POLICY RESEARCH WORKING PAPER 3 069 International Migration, Remittances, and the Brain Drain A Study of 24 Labor-Exporting Countries Richard H. Adams, Jr. The World Bank R Poverty Reduction and Economic Management Network W Poverty Reduction Group June 2003 | POLICY RESEARCH WORKING PAPER 3069 Abstract While the level of international migration and the OECD have a secondary (high school) education or remittances continues to grow, data on international higher. migration remains unreliable. At the international level, * While migrants are well-educated, international there is no consistent set of statistics on the number or migration does not tend to take a very high proportion skill characteristics of international migrants. At the of the best educated. For 22 of the 33 countries in which national level, most labor-exporting countries do not educational attainment data can be estimated, less than collect data on their migrants. 10 percent of the best educated (tertiary-educated) Adams tries to overcome these problems by population of labor-exporting countries has migrated. constructing a new data set of 24 large, labor-exporting * For a handful of labor-exporting countries, countries and using estimates of migration and international migration does cause brain drain. For educational attainment based on United States and example, for the five Latin American countries OECD records. He uses these new data to address the (Dominican Republic, El Salvador, Guatemala, Jamaica key policy question: How pervasive is the brain drain and Mexico) located closest to the United States, from labor-exporting countries? migration takes a large share of the best educated. This Three basic findings emerge: finding suggests that more work needs to be done on the * With respect to legal migration, international relationship between brain drain, geographical proximity migration involves the movement of the educated. The to labor-receiving countries, and the size of the vast majority of migrants to both the United States and (educated) population of labor-exporting countries. This paper-a product of the Poverty Reduction Group, Poverty Reduction and Economic Management Network-is part of a larger effort in the network to better understand how international migration and remittances affect developing countries. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Nelly Obias, room MC4-834, telephone 202-473-1986, fax 202-522-3283, email address nobias@worldbank.org. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at radams@worldbank.org. June 2003. (29 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the excbange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers cany the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by Partnerships, Capacity Building, and Outreach International Migration, Remittances and the Brain Drain: A Study of 24 Labor-Exporting Countries* Richard H. Adams, Jr. PRMPR World Bank 1818 H Street, NW Washington, DC 20433 Phone: 202-473-9037 E-Mail: radams(aworldbank.org *I would like to thank Maureen Lewis and Zafiris Tzannatos for useful comments on an earlier draft. Within the last decade an increasing amount of attention has focused on the relationship between international migration, brain drain, and economic growth. Since education has often been cited as a major determinant of long-term economic growth, conventional wisdom has typically argued that the international migration of people endowed with a high level of human capital - the so-called "brain drain" - is detrimental for the country of emigration.' According to this argument, the large-scale departure of highly-educated workers from developing countries tends to depress income levels and long-run economic growth rates in the developing world. This conventional view, however, has recently been challenged by the following argument. In a developing economy with a limited growth potential, the return to human capital is likely to be low. This in turn would lead to limited incentive to acquire education, which is seen as the engine of economic growth. However, since the world at large values education, allowing migration of the "best and brightest" from a developing country may actually increase the incentive to acquire education. Since only a small faction of educated people in a specific country would migrate, this would encourage the average level of education of the remaining population to rise. Deciding which of these arguments is most accurate is difficult, given the paucity of available information on the level and characteristics of international migration. For example, at the international level, there is no consistent or reliable system of data on either the number or skill characteristics of international migrants. Moreover, at the ' By "brain drain" this study does not mean the migration of engineers, physicians or other very highly skilled professions, but simply, the migration of more than 10 percent of the tertiary-educated population of a particular labor-exporting country. 2 national level, labor-exporting countries do not typically keep track either of the number or the skill characteristics of migrants. And while some labor-receiving countries do keep track these issues, they often use different definitions of immigration. As a result of these data problems, a host of key policy questions remain unanswered. Exactly how pervasive is the brain drain? Which countries or regions of the developing world are most affected? Does international migration deprive labor-exporting countries of a sizeable fraction of their "best and brightest," or are their numbers too small to worry about? This paper proposes to answer these, and similar questions, by developing a new data set composed of 24 large, labor-exporting countries.2 This data set includes all those developing countries which received more than $500 million dollars in official worker remittances in the year 2000.3 To ensure representativeness, the data set includes countries drawn from each major region of the developing world: Latin America and the Caribbean, Middle East and North Africa, Europe and Central Asia, East Asia, South Asia and Sub-Saharan Africa. The paper uses this new data set to do the following three tasks. First, in order to provide some perspective on the importance of international migration, the paper shows how the level of worker remittances received by these 24 labor-exporting countries has increased over time. Second, using data collected by the two main labor-receiving 2The 24 study countries include: Brazil, Colombia, Dominican Republic, El Salvador, Guatemala, Janaica, Mexico, Peru (Latin America and Caribbean); Egypt, Morocco, Tunisia (Middle East and North Africa); Albania, Armenia, Croatia, Turkey (Europe and Central Asia); China, PR, Indonesia, Philippines (East Asia); Bangladesh, India, Pakistan, Sri Lanka (South Asia); and Nigeria and the Sudan (Sub-Saharan Africa). 3The source of all data on official worker remittances in this study is: International Monetary Fund, Balance of Payments Statistics Yearbook. 3 regions in the world (United States and OECD),4 the paper provides estimates of the total level of migration from each of the 24 study countries. Finally, the paper uses data from the 2000 U.S. Population Census and estimates from the 2001 OECD Continuous Reporting System on Migration to show the stock of migrants by educational level (primary, secondary and tertiary)5 in the two main labor-receiving regions. These data also show the fraction of the population in each educational category in each labor- exporting country that has migrated to the United States or the OECD.6 The latter estimates, which are quite rough and in need of further refinement, provide some sense of the magnitude of the brain drain from each of the 24 labor-exporting countries. Results from the study show that international migrants are well-educated: 67 percent of all immigrants to the United States and 88 percent of those to the OECD have a secondary (high school) education or higher. While these figures do not include the large numbers of illegal (and presumably less educated) international migrants, it does appear that uneducated individuals have limited access to legal international migration. International migrants also tend to be much better educated that the rest of the population of their country of origin. However, in terms of actual brain drain on their country of origin, international migration does not seem to take a very high proportion of the best educated (tertiary). For example, with respect to migration to the United States. for 14 of 4Unfortunately, no data are available on the level of migration to the third most important labor-receiving region in the world: the Arab Gulf The OECD includes 30 member countries: Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan, Korea, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovak Republic, Spain, Sweden, Switzerland, Turkey, United Kingdom, and United States. 5 Primary education is defined as 0 to 8 years of schooling; secondary is 9 to 12 years; and tertiary is 13 years and above. 6For the OECD, estirates of the fraction of population in each educational category from each study country are taken from Carrington and Detragiache (1998). 4 the 20 countries for which data exist, international migration takes less than 10 percent of the population with a tertiary education. Migration to the OECD results in a slightly higher degree of braip drain, but even here international migration takes more than 10 percent of the tertiary-educated population from only 5 of 13 labor-exporting countries. The balance of the paper is organized as follows. To provide perspective on the growing importance of international migration, Section II presents data on changes in the level of worker remittances received in the 24 countries over the last twenty years. Section III explains the methodology used for estimating the level of migration to the United States and the OECD. Section IV applies this methodology to estimate the extent of brain drain to the United States by analyzing migration rates and the educational composition of U.S. migrants. Estimates of the level of brain drain to the OECD are presented using similar methods in Section V. Section VI concludes. II. Official Remittances from International Migration To gain a proper perspective on the importance of international migration, it is best to begin with the most visible product of international migration: remittances received. Not only are remittances critical to the foreign exchange position of many labor-exporting countries, but they are also vital to the consumption and investment behavior of migrant households themselves. Perhaps because of their importance to both labor-exporting countries and households, remittances tend to be the best measured and recorded aspect of the migration experience. For instance, the International Monetary Fund (IMF) keeps annual records of the amount of worker remittances received by each 5 labor-exporting country.7 No other organization - national or intemational - collects and publishes annual records on any other part of the international migration process. It should, however, be noted that the IMF only reports data on official worker remittance flows, that is, remittance monies which are transmitted through official banking channels. Since a large (and unknown) proportion of remittance monies is transmitted through informal, unofficial channels, the level of remittances recorded by the IMF underestimates the actual flow of remittance monies retuming to labor-exporting countries. For instance, a recent IMF study (El-Qorchi, Maimbo and Wilson, 2002) estimated that informal transfers of remittance monies could amount to $10 billion per annum.8 Despite these problems, Table 1 shows the level of official worker remittances received by each of the 24 study countries over the twenty-year period, 1981 to 2000. Total official remittances - measured in real terms - for the 24 labor-exporting countries now amount to over $36 billion per annum. Among the study countries, the three largest recipients of official worker remittances in 2000 are: India ($7,994 billion), Mexico ($5,816 billion) and Turkey ($4,035 billion). Mexico is also one of the countries recording the largest percentage rate of increase in remittances over the twenty-year period. It is interesting to note that that each of the three largest remittance-receivers is sending migrants to a different labor-receiving region. Although data are scanty, India is 7The IMF records amnual flows in international remittances in its publication, Balance of Payments Statistics Yearbook. 8 Focusing on the hawala system of informal transfer, this IMF study of 15 developing countries (2002: 64) estimated that about $35 billion per annum of remittance monies was transmitted through informal channels in the early 1980s, falling to $10 billion per annum in more recent years. The decline in the level of informal transfers was attributed to the disappearance of the black market exchange premiums in many developing countries during the 1990s. 6 probably sending most of its migrants to the OECD and the Arab Gulf. Because of its proximity to the United States, Mexico is sending most of its migrants to its northern neighbor. Turkey, the third largest remittance-receiver, is sending most of its migrants to the OECD (especially Germany). Table 2 provides another way of looking at the flow of official remittances. Here the data on worker remittances for the 24 countries are disaggregated by region of the world. Because of the importance of the United States as a labor-receiving region, Latin America and the Caribbean is the largest recipient of official remittances. This region of the world also recorded the highest percentage rate of increase in official worker remittances over the last twenty years. For the 24 study countries as a whole, Table 2 reveals that official worker remittances have increased at a strong 3.86 percent per year in real terms. To put this figure into perspective, Table 3 compares the annual rate of growth in official remittances for the 24 study countries with that of gross domestic product (GDP) for various regions of the developing world. For the twenty-year period, official remittances grew at a faster annual rate than did GDP in 119 low and lower middle-income developing countries (3.86 versus 1.61 percent per year). In two of the six regions of the developing world (Latin America and the Caribbean, Sub-Saharan Africa), official remittances also grew at a faster annual rate than did GDP. 7 III. Methodology for Estimating Levels of Migration Since few, if any, of the major labor-exporting countries keep accurate records on either the number or the educational characteristics of migrants, it is necessary to estimate these variables by using data collected by the main labor-receiving countries. For the purposes of this paper, the main labor-receiving countries (regions) include two: United States and the OECD. The third large labor-receiving region - the Arab Gulf- does not publish any data on the number or characteristics of migrants, and thus will not be included in this study. lIIa. Estimating Migration to the United States Following the pioneering methodology of Carrington and Detragiache (1998),9 this study employs four steps to estimate migration rates and the educational attainment of migrants to the United States. The first step is to use data from the newly released files of the 2000 U.S. Population Census on the "place of birth for the foreign-born population." Fortunately, these data are disaggregated by country of birth for about 50 specific countries. However, it is not clear whether all of the "foreign-bom" population are, in fact migrants. For example, a person born in Mexico and brought to the United States as an infant would probably not consider himself as a migrant. Moreover, it is also not clear how many of those who enter the United States illegally are, in fact, included in the "foreign- born" population figures. As some observers have suggested, the U.S. Census data may be grossly undercounting the actual migrant population that is living - legally or illegally 9 While the Carrington and Detragiache (1998) study is based upon data from the 1990 U.S. Population Census, this study uses newly released data from the 2000 U.S. Census. 8 - in the United States.10 Since illegal migrants are likely to be less educated than legal migrants, this may in turn lead to an overestimate of the average level of migrant education. This source of bias, however, should not seriously distort our estimates of the migration rate of individuals with tertiary education, since those with higher levels of education are more likely to be in the United States on a legal basis. The second step is to estimate from the 2000 U.S. Census data the number of "foreign born" from each country who are 25 years of age or older. This is necessary to maximize the comparability of the immigrant population from each country with the Barro and Lee (2000) data set, which measures the educational attainment of the population over the age of 25 in each country. The third step is to calculate for each population from each labor-exporting country the number of migrants in three specific educational categories: primary or less (0-8 years of schooling), secondary (9-12 years of schooling) and tertiary (13 or more years of schooling). For the United States, this is done by using data from a special tabulation from the 2000 U.S. Census done by the U.S. Census Bureau. This special tabulation shows by county of origin the educational attainment level of the foreign born population living in the United States in the year 2000. The final step is to use the Barro and Lee educational attainment data set to compute migration rates for each labor-exporting country for the same three educational categories mentioned above. These calculations enable us to show what fraction of each country's educational group has migrated to the United States. 10 In 1992 the stock of illegal immigrants in the United States was estimated at 3.4 million, or about 16 percent of the stock of the "foreign-born" population (Borjas, 1995). 9 Since this study relies heavily upon the educational attainment data contained in Barro and Lee (2000), it is useful to describe this data set in some detail. Barro and Lee base their estimates of the educational attainment of the population in various countries on either the most recent census data or on historical schooling enrollment figures. In cases where enrollment figures are used, estimates of the current stock of education are constructed using a perpetual inventory method. Of course, various factors (including migration) may distort estimates based on historical enrollment patterns. Nevertheless, the Barro and Lee data set represents the best available set of estimates of educational attainment for a broad cross-section of developing countries. 11Ib. Estimating Migration to the OECD For OECD countries, data on migration flows by country are collected and published in Trends in International Migration: Annual Report (2001). Unfortunately, these data are not as comprehensive as the U.S. Census data, and they differ from the United States data in at least three key ways. First, and most importantly, the data for OECD countries do not report the educational attainment of migrants. For example, no data are available on the educational distribution of Turkish migrants to Germany. For the lack of alternative, we will assume that the educational distribution of migrants from each labor-exporting country to the OECD is the same as that for the United States.'1 In other words, if 40 percent of Turkish migrants to the United States have a tertiary education, we will assume that the same percentage of Turkish migrants to the OECD have this level of education. This procedure is most likely to be accurate for those labor-exporting "This is the same strategy used by Carrington and Detragiache (1998) in their study of brain drain. 10 countries which send similar numbers of migrants to both the United States and the OECD. However, for some labor-exporting countries - like Tunisia or Turkey - which send much larger numbers of migrants to the OECD than to the United States, this procedure is likely to be quite problematic. For instance, our procedure will impute to the German Turks the same high level of education as is found in their United States counterparts. This may, in turn, lead to a gross overestimate of the brain drain from a country like Turkey. The second problem with the OECD data lies in its different way of classifying immigrants. Since United States-born children of immigrants have US citizenship, the United States defines an immigrant as a person who was born abroad to non-US citizens. Most OECD countries, however, follow an ethnicity-based definition of immigration status. This method classifies a person on the basis of the ethnicity of the parent, rather than on place of birth. Thus, a child of Turkish parents bom in Germany is typically classified as an immigrant. This different way of classifying immigrants has the net effect of increasing the stock of immigrants in any particular OECD country, and perhaps biasing our estimates of the educational status of "migrants" who were actually bom, raised and educated in that OECD country. The third difference between the OECD and the United States data has to do with the number of labor-exporting countries recorded. As noted above, the 2000 U.S. Census data can be used to count the number of migrants from about 50 different labor-exporting countries. By contrast, for each OECD country, OECD data only record the number of migrants from approximately the top fifteen-labor exporting countries. On the one hand, this might not be a significant problem for large labor-exporting countries, like Turkey. 11 However, for smaller labor-exporting countries, like Brazil or Sri Lanka, the actual number of migrants to any particular OECD country might not be recorded. This makes it difficult to accurately compare and contrast migration rates between different labor- exporting countries. IV. Brain Drain to the United States: Migration Rates and Educational Levels This section presents estimates of immigrants in the United States by educational attainment, both in absolute number and as a fraction of the individuals in the labor- exporting country with the same level of educational attainment. As explained in the previous section, following the reporting conventions of the U.S. Census Bureau, these figures consider as immigrants all "foreign-born" individuals residing in the United States. Thus, these figures may well include substantial numbers of people who are not usually regarded as either migrants or as part of the brain drain, such as individuals who migrated to the United States as children and have few ties with their original country of birth. Table 4 presents immigration data to the United States for all 24 study countries. It is striking to note that one single country - Mexico - dominates migration to the United States. In our sample about 50 percent of all immigrants in the United States come from Mexico!12 The second largest source of immigrants is the China, PR, which supplies less than 10 percent of total migrants. Table 4 shows the key role that education plays in migration. In the sample as a whole, about 67 percent of all immigrants have a secondary education or higher. This 12According to the 2000 U.S. Population Census, of the total foreign-born populationim the United States (31,107,000), Mexico accounts for 29.5 percent (9,177,000) 12 figure is even higher if the very large number of poorly educated immigrants from one single country (Mexico) are excluded. Excluding Mexico from the sample, 83 percent of all migrants to the United States have a secondary education or higher. As the case of Mexico suggests, the educational attainment level of migrants varies considerably from country to country. Focusing on those immigrants with the highest level of education (tertiary), the share of immigrants to the United States with a tertiary education varies from a low of 14 percent (Mexico) to a high of 80 percent (India). In general, there is a strong tendency for the level of educational attainment to vary inversely with distance from the United States. Latin American and Caribbean countries generally produce a lower share of migrants with a tertiary education, while those countries which are more distant (particularly in South Asia) produce the highest share of tertiary-educated migrants.'3 This phenomenon may reflect the impact of migration costs (both financial and time). That is, prospective migrants with low levels of education in countries which are close to the United States are better able to afford the costs of legal (and illegal) migration than similar prospective migrants in more distant countries. Table 5 presents estimated migration rates by educational category for each labor- exporting country. Focusing on the highest (tertiary) level of education, the results suggest that - for most countries -- migration to the United States is not causing much brain drain. For 14 of the 20 countries for which data exist, less than 10 percent of those with a tertiary education have migrated to the United States. This result even holds for China, PR, which is the third-largest source of immigrants to the United States (see Table 13 According to Table 4, while the share of tertiary-educated migrants from South Asia is 76.5 percent, the share of such migrants from Latin America and the Caribbean is only 19.8 percent. 13 4). In China, PR, only 2 percent of those with a tertiary education have migrated to the United States. However, for a handful of countries, Table 5 reveals that migration to the United States does cause brain drain. For the five Latin American countries (Dominican Republic, El Salvador, Guatemala, Jamaica and Mexico) located closest to the United States, migration takes a large share of the best educated. While 16 percent of those with a tertiary education in Mexico migrate to the United States, over 360 percent of those with this level of education in Jamaica migrate. The very high migration rate of the best educated from Jamaica means that the number of migrants with a tertiary education from this country actually exceeds the number of local residents with this level of education. In other words, the estimated figure for tertiary-educated migrants from Jamaica must somehow overstate the true size of the brain drain from that country. The second theme in Table 5 is that low-skilled migration to the United States is not very important for most labor-exporting countries. There are only two countries (El- Salvador and Mexico) for which the migration rate for those with a primary school or less education is greater than 10 percent. The highest figure for primary school migrants is that of El Salvador (12.4 percent), followed by Mexico (10.8 percent). While the migration data in Table 5 do not include the sizeable (and unknown) number of illegal migrants in the United States, who probably have much lower levels of education, the results for legal migration seem clear. Legal migration to the United States involves the movement of better educated people, people who are probably more educated than those who remain at home. 14 The final finding in Table 5 is that total migration rates to the United States are not very high for most countries. Total migration rates exceed 10 percent for only four countries: Dominican Republic, El Salvador, Jamaica and Mexico. All four of these countries are in Latin America, and three of them have fairly small populations.14 It should come as no surprise that small, Latin American countries have high average rates of migration because this is a consequence of United States immigration policy. By setting annual limits on the numbers of migrants that can come from each labor-exporting country, regardless of size, United States policy in effect favors lightly-populated countries in the Western Hemisphere. A limit of 20,000 legal migrants per year per country is much more of a binding constraint for Pakistan than it is for Jamaica, and this is one reason why Jamaica has the highest total migration rate (33.3 percent) in the table. V. Brain Drain to the OECD: Migration Rates and Educational Levels Table 6 presents data for 13 of the 24 study countries for which immigration data to the OECD exist. As in the case of immigration to the United States, one country dominates: Turkey. In our sample about 40 percent of all immigrants to the OECD come from Turkey. The second largest source of immigrants in our sample is China, PR, which supplies about 15 percent of total migrants. Two North African countries -- Morocco and Tunisia -- combine together to produce about 15 percent of all migrants to the OECD. Table 6 shows that education plays an even more important role in migration in the OECD than in the United States. While 67 percent of all immigrants to the United 14 The Dominican Republic, El Salvador and Jamaica each have a total (2000) population of less than 10 million. 15 States have a secondary education or higher, fully 88 percent of all immigrants to the OECD have this level of education. Moreover, of these secondary-educated immigrants to the OECD, over two-thirds -- 69 percent -- have a tertiary education. However, it is important to emphasize that these estimates of the number of educated migrants probably overestimate the true extent of the brain drain to the OECD for two reasons. First, levels of educational attainment for migrants to the OECD must be estimated, rather than observed as in the United States, because the OECD does not collect data on the educational characteristics of immigrants. For example, since the educational distribution of Turkish migrants to the OECD must be estimated on the basis of U.S. data, and the educational distribution of these OECD migrants might be quite different from that of their United States-bound compatriots, the figures in Table 7 might over-estimate the extent of the brain drain. Second, as in the United States, published data on the number of immnigrants to the OECD exclude the large (and unknown) number of illegal immigrants to these countries, who are likely to have lower levels of education. For instance, it is likely that a large number of illegal immigrants from North Africa to the OECD have low levels of education, simply because rates of educational attainment in these North African countries are low.15 Table 7 presents migration rates by educational category for each labor-exporting country. Focusing on the tertiary level of education, the available data suggest that migration to the OECD causes more of a brain drain than it does in the United States. For 5 of the 13 countries for which data exist, more than 10 percent of those with a tertiary education have migrated to the OECD. For these five countries, the rates of 15 For example, in Morocco and Tunisia about 76 percent of the population over age 25 has a primary school education or less. 16 migration for those with a tertiary education range from 16.5 percent (Sri Lanka) to 95.8 percent (Jamaica). The latter country - Jamaica - must be suffering from a particularly high degree of brain drain, because it has the highest migration rate for tertiary-educated people to both the OECD and the United States (see Table 5). The data in Table 7 suggest that migration to the OECD differs from that to the United States in another important way. While migration to the United States tends to take a high percentage of tertiary-educated people from neighboring (Latin American) countries, in the OECD the link between migration, education and geography is not so obvious. While three of the countries (Morocco, Tunisia and Turkey) with the highest rates for tertiary-educated migration are located close to the OECD, two other prominent countries (Jamaica and Sri Lanka) are not. More work needs to be done to identify the reasons why Jamaica and Sri Lanka send such a high proportion of their "best and brightest" to the OECD. The final point to be noted in Table 7 is that total migration rates to the OECD are lower than those to the United States. While four labor-exporting countries have total migration rates in excess of 10 percent to the United States, no country in the OECD has a total migration rate above that mark. Jamaica has the highest overall migration rate to the OECD, at 8.7 percent. Again, the combination of small population and a high degree of migration abroad lead Jamaica to record a higher migration rate than such large, labor- exporting countries, like Turkey and the Philippines. 17 VI. Conclusion While the level of intemational migration and remittances continues to grow between countries, data on the whole process of international migration remains fragmented, scattered and unreliable. At the international level, there is no consistent set of statistics on either the number or skill characteristics of international migrants. At the national level, labor-exporting countries do not typically keep track either of the number or the skill characteristics of migrants. And while some labor-receiving countries do keep track these issues, they often miss the large numbers of migrants who enter their borders on an illegal basis. Using a new data set of 24 large, labor-exporting countries, and employing estimates of migration and educational attainment based on United States and OECD records, this paper has tried to overcome these data problems to answer such policy questions as: How pervasive is the brain drain from labor-exporting countries? Which countries or regions of the developing world are most affected? Does international migration deprive labor-exporting countries of a sizeable fraction of their "best and brightest," or are their numbers too small to worry about? Five conclusions to these, and similar questions, emerge from this paper. While these conclusions are often based on estimates of migration which are rough and in need of further refinement, they are still suggestive. The first finding is that the level of international migration continues to grow annually. While no accurate, time-series data exist on changes in the total level of international migration, the International Monetary Fund (IMF) does keep annual records on the level of official remittances sent home by migrant workers. Over the last twenty 18 years, these records show that the official worker remittances have increased at a steady 3.86 percent per year for the 24 study countries.16 This annual rate of increase for official worker remittances is higher than that recorded for the annual rate of GDP growth (1.61 percent per year) in 119 low and lower middle-income developing countries over the last twenty years. The second finding from this study is that with respect to legal migration, international migration definitely involves the movement of the educated. In both the United States and the OECD, individuals with a primary education account for less than 30 percent of total immigrants. The vast majority of immigrants to both regions (67 percent in the United States and 88 percent in the OECD) have a secondary (high school) education or higher.17 From an educational standpoint, international migrants represent an elite that is much better educated than the rest of the population in their country of origin. The third finding is that while migrants are well-educated, international migration does not tend to take a very high proportion of the best educated (tertiary) population in labor-exporting countries. For 22 of the 33 countries in which educational attainment data can be estimated, 18 less than 10 percent of the tertiary-educated population of labor- exporting countries has migrated to the United States or the OECD. Part of the reason for this finding is that large labor-exporting countries are also typically large population 16 If the figures for remittance flows which occur outside of official banking channels could be included, it is likely that total remittances - official and unofficial - would have increased at an even faster rate over the last twenty years. " As emphasized in the paper, these figures do not include the large numbers of illegal (and presumably less educated) immigrants to the United States or the OECD. 18 Table 5 presents 20 countries with educational attainment data, and Table 7 presents 13 countries. 19 countries, which have a substantial number of tertiary-educated people. For example, one of the largest labor-exporters in this study - China, PR - has such a large pool of tertiary-educated people that international migration has a relatively small proportional impact on the domestic labor-market for the "best and brightest".'9 The fourth result follows from the preceding. For a handful of labor-exporting countries, international migration does cause brain drain. For example, for the five Latin American countries (Dominican Republic, El Salvador, Guatemala, Jamaica and Mexico) located closest to the United States, migration takes a large share of the best educated. Sixteen percent of those with a tertiary education in Mexico migrate to the United States, and over 360 percent of those with this level of education in Jamaica migrate. With respect to the OECD, international migration takes more than 10 percent of those with a tertiary education from five countries: Jamaica, Morocco, Tunisia, Turkey and Sri Lanka. Three of these large brain drain countries are located close to the OECD, while two of them (Jamaica and Sri Lanka) are not. The final finding concerns promising directions for future work on the issue of brain drain. Results from this paper suggest that there is a close relationship between brain drain, geographical proximity to labor-receiving countries, and the size of the population (especially the educated population) of labor-exporting countries. In many cases lightly-populated countries located close to either the United States or the OECD suffer the highest rates of brain drain for tertiary-educated people. Good examples of this phenomenon include Jamaica, El Salvador and Tunisia, each of which have populations of less than ten million. More work needs to be done to clarify the relationship between 19 For exarmple, China, PR has an estimated number of 20,562,000 people who have completed a tertiary education. Of this pool of highly-educated people, about 2.2 percent are immigrants in the United States and about 1.4 percent are immigrants in the OECD. 20 brain drain, geography and population, and to identify possible policy programs for ameliorating the possibly adverse impact of brain drain on these countries. 21 References Barro, Robert and Lee, Jong-Wha. 2000. Data Set on Educational Attainment of the Total Population Aged 25 and Over. Available on the web site of the Center for International Development at Harvard University (www.ksg.harvard.edu/CID). Borjas, George. 1995. "The Economic Benefits from Immigration," Journal of Economic Perspectives. No. 9, pp. 3-22. Carrington, William and Detragiache, Enrica. 1998. "How Big is the Brain Drain?," IMEF Working Paper WP/981102, International Monetary Fund, Washington, DC, July 1998. Intemational Monetary Fund (IMF). (Various). Balance of Payments Statistics Yearbook. Washington, DC. Organization for Economic Cooperation and Development (OECD). 2001. Trends in International Migration: Annual Report. Paris, France. El-Qorchi, Mohammed, Maimbo, Samuel and Wilson, John. 2002. "The Hawala Informal Funds Transfer System: An Economic and Regulatory Analysis." Unpublished paper, August 2002. United States, Census Bureau. 2000. Population Census. Washington, DC. United States, Census Bureau. 2003. Special Tabulation on Educational Attainment of Foreign-Born Population in the U.S. Based on 2000 Population Census. Washington, DC. World Bank. 2000. World Development Report. 2000/01. Washington, DC. 22 Table 1. Official Worker Remittances Received in 24 Selected Labor-Exporting Countries, 1981 - 2000 (in millions of US Dollars; Real Terms) Albania Armenia Bangladesh Brazil China, PR Colombia Croatia Don. Republic Guatemala Egypt El Salvador India 1981 -- -- 549 409 - 140 - 259 129 3101 70 3260 1982 -- -- 552 204 774 101 - 271 90 3885 123 3742 1983 - -- 982 167 638 90 - 278 42 5608 142 4066 1984 -- -- 733 195 465 104 - 300 40 5812 167 3361 1985 -- - 711 164 255 149 - 342 27 4548 178 3494 1986 -- -- 800 140 289 546 -- 312 44 3484 193 3110 1987 -- -- 1002 157 222 825 - 366 51 4830 224 3568 1988 -- - 984 117 166 577 - 372 55 4857 250 2975 1989 -- -- 931 261 93 564 - 368 85 5229 250 3176 1990 -- - 908 833 145 569 - 367 125 6446 417 2743 1991 - - 860 1414 232 906 - 368 155 4536 522 3664 1992 163 -- 990 2278 282 684 --. 375 203 6629 746 3140 1993 290 - 1062 1846 327 791 - 381 216 5974 833 3686 1994 272 -- 1183 1885 406 993 349 778 270 3775 994 5944 1995 385 12 1202 2891 350 739 506 795 358 3226 1061 6139 1996 485 11 1306 1813 1624 617 585 888 364 3018 1053 8212 1997 253 8 1448 1257 4198 624 497 1033 387 3509 1138 9775 1998 422 9 1495. 900 230 451 486 1239 427 3150 1250 8837 1999 326 14 1643 1088 351 839 415 1389 426 2959 1256 10064 2000 470 14 1732 985 492 989 470 1494 498 - 2523 1549 . 7994 Annual Percent Change, - - 5.1 8.2 (-3.9) 12.0 -- 10.1 10.1 (-2.2) 15.8 5.4 1981/83 1998/00 -- Table continued on next page -- 23 Table 1 (contd) Official Worker Remittances Received in 24 Selected Labor-Exporting Countries, 1981 - 2000 (in millions of US Dollars; Real Terms) Indonesia Jamaica Mexico Morocco Nigeria Pakistan Peru Philippines SriLanka Sudan Tunisia Turkey 1981 - 89 183 1440 11 2923 231 360 469 - 507 3540 1982 -- 106 155 1214 19 3690 240 342 456 208 532 3060 1983 15 59 169 1311 12 4476 315 257 449 420 514 2165 1984 78 38 259 1278 10 3768 231 86 441 418 465 2650 1985 86 130 245 1369 14 ,3575 189 157 415 369 383 2427 1986 98 75 250 1943 7 3385 208 226 453 157 502 2271 1987 115 79 277 2127 4 2908 241 282 469 185 651 2708 1988 127 87 269 1678 3 2400 203 500 461 279 701 2288 1989 205 144 2180 1642 12 2468 290 442 440 512 600 3726 1990 194 159 2906 2339 12 2329 288 305 467 72 642 3785 1991 145 152 2701 2226 74 1724 353 368 494 50 587 3154 1992 248 171 3334 2356 61 1700 267 342 595 135 576 3266 1993 365 197 3514 2006 836 1518 332 328 666 79 470 3079 1994 461 470 3572 1878 565 1798 485 455 735 110 646 2700 1995 651 582 3673 1970 804 17i2 600 432 790 346 680 3327 1996 773 618 4103 2103 920 1247 579 552 808 214 715 3441 1997 688 609 4618 1797 1822 1620 603 1003 875 394 650 3984 1998 896 618 5260 1879 1471 1095 605 190 936 642 671 5007 1999 1014 627 5405 1772 1198 911 614 93 962 607 696 4143 2000 1053 698 5816 1912 - 951 635 111 1010 564 619 4035 Annual Percent Change, 27.9 12.7 22.7 2.0 30.7 -7.5 5.2 (-130.0) 4.5 3.9 1.5 2.4 1981/83 1998/00 Note: Real figures calculated by deflating nominal figures by US Consumer Price Index (1995 = 100). Data record only those worker remittances which enter the official banking system. Sources: International Monetary Fund, Balance of Payments Statistics Yearbook (various issues). 24 Table 2. Official Worker Remittances Received in 24 Selected Labor-Exporting Countries, Classified by Region of the World, 1981 - 2000 (in millions of US Dollars; Real Terns) Eastem Europe Latin America Middle East Sub-Saharan East Asia and Central Asia and Caribbean and North Africa South Asia Africa World 1981 360 3540 1510 5048 7201 11 17,670 1982 1116 3060 1290 5631 8440 227 19,764 1983 910 2165 1262 7433 9973 432 22,175 1984 629 2650 1334 7555 8303 428 20,899 1985 498 2427 1424 6300 8195 383 19,227 1986 613 2271 1768 5929 7748 164 18,493 1987 619 2708 2220 7608 7947 189 21,291 1988 793 2288 1930 7236 6820 282 19,349 1989 740 3736 4142 7471 7015 524 23,628 1990 644 3785 5664 9427 6447 84 26,051 1991 745 3154 6571 7349 6742 124 24,685 1992 872 3429 8058 9561 6425 196 28,541 1993 1020 3369 8110 8450 6932 915 28,796 1994 1322 3321 9447 6299 9660 675 30,724 1995 1433 4230 10,699 5876 9843 1150 33,231 1996 2949 4522 10,035 5836 11,573 1134 - 36,049 1997 5889 4742 10,269 5956 13,718 2216 42,790 1998 1316 5924 10,750 5700 12,363 2113 - 38,166 1999 1458 4898 11,644 5427 13,580 1805 38,812 2000 1656 4988 12,664 5054 11,687 564 36,613 Annual Percent Change, 3.7 3.5 13.5 (-0.7) 2.3 11.8 3.86 1981/83 1998/00 Note: Real figures calculated by deflating noominal figures by US Consumer Price Index (1995 = 100). Data record only those worker remittances which enter the official banking system. The 24 countries included in the table are listed in Table 1. Sources: International Monetary Fund, Balance of Payments Statistics Yearbook (various issues). 25 Table 3. Comparing Annual Rates of Growth of Official Worker Remittances and Gross Domestic Product (GDP) by Region of the World, 1981/83 to 1998/00 Annual Percent Change. 1981/83 to 1998/00 Region Official Worker Gross Domestic Remittances' Product (GDP)2 East Asia 3.7 6.0 Eastem Europe and 3.5 Central Asia Latin America and 13.5 0.8 Caribbean Middle East and (-0.7) 0.2 North Africa South Asia 2.3 3.3 Sub-Saharan Africa 11.8 (-0.8) Low and Lower Middle Income Developing 3.86 1.61 Countries3 Notes: 'Official worker remittances are measured in real terms by deflating nominal figures by US Consumer Price Index (1995 =.100). Official worker remittances include remittances received by the 24 study countries, and include only remittances which enter the official banking system. The 24 study countries are listed in Table 1. 2Gross domestic product is measured on a per capita basis using constant 1995 US dollars. No GDP data are available for Eastem Europe and Central Asia in 1981/83. 3For official worker remittances, the category "low and lower middle-income developing countries" includes the 24 study countries; for gross domnestic product, the category "low and lower middle-income developing countries" includes 119 countries, which are classified by the World Bank (2000) as falling into this category. Sources: Remittance data: International Monetary Fund, Balance of Pavments Statistics Yearbook (various issues). - GDP data: World Bank, World Development Indicators database (2002). 26 Table 4. Number of Immigrants (Age 25 and older) to the United States by Level of Educational Attainment, 2000 Total Educational Level Country Immigrants Primary or less Secondary Tertiary East Asia China, PR 846,780 173,545 217,185 456,050 Indonesia 53,170 1,460 12,065 39,645 Philippines 1,163,555 90,200 228,955 844,400 Eastem Europe, Central Asia Albania 25,785 3,540 12,400 9,845 Armenia 44,380 3,815 17,975 22,590 Croatia 35,455 6,725 14,350 14,380 Turkey 64,780 8,905 18,090 37,785 Latin America, Caribbean Brazil 154,250 14,005 56,010 84,235 Colombia 402,935 53,485 163,415 186,035 Dom. Republic 527,520 155,685 226,270 145,565 El Salvador 619,185 255,170 257,455 106,560 Guatemala 341,590 146,515 127,860 67,215 Jamaica 449,795 36,430. 209,710 203,655 Mexico 6,374,825 3,081,310 2,398,000 895,515 Peru 220,815 16,965 87,085 116,765 Middle East, North Africa Egypt 96,660 3,480 18,010 75,170 Morocco 29,670 1,625 8,900 19,145 Tunisia 5,555 390 1,625 3,540 South Asia Bangladesh 69,180 6,000 20,095 43,085 India 836,780 41,185 127,540 668,055 Pakistan 165,425 11,630 43,365 110,430 Sri Lanka 21,820 495 5,695 15,630 Sub-Saharan Africa Nigeria 109,160 2,630 15,910 90,620 Sudan 12,730 960 3,715 8,055 Total 12,671,800 4,116,150 4,291,680 4,263,970 Notes: Immigrants defined as foreign bom population in the United States age 25 years or over. Primary education or less corresponds to 0-8 years of schooling ; secondary to 9-12 years of schooling, and tertiary to more than 12 years of schooling Source: Special tabulation from U.S. Census Bureau based on 2000 U.S. Population Census. 27 Table 5. Migration Rates to the United States by Level of Educational Attainment, 2000 Educational Level Country Total Primary or less Secondary Tertiary East Asia China, PR 0.1 0.1 0.2 2.2 Indonesia 0.1 0.1 0.1 0.7 Philippines 3.6 0.6 2.2 11.7 Eastern Europe, Central Asia Croatia 1.2 0.4 1.2 4.7 Turkey 0.2 0.1 0.4 1.3 Latin America, Canbbean Brazil 0.2 0.1 0.5 1.1 Colombia 2.1 0.4 4.0 9.9 Dom. Republic 12.9 5.3 42.4 24.8 El Salvador 24.3 12.4 114.8 39.5 Guatemala 7.6 3.8 29.9 25.8 Jamaica 33.3 4.7 40.9 367.6 Mexico 13.3 10.8 17.2 16.5 Peru 1.8 0.3 2.5 4.2 Middle East, North Africa Egypt 0.3 0.1 0.2 2.3 Tunisia 0.1 0.1 0.2 .1.3 South Asia Bangladesh 0.1 0.1 0.3 2.3 India 0.2 0.1 0.2 2.8 Pakistan 0.3 0.1 0.5 6.4 SriLanka 0.2 0.1 0.1 5.6 Sub-Saharan Africa Sudan 0.1 0.1 0.3 3.4 Notes: No educational attainment data available on: Albania, Armenia, Morocco and Nigeria. Imniigrants defined as foreign born population in the United States, age 25 years or over. Primary education or less corresponds to 0-8 years of schooling ; secondary to 9-12 years of schooling, and tertiary to more than 12 years of schooling. Source: Author's calculations using data from special tabulation from U.S. Census Bureau based on 2000 U.S. Population Census, and the Barro-Lee data set (2000) on educational attainment. 28 Table 6. Number of Immigrants (Age 25 and older) to the OECD by Level of Educational Attainment, 2000 Total Educational Level Country Immigrants Primary or less Secondary Tertiary East Asia China, PR 722,400 148,029 185,295 389,076 Indonesia 142,540 3,910 32,347 106,283 Philippines 356,134 27,604 70,079 258,451 Eastem Europe, Central Asia Turkey 1,913,782 263,078 534,429 1,116,275 Latin America, Caribbean Brazil 176,519 16,026 64,097 96,396 Jamaica 117,199 9,483 54,647 53,069 Middle East, North Africa Morocco 560,658 30,706 168,179 361,773 Tunisia 142,828 10,027 41,782 91,019 Egypt 20,373 733 3,796 15,844 South Asia Bangladesh 44,417 3,852 12,902 27,663 India 375,283 18,471 57,199 299,613 Pakistan 85,668 6,022 22,458 57,188 Sri Lanka 64,143 1,455 16,741 45,947 Total 4,721,944 539,396 1,263,951 2,918,597 Notes: No data available on number of immigrants from: Albania, Armenia, Colombia, Croatia, Domr Republic, El Salvador, Guatemala, Mexico, Nigeria, Peru and Sudan. Immigrants defined as immigrant or foreign bom population by individual countries in the OECD, age 25 years or over. Primary education or less corresponds to 0-8 years of schooling ; secondary to 9-12 years of schooling, and tertiary to more than 12 years of schooling. For a list of OECD countries, see footnote (4). Source: Author's calculations from OECD, Trends in Intemational Migration: Annual Report (2001). 29 Table 7. Migration Rates to the OECD by Level of Educational Attainment, 2000 Educational Level Country Total Primnary or less Secondary Tertiary East Asia China, PR 0.1 0.1 0.1 1.9 Indonesia 0.1 0.1 0.2 2.0 Philippines 1.1 0.2 0.7 3.6 Eastern Europe, Central Asia Turkey 5.7 1.0 11.5 39.1 Latin America, Caribbean Brazil 0.2 0.1 0.6 1.3 Jamaica 8.7 1.2 10.6 95.8 Middle East, North Africa Morocco 4.0 0.3 6.9 43.5 Tunisia 3.1 0.3 5.2 33.3 Egypt 0.1 0.1 0.1 0.5 South Asia Bangladesh 0.1 0.1 0.2 1.5 India 0.1 0.1 0.1 1.3 Pakistan 0.1 0.1 0.3 3.3 Sri Lanka 0.7 0.1 0.4 16.5 Notes: Immigrants defined as immigrant or foreign born population by individual countries in the OECD, age 25 years or over. Primary education or less corresponds to 0-8 years of schooling; secondary to 9-12 years of schooling, and tertiary to more than 12 years of schooling. For a list of OECD countries, see footnote (4). Source: Author's calculations from OECD, Trends in International Migration: Annual Report (2001), and the Barro-Lee data set (2000) on educational attainment. Policy Research Working Paper Series Contact Title Author Date for paper WPS3043 The Incentive-Compatible Design Thorsten Beck May 2003 K. Labrie of Deposit Insurance and Bank 31001 Failure Resolution: Concepts and Country Studies WPS3044 Impregnated Nets Cannot Fully Mead over May 2003 H. Sladovich Substitute for DDT: Field Bernard Bakote'e 37698 Effectiveness of Malaria Prevention Raman Velayudhan in Solomon Islands Peter Wilikai Patricia M. Graves WPS3045 Causes and Consequences of Civil Klaus Deininger May 2003 M. Fernandez Strife: Micro-Level Evidence from 33766 Uganda WPS3046 Migration, Spillovers, and Trade Ross Levine May 2003 E. Khine Diversion: The Impact of Sergio L. Schmukler 37471 Internationalization on Stock Market Liquidity WPS3047 Comparing Mortgage Credit Risk Robert Buckley May 2003 0. Himid Policies: An Options-Based Approach Gulmira Karaguishiyeva 80225 Robert Van Order Laura Vecvagare WPS3048 Targeted Transfers in Poor Countries: Martin Ravallion May 2003 P. Sader Revisiting the Tradeoffs and Policy 33902 Options WPS3049 Hidden Impact? Ex-Post Evaluation Shaohua Chen May 2003 P. Sader of an Anti-Poverty Program Martin Ravallion 33902 WPS3050 The Changing Financial Landscape: Wafik Grais May 2003 R. Vo Opportunities and Challenges for Zeynep Kantur 33722 the Middle East and North Africa WPS3051 Regional, Multilateral, and Unilateral Glenn W. Harrison May 2003 P. Flewitt Trade Policies of MERCOSUR for Thomas F. Rutherford 32724 Growth and Poverty Reduction in David G. Tarr Brazil Angelo Gurgel WPS3052 Long-Run Impacts of China's WTO Kym Anderson May 2003 S. Lipscomb Accession on Farm-Nonfarm Income Jikun Huang 87266 Inequality and Rural Poverty Elena lanchovichina WPS3053 Economic Impacts of China's Elena lanchovichina May 2003 S. Lipscomb Accession to the World Trade William Martin 87266 Organization WPS3054 The New Comparative Economics Simeon Djankov May 2003 G. Sorensen Edward Glaeser 37088 Rafael La Porta Florencio Lopez-de-Silanes Andrei Shleifer WPS3055 The Global Growth of Mutual Funds Deepthi Fernando May 2003 A. Yaptenco Leora Klapper 31823 Vlktor Sulla Dimitri Vittas Policy Research Working Paper Series Contact Title Author Date for paper WPS3056 Avoiding the Pitfalls in Taxing Patrick Honohan May 2003 A. Yaptenco Financial Intermediation 31823 WPS3057 Teaching Adults to Read Better and Helen Abadzi May 2003 H. Abadzi Faster: Results from an Experiment 80375 in Burkina Faso WPS3058 Working for God? Evaluating Service Ritva Reinikka May 2003 H. Sladovich Delivery of Religious Not-for-Profit Jakob Svensson 37698 Health Care Providers in Uganda WPS3059 Do Capital Flows Respond to Risk Cesar Calder6n May 2003 T. Tourougui and Return? Norman Loayza 87431 Luis Serven WPS3060 World Market Integration through the Rui Albuquerque May 2003 T. Tourougui Lens of Foreign Direct Investors Norman Loayza 87431 Luis Serven WPS3061 Reciprocity in Free Trade Caroline Freund May 2003 P. Flewitt Agreements 32724 WPS3062 Labor Effects of Adult Mortality in Kathleen Beegle May 2003 E. de Castro Tanzanian Households 89121 WPS3063 Trade Liberalization, Firm Paolo Epifani May 2003 P. Flewitt Performance, and Labor Market 32724 Outcomes in the Developing World: What Can We Learn from Micro-Level Data? WPS3064 Trade Policy, Trade Volumes, and Ana M. Fernandes May 2003 A. Fernandes Plant-Level Productivity in Colombian 33983 Manufacturing Industries WPS3065 Ghost Doctors: Absenteeism in Nazmul Chaudhury May 2003 H. Sladovich Bangladeshi Health Facilities Jeffrey S. Hammer 37698 WPS3066 The Potential bemand for and Shreelata Rao Seshadri May 2003 H. Sladovich Strategic Use of an HIV-1 Vaccine P. Subramaniyam 37698 in Southern India Prabhat Jha WPS3067 The Mini-Integrated Macroeconomic Pierre-Richard Agenor May 2003 M. Gosiengfiao Model for Poverty Analysis: A Framework 33363 for Analyzing the Unemployment and Poverty Effects of Fiscal and Labor Market Reforms WPS3068 Poverty and Economic Growth in Heba El-laithy June 2003 P. Sader Egypt, 1995-2000 Michael Lokshin 33902 Arup Banerji