broader involvement of black economists in discussions of race and public policy: a plea for a reconceptualization of race and power in economic theory’, in J.B Stewart (ed.), African Americans and Post-­Industrial Labor Markets, New Brunswick, NJ: Transactions Zimmermann, Klaus F (1996), ‘European migration: push and pull’, Supplement to The World Bank Economic Review and The World Bank Research Observer, 10 (1995), 313–42, reprinted in International Regional Science Review, 19 (1996), 95–128; also in Klaus F Zimmermann and Thomas Bauer (eds) (2002), The Economics of Migration, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing, vol 1, pt I, pp 70–99 Zimmermann, K.F., A.F Constant and L Gataullina (2009), ‘Naturalization proclivities, ethnicity and ­integration’, International Journal of Manpower, 30 (1–2), 70–82 15  Interethnic marriages and their economic effects* Delia Furtado and Stephen J Trejo 1  Introduction Interethnic marriage rates have often been used as a proxy for the extent of assimilation by immigrant groups (Pagnini and Morgan, 1990; Qian and Lichter, 2007) Sometimes referred to as ‘the final stage of assimilation’ (Gordon, 1964), marriages between immigrants and natives simultaneously measure immigrants’ views of the host society and natives’ views of the foreign born Moreover, interethnic marriages can have direct impacts on the participants, including the children produced by these unions Immigrants that resemble natives are more likely to intermarry, but in sharing a life with a native, immigrants may become even more similar to natives Closely related to spouse selection and assimilation is ethnic identity, which itself can have important implications for how researchers measure intergenerational integration.1 This chapter selectivity surveys recent research on these issues by economists and other social scientists Section discusses the causes of intermarriage, differentiating between determinants related to direct preferences for ethnic endogamy, indirect preferences, and opportunity structures Section examines the economic consequences of intermarriage, focusing on the empirical methods used to identify causal effects Section discusses the strong links between intermarriage and ethnic identification, and the potential problems that arise for tracking the socioeconomic progress of later-­generation descendants of immigrant groups A final section summarizes and concludes 2  Determinants of Interethnic Marriage In his pioneering work on the economics of marriage, Becker (1973) develops a model of household formation whereby the marriage market generates couples that match on traits which are complements in the production of household goods Conceptualizing these household goods as companionship, healthy and happy children, and quality of meals, for example, he cites education, religion and race as examples of traits which are likely to be complements in production In Lam’s (1988) model of marriage, the gains from marriage result from the joint consumption, as opposed to production, of household public goods Since many of the commodities produced within families are also jointly consumed within families, it is optimal for marriages to form between people with similar demands for these goods Because ethnic backgrounds of spouses are likely to be complements in the production of ethnicity-­related household public goods (such as vacations to the homeland and ethnic meals), both Becker’s and Lam’s models predict marriage market matching based on ethnic background For similar reasons, spouse-­searchers may also find it optimal to match on education, age, language and religion, for example In Becker’s model, couples 276 Interethnic marriages and their economic effects  ­277 are formed in a manner which maximizes aggregate surplus in the marriage market However, in a world with search costs, optimal matches not always occur, forcing marriage market participants to make decisions about the characteristics of spouses they value most Moreover, given the spatial distribution of these traits and the fact that marriage markets tend to be local, ultimate marriage patterns will also depend on who belongs to particular marriage markets Thus, the determinants of intermarriage can be classified into three main categories Starting with the most obvious, some characteristics are suggestive of stronger preferences for ethnic goods consumed within families Recently arrived immigrants, for example, are likely to value shared ethnicity with a spouse more than people whose families have been in the host country for several generations The second category relates to preferences for spouse characteristics which are not in themselves about ethnicity, but happen to be more or less common among co-­ethnics For example, high endogamy rates among vegetarian Indians may be driven by tastes for marrying vegetarians, which are relatively more common among Indians, as opposed to tastes for marrying other Indians per se The final category does not concern preferences at all but is instead related to opportunity If because of ethnic residential segregation, spouse-­searchers are more likely to meet potential spouses with the same ethnic background, then high endogamy rates may result even if marriage market participants are randomly matched with each other.2 Several researchers have empirically examined the determinants of ethnic intermarriage.3 Comparisons between studies are not perfectly straightforward because they use different samples, different definitions of intermarriage and include different control variables In fact, while the term ‘immigrant’ most often refers to the foreign born, in some countries, nationality determines immigrant status so that people who were born and raised in a country may be considered foreign if they are not citizens of that country Even studies relying on US Census data are not directly comparable because questions asked in the Census varied from year to year For example, Furtado (2012) uses 1970 data on native-­born males with two immigrant parents and classifies a marriage as endogamous if a spouse has at least one parent born in the country of birth of the male’s father Chiswick and Houseworth (2011) use 1980 data on immigrants who arrived in the US before marriage and define ethnic endogamy based both on country of birth and ancestry Furtado and Theodoropoulos’ (2011) analysis of 2000 data uses immigrants who arrived in the US before the age of 18 and the native born who identify with a particular ancestry The authors classify a marriage as endogamous if both spouses identify with the same ancestry Despite the different samples and variable definitions, several results appear very robust across studies Preferences for Marrying within Ethnicity We start by considering those characteristics which are likely to directly impact people’s preferences for marrying a co-­ethnic Recently arrived immigrants are more likely to marry endogamously as are immigrants who arrived as older adults (Chiswick and Houseworth, 2011) Relative to the foreign born, the native born with two immigrant parents are more likely to marry outside of their ethnicity and the native born with only one foreign-­born parent are even more likely to marry out (Kalmijn and Van Tubergen, 278   International handbook on the economics of migration 2010) Individuals who identify with multiple ancestries are also likely to marry exogamously (Chiswick and Houseworth, 2011) Given that parents tend to prefer endogamous marriages for their children (Kalmijn, 1998), people who are more tied to the family home are more likely to marry within ethnicity Not surprisingly therefore, people who marry young are more likely to marry endogamously, while second and higher-­ order marriages are less likely to be endogamous (Chiswick and Houseworth, 2011).4 There are also certain characteristics which tend to make people more open to marry­ ing outside of their ethnic group Participation in the military is associated with more exogamous marriages (Chiswick and Houseworth, 2011; Furtado and Theodoropoulos, 2011), potentially because the military forces soldiers to leave their potentially homogeneous home environments and interact with people from various racial and ethnic groups (Fryer, 2007) Certain features of people’s ethnic backgrounds may also make them more accepting of outsiders Measuring globalization using a combination of several variables – the sum of imports and exports of books, the number of Ikea stores per capita, and the number of McDonalds per capita – Kalmijn and Van Tubergen (2010) find that, indeed, people from more globalized countries are more likely to marry natives Moreover, even when they marry immigrants or second-­generation immigrants, their spouses are less likely to share their ethnic background Males from countries with more ethnic heterogeneity (as measured by Alesina et al., 2003) are also more likely to marry natives than to marry immigrants or second-­generation immigrants from the same origin country (Kalmijn and Van Tubergen, 2010) Preferences for Other Characteristics Turning next to the second category of intermarriage determinants, we consider the role of preferences for characteristics which are not directly related to ethnicity but happen to be correlated with ethnic background In his seminal paper on the economics of marriage, Becker (1973) cites religion as an example of a trait which is likely to be complementary in the production of household goods Religious practices in general and especially the religious upbringing of children can be viewed as household public goods, making it optimal, according to Lam’s theory, for marriage market participants to match with someone with the same religion Empirically, Lehrer (1998) and, more recently, Sherkat (2004) find evidence that marrying someone of the same faith is important for spouse-­searchers, especially those in certain religions Chapter 18 in this volume specifically considers the religiosity of immigrants Given the relationship between religion and ethnic background – Italians tend to be Catholic, Israelis Jewish, and Iranians Muslim – a preference for marrying someone of the same faith may increase the likelihood of marrying a co-­ethnic Empirically examining this hypothesis is difficult because the most often used datasets for studying ethnic intermarriage, at least in the US, not contain information on religious background Noting that the US is predominantly Christian, Kalmijn and Van Tubergen (2010) start by considering the impact of the percentage of a person’s home country that is Christian on the probability of marrying within ethnicity Proxying for a person’s religion using the dominant religion in that person’s country of origin, they also construct the percentage of respondents in a state who have the same religion after subtracting the number of Interethnic marriages and their economic effects  ­279 people from the same country of origin As might be expected, a Christian background is associated with fewer ethnic intramarriages In addition, religious similarities to other ethnic groups increase the likelihood of marrying a first-­or second-­generation ­immigrant from a different country of origin Marital preferences need not arise from optimality conditions in models of household production or consumption Tastes for spouse characteristics may simply reflect social norms For example, Belot and Fidrmuc (2010) study gender asymmetries in the propensity to marry-out, noting especially that while black males are significantly more likely to marry whites than black females, the opposite is true for Chinese They present evidence suggesting that preferences for taller husbands than wives, in combination with differences in height distributions across ethnic groups, can explain a significant portion of the gender asymmetries in intermarriage rates Opportunity for Endogamous Marriages Regardless of people’s preferences for whom to marry, a prerequisite for marriage, at least in modern-­day society, is that marriage market participants first meet Independent of preferences, ultimate endogamy patterns will be heavily influenced by the availability of potential partners with the same ethnic background People from immigrant groups that are not highly represented in the host country will find it more difficult to encounter co-­ethnics in the everyday course of their lives Thus, group size is likely to be an important determinant of intermarriage In addition, because marriage markets are not national, groups with diffuse settlement patterns are likely to have lower endogamy rates than groups with concentrated settlement patterns It is difficult to determine the precise size of spouse-­searchers marriage markets Given tendencies for ethnic groups to self-­segregate, defining a marriage market which is too large is likely to underestimate people’s opportunities for endogamous marriages On the other hand, because people with stronger ethnic preferences are more likely to choose to live in ethnic enclaves, defining marriage markets which are too small can confound the roles of ethnic preferences and opportunity It may not be surprising, therefore, that researchers choose different levels of geography on which to construct availability measures Furtado (2012) constructs marriage markets based on county group, the smallest identifiable geographic area publically available from the 1970 US Census Micro Data Furtado and Theodoropoulos (2011) use metropolitan statistical areas (MSAs) which are larger than the 2000 equivalent of county groups Presumably even more concerned that using small geographic areas confounds the role of preferences and opportunity, Kalmijn and Van Tubergen (2010) use states Drawing on information from metropolitan areas and states, Chiswick and Houseworth (2011) construct measures of opportunity using data from the Census year around which people were most likely making marriage decisions Thus, marriage market variables for 18–27-­year-­olds in 1980 are constructed using 1980 Census data while variables for 46–64-­year-­olds in 1980 are constructed using the 1960 Census Regardless of how they are defined, these studies consistently find that more contact opportunities with co-­ethnics are associated with stronger tendencies to marry a co-­ethnic Sex ratios are another important factor in determining the likelihood of ­encountering 280   International handbook on the economics of migration an acceptable same-­ethnicity spouse Chiswick and Houseworth (2011) construct availability ratios as the number of males of appropriate age from the same ethnic group living in a geographic area divided by a corresponding number of females, taking into account the fact that husbands are typically two years older than wives While they find that sex ratios are important predictors of endogamy, Kalmijn and Van Tubergen (2010) fail to find a similar role for sex ratios defined in a slightly different way We conclude, therefore, that the estimated effect of sex ratios on endogamy patterns is not as robust as the estimated effect of group size This may be because variation in sex ratios is driven by some characteristics of the geographic area, such as labor market opportunities, which affect intermarriage patterns directly Characteristics that Affect Endogamy through Multiple Mechanisms Some traits not fit perfectly into only one of the three categories above but instead affect marriage decisions through multiple avenues Furtado (2012) and Furtado and Theodoropoulos (2011) present evidence suggesting that education affects endogamy decisions through all three mechanisms discussed above By what they call the cultural adaptability effect, schooling makes people more accepting of cultural differences in spouses resulting in a decreased likelihood of marrying within ethnicity By the enclave effect, schooling increases the probability of leaving ethnic enclaves, potentially to acquire schooling or because education is associated with more geographically dispersed labor markets (Wozniak, 2010) With fewer opportunities for encountering co-­ethnics, people with more schooling are less likely to marry endogamously Finally, the assortative matching effect starts with the premise that people have a preference for marrying someone with a similar level of education, as suggested theoretically in Becker (1973) and Lam (1988) and empirically in Schwartz and Mare (2005) Thus, the effect of schooling on ethnic endogamy should depend on the distribution of education by ethnic group More specifically, education should decrease the probability of marrying within ethnicity for people in low education groups but increase that probability for people in high education groups A spouse search model developed in Furtado (2006) demonstrates how these mechanisms operate in a theoretical perspective, while, using 1970 US Census data on second-­generation immigrants with two foreign-­born parents, Furtado (2012) shows that, controlling for the enclave effect, there is no empirical support for the cultural adaptability effect but the assortative matching effect seems to be an important mechanism through which schooling affects marriage decisions Using more recent data on the foreign born who arrived before the age of 18 and natives who identify with a particular ancestry, Furtado and Theodoropoulos (2011) find evidence for all three mechanisms Measuring assortative matching on education in different ways, Chiswick and Houseworth (2011) as well as Kalmijn and Van Tubergen (2010) also find evidence consistent with both cultural adaptability and assortative matching effects of education Host country language acquisition is likely to affect endogamy patterns in similar ways However, while language proficiency is associated with weaker propensities to marry within ethnicity (Furtado, 2012; Furtado and Theodoropoulo, 2011), interpretation of this relationship is difficult given that people with stronger ethnic attachments are more likely to marry endogamously and less likely to learn the host country language Interethnic marriages and their economic effects  ­281 Chiswick and Houseworth (2011) address this issue by examining the effect of linguistic distance of the immigrant’s mother tongue from English Although they generally find that people whose native languages are farther linguistically from English are less likely to marry-­out, women who speak the languages furthest away from English, Korean and Japanese, are in fact more likely to intermarry, suggesting that something besides language is driving results Using a potentially more exogenous source of variation in identifying the role of language, Bleakley and Chin (2010) show that English proficiency does indeed have a negative causal effect on the probability that immigrants marry someone with the same country of birth.5 These studies are not able to disentangle the mechanisms through which host country language acquisition affects endogamy patterns, but Kalmijn and Van Tubergen (2010) show that an increase in the number of people from other countries that speak the same language increases the likelihood of marrying immigrants or second-­generation immigrants from different origin countries relative to marrying natives This result is certainly suggestive of an assortative matching on language effect 3  Labor Market Effects of Interethnic Marriage The previous section showed that there are systematic differences between people who choose to marry within their ethnicity and people who choose to marry-­out Some of these characteristics may indeed be associated with labor market outcomes, but surely marriage outcomes also depend on random encounters and idiosyncratic personality traits A natural question then is whether, conditional on a person’s preferences and characteristics, spouse’s ethnicity has any impact on labor market and educational outcomes Analyzing these effects of intermarriage is interesting in its own right, but more importantly, spouse ethnicity can be viewed as a proxy for whether a person’s social circle is comprised mostly of co-­ethnics While the papers discussed in the previous section emphasized the distinction between marrying within ethnicity versus outside of ethnicity, the literature on the economic consequences of intermarriage focus on the distinction between marrying an immigrant and marrying a native This section examines this literature on the labor market impacts of marriage to a native and by extension, association with natives more generally There are several reasons why marriage choice may influence labor market outcomes of immigrants While it is true that immigrants fluent in the host country’s language are more likely to marry outside of their ethnicity (Bleakley and Chin, 2010), marrying a native is likely to further improve an immigrant’s language abilities A similar story can be told with respect to knowledge of, and comfort with, the host country’s customs and social norms As immigrants become socially indistinguishable from natives, they are likely to become more successful in the labor market Although by definition, marriages involve only two people, in practice, people typically acquire new friends and acquaintances as they start romantic relationships Given that personal connections play a central role in job acquisitions (Ioannides and Loury, 2004), social circle members acquired through marriage may be important in determining the jobs that immigrants get It is reasonable to assume that new network members are relatively more likely to be native born if an immigrant marries a native, and ­relatively 282   International handbook on the economics of migration more likely to be an immigrant if he marries another immigrant Given that natives are more likely to be employed and tend to have higher wages (Larsen, 2004), social connections to natives are likely to expose immigrants to better labor market opportunities One last mechanism through which marriage may affect wages and employment rates is purely institutional For undocumented immigrants, marriage to a native may bring with it the legal right to work in the host country Thus, immigrants who may have been limited to under-­the-­table work can gain access to a broader range of higher paying more formal jobs Within a variety of studies using different samples in different contexts, ordinary least squares (OLS) estimates unambiguously point to a positive labor market impact of marriage to a native In the seminal paper in this literature, Meng and Gregory (2005) find that intermarried immigrants earn higher wages than endogamously married immigrants even when controlling for observable measures of human capital such as schooling, English proficiency, and years since migration Similar relationships were found in Denmark (Çelikaksoy, 2007), France (Meng and Meurs, 2009), the Netherlands (Gevrek, 2009) and the US (Chi, 2010; Kanterevic, 2004) when not controlling for ­selection into different types of marriages It is difficult to interpret these results because the immigrants that marry natives are likely to have different unobservable characteristics than the immigrants that marry other immigrants If the immigrants that marry natives are more assimilated, in ways not captured by the assimilation controls in the models, then ordinary least squares analyses will overstate the labor market returns to marrying natives On the other hand, if conditional on the measures of assimilation and human capital in the models, the immigrants that marry natives have worse unobservable characteristics, then OLS may actually underestimate the labor market returns to marrying a native To address the endogeneity issue, most papers in this literature take an instrumental variables (IVs) approach, using marriage market conditions to instrument for marriage to a native Meng and Gregory (2005) use two variables, specifically the number of opposite sex individuals from the person’s age-­ethnicity-­religion group divided by the total number of opposite sex individuals and sex ratios within those age-­ethnicity-­religion cells, to instrument for intermarriage They find that IV estimates are larger in magnitude than OLS estimates pointing to negative selection on observables into marrying a native Meng and Meurs’ IV is the number of opposite sex individuals of the same sex, ethnicity and age living in the same region divided by the total number of opposite sex individuals of the same age and living in the same region Despite using region instead of religion in constructing their IV, Meng and Meurs’ (2009) results are similar to those in Meng and Gregory (2005) In contrast, Kantarevic (2004) finds that after selection is taken into account, immigrants in the US who marry natives not have higher wages than the immigrants who marry other immigrants However, also using US data,6 Furtado and Theodoropoulos (2010) find that marrying a native increases employment rates of immigrants, especially those with the lowest levels of education If marriage to a native sufficiently increases employment rates of low-­wage immigrants, then selection into the labor force may explain why researchers have failed to find wage effects of marrying a native in the US Furtado and Theodoropoulos (2010) go on to explore the mechanisms through which marriage to a native affects immigrant employment With cross-­sectional data, it is not Interethnic marriages and their economic effects  ­283 possible to distinguish between language abilities improving after marrying a native or immigrants fluent in English being more likely to marry natives However, they show that adding measures of assimilation, such as English fluency and residence in ethnic enclaves, to the baseline model only decreases the estimated coefficient on marriage to a native by a rather small amount, suggesting that there must be other mechanisms through which marrying a native increases the employment probabilities of the foreign born To examine whether changes in legal status can explain the employment premium of marriage to a native without information on immigrants’ past legal statuses, Furtado and Theodoropoulos (2010) compare the intermarriage returns for immigrants who have characteristics that are most common among undocumented immigrants, such as low levels of education and coming from Mexico and Central America, to those for immigrants without these characteristics They also compare the returns to marrying a native for Puerto Ricans and Mexicans These two groups are similar in terms of language, culture and average education but while a majority of undocumented immigrants in the US are Mexican (Passel and Cohn, 2009), Puerto Ricans all have the legal right to work in the US Furtado and Theodoropoulos also compare the returns from marrying a native to the returns from cohabiting with a native; this comparison is interesting because marriage confers legal status whereas cohabitation does not These tests suggest that legal status may be part of the explanation for why marriage to a native increases the employment rate of immigrants, but it is unlikely to be the entire explanation.7 To explore the role of networks, Furtado and Theodoropoulos (2010) formulate and test a series of hypotheses, based on the network literature, which should hold true if, indeed, networks can explain why immigrants married to natives are more likely to be employed For example, they find that the largest returns are for immigrants with the lowest levels of education, that is, those who are most likely to find jobs through networks gained through marriage They also find the returns are smallest, or even nonexistent, for immigrants living in ethnic enclaves This makes sense in that co-­ethnics are more likely to have information about job openings in these ethnic neighborhoods than natives Consistent with the idea that native contacts are better able to aid immigrants in finding wage employment than self-­employment, the authors find no evidence that marriage to a native increases the probability of self-­employment In a separate analysis, Georgarakos and Tatsiramos (2009) also find that marriage to a native decreases the probability that immigrants start a business but increases the survival rate of the ­businesses they start Although the intermarriage literature generally points to a positive impact of intermarriage on labor market success, the validity of all of these results rests on the assumption that the instruments are not correlated with unobserved determinants of wages or employment rates IV estimates of the returns to marrying a native may overestimate the true relationship if immigrants living in areas with many co-­ethnics (or in areas with more males per female) have worse unobservable characteristics or opportunities Although this is inconsistent with immigrants migrating to areas with better labor market conditions – a typical concern in the immigration literature – it is consistent with the more assimilated and ambitious immigrants leaving their ethnic enclaves To address this concern, several new papers have taken a person fixed effects approach using panel data (Nekby, 2010; Nottmeyer, 2010) Controlling for all person-­specific 284   International handbook on the economics of migration characteristics which remain stable throughout a person’s lifetime, such as unobserved human capital, they examine whether immigrant wages are higher post-­marriage when they marry natives as opposed to other immigrants Using data from Germany, Nottmeyer (2010) finds that, if anything, foreign-­born males receive a short-­term boost in earnings shortly after marrying other immigrants and no statistically significant growth in earnings post-­marriage regardless of whether they marry immigrants or natives Similarly, Nekby’s (2010) analysis of Swedish data points to pre-­marriage wage growth for immigrants that eventually marry natives but no wage growth post-­marriage Married immigrant males in her study earn higher wages than single males, but the marriage premium is similar regardless of whether they marry immigrants from the same country of origin, immigrants from a different country of origin, or natives One interpretation of these findings is that the positive estimated labor market returns to intermarriage found in other studies are an artifact of either selection into marrying a native or instrumental variables which not satisfy the necessary exclusion restrictions Another interpretation is that the person fixed effects approaches are only able to identify the returns to the change in marital status itself as opposed to association with natives more generally Given the relatively short panels in both papers, wage comparisons are made just before and just after marriage Since a great deal of communication between spouses occurs before marriage, it is unlikely that this empirical strategy will pick up many of the returns arising from improvements in host country language fluency or knowledge of social norms Similarly, because social circles of husband and wife are likely joined before the actual wedding ceremonies, many of the returns arising from social networks will not be identified using this approach The panel data approach is able to identify any institutional gains marrying a native Thus, as long as any required paperwork is processed quickly, the gains to marrying a native as a result of acquiring the legal right to work in a country would be cleanly identified using a panel approach Under this interpretation, however, it should not be surprising that both Nekby and Nottmeyer find no labor market returns to marrying a native given that Furtado and Theodoropoulos (2010) find only weak evidence that legal status can explain the employment related returns to marrying a native Given the difficulties with both the instrumental variables and person fixed effects approaches to identifying the returns to marrying a native, it seems that natural experiments are in order Unfortunately, it is difficult to find situations generating a change in the likelihood of a particular type of marriage with no impact on labor market opportunities directly To our knowledge, no such natural experiment has been studied with respect to wages or employment rates, but Van Ours and Veenman (2010) exploit one such experiment in their study of educational outcomes of children from interethnic marriages Using exogenous variation resulting from the virtually random allocation of Moluccan immigrants across towns and villages at arrival in the Netherlands, Van Ours and Veenman (2010) find that children with a Moluccan father and a Dutch mother have higher educational attainments than children with either two Moluccan parents or a Moluccan mother and Dutch father The authors conclude that because mothers play a more central role in raising children, it is the origin of the mother that matters for children’s educational attainment, and familiarity with the Dutch educational system make Dutch mothers especially beneficial Interethnic marriages and their economic effects  ­285 A remaining question, however, is whether results regarding Moluccan intermarriage in the Netherlands are generalizable Using 2000 US Census data, Furtado (2009) explores the relationship between marriage to a native and educational outcomes of children from these marriages Exploiting the fact that parental marriage decisions are made at different times and often different places than 16–17-­year-­old children’s high school dropout decisions, Furtado instruments for parental marriage decisions using the size of the foreign-­born population in the child’s birth state in 1980 (just before the children in her sample were born) but controls for the foreign-­born population in the child’s state of residence in the year 2000 (just after dropout decisions are made) She finds that without instrumenting, native-­born children with two foreign-born parents are more likely to drop out of high school than native-­born children with one native-­born parent After instrumenting for parental marriage type, however, it is the native-­born children with two immigrant parents that are least likely to drop out of high school In addition, models which control for endogeneity not yield gender differences in the estimated effect of marriage to a native We conclude from this review that marrying a native, and by extension association with natives more generally, is generally associated with more labor market success of immigrants This result is robust to using data from different countries and constructing the instrumental variables with slightly different functional forms Positive estimated effects of marriage to a native, however, are not universal Kantarevic (2004) does not find any evidence that intermarriage increases wages for immigrants in the US after controlling for selection into marrying a native Using person fixed effects approaches, Nekby (2010) and Nottymeyer (2010) also find no evidence of a return to marrying a native in Sweden and Germany respectively With respect to educational outcomes of children from interethnic marriages, Van Ours and Veenman’s (2010) results conflict with those in Furtado (2009) In the end, it is difficult to determine whether any conflicting conclusions are a result of the different identification strategies or different contexts We conclude, therefore, with a call for research which is able to reconcile the conflicting results in the literature 4  Intermarriage and Ethnicity Sociologists have long recognized the strong links between intermarriage and ethnic attachments The links are complex and causality runs in both directions On the one hand, ethnic preferences influence the intensity with which individuals seek out co-­ ethnics as marriage partners, and ethnic preferences also help determine how readily individuals are accepted as potential mates by those belonging to other groups As a result, frequent intermarriage is one of the strongest signals of social assimilation by an ethnic group (Alba and Nee, 2003; Gordon, 1964) On the other hand, interethnic marriage complicates the ethnic origins of the resulting children, which can weaken the ethnic attachments of these children and of those in subsequent generations of the family tree (Perlmann and Waters, 2007; Waters, 1990) After a few generations in the US, so much intermarriage had taken place among the descendants of European immigrants who arrived in the late 1800s and early 1900s that most white Americans could choose among multiple ancestries or ethnic identities (Alba, 1990; Hout and Goldstein, 1994; 286   International handbook on the economics of migration Waters, 1990) For such individuals, ethnicity became subjective, situational and largely symbolic, and the social boundaries between these ethnic groups were almost completely erased Consequently, intermarriage has been a fundamental source of ethnic flux and leakage in American society (Hout and Goldstein, 1994; Lieberson and Waters, 1988; Perlmann and Waters 2007) In recent years, economists have shown increasing interest in issues related to ethnic identification and intermarriage An emerging literature within economics explicitly recognizes the complexity of ethnic identification and has begun to investigate the consequences of this complexity for labor market outcomes and policy In particular, economic models emphasize the potential endogeneity of identity and suggest mechanisms through which ethnic identification could be associated with both observed and unobserved characteristics of individuals and groups.8 A related strand of economic research focuses on developing nuanced measures of ethnic identity and the insights gained from analysis of these new measures (Constant and Zimmermann, 2008, 2009; Zimmermann, 2007) To date, most empirical work in the relevant economics literature has focused on foreign-­born immigrants, but some research has begun to analyze the native-­born second generation (Constant et al., 2009; Nekby and Rödin, 2010; Nekby et al., 2009) An important paper by Bisin and Verdier (2000) provides a useful economic framework for thinking about endogenous decisions regarding marriage and the socialization of children, and how such decisions influence the evolution of ethnic attachments across generations A parent hopes that his or her children will adopt the parent’s ethnic or cultural traits A parent can increase the chances of this happening by marrying a ­co-­ethnic and by exerting greater effort in socializing his or her children, but the parent faces increasing marginal costs in searching more intensively for a co-­ethnic spouse and also in exerting greater socialization effort Therefore, members of larger ethnic groups have less incentive to search intensively for a co-­ethnic spouse and to exert socialization effort, because being from a larger group improves your chances of finding a co-­ethnic spouse without searching very hard, and being from a larger group also raises the odds that your children are socialized in the preferred way with little effort on your part (because the socializing influence of ‘society at large’ on children is more likely to reflect the ethnic and cultural traits of larger groups) As a result, minority ethnic groups not vanish across generations through assimilation into the majority, because in equilibrium smaller groups have stronger incentives to marry-­in and to socialize their children In this way, minority groups can persist indefinitely, despite intermarriage rates that, when linearly extrapolated from the initial few generations, would suggest a relatively rapid extinction Instead, endogenous decisions regarding how hard parents work at inculcating ethnic identity among their children produces nonlinear assimilation across generations and the survival of small yet distinct ethnic groups Bisin et al (2004) provide an interesting empirical application of this model to the dynamics of religious populations within the US Duncan and Trejo (2007, 2009, 2011, 2012) argue that selective intermarriage and the resulting ‘ethnic attrition’ can generate potentially serious problems for tracking the socioeconomic progress of later-­generation descendants of US immigrant groups Because of data limitations, research on the US-­born descendants of Hispanic and Asian immigrants typically must identify the populations of interest using subjective measures of racial/ethnic identification rather than arguably more objective measures based on the Interethnic marriages and their economic effects  ­287 countries of birth of the respondent and his ancestors (Duncan et al., 2006, Snipp and Hirschman, 2004) In particular, this approach is typically the only feasible option for studies that seek to examine long-­term integration by distinguishing immigrant descendants in the third and higher generations (Blau and Kahn, 2007; Borjas, 1994; Farley and Alba, 2002; Smith, 2003; Trejo, 1997, 2003) A potential problem with this approach is that assimilation and intermarriage can cause ethnic attachments to fade across generations (Alba, 1990; Perlmann and Waters, 2007; Waters, 1990), and therefore subjective measures of racial/ethnic identification might miss a significant portion of the later-­ generation descendants of immigrants Furthermore, if such ethnic attrition is selective on socioeconomic attainment, then it can distort assessments of integration and generational progress For the specific case of Mexican Americans, Duncan and Trejo (2007, 2009, 2011) demonstrate the salience of these issues and elucidate the linkages between intermarriage, generational complexity and ethnic identification Analyzing microdata from the Current Population Survey (CPS) for children living with both parents, Duncan and Trejo (2011) compare an objective indicator of Mexican descent (based on the countries of birth of the child, his parents, and his grandparents) with the standard subjective measure of Mexican identification (based on the response to the Hispanic origin question) Immigrant generations turn out to be quite complex, and this complexity is closely related to children’s subjective Mexican identification For example, only 17 percent of third-­generation Mexican children have a majority of their grandparents born in Mexico Moreover, third-­generation children are virtually certain of being identified as Mexican if three or four grandparents were born in Mexico, whereas rates of Mexican identification fall to 79 percent for children with two grandparents born in Mexico and 58 percent for children with just one Mexican-­born grandparent Overall, about 30 percent of third-­generation Mexican children are not identified as Mexican by the Hispanic origin question in the CPS, and this ethnic attrition is highly selective In particular, the high school dropout rate of third-­generation Mexican youth (ages 16 and 17) is 25 percent higher when the sample is limited to those youth subjectively identified as Mexican This research suggests that ethnic attrition is substantial among third-­generation Mexicans and could produce significant downward bias in standard measures of attainment which rely on subjective ethnic identification rather than objective indicators of Mexican descent Do these findings necessarily mitigate concerns that Mexican Americans are experiencing markedly less intergenerational progress than other US immigrant groups (Huntington, 2004; Perlmann, 2005)? Duncan and Trejo (2011) show that available data are likely to understate the socioeconomic achievement of later-­generation Mexican Americans, but what does this imply about their standing relative to other immigrant groups? Given that intermarriage is the primary source of this bias, we might expect similar or larger biases for other immigrant groups, because most other groups exhibit intermarriage rates at least as high as those of Mexicans (Lichter and Qian, 2005; Lieberson and Waters, 1988) If the direction of the bias is the same for all groups, then appropriate corrections could produce no improvement or even deterioration in the relative position of Mexican Americans To address this issue, Duncan and Trejo (2012) investigate selective ethnic attrition for a wide range of national origin groups from important Hispanic (Mexico, 288   International handbook on the economics of migration Puerto Rico, Cuba, El Salvador and the Dominican Republic) and Asian (China, India, Japan, Korea and the Philippines) source countries Their findings suggest that ethnic attrition generates measurement biases that vary across national origin groups in direction as well as magnitude, and that correcting for these biases is likely to raise the socioeconomic standing of the US-­born descendants of most Hispanic immigrants relative to their Asian counterparts Like Mexicans, Puerto Ricans are an Hispanic group that shows signs of intergenerational stagnation, and the extent and selectivity of ethnic attrition seems roughly similar for US-­born Puerto Ricans as for Mexican Americans The selectivity of ethnic attrition is reversed, however, for Asian-­American groups with comparatively high levels of education, such as US-­born Chinese, Japanese, Koreans and Indians Among the descendants of immigrants from these Asian countries, those with fewer years of schooling are less likely to retain an Asian identification, which suggests that ethnic attrition inflates standard measures of socioeconomic attainment for later-­generation Asian Americans Note that these patterns are broadly consistent with the ‘assortative matching effect’ in Furtado’s (2006, 2012) model of interethnic marriage, which predicts that members of high-­attainment groups who intermarry should be negatively selected in terms of attainment, whereas the corresponding selectivity should be positive for intermarried members of low-­ attainment groups.9 5  Conclusion Among academics as well as policy analysts, there is a general appreciation for the association between the social integration of immigrants and their economic assimilation From a theoretical perspective, causality is likely to run in both directions, which makes it difficult to identify empirically the underlying mechanisms Another empirical issue is that, while there are many standard ways to measure economic assimilation, measures of the degree to which immigrants and their children interact with the host society are not as readily available This chapter selectively surveyed the recent economic literature on one particular measure of immigrants’ social integration: interethnic marriage We started by considering the determinants of intermarriage, separating them into those factors which are likely to affect preferences for endogamy directly and those which reflect the availability of desirable same-­ethnicity potential spouses residing within close geographic proximity We then turned to an examination of the labor market effects of interethnic marriage Most studies find beneficial effects for immigrants who marry natives rather than other immigrants, even after accounting for the endogeneity of cross-­nativity marriage, but this finding is not universal In discussing this literature, we offered several ways to interpret seemingly conflicting results Ultimately, however, further research is needed – ideally, research that explores several different sources of exogenous variation in intermarriage within similar contexts – before we can more definitively determine how and why intermarriage affects economic outcomes Finally, we described some of the emerging economic research on the links between intermarriage and ethnic identity Although this literature is still in its infancy, theoretical work in this area provides important insights into the mechanisms through which ethnicity and culture are transmitted across generations, and related empirical work Interethnic marriages and their economic effects  ­289 refines our understanding of ethnic identity and its economic effects and reveals the complexity of immigrant generations Future work in this area holds much promise NOTES * We would like to thank Klaus F Zimmermann, Amelie F Constant, and two anonymous referees for helpful comments and suggestions See Chapter in this volume for a discussion of the relationship between assimilation and ethnic identity Many researchers have classified the determinants of intermarriage along these lines In her study of black–white intermarriage, Wong (2003) refers to mating taboos, courting opportunities, and differences in endowments In Furtado (2012) and Furtado and Theodoropoulos (2011), education affects ethnic endogamy patterns through the cultural adaptability effect, the assortative matching effect, and the enclave effect Kalmijn and Van Tubergen (2010) refer to the first category as cultural and refer to the second and third categories jointly as structural In empirically examining the determinants of intermarriage, researchers must assume, either explicitly or implicitly, that the immigrants in their samples had the opportunity to choose between marrying within ethnicity or outside of ethnicity We note, however, that some immigrants are only allowed into a host country as a result of whom they marry To address this issue, researchers typically only consider the marriage decisions of the native born or the foreign born who arrived in a host country before marriageable age Although Kalmijn and Van Tubergen (2010) not consider age at marriage directly, they find that early marriage customs in a person’s country of origin are associated with increases in the likelihood of endogamy They exploit the fact that because of developments in the brain, language acquisition is significantly more difficult after a certain age In specifying his IV, Kantarevic (2004) takes the percentage of a person’s ethnic group that lives in that person’s state and divides that by the percentage of natives that live in the state Measures are constructed using data on unmarried individuals of the opposite sex Furtado and Theodoropoulos (2010) use the proportion of people in the person’s age group living in that person’s MSA that is foreign born as well as the sex ratio in the person’s country of origin-­age group While Kantarevic’s specifications controlling for selection not yield statistically significant results, Furtado and Theodoropoulos’ (2010) OLS and IV results are not statistically different from each other Furtado and Theodoropoulos (2010) demonstrate that Puerto Ricans who marry natives are more likely to be employed than Puerto Ricans who marry immigrants Chi and Drewianka (2011), however, find that Puerto Ricans receive no wage gain from marrying a native, whereas the corresponding wage gain for Mexicans is 30 percent Examples include Akerlof and Kranton (2000), Bisin and Verdier (2000), Bisin et al (2004), Austen-­Smith and Fryer (2005), Darity et al (2006), Manning and Roy (2010) and Bisin et al (2011) Bisin and Verdier (2011) and Chapter 14 in this volume survey some of the relevant literature As discussed earlier, Furtado’s model emphasizes how the supplies of potential spouses vary with ethnic-­ specific schooling distributions in marriage markets where individuals hope to match on both education and ethnicity A college-­educated Mexican American, for example, may choose to intermarry because of the relative scarcity of other Mexican ethnics with a college degree Asian Americans tend to be overrepresented on college campuses, however, so for these groups it may instead be the less-­educated individuals that face a more difficult time finding co-­ethnics to marry within their education group Consequently, this model predicts that members of high-­education groups who intermarry should be negatively selected in terms of education, whereas the selectivity should be positive for intermarried members of low-­education groups Because intermarriage is a fundamental source of ethnic attrition, the differences across groups in intermarriage selectivity predicted by Furtado’s model can generate corresponding differences in the selectivity of ethnic attrition references Akerlof, G.A and R.E Kranton (2000), ‘Economics and identity’, Quarterly Journal of Economics, 115 (3), 715–53 290   International handbook on the economics of migration Alba, Richard D (1990), Ethnic Identity: The Transformation of White America, New Haven, CT: Yale University Press Alba, Richard D and Victor Nee (2003), Remaking the American Mainstream: Assimilation and Contemporary Immigration, Cambridge, MA: Harvard University Press Alesina, A., A Devleeschauwer, W Easterly, S Kurlat and R Wacziarg (2003), ‘Fractionalization’, Journal of Economic Growth, (2), 155–94 Austen-­Smith, D and R.G Fryer, Jr (2005), ‘An economic analysis of “acting white”’, Quarterly Journal of Economics, 120 (2), 551–83 Becker, G.S (1973), ‘A theory of marriage: part I’, Journal of Political Economy, 81 (4), 813–46 Belot, M and J Fidrmuc (2010), ‘Anthropometry of love: height and gender asymmetries in interethnic marriages’, Economics & Human Biology, (3), 361–72 Bisin, A and T Verdier (2000), ‘Beyond the melting pot: cultural transmission, marriage, and the evolution of ethnic and religious traits’, Quarterly Journal of Economics, 115 (3), 955–88 Bisin, Alberto and Thierry Verdier (2011), ‘The economics of cultural transmission and socialization’, in Jess Benhabib, Alberto Bisin and Matthew O Jackson (eds), Handbook of Social Economics, vol 1A, Amsterdam: North Holland, pp. 339–416 Bisin, A., E Patacchini, T Verdier and Y Zenou (2011), ‘Ethnic identity and labour market outcomes of immigrants in Europe’, Economic Policy, 26 (65), 57–92 Bisin, A., G Topa and T Verdier (2004), ‘Religious intermarriage and socialization in the United States’, Journal of Political Economy, 112 (3), 615–64 Blau, Francine D and Lawrence M Kahn (2007), ‘Gender and assimilation among Mexican Americans’, in George J Borjas (ed.), Mexican Immigration to the United States, Chicago, IL: University of Chicago, Press, pp. 57–106 Bleakley, H and A Chin (2010), ‘Age at arrival, English proficiency, and social assimilation among U.S immigrants’, American Economic Journal: Applied Economics, (1), 165–92 Borjas, G.J (1994), ‘Long-­run convergence of ethnic skill differentials: the children and grandchildren of the great migration’, Industrial and Labor Relations Review, 47 (4), 553–73 Çelikaksoy, A (2007), ‘A wage premium or penalty: an analysis of endogamous marriage effects among the children of immigrants?’, Danish Journal of Economics (Nationaløkonomisk Tidsskrift), 145 (3), 288–311 Chi, Miao (2010), ‘Intermarriage and the economic assimilation of immigrants in the United States’, Job Market Paper, University of Wisconsin, Milwaukee, Department of Economics Chi, Miao and Scott Drewianka (2011), ‘How much is a Green Card worth? Evidence from Mexican and Puerto Rican men who marry women born in the U.S.’, available at: http://globalnetwork.princeton.edu/ publications/fulltext/2.pdf (accessed 11 January 2012) Chiswick, B and C Houseworth (2011), ‘Ethnic intermarriage among immigrants: human capital and assortative mating’, Review of Economics of the Household, (2), 149–80 Constant, A.F and K.F Zimmermann (2008), ‘Measuring ethnic identity and its impact on economic behavior’, Journal of the European Economic Association, (2–3), 424–33 Constant, A.F and K.F Zimmermann (2012), ‘Ethnosizing immigrants’, Journal of Economic Behavior and Organization, 69 (3), 274–87 Constant, Amelie F., Olga Nottmeyer and Klaus F Zimmermann (2012), ‘Cultural integration in Germany’, in Yann Algan, Alberto Bisin, Alan Manning and Thierry Verdier (eds), Cultural Integration of Immigrants in Europe, Oxford: Oxford University Press, pp 69–124 Darity, Jr, W.A., P Mason, and J.B Stewart (2006), ‘The economics of identity: The origin and persistence of racial Norms’, Journal of Economic Behavior and Organization, 60 (3), 283–305 Duncan, Brian and Stephen J Trejo (2007), ‘Ethnic identification, intermarriage, and unmeasured progress by Mexican Americans’, in George J Borjas (ed.), Mexican Immigration to the United States, Chicago, IL: University of Chicago Press, pp. 227–69 Duncan, Brian and Stephen J Trejo (2009), ‘Ancestry versus ethnicity: the complexity and selectivity of Mexican identification in the United States’, in Amelie F Constant, Konstantinos Tatsiramos and Klaus F Zimmermann (eds), Ethnicity and Labor Market Outcomes, Research in Labor Economics, vol 29, Bradford: Emerald Group, pp. 31–66 Duncan, B and S.J Trejo (2011), ‘Intermarriage and the intergenerational transmission of ethnic identity and human capital for Mexican Americans’, Journal of Labor Economics, 29 (2), 195–227 Duncan, Brian and Stephen J Trejo (2012), ‘The complexity of immigrant generations: implications for assessing the socioeconomic integration of Hispanics and Asians’, IZA Discussion Paper No 6276, Institute for the Study of Labor (IZA), Bonn Duncan, Brian, Joseph V Hotz and Stephen J Trejo (2006), ‘Hispanics in the U.S labor market’, in Marta Tienda and Faith Mitchell (eds), Hispanics and the Future of America, Washington, DC: National Academies Press, pp. 228–90 .. .INTERNATIONAL HANDBOOK ON THE ECONOMICS OF MIGRATION Acclaim for the International Handbook on the Economics of Migration ‘Constant and Zimmermann have assembled a collection of essays... 13 14   International handbook on the economics of migration 2  THE ECONOMICS OF ASSIMILATION Starting with the pioneering work of Chiswick (1978) on the assimilation of immigrant men in the United... evidence on the impact of migration on elderly parents After discussing the identification issues involved in the estimation, the chapter reviews the literature on the effects of migration on the