Research Proposal in response to the Call Migration and Labor Market Outcomes in Sending and Southern Receiving Countries

Davis, November 30, 2008 Research Proposal in response to the Call Migration and Labor Market Outcomes in Sending and Southern Receiving Countries Title: Brain Drain, Return Migrations and South-South Migration: impact on labor markets and on Human capital Project Coordinator: Giovanni Peri (University of California, Davis and Institute of Governmental Affairs) Principal Investigators: Christian Dustmann (University College London and CReAM), Frédéric Docquier (Université Catholique de Louvain), Giovanni Facchini (University of Milano and Centro Studi Luca D Agliano), Giovanni Peri (UC Davis and IGA) Other Participating Researchers: Based in Receiving/Northern Countries: Michel Beine (Univ of Luxembourg); Alok Bhargava (Houston University, USA); Susana Iranzo (Universitat Rovira i Virgili, Spain) Abdeslam Marfouk (Universite Libre de Bruxelles, Belgium); Anna Maria Mayda (Georgetown University, USA); Karin Mayr (University of Linz, Austria); Maria Pia Mendola (University of Milano-Bicocca, Italy); Josep Mestres (University College London and CReAM, UK); Hillel Rapoport (Bar Ilan University, Israel); Yoram Weiss (Tel Aviv University and CReAM, Israel). Based in Sending/Southern Countries Daniel Chiquiar (Bank of Mexico); Ines Raimundo (South African Migration Project, Mozambique). Introduction: The call for proposals by the World Bank identifies some fundamental issues and trends in international migrations. It calls for researchers to improve their understanding of migration with particular attention to developing countries and the eventual goal of producing sound, 1

research-based policy prescriptions. International migrations are not (and have never been) a one way-street. Returning migrants are a potentially valuable asset for the countries of origin. Examples include highly skilled engineers returning to Mumbai from Silicon Valley or Romanian construction workers returning to Bucharest from Barcelona. Despite this fact, large-scale emigration in particular of the highly skilled (and the associated brain drain) has usually been considered a threat for the development of poor countries. At the same time, immigrants to less developed countries (either present as legal guest workers or as illegal workers) could also be valuable to the receiving country. The extent of their economic integration, the probability of their return, the amount of remittances sent home and the effect on the sending and receiving countries are less known and documented vis-à-vis migration to rich countries due to the lack of systematic data. Our research proposal combines and coordinates the research agenda of sixteen academic economists based in highly regarded institutions in four different immigration-receiving countries, interacting with two different institutions in immigration-sending countries. Combining principal investigators, collaborators and supporting institutions the projects includes people from ten different countries. The composition of the team includes senior researchers with a well established agenda of research and proven ability to produce high quality contributions as well as young scholars and scholars from industrializing countries. The team will thus draw on an international network of reputable economists, with large visibility in the policy community and among international organizations. The objective is to produce five to six high quality publications, ultimately destined to top academic journal and to be presented in high profile conferences and relevant policy groups. At the same time, the research conducted will form an important basis for the public and policy debate. Research Objectives and Structure of the Team Several of the authors involved in the current proposal have an extensive track record of research on migration, in areas like return migration, brain drain, brain gain and south-south migrations. Our research aims at connecting leading researchers in the US, in Europe and establishing contacts and collaborations with researchers in immigrant-sending countries. We will combine research expertise on the themes of migration, return, brain drain and remittances from the sending country point of view and on south-south migration. We will, 2

through intensive interaction and collaboration, pursue the proposed research agenda and draw policy implications. The team participating to the project will be organized in four sub-team each based at one of the four participating institutions (University College London, Universite Catholique de Louvain, University of Milano and University of California, Davis). Each one has a research agenda, described below in more detail. Each sub-team is composed of a principal investigator (senior professor) and several collaborators in the participating institution and research partners in other prestigious institutions. Specific Research Themes The research activities will be carried out simultaneously in each of the four institutions, following the plans outlined in more detail, by each sub-team, in the rest of the proposal. Exploiting their specialization and their expertise each group will focus on a sub-set of topics related to the measurement and analysis of migration, its impact on the sending countries, return, brain drain, remittances and south-south migration. More specifically, the four groups will operate as follows: 1) The research group directed by Frederic Docquier will focus on the quantification of the brain drain, the potentially related brain gain with a specific project on the medical brain drain. 2) The research group directed by Christian Dustmannn will focus on migrants behavior in terms of decisions to return, savings and remittances and looks at the implications of these decisions for immigration policies. 3) The research group directed by Giovanni Facchini will investigate the impact of immigration on the labor markets of Southern receiving countries, with an emphasis on South Africa. 4) The research group directed by Giovanni Peri will focus on the impact of skilled migration from Eastern Europe to Western Europe and on measuring the premium and the economic relevance of return migrants from the US, particularly to Mexico. A comprehensive description of each research project, including motivations, objectives, methodology and Policy implications are detailed below. 3

Research Activity and Dissemination The team will produce five to six high quality academic papers and present them in high profile international conferences. Some of the conferences would be directed more specifically to high quality researchers and academics in the area of international migration while others could be aimed at attracting a high profile audience of policy-makers, policyadvisors and professionals in international migration organizations. The research group is interested and available to help organize one research conference/ research workshop to bring together members of the team and other high profile researchers in the area of migration. Obviously the involvement of the World Bank in the phase of dissemination and presentation of the results is important and we look forward to collaborate to best insure the success of the initiative. Proposed use of the Grant Money As the core part of the proposal is producing high quality academic research based in the four participating institutions we envision the use of Grant Money to foster this objective. Of the $ 500,000 we project each participating institutions to receive $100,000 (for a total of $400,000) to be used as research money for the principal investigator and his collaborators. These funds will be mainly used to pay for summer money and teaching buy-outs of the investigators, for hiring research assistants and data managers, to purchase data, software and equipment needed for the research, and to pay travel expenses to the collaborators. The remaining $100,000 would be used to organize and promote one international conference/workshop especially if we can coordinate with the World Bank logistic organization. 4

Specific research project of the group based in Université Catholique de Louvain Principal investigator: Frédéric Docquier Composition of the research Group: Abdeslam Marfouk (Université Libre de Bruxelles), Michel Beine (University of Luxembourg), Hillel Rapoport (Bar Ilan University) and Alok Bhargava (Houston University). Three RAs from the University of Louvain will also be involved: Luca Marchiori, Aysegul Kayaoglu, Sara Salomone. Additional RAs will have to be recruited. The research of this group focuses on the quantification of the brain drain, the potentially related brain gain with a particular focus on medical brain drain. Its goal is to perfect the collection and organization of data on international flows of migrants, use them to produce a complete picture of gains and costs of skilled migration and extend it to cover medical brain drain. Project 1: Extension and analysis of the brain drain data sets. (Frédéric Docquier, Michel Beine, Abdeslam Marfouk and Sara Salomone) Background Building on a previous work (Docquier and Marfouk, 2006), we have recently updated the database using new sources, homogenizing 1990 and 2000 definitions of what a migrant is, and introducing the gender breakdown. In Docquier, Lowell and Marfouk (2007), we provided new stocks and rates of emigration by level of schooling and gender for 195 countries for the years 1990 and 2000. In these data sets, the set of receiving countries is restricted to OECD nations. Generally speaking, the skill level of immigrants in non-oecd countries is expected to be very low, except in a few countries such as South Africa, the six member states of the Gulf Cooperation Council, some Eastern Asian countries. Our first objective is to expand the coverage of the data set by collecting immigration data from about 30 non-oecd countries. Up to now, we are certain that census data are available from 10 non-oecd EU countries or candidates (Bulgaria, Cyprus, Estonia, Latvia, Lithuania, Malta, Romania, Slovenia, Croatia and Macedonia), 3 Asian countries (Singapore, Israel and the Philippines), 5 African countries (South Africa, Rwanda, Uganda, Kenya and Ivory Coast) and that estimates based on labor force surveys can be obtained for the Persian Gulf countries (Saudi Arabia, Bahrain, Kuwait, Oman, Qatar and United Arab Emirates). Hence, we will be 5

able to characterize the size and structure (by educational attainment and gender) of bilateral migration stocks and flows from 195 origin countries to about 60 host countries (i.e. 11,700 pairs of countries). Objectives A first objective is to estimate the structure of immigration in the remaining countries of the world (starting from the bilateral matrix of Parsons et al. (2007), which does not include any gender or education breakdown). We should then be able to produce comprehensive 195x195 bilateral matrices of migration stocks by education level and gender for 1990 and 2000. A second objective is to use the 195x60 data set to study the dynamics of migrants diasporas (defined as the stock of people born in a country i and living in another country j). The evolution of diasporas is the outcome of a complex combination of self-selection (endogenous decisions of heterogeneous individuals to leave their country) and out-selection mechanisms (host-country decisions to accept immigrants, reflected in their immigration policies). How do characteristics of existing diasporas (size, education level, gender structure) affect the size and structure of current migration flows? How do existing diasporas impact on the effectiveness of immigration policy reforms? In particular, would more selective immigration policies generate important effects on the structure of EU immigration flows? Using both stock and flow data, we will model the evolution of the size, education level and gender structure of bilateral diasporas. We will also test for the existence of counter-migration flows (does the existing diaspora from i to j induce a migration flow from j to i?), a concept which has been largely disregarded in the literature, and which could attenuate the severity of the observed brain drain. Methodology Our empirical specification will have to take into account the following challenging issues. First, given the large number of zeros in bilateral stocks and flows (33 percent of zeros in bilateral migration flows to OECD and more than 60 percent of zeros in bilateral migration flows to poor countries), we have to use a two-step estimation method to avoid selection biases in the sample. Second, although we will control for country of origin and destination fixed effects, the sizes of the lagged diaspora and migration flow can be linked to some unobserved bilateral shocks contained into the error term. The endogeneity problem has to be addressed by instrumenting the lagged diaspora size. Finally, when dealing with the gender structure of migration flows, we have to account for the simultaneity of men and women s migration decisions. 6

Project 2: Integrated general equilibrium analysis. (Frédéric Docquier, Luca Marchiori and Hillel Rapoport) Background: The Docquier-Marfouk s database gave rise to a number of empirical studies on the determinants and consequences of the brain drain. In particular, it allowed many scholars to estimate some elasticities related to the feedback effects of the brain drain on sending countries. Nevertheless, there is no study in the literature that analyzes how all the direct and feedback effects interact. Objective Our objective is to combine the major results of the empirical and theoretical literatures in a unified general equilibrium model of the world economy. We will use an overlapping-generation structure with skilled and unskilled workers and divide the world in ten regions. The model will account for the major mechanisms through which the brain drain affects developing regions: the emigration of young individuals deteriorates the dependency ratio; skilled migration stimulates (resp. deteriorates) the marginal productivity of physical capital in the North (resp. in the South). This leads to a reallocation of physical capital; the skilled diaspora abroad generate network externalities in terms of foreign direct investments and diffusion of knowledge; new migrants remit a fraction of their income. Their propensity to remit depends on their level of education; increased migration prospects for the skilled stimulate the expected returns to schooling and generate more human capital formation ex-ante at origin. Using this general equilibrium model, we will provide a first global evaluation of the brain drain impact on economic activity, income, inequality in source and destination regions. We will also disentangle the relative impact of each specific mechanism and conduct robustness checks on key assumptions. This is important since there is still a lot of uncertainty about the magnitude of some feedback effects, an uncertainty arising from the fact that causality is hard to establish in cross-section empirical studies. Methodology 7

A first version of the model is already available and has been used in a couple of numerical exercises. Compared to the existing version, our goals are the following Refining the calibration of the propensities to remit by region of origin and level of education. The main data source to be used is the recent database (not yet in the public domain) constructed by Jimenez-Martin, Jorgensen and Labeaga (2007) on the volume and geographical breakdown of remittances from the 27 member countries of the European Union. This data set will be merged to the one described in the first section. Introducing microfoundations for migration and fertility choices. Developing methods generating country-specific predictions (rather than regional predictions). This final version of the model will be used to evaluate the relative contribution of the mechanisms studied in the new brain drain literature as well as to simulate the implications of various policy reforms. Project 3: Medical brain drain. (Frédéric Docquier, Alok Bhargava, Aysegul Kayaoglu) Background Recently, several authors such as Stark et al. (1997, 1998), Vidal (1998), Mountford (1999), Beine et al. (2001) or Stark and Wang (2002) argued that migration prospects may foster educational investments in developing countries, thus making it possible for a brain drain to be beneficial to the source country (i.e., the country may end up with a higher level of human capital after emigration is netted out). From a macroeconomic perspective, cross-section empirical studies by Beine et al. (2001, 2008) confirm that migration prospects have a positive and significant impact on human capital formation. Depending on the magnitude of the migration rate and the initial human capital stock, the ex post response on residents (after migration is netted out) can be either positive or negative. Such a mechanism is very unlikely to be observed in the medical sector, due to limited training capacities (especially in the short and medium terms). With a few exceptions (such as the Philippines and other middle-income countries), the training capacities are very limited. This is the case of many sub-saharan African countries. In these countries, the medical brain drain is a pure loss. 8

Objectives We will improve the data set on medical brain drain described in Docquier and Bhargava (2007) by extending the number of host countries and harmonizing the definition of what a migrant is across host countries; we will collect data on medical training capacities in developing countries. We will test for the incentive mechanism when capacity constraints are factored in. We will simulate the impact of educational policies (increasing training capacities), wage policies and migration policies on the supply of healthcare professionals in developing countries. Methodology Our empirical strategy will be similar to Beine et al (2008) except that we will take advantage of the panel structure of the data. Our empirical exercise will allow us to quantify the impact of training capacities and medical brain drain on the number of physicians per 1,000 people at origin. 9

Specific research Project of the group based at University College London and CReAM Principal Investigator: Christian Dustmann Composition of the research group: Josep Mestres (CReAM), Yoram Weiss (Tel Aviv University and CReAM, Israel). Two research assistants will be hired within the project. This group focuses on migrants behavior in terms of decisions to return, savings and remittances and looks at the implications of these decisions for immigration policies. Project 1: Return migration, Human Capital Accumulation, and the Brain Drain (Christian Dustmann and Yoram Weiss) Background Mobility of workers across national borders responds not only to the return to skills, but also to the opportunity and efficiency of skill acquisition. This last aspect has been largely overlooked in the literature that analyzes the causes and forms of migration. Skills are often acquired more cost effectively in countries where they do not get the highest return. Thus, individuals may choose to acquire skills that are highly sought after in countries other than their current residence. Student migrations are an example, with some countries having established themselves as learning centers that provide educational services above those demanded domestically. Movement of individuals across national borders to acquire skills where they can be acquired more efficiently, and to sell these skills where their return is highest leads to forms of migration that are different from the classic permanent migration type, inducing return migration and chain migrations. At the same time, these movements shift skills from one country to another, with implications for aggregate human capital formation, brain drain and brain gain, and growth. The tendency of individuals to acquire skills for the lowest price, and to sell them for the highest price, thus links the literature on causes and forms of migration with issues like brain drain and brain gain. Objectives: In this first project we develop a model that explains migrations as decisions that respond to where human capital can be acquired more efficiently and where the return to human capital is highest. This framework allows us to analyze flexible migration patterns, like return- and 10

chain migrations.1 Further, we will be able to rationalize patterns of return migration that are not easily compatible with a one-skill model as in Borjas and Bratsberg (1996), thus offering important new insights in issues like the selection of return immigrants. Our model has also implications for the debate on brain drain and brain gain. In Borjas and Bratsberg (1996) a brain drain issue arises when the price of skills is higher abroad, and may be amplified by those who return being the less able among those who left. However, if one considers the additional human capital accumulated abroad, the brain drain can be mitigated. In this case, it is possible to get a "brain gain", where those who return with their augmented local skills are locally more productive than those who left. If the proportion of those who return is large enough, aggregate output may even increase. Methodology: The basic framework is one in which a worker possesses two distinct skills that can be augmented by learning by doing. These skills command a different implicit price in different countries. The rate of accumulation is also different in different countries. Thus a person may move to a country where her skills grow fast and then apply these skills in a different country where the skills command a high price. Our model is a dynamic Roy model in which skills vary over time. We discuss one dynamic force: the differential learning rates in different countries. In contrast, to the static Roy model in which alternatives are characterized by the prices of skills only, our model specifies each alternative in terms of price and learning opportunities. Contribution: The project will address the important issue of return migration in an extended theoretical setting that explains migrations as decisions that respond to where human capital can be acquired more efficiently, and where the return to human capital is highest. Our analysis contributes to the literature on the selection of immigrants and return migrants, by offering a richer framework that may help to accommodate selections of emigrants and return migrants that are not immediately compatible with the one-dimensional skill model. Our analysis has also implications for the debate on brain drain and brain gain. In the two skills model presented here, return migration can lead to a mitigation of the brain drain, or even the creation of a "brain gain", where those who return are more productive at home than those who left. The work will offer a framework for the analysis of selective return migration and brain drain that can accommodate a far larger variety of empirical findings. It is important for our understanding of patterns of migration and labor market outcomes of immigrants in both 1 A recent OECD publication (OECD 2008, Table III.1) reports high incidence of return migration: The average out-migration rate after 5 years ranges from 28% for the Netherlands to 60% for Ireland. 11

sending- and receiving countries. Work will complement research of the Davis team and the Louvain team on return migration and the brain drain. Output will be one research paper, to be submitted to a top journal in economics. Project 2: Return migration, Remittances, and asset accumulation (Christian Dustmann and Josep Mestres) Background The theoretical determinants of how much and when to remit are diverse, ranging from altruism to insurance motives, but there is little empirical evidence on how the size and nature of remittances vary with return intentions and how remittances are used. Further, we know little about how return plans affect the way immigrants invest into assets and housing stock in home- and host country. Understanding this seems to be of key importance, in particular in view of the growing evidence that return migrations are far more widespread than previously thought. Objectives Our objective is to study jointly the remittance and the savings behavior of migrants. Further, it will investigate the way home ownership and asset accumulation interacts, and how this is affected by return plans. Our data is panel data, and allows us to disaggregate remittances by the use done either as family support, savings at home or other. This will allow us to address the problem of household fixed effects, correlated with remittance behavior and return plans alike. Analysis will assess the magnitude of the effect of return plans on remittances, as well as the way return plans lead to re-locating of savings and housing stock across countries. Both issues are extremely important for the evaluation of immigration policies, in particular in view of renewed debate about temporary migration schemes, and strengthening evidence that return migration is more widespread than previously thought. Methodology To structure the empirical analysis, we will first build a simple theoretical model that incorporates three reasons for remittances: remittances as an intra-family allocation across national borders, remittances as a way to accumulate savings in the home country, and remittances as an insurance motive. We will then develop an empirical model which is to be estimated using panel data on both remittances, and immigrant household return plans. This will help isolating the causal effect of return plans, through inclusion of household fixed effects. To address the problem of endogeneity through past remittances affecting future 12

return plans, we use an IV estimator in the tradition of Arellano and Bond (1991), where past return plans are used as instruments for current plans. Contribution Based on a unique panel data set for immigrants, we will give a detailed account of the various forms of remittances, and how they relate to individual and household characteristics. As we discuss above, remittances and savings are often conceptually not distinguishable, and we study remittances alongside savings. Our analysis will distinguish between remittances for family support, investments, and other remittances; on savings, we will distinguish between savings in the home country and savings in the host country. We will then proceed with analysis that distinguishes between immigrants who consider their migration as temporary and immigrants who consider their migration as permanent, allowing precise assessment how permanence of migration affects the various remittance flows. We will further investigate asset holdings of immigrants, and how immigrants distribute assets between the home - and the host country again an area where empirical evidence is extremely scarce. In co-operation with Chinese and Australian partners the empirical part of this study could use the collected panel data set for rural migrants to urban areas in China. The first wave of this survey has been collected early this year, and analysis will start soon. Our analysis here will again distinguish between permanent and temporary migrations. We aim at two research papers to be published in (top) field journals. 13

Specific research Project of the group based at University of Milan and Centro Studi Luca D Agliano Principal Investigator: Giovanni Facchini Composition of the Research Group: Anna Maria Mayda (Georgetown University) and Mariapia Mendola (University of Milan-Bicocca). To carry out the project at least two research assistant will be recruited. A group of scholars working on migration in a Southern African migration-sending country will also be involved in the project. The group will expand an existing collaboration with the Center for Population Studies and Centro de Estudios Africanos at the Eduardo Mondlane University in Maputo in Southern Mozambique. Our local partner, Ines Raimundo, is a professor at the Faculty of Arts and Social Sciences and the national coordinator of the South African Migration Project (SAMP) for Mozambique. This group will investigate the impact of immigration on the labor markets of Southern receiving countries, considering the interesting case of South Africa, the destination of substantial regional migration flows throughout its history and one of the most significant migration-receiving countries in the South (Ratha and Shaw, 2007). Project: The Labor Market Impact of Immigration to South Africa (Giovanni Facchini, Anna Maria Mayda and Mariapia Mendola) Background: With the fall of the apartheid regime, South Africa has been struggling to redefine its migration policy, but very little systematic evidence on the characteristics and implications of migration for the receiving country is available to inform policy makers. International migration from neighboring countries is a longstanding feature of the South African labor market, dating back to the mid-nineteenth century. Southern African intraregional migration was probably the single most important factor tying together the various colonies and countries of the sub-continent into a single regional labor market during the twentieth century (Wilson, 1972; Rogerson, 1995). Yet, entrenched patterns of migration have undergone major changes in the last fifteen years and the greater magnitude of intra-regional mobility has made South Africa a very interesting case study as a South-South migration receiving country. 14

South Africa has been the destination of substantial migration flows since the colonial period. Male labor migration to the mines and commercial farms from almost every other country in the region (especially modern-day Lesotho, Zimbabwe, Mozambique) was the most enduring form of legal cross-border labor migration. Significant poverty and unemployment rates in source countries have pushed also undocumented migrants to cross the border with South Africa (Crush et al. 2005). The end of the apartheid in the early 90s and the ensuing integration of South Africa in the Southern African Development Community (SADC) produced new opportunities for crossborder mobility and new incentives for temporary migration. For legal labor immigrants, official figures report that in 2005 the existing stock is was just above 1.1 million individuals, or 2.3% of the population (Ratha and Xu 2008). Yet, several observers suggest that also the number of illegal immigrants for employment is substantially larger, making the phenomenon a key policy issue. Objectives The recent migration policy debate in the country has been characterized by strong xenophobic feelings, which have been recently documented both in the press (McDonald and Jacobs 2005), as well as in opinion surveys. As in other destination countries, immigrants are blamed by natives for disrupting the local labor market, lowering wages and draining public resources ( Give them a Better Life The Economist, May 22, 2008). However, very little is known on the main drivers of the strong xenophobic feelings of the population as no systematic evidence exists on the labor market effects of immigration in South Africa. The purpose of this project is to provide a systematic assessment of the labor market effects of immigration in South Africa, using the recently released data contained in the 1996 and 2001 rounds of the international IPUMS. The extent to which immigration affects the economic opportunities of native workers has been a central issue in the ongoing global migration debate. The main theoretical argument is that immigration increases labor supply and thereby is likely to put downward pressure on natives' wages. The assumptions of this argument are that the labor demand curve is downward sloping and immigrants and natives are substitutes in employment. Most papers have tested this hypothesis for the US and European labor markets, where migration inflows are significant and data are available. Yet, the literature has found mixed evidence on the 15

labor market impact of immigration (see, for example, Borjas 1999, Card 2005, Ottaviano and Peri 2008). The pioneering work by Card (1990) on the Mariel boatlift in Florida uses the spatial correlation approach, i.e. the variation in the number of immigrants across different cities, to show that there is a weak (if any) impact of immigration on real wages and employment in the US. On the other hand, using data on the US national labor market and defining labor markets according to a worker s education and labor-market experience, Borjas (2003) finds evidence suggesting that immigration depresses wages for low-skilled US workers (see Hanson 2008 for a review of the literature). There are a number of conceptual and methodological problems to be considered while estimating the labor market impact of immigration - that can also help explaining the contrasting findings obtained in the literature. The concern with the national-level approach is that it might confound immigration with other labor market shocks that have hurt low-skilled workers - for example skill-biased technological change (Card 2005; Lewis, 2005). Thus, to avoid spurious correlations it is crucial to control for local demand shocks that raise wages and attract in-migrants, or relying on a robust (quasi-experimental) migration identification strategy. Yet, also the strategy of estimating the impact of immigration by comparing labor market outcomes across cities is not without concerns. First, an increase in the fraction of immigrants in a city does not necessarily raise the supply of low-skilled labor, since natives may move out in response to immigrant inflows; second, in the long run, an immigration induced increase in the supply of labor to a particular city can be diffused across the economy by intercity trade (Borjas, 1994, 2003). The last, but not less important concern, is that immigrant and native workers may be imperfect substitutes in production, such that net foreign labor inflows do not hurt native workers (and may actually help them) (Ottaviano and Peri, 2008). While the ongoing methodological debate suggests that there is ample scope to learn more on the labor market impact of immigration in the North, the economic analysis of the impact of migration between countries with contiguous borders in the South is in many ways an unexplored question. This is remarkable as South-South migration is nearly as large as South- North migration (half of the migrants from developing countries reside in other developing countries; Ratha and Shaw, 2008) and the debate on the effects of labor migration in Southern 16

countries, as we already mentioned, is not less heated. The purpose of this project is to start filling this gap in the literature, by carrying out what is to the best of our knowledge the first systematic analysis of the labor market consequences of migration to South Africa. Data and Methodology To carry out our analysis, we will use data from the 1996 and 2001 rounds of the South African Census, which are part of the Integrated Public Use Microdata Series International (IPUMSI) data set (https://international.ipums.org/international/index.html). The density of the IPUMSI South African data file is 10%. 2 The census data provides information at the individual level on a number of variables, such as the province and magisterial district of residence, the age, sex, race, religion and marital status and most importantly for our analysis the nativity 3, country of birth and citizenship of the respondent. The data set also contains information on the educational attainment, years of schooling, employment status, occupation, industry and total income of the individual. A starting point in our analysis will be the recognition of the high heterogeneity in the composition of South Africa immigrants. We will use individual level data in order to first investigate the characteristics and labor market outcomes of migrants to South Africa from different African countries. We will then explore the labor market effects of immigration, and in particular we will exploit the heterogeneous spatial distribution of the immigrants across magisterial districts in South Africa, and we will also take advantage of the distribution of the immigrants across skill and labor market experience. Implications Obtaining a better measurement of the labor market implications of international migration in South Africa will help policy makers designing better policies towards international factor mobility in the region. Indeed, the lack of good understanding of the phenomenon is likely to have kept the explicit statements in support of greater freedom of movement contained in the SADC agreement falling short from effectively easing constraints on migration. 2 Notice that, in the 1996 round of the South African census, 19 districts in the Eastern Cape province are not organized into households, i.e. 1.3% of the file. 3 Nativity indicates whether the person was native- or foreign-born. 17

Specific research Project of the group based at University of California Davis and Institute of Governmental Affairs Principal investigator: Giovanni Peri Composition of the research group: Susana Iranzo (Universitat Rovira i Virgili, Spain), Karin Mayr (University of Linz, Austria). We also use the collaboration of one Research- Assistant at UC Davis, Gregory Wright, plus we would need to hire one more research assistant also at UC Davis within the project. Collaboration between Giovanni Peri and Daniel Chiquiar of the Bank of Mexico is an essential part of the second project. This group focuses on two themes: first we analyze the impact of skilled migration from Eastern Europe to Western Europe accounting for the increasing trade integration and return migration. Second we measure the premium and the economic relevance of return migrants from the US, particularly to Mexico. Project 1: Skilled migration, returns and spillovers, the impact on Eastern Europe (Giovanni Peri, Susana Iranzo and Karin Mayr) Motivation As described by the research projects of the other groups there is an ongoing debate about the costs and potential benefits from loosing highly educated workers to international migration. Within this context an interesting natural experiment is about to take place when the European Union eliminates the transitory clauses limiting labor mobility between the Eastern to the Western Countries. In fact, since 1990 a large flow of educated Eastern European to the West has already taken place. Two important features accompany this potential brain drain from East to West and make our analysis of this case particularly interesting, from a theoretical and empirical point of view: 1) Several people only migrated temporarily from Eastern Europe and have already returned to their countries of origin. The OECD International Migration Outlook (2008), for instance, reports that almost 50% of the initial migrants to Western European countries are not any longer in the country after five years. The return to Eastern Europe of temporary migrants, often with some saved assets, entrepreneurial experience and accumulated skills to their countries of origin can constitute a relevant gain for those countries. 18

2) The increased integration of labor markets between East and West has been preceded by capital and goods market integration. Trade and investment have flown in abundance between East and West so that an increasingly integrated market can spread some of the productivity gain of highly educated migrants back to their country of origin. Analyzing migration and particularly the effects of skilled migration in a context in which return-migration and trade can produce beneficial spillovers from the migrants to their country of origin is the goal of this project. Aims and objectives Our goal is to incorporate into the analysis of skilled migration, particularly from Eastern Europe, the two factors mentioned above. The impact of Brain-Drain on the sending countries can be very different accounting for them. If returnees bring back valuable skills and if improved trade ties allow the increase productivity of emigrates to the West to benefit also the East then emigration of skilled workers can benefit also workers in the countries of origin. While other groups are working on related potential brain gain phenomena both from a theoretical and empirical point of view we focus on the case of East-West Europe as the one potentially best fit to exploit those gains. Methodology and Contributions: With Susana Iranzo I am developing a model of migration and trade between Eastern and Western Europe. The two economies have technological differences that imply that highly skilled workers are particularly productive in the West, and they have workers of different schooling levels. Using the model and data on income per person, trade and migration between East and West we analyze the gains from opening to trade (that already occurred) and we analyze the gains and losses from removing barriers to labor mobility. The model incorporate gains from trade due to specialization and increased varieties, and gains from migration accruing to migrants, who locate where they are more productive. The interesting feature of this analysis is that the increased productivity of highly skilled migrants to the West would generate more trade, a larger number of varieties traded and gains for the Eastern European Countries. Trade of goods and services are channels of diffusion of positive spillovers from higher productivity of the more educated. With the use of data and simulation we will quantify these gains. In a related project with Karin Mayr we inquire into the brain gain effect of return, estimating it quantitatively between Eastern and Western Europe. If experience abroad 19

generates useful skills that produce a wage premium upon return, then the positive incentive effects of international mobility can be achieved also with temporary migration. We find that given a sufficient wage premium for returnees, the net effect of skilled mobility on average human capital and wages of sending countries is likely to be positive. Again we use data on East-West migration and estimates of wage premia of returnees to quantify the extent of this possible brain gain. Policy implications The interesting policy implications to be derived from this analysis and simulations concern the design of temporary migration policies that are most effective in terms of the sending country welfare. Temporary immigration permits (for highly educated professionals and some workers with specific skills) could be the most viable and effective way to benefit sending as well as receiving countries. The incentive effects of such permits, their optimal design and the potential risks of enforcement are the issues that we intend to analyze within this theoretical framework. At the same time a policy of increased market integration between eastern and Western Europe seems the best way to ensure that the benefits produced by skilled migrants diffuse to the sending countries as well. Second project: Return migration and Premium to Returnees (Giovanni Peri and Daniel Ciquiar) Motivation From the empirical point of view there is anecdotal evidence that the return channel may already be very important for some countries and be picking up momentum. For instance, the recent article "Brain gain for India as elite return" (The Observer, April 20, 2008) claims that the very best IT engineers are now the returnees to India and "The Return of the Boat People" (The Economist, April 24th, 2008) emphasizes the positive role in entrepreneurship and the provision of skills of returnees to Vietnam. Moreover recent studies on returnees to Eastern Europe (e.g. Iara, 2008) have shown the existence of a significant return premium. The question is whether the temporary migration experience enhances the skills of returnees and puts them at an advantage towards other workers and earns them a premium. Aims and objectives Our goal is to use micro-data from the U.S. census (1990, 2000 and American Community Survey 2001-2007) and from the sending country censuses to obtain a clearer picture of the 20

return migration phenomenon. The individual data for the Mexican Census (1990 and 2000), for instance, contain information on whether a person has lived abroad. This information can be used to construct measures of the prevalence and intensity of return migration as well as to quantify the existence of a return premium, namely whether returnees earn more than similar workers who did not have experience abroad. Moreover with the cooperation of Daniel Chiquiar at the Bank of Mexico who has access to a unique household dataset from Mexico we can use microdata on a panel of Mexican families to analyze the economic impact of returnees. Methodology and Contribution The difficult part to identify a return premium is to distinguish it from a premium to the skills of the returnee as international migrants are usually selected and different from workers who do not migrate. With micro data on the wage earning of Mexican migrants to the US and on the wage of returnees in Mexico we can use some external shocks (such as recessions in the US) that caused larger flows of returnees to analyze whether such differentials can genuinely be interpreted as premium for the skills acquired abroad (language, entrepreneurial experience). At the same time, with those data, we would like to understand what affects the return probability of workers. Policy Implications Measuring the return to skills acquired abroad and their contribution to the economy of the sending country has important implications in designing incentives for returning immigrants. One could also use the estimates to predicting the likely magnitudes of return-migration as sending countries become richer and migrants become older. 21

References Arellano, M. & Bond, S. (1991), `some tests of specification for panel data: Monte Carlo evidence and an application to employment equations', The Review of Economic Studies 58(2), 277{297. Barghava A., F. Docquier (2008), HIV prevalence and migration of healthcare staff in Africa World Bank Economic Review, in press. Batista C., A. Lacuesta and P. Vicente (2007), Brain Drain or Brain Gain: Micro Evidence from an African Success Story, IZA discussion paper # 3035, Bonn, September 2007. Beine M., F. Docquier and H. Rapoport (2001), Brain Drain and Economic growth: Theory and Evidence, Journal of Development Economics 64, 275-289. Borjas, G. J. (1999) The Economic Analysis of Immigration. In Orley C. Ashenfelter and David Card, eds., Handbook of Labor Economics, Amsterdam: North-Holland, pp. 1697-1760. Borjas, G.J. (2003), The Labor Demand Curve is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market'' Quarterly Journal of Economics 1335-1374. Borjas, G.J. and B. Bratsberg (1996), "Who Leaves? The Outmigration of the Foreign- Born," The Review of Economics and Statistics, 78, 65-76. Card D. (1990), The Impact of the Mariel Boatlift on the Miami Labor Market Industrial and Labor Relations Review 43. Card, D. (2005). Is the New Immigration Really so Bad? Economic Journal, 115(507): 300-323. Crush, J., V. William and S. Peberdy (2005) Migration in Southern Africa Paper prepared for the Policy Analysis and Research Programme of the Global Commission on International Migration. Crush, J., and V. Williams (2004) Regionalizing International Migration: Lessons for SADC. Migration Policy Brief 11, Southern African Migration Project, Queens University, Kingston, Ontario. Docquier F. and A. Bhargava, A New Panel Data Set on Physicians Emigration Rates (1991-2004) Mimeo - last update: April 2007. Docquier F., B.L. Lowell and A. Marfouk (2007), A Gendered Assessment of the Brain Drain IZA Discussion paper 3235. 22

Docquier F. and A. Marfouk (2006), Measuring International Migration by Educational Attainment, 1990-2000, in Caglar Ozden and Maurice Schiff, eds., International Migration, Remittances, and the Brain Drain. The World Bank and Palgrave McMillan, Washington, DC, 151-200. Docquier F. and H. Rapoport (2008), Brain Drain and Human Capital Formation in Developing Countries: Winners and Losers, The Economic Journal 118 (528), 631-652. Dustmann C. and Y. Weiss (2008), Return Migration: Theory and Empirical Evidence for the UK British Journal of Industrial Economics, Vol. 45 Issue 2, pp. 236-256, 2007. Dustmann C. and O. Kirchkamp 2002, The Optimal Migration Duration and Economic Activities after Re-Migration Journal of Development Economics, 67, 351-372, 2002. Hanson, G. (2008), The economic consequences of international migration of labor, mimeo, UCSD. Iara Anna (2008) Skill Diffusion by Temporary Migration? Vienna Institute for Economic Studies, Working Paper # 46. Jimenez-Martin S., N. Jorgensen and J.M. Labeaga (2007) The volume and geography of remittances from the EU, Report for the European Commission, September 2007. Lucas, R.E.B. (1987) Emigration to South Africa s mines American Economic Review, 77, 3: 313-30. McDonald, D.A. and S. Jacobs (2005), (Re)writing xenophobia: Understanding press coverage of cross border migration in Southern Africa Journal of African Studies 23: 295-325. Mountford A. (1997), Can a brain drain be good for growth in the source economy? Journal of Development Economics, 53, 2: 287-303. Ottaviano G.I.P. and G. Peri (2008), Immigration and National Wages: Clarifying the Theory and the Empirics NBER Working Paper 14188, July 2008. Parsons C.R., R. Skeldon, T.L. Walmsley, and L.A. Winters (2007), Quantifying International Migration: A Database of Bilateral Migrant Stocks World Bank Policy Research Working Paper No. 4165. Ratha, D. and W. Shaw (2007), South-South Migration and Remittances, Development Prospects Group World Bank, Washington DC. Ratha D. and Z. Xu (2008), Migration and Remittances Factbook The World Bank. Rogerson, C. M. (1995) Forgotten places, abandoned places: Migration research issues in Southern Africa in J. Baker and T.A. Aina The Migration Experience in Africa, Nordika Afrikainstitutet 23