Returns to Investments in Human Capital, within and across Countries Preliminary and Incomplete

Transcription

1 Returns to Investments in Human Capital, within and across Countries Preliminary and Incomplete Rafael de Hoyos, John Kennan, and Rebecca Lessem World Bank, University of Wisconsin-Madison and NBER, Carnegie-Mellon University October Introduction There are vast income differentials across countries, implying that there are strong incentives for individuals from poorer countries to move to richer countries. If these wage differentials are associated with productivity, the strict immigration restrictions that exist in most countries lead to large efficiency losses. But relatively little is known about how cross-country wage differences relate to differences in human capital endowments (as opposed to cross-country differences in the productivity of human capital). The International Income Distribution Data Set (I2D2), which has information on wages and characteristics of individual workers from many less developed countries, is a rich dataset that has the information needed to study these relationships. In particular, we can use the I2D2 data to calculate the wages that similar people earn in different countries. This relates to two important questions that are distinct but closely related. The first question is the extent to which adjustments for cross-country differences in human (and physical) capital endowments can account for differences in income levels (the so-called levels accounting question). The second question, which has received much less attention, is the extent to which the incentive to emigrate from poorer countries is related to differences in individual human capital endowments: if immigration restrictions are relaxed, what is the likely skill mix in the marginal migration flow? Lagakos, Moll, Porzio, and Qian (2015) show that when life-cycle wage profiles are estimated for lessdeveloped countries, the returns to experience are much lower than the returns estimated using data from developed countries. The data set used in this study includes 17 countries, but many of these are developed countries, and no African countries are included. The analysis covers only male, full-time private sector workers who earn wages (measured as labor earnings divided by the number of hours worked). The I2D2 dataset covers many large countries that are not included in the Lagakos, Moll, Porzio, and Qian (2015) study (e.g. Nigeria, Ethiopia, Uganda and Nepal). Our project aims to provide a broader empirical analysis of cross-country differences in the determinants of labor productivity, quantifying the returns to education and experience in different less-developed countries. We use wage regressions to determine how demographic characteristics affect wages in each country. The incentives to migrate depend greatly on the returns to education and experience in each country. We use the I2D2 data to study this in detail for each country, finding that sometimes wage paths do not follow the standard Ben Porath predictions. We can then use this information to calculate the wage gains of moving to a richer country for people with different characteristics, such as education, age, and home country. We can then test if the incentive to move is greater for more skilled workers. A dynamic migration model would 1

2 predict that younger workers have the most to gain from migration, in that they have more years to earn the higher wages. However, changes in the age profiles across countries could provide additional differences in migration patterns. For example, older workers in countries with low returns to experience will have a greater incentive to move than those in countries with high returns to experience. 2 Human Capital Investments Following Ben-Porath (1967) and Heckman, Lochner, and Taber (1998), assume that there is a human capital investment technology described by the function Q = AI α H β where H is the current level of human capital, I is the fraction of time allocated to investment, and Q is the quantity of new human capital produced. Time can instead be allocated to production of a consumption good, and the marginal product of human capital is the wage, w. Then the cost of producing new capital is forgone earnings, wih, so the cost function is C (Q) = w ( ) 1 Q α H 1 β α A and marginal cost is C (Q) = w α A 1 α Q 1 α 1 H 1 β α The value of a unit of human capital, for someone aged a, is wv (a)where a) 1 e ρ(t v (a) = ρ where ρ = r + δ, and r is the interest rate, and δ is the depreciation rate, and T is the retirement age. The optimal level of investment is then determined by the point where the value is equal to the marginal cost, so that Then Also so I = = 1 α A 1 α Q 1 α 1 H 1 β α = v (a) Q 1 α 1 = αa 1 α v (a) H β α 1 I = Q 1 α A 1 α H β α (αa 1 β α v (a) H 1) 1 1 α α A 1 α H β α (αav (a) H β 1) 1 1 α α H β α 2

3 provided that I 1. Earnings are given by Y = w (1 I) H or which reduces to Y = wh w (αa 1 β α v (a) H 1) 1 1 α α A 1 α H 1 β α Y = wh w (αav (a)) 1 1 α H β α 1 α If α = β, as was assumed by Ben-Porath (1967), investment does not depend on the current level of human capital, so that earnings are given by Y = wh (αav (a)) 1 1 α The level of human capital at age a is found by solving the differential equation Ḣ = Q δh If α = β, then Ḣ = A 1 1 α (αv (a)) α 1 α δh The solution of this equation is Thus the age-earnings profile is given by H (t) = H 0 e δt + A 1 1 α α α 1 α Y (a) = wh 0 e δa + wa 1 1 α α α 1 α ˆ a 0 ˆ t 0 e δ(t s) v (s) α 1 α ds e δ(a s) v (s) α 1 α ds w (αav (a)) 1 1 α Figure 1 shows the age earning profiles for different sets of parameter values. 2.1 Age-Earnings Profiles in Transition Economies For countries governed by communist regimes before the breakup of the Soviet Union in 1991, the cross-section age-earnings profiles tend to have a distinctive shape. We will argue that this finding can be rationalized using the Ben-Porath model. Consider an individual who starts at age a 0 with human capital H 0 in a regime where wages are fixed meaning that investments in human capital yield no return. Then at age a 1 there is an unexpected regime change, and subsequently wages are determined as in the Ben-Porath model. Human capital at the transition point is given by H 1 = H 0 e δ(a1 a0) Now consider the cross-section age-earnings profile at date t 2, given that the regime change occurred at date t 1. A person aged a a 0 + t 2 t 1 at date t 2 has lived in the old regime from age a 0 to age a (t 2 t 1 ), 3

4 so at the transition this person had human capital H 1 = H 0 e δ(a (t2 t1) a0) Earnings at age a are then given by ˆ a Y (a) = wh 1 e δ(t2 t1) + wa 1 α 1 α α 1 α a (t 2 t 1) Suppose t 2 = 2007, and t 1 = 1991, and a 0 = 20, and T = 60. e δ(a s) v (s) α 1 α ds w (αav (a)) 1 1 α This result is illustrated in Figure 2. The figure plots two age-earnings profiles. One assumes a stationary economy, as in Figure 1, so that the profile in the cross-section is the life-cycle profile for each individual. The second (blue) plot describes the actual cross-section profile measured 16 years after a transition to a market economy. Individuals younger than 36 entered the labor market after the transition, and so they follow the stationary profile. But older workers had no incentive to invest in human capital when they were young, and their incentive to invest after the transition is relatively weak, especially for those closer to retirement. Thus the cross-section profile has a kink at age 36, and the earnings of older workers are far below the level that will be achieved by the younger generation. 3 Data description The International Income Distribution Data Set (I2D2) created by the World Bank is a global harmonized survey database. To create the I2D2, individual surveys were collected from a wide range of countries from many years, and then combined into a single database with consistent variable definitions and classifications. The data are based on regional surveys, sometimes using publicly available data and sometimes using internal World Bank surveys. It includes countries across the world, covering a wide distribution of income levels from across the world. Table 1 shows how many countries in our data are from each region of the world, and Table 2 shows how many countries are in each income level. 1 The I2D2 provides a lot of information on each individual surveyed. For the purposes of this research, where we are interested in the determinants of wage outcomes, we have individual level information on demographics, education, labor force participation, and outcomes. The I2D2 has data from a wide range of countries. In this paper, we are interested in studying the determinants of wages for individuals at a specific point in time. Therefore we pick the data closest to the year 2007 for each country to compare wages at this time. Appendix A shows the year of data used for each country. Wages are reported in different units for each country, so we convert to monthly wages for each observation. Wages are also reported in local currency, so we convert to 2007 US dollars using PPP exchange rates. We first calculate the average wage in each country, weighted by the sampling weights in the I2D2. Appendix B shows the average wages by gender and education for country in our sample, splitting the countries by geographic regions. 2 As expected, we see wide variation in average wages, showing the productivity 1 These are divided according to World Bank classifications. Low income is a GNI per capita of less than $1045 in Lower middle income is a GNI between $1,045 and $4125. Upper middle income is a GNI between $4,125 and $12,736. High income is a GNI greater than $12, For some countries, we have a small number of observations for certain education groups. In these cases, the sample mean is not reported. 4

5 differences across countries that generate strong incentives for migration. To make sure these numbers seem reasonable, we next compare GDP per capita for each country to the average wages in the I2D2 data. 3 This is shown in Figure 3, displaying a strong correlation between GDP per capita and average monthly income in the I2D2 data. We next look at the distribution of education in each country. We report the fraction of the sample that has no education, primary education, secondary education, and post-secondary education. This is in Tables 3-7. Next we look at female and male labor force participation rates in each country. This is reported in Tables We are also interested in how female labor force participation changes with age. This is plotted for each country in Figures Determinants of wages 4.1 Returns to education We are interested in how the returns to education vary across countries. This tells us about the returns to skills at different productivity countries, as well as informs us about the people most likely to migrate in the case of open borders. For example, in a low income country with strong returns to education, we would see fewer highly skilled migrants than from a similar income country with lower returns to education. We use calculate the average wage for a man with primary education and a man with secondary education, assuming he is years old. We take the ratio of the wage in the higher educated group relative to the wage in the lower educated group. In Figure 9 we plot the wage ratio on the y-axis and the average wage for the low educated group on the x-axis. We see a trend that countries with higher wage levels have lower returns to skills. There is wide variation in the returns to education. The bulk of the data has a wage ratio between 1 and 2.5. In some countries, people with secondary education have approximately the same wages as those with just a primary education. However, in a large mass of these countries, we also see that people with a secondary education earn 2.5 times as much as people with a primary education. When thinking about incentives to migrate to wealthier countries, we expect that educated people in countries with high returns to education will have less of an incentive to migrate than educated people in countries with low returns to education. This implies a different skill mix of immigrants across countries. 4.2 Returns to experience Next we look at the returns to experience in different countries. This again informs us about who has the greatest incentive to migrate to wealthy countries. In developing countries with strong returns to experience, we would be more likely to see younger immigrants, at least in the context of a static model. If returns to experience are low, then young and old workers earn similar wages at home, meaning that the incentive to migrate for young versus old workers only depends on the returns to experience in destination countries. To study the returns to experience, we calculate the wages of a and year old man with a secondary education in each country. We plot the ratio of the wages for the older to younger workers on the y-axis, and the younger wages on the x-axis. This is shown in Figure 10. Unlike Lagakos, Moll, Porzio, and Qian (2015), we do not see a clear trend between income levels and the returns to experience. There 3 GDP per capita data taken from the World Development Indicators. 5

6 is wide dispersion in the returns to experience. Most countries have a wage ratio between 1 and around 3. This means that in many countries, people earn similar wages around age 25 and age 40. However, in many countries, we see people earning up to 300% higher wages around age Age-earnings profiles We next aim to quantify the shape of wage profile in each country. To do this, we run wage regressions to determine the shape of the age earnings profiles. We split the sample by country, gender, and education. We denote a person as i, their country as j, the time as t, their gender as g, and their education as e. The regressions are specified as follows: w ijt = β jge 0 + β jge 1 a ijt + β jge 2 a 2 it + β jge 3 a 3 it + β jge 4 a 4 it + β jge 5 u it + ε ijt where w is a person s wage (in levels), a is age, and u is a dummy variable indicating whether or not a person is living in an urban area. 4 Because we run the regressions by gender and education status, the returns to age vary with these factors. The regressions are weighted using the survey weights in the I2D2 data. Looking at men who have a primary education, we show the initial (age 22) earnings, the age of maximum earnings, the maximum earnings, and a measure of the concavity of the wage function. To calculate this last measure, we first plot the expected earnings over the life cycle for a man with this constant set of characteristics. We then take the midpoint of the line between lowest and highest earnings, and calculate the vertical distance between this line and the age-earnings plot. These numbers are in Tables Again, we see a wide dispersion across countries, indicating that the returns to experience vary substantially. Comparing the concavity measure across income groups, we see that this is much higher for developed than developing countries. Furthermore, in some countries, the concavity measure is negative, indicating that the age profiles are not concave in all countries. As another measure of returns to experience, we look at the age profiles in each country. This allows us to graphically see the concavity of the wage earnings profile in each country, again showing that we do not always see concave age earnings profiles. These profiles are shown in Figure 11 for West Africa, Figure 12 for East and South Africa, Figure 13 for Central Africa, Figure 14 for South Asia, Figure 15 for South America, Figure 16 for Central America, Figure 17 for former Soviet countries, Figure 18 for Western Europe, Figure 19 for the Middle East and North Africa, Figure 20 for East Asia and the Pacific, and Figure 21 for North America. 4.4 Alternative specification We can also combine all of the data and run a wage regression for all countries. separately for each gender and education group. This is specified as follows: As before, we do this w ijt = β ge 0 + βge 1 a ijt + β ge 2 a2 it + β ge 3 a3 it + β ge 4 a4 it + β ge 5 y j + t j + k I(region j = k)β ge k + ε ijt In this regression, we control for age, as before. We also control for home country GDP per capita, which is denoted as y j. Since surveys are done in different years across countries, we also control for the year of the 4 The urban variable is not available in all surveys. 6

7 survey t j. We also include region fixed effects, where I is an indicator function that equals 1 if country j is in region k and 0 otherwise, and β ge k is the fixed effect for region k. We use World Bank classifications to divide countries into geographic regions. 5 We also reported this with country fixed effects. The results of these regression are in Tables Wage gains from migration We use the results from these wage regressions to understand the wage gains from migration, focusing on men living in developing countries. We look at this for different education and age levels. In particular, we look at people with a primary and a secondary education, at ages 25 and 40. For each country and group, we calculate the average wage in the home country (for ages and ages 38-42). The data show differences in the returns to education and experience across countries. This implies that immigration incentives vary not only across countries, but also across education and experience groups within countries. To calculate expected wages in the US, we run a wage regression using ACS data, controlling for age, education, gender, years in the US, region of the world a person immigrated from, and home country GDP per capita. The regression results are in Table 22. We then use the regression results to calculate the expected wages an immigrant s first year in the US, so we are setting years experience equal to 0 for each potential immigrant. We report the ratio of US wages to home wages. This is in Table We see large wage gains from migration, and these vary substantially by country, education, age level. References Ben-Porath, Y. (1967): The production of human capital and the life cycle of earnings, The Journal of Political Economy, pp Heckman, J. J., L. Lochner, and C. Taber (1998): Explaining Rising Wage Inequality: Explorations with a Dynamic General Equilibrium Model of Labor Earnings with Heterogeneous Agents, Review of Economic Dynamics, 1, Lagakos, D., B. Moll, T. Porzio, and N. Qian (2015): Life-Cycle Wage Growth Across Countries, unpublished ( 5 The regions are East Asia and the Pacific, Europe and Central Asia, Latin America and the Caribbean, Middle East and North Africa, North America, South Asia, and Sub-Saharan Africa. 7

8 6 Tables and figures Table 1: Geographic distribution of I2D2 countries Region Number of countries East Asia and the Pacific 9 Europe and Central Asia 31 Latin America and the Caribbean 20 Middle East and North African 7 North America 2 South Asia 6 Sub-Saharan Africa 27 Table 2: Income distribution of I2D2 countries Income levels Number of countries Low income countries 19 Lower middle income countries 33 Upper middle income countries 25 High income countries 25 Low income is a GNI per capita of less than $1045 in Lower middle income is a GNI between $1,045 and $4125. Upper middle income is a GNI between $4,125 and $12,736. High income is a GNI greater than $12,736. 8

9 Table 3: Education distribution, Sub-Saharan Africa Education None Primary Secondary Post-secondary West Africa Burkina Faso Cote d Ivoire Gambia Mali Mauritania Nigeria Senegal Sierra Leone Togo East/South Africa Botswana Comoros Ethiopia Kenya Lesotho Madagascar Malawi Mauritius Mozambique Swaziland Uganda Central Africa Burundi Cameroon Chad Congo, Dem. Rep Gabon Rwanda

10 Table 4: Education distribution, Asia Education None Primary Secondary Post-secondary South Asia Afghanistan Bangladesh India Nepal Sri Lanka East Asia and the Pacific China Indonesia Lao PDR Mongolia Papua New Guinea Philippines Solomon Islands Thailand Timor-Leste Table 5: Education distribution, Americas Education None Primary Secondary Post-secondary South America Brazil Colombia Ecuador Guyana Paraguay Peru Uruguay Central America Bahamas Belize Costa Rica Dominican Republic Haiti Honduras Jamaica Mexico Nicaragua Panama North America Canada United States

11 Table 6: Education distribution, Europe Education None Primary Secondary Post-secondary Former Soviet countries Armenia Croatia Estonia Georgia Kyrgyz Republic Latvia Macedonia, FYR Moldova Russian Federation Slovak Republic Slovenia Ukraine Western Europe Austria Belgium Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Table 7: Education distribution, Middle East and North Africa Education None Primary Secondary Post-secondary Egypt, Arab. Rep Jordan Morocco Turkey West Bank and Gaza Yemen, Rep

12 Table 8: Labor force participation rates, Sub-Saharan Africa Men Women West Africa Burkina Faso Cote d Ivoire Gambia Mali Mauritania Nigeria Senegal Sierra Leone Togo East/South Africa Botswana Comoros Ethiopia Kenya Lesotho Madagascar Malawi Mauritius Mozambique Swaziland Uganda Central Africa Burundi Cameroon Chad Congo, Dem. Rep Gabon Rwanda Table 9: Labor force participation rates, Asia Men Women South Asia Afghanistan Bangladesh India Nepal Sri Lanka East Asia and the Pacific China Indonesia Lao PDR Mongolia Papua New Guinea Philippines Solomon Islands Thailand Timor-Leste

13 Table 10: Labor force participation rates, Americas Men Women South America Brazil Colombia Ecuador Guyana Paraguay Peru Uruguay Central America Belize Costa Rica Dominican Republic Haiti Honduras Jamaica Mexico Nicaragua Panama North America Canada United States

14 Table 11: Labor force participation rates, Europe Men Women Former Soviet countries Armenia Croatia Estonia Georgia Kyrgyz Republic Latvia Macedonia, FYR Moldova Russian Federation Slovak Republic Slovenia Ukraine Western Europe Austria Belgium Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Table 12: Labor force participation rates, Middle East and North Africa Men Women Egypt, Arab. Rep Jordan Morocco Turkey West Bank and Gaza Yemen, Rep

15 Table 13: Age profile, Sub-Saharan Africa Earnings Age 22 Max Age at max Concavity West Africa Burkina Faso Cote d Ivoire Gambia Mali Mauritania Nigeria Senegal Sierra Leone Togo East/South Africa Botswana Comoros Ethiopia Kenya Lesotho Madagascar Malawi Mauritius Mozambique Swaziland Uganda Central Africa Burundi Cameroon Chad Congo, Dem. Rep Gabon Rwanda Based on monthly wage regression results, for urban men with a secondary education. The concavity measure is the vertical distance from the age-earnings plot to the midpoint of the line between minimum and maximum earnings. 15

16 Table 14: Age profile, Asia Earnings Age 22 Max Age at max Concavity South Asia Afghanistan Bangladesh India Nepal Sri Lanka East Asia and the Pacific China Indonesia Lao PDR Mongolia Papua New Guinea Philippines Solomon Islands Thailand Timor-Leste Based on monthly wage regression results, for urban men with a secondary education. The concavity measure is the vertical distance from the age-earnings plot to the midpoint of the line between minimum and maximum earnings. Table 15: Age profile, Americas Earnings Age 22 Max Age at max Concavity South America Brazil Colombia Ecuador Guyana Paraguay Peru Uruguay Central America Bahamas Belize Costa Rica Dominican Republic Haiti Honduras Jamaica Mexico Nicaragua Panama North America Canada United States Based on monthly wage regression results, for urban men with a secondary education. The concavity measure is the vertical distance from the age-earnings plot to the midpoint of the line between minimum and maximum earnings. 16

17 Table 16: Age profile, Europe Earnings Age 22 Max Age at max Concavity Former Soviet countries Armenia Croatia Estonia Georgia Kyrgyz Republic Latvia Macedonia, FYR Moldova Russian Federation Slovak Republic Slovenia Ukraine Western Europe Austria Belgium Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Based on monthly wage regression results, for urban men with a secondary education. The concavity measure is the vertical distance from the age-earnings plot to the midpoint of the line between minimum and maximum earnings. Table 17: Age profile, Middle East and North Africa Earnings Age 22 Max Age at max Concavity Egypt, Arab. Rep Jordan Morocco Turkey West Bank and Gaza Yemen, Rep Based on monthly wage regression results, for urban men with a secondary education. The concavity measure is the vertical distance from the age-earnings plot to the midpoint of the line between minimum and maximum earnings. 17

18 Table 18: Pooled regression, no education Women Men (1) (2) (3) (4) Year of survey (8.01) (24.68) Age (3425.7) (3566.3) (6323.1) (6308.0) Age (5586.9) (5763.0) (9793.3) (9759.6) Age (3751.5) (3858.4) (6455.1) (6426.6) Age (892.3) (917.6) (1537.9) (1529.6) Home country GDP ( ) ( ) East Asia and the Pacific (46.49) (57.09) Europe and Central Asia (22.66) (28.34) Latin America and the Caribbean (18.26) (31.92) Middle East and North Africa (250.0) (207.3) South Asia (24.7) (33.63) Constant ( ) (749.1) ( ) (1447.1) Country fixed effects No Yes No Yes Observations Adjusted Rˆ Standard errors in parentheses. Age divided by

19 Table 19: Pooled regression, primary education Women Men (1) (2) (3) (4) Year of survey (1.801) (9.785) Age (940.5) (893.3) (1845.7) (2009.8) Age (1545.3) (1473.0) (2966.5) (3202.1) Age (1059.3) (1013.5) (1982.3) (2135.1) Age (258.7) (248.3) (472.5) (508.3) Home country GDP ( ) ( ) East Asia and the Pacific (41.26) (45.97) Europe and Central Asia (24.48) (26.76) Latin America and the Caribbean (33.15) (31.6) Middle East and North Africa (45.66) (184.4) South Asia (37.04) (34.41) Constant (3567.7) (188.0) ( ) (436.1) Country fixed effects No Yes No Yes Observations Adjusted Rˆ Standard errors in parentheses. Age divided by

20 Table 20: Pooled regression, secondary education Women Men (1) (2) (3) (4) Year of survey (56.89) (2.024) Age (1618.1) (1708.9) (943.8) (928.7) Age (1870.5) (3996.3) (1586.9) (1555.0) Age (1531.1) (3574.5) (1127.7) (1102.2) Age (528.2) (1059.3) (287.5) (280.5) Home country GDP ( ) ( ) East Asia and the Pacific (91.5) (18.19) Europe and Central Asia (80.52) (17.54) Latin America and the Caribbean (417.6) (22.33) Middle East and North Africa (53.1) (29.98) South Asia (159.5) (21.63) Constant ( ) (262.8) (4065.7) (197.5) Country fixed effects No Yes No Yes Observations Adjusted Rˆ Standard errors in parentheses. Age divided by

21 Table 21: Pooled regression, post-secondary education Women Men (1) (2) (3) (4) Year of survey (9.21) (12.95) Age (3374.8) (3467.2) (4706.2) (4612.0) Age (5479.0) (5648.4) (7410.6) (7211.4) Age (3809.0) (3935.8) (4989.2) (4830.8) Age (955.0) (987.9) (1215.4) (1173.0) Home country GDP ( ) ( ) East Asia and the Pacific (51.37) (68.19) Europe and Central Asia ) (38.98) Latin America and the Caribbean (85.18) (59.21) Middle East and North Africa (100.8) (146.4) South Asia (87.72) (44.06) Constant ( ) (769.9) ( ) (1065.5) Country fixed effects No Yes No Yes Observations Adjusted Rˆ Standard errors in parentheses. Age divided by

22 Table 22: ACS wage regression (1) (2) Age (2098.2) (2098.5) Age (3416.5) (3416.8) Age (2358.4) (2358.6) Age (585.4) (585.4) Male (10.69) (10.69) Primary education (28.66) (58.11) Secondary education (28.87) (57.15) Post-secondary education (29.91) (57.51) Home country GDP per capita ( ) ( ) Primary education*gdp ( ) Secondary education*gdp ( ) Post-secondary education*gdp ( ) Years in US (0.574) (0.574) Constant (460.3) (463.5) Observations R Home region fixed effects Yes Yes Standard errors in parentheses. Using only data on immigrants from the ACS. Age divided by

23 Table 23: Ratio, US wages to home wages, Sub-Saharan Africa Primary education Secondary education Age 25 Age 40 Age 25 Age 40 West Africa Burkina Faso Cote d Ivoire Gambia Mali Mauritania Nigeria Senegal Sierra Leone Togo East/South Africa Botswana Comoros Ethiopia Kenya Lesotho Madagascar Malawi Mauritius Mozambique Swaziland Uganda Central Africa Burundi Cameroon Chad Congo, Dem. Rep Gabon Rwanda Wages in the US are calculated using wage regressions from ACS data on immigrants. Home wages calculated using means from the I2D2 data. 23

24 Table 24: Ratio, US wages to home wages, Asia Primary education Secondary education Age 25 Age 40 Age 25 Age 40 South Asia Afghanistan Bangladesh India Nepal Sri Lanka East Asia and the Pacific China Indonesia Lao PDR Mongolia Papua New Guinea Philippines Solomon Islands Thailand Timor-Leste Wages in the US are calculated using wage regressions from ACS data on immigrants. Home wages calculated using means from the I2D2 data. Table 25: Ratio, US wages to home wages, Americas Primary education Secondary education Age 25 Age 40 Age 25 Age 40 South America Brazil Colombia Ecuador Guyana Paraguay Peru Uruguay Central America Bahamas Belize Costa Rica Dominican Republic Haiti Honduras Jamaica Mexico Nicaragua Panama North America Canada United States Wages in the US are calculated using wage regressions from ACS data on immigrants. Home wages calculated using means from the I2D2 data. 24

25 Table 26: Ratio, US wages to home wages, Europe Primary education Secondary education Age 25 Age 40 Age 25 Age 40 Former Soviet countries Armenia Croatia Estonia Georgia Kyrgyz Republic Latvia Macedonia, FYR Moldova Russian Federation Slovak Republic Slovenia Ukraine Western Europe Austria Belgium Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Wages in the US are calculated using wage regressions from ACS data on immigrants. Home wages calculated using means from I2D2 data. Table 27: Ratio, US wages to home wages, Middle East and North Africa Primary education Secondary education Age 25 Age 40 Age 25 Age 40 Egypt, Arab Rep Jordan Morocco Turkey West Bank and Gaza Yemen, Rep Wages in the US are calculated using wage regressions from ACS data on immigrants. Home wages calculated using means from the I2D2 data. 25

26 Figure 1: Examples of Ben-Porath Age-Earnings Profiles Figure 2: Age-Earnings Profiles in a Transition Economy 26

27 Figure 3: Average wages in I2D2 and GDP per capita I2D2 monthly income JOR BHS SVN PRT WBG HRV BLZ JAMCRITUR SWZ BRA PAN EST PNG GMB LSO MUS LAO NGA MNG THA URY RUS SVK ECU HND PRYCOL DOM MKDMEX GAB LVA COM BDI ETH TCD NIC YEM MAR SEN KEN BWA MWI MOZ TGO SLE AFG UGA SLB NPL BFA CMR CIV MRT PHL GUY PER EGY TMP HTI MDA ZAR RWA MLI MDG BGDCHN INDIDN LKA UKR KGZ GEO ARM DEU NLD CAN FIN ITABEL FRA AUT ESP GRC USA IRL GDP per capita, WDI LUX GDP per capita taken from the World Development indicators. Average wages is the average monthly income in each country in the I2D2 data. Blue countries are classified as low income, gold countries are classified as lower middle income, and pink countries are classified as upper middle income. 27

28 Figure 4: Female labor force participation rates, Sub-Saharan Africa (a) West Africa (b) East/South Africa (c) Central Africa Figure 5: Female labor force participation rates, Asia (a) South Asia (b) East Asia and the Pacific 28

29 Figure 6: Female labor force participation rates, Americas (a) South America (b) Central America (c) North America Figure 7: Female labor force participation rates, Middle East and North Africa 29

30 Figure 8: Female labor force participation rates, Europe (a) Western Europe (b) Former Soviet countries Figure 9: Returns to education Ratio of male wages, Secondary to primary education BDI BFA RWA PNG KEN MWI UGAGAB CMR THA IND HND BRA CHN ETH MOZ MLI CIV MNG SWZ HTI BGD PRY LSO LUX TGO PHLNGA LKA NPL COL MEX URY BWAMRT MUS BLZ PERPAN MAR PRT USA ECU TUR ZAR COM JAM SLE NIC CRI HRV FIN IDN TCD DOM GRC KGZ LAO SEN AFG SLB GUY YEMKD FRA BEL ITA IRL WBG SVN ESP JOR TMP UKREGY BHS GMB MDG CAN NLD Primary education wages X-axis is average monthly wages in each country for a 40 year old male with a primary education. Y-axis is the ratio of wages for this person with a secondary education relative to a primary education. Blue countries are low-income, gold countries are lower middle income, and pink countries are upper middle income countries. 30

31 Figure 10: Returns to experience Ratio of wages, age 40 to age ARM MLI BFA BWA BDI RWA SWZ GAB MAR THA DEU CIV MDG MWI UGA CHN ETH MOZ CMR IND BRA SLB PRT SEN LUX CAN MRT TCD KENGA MUS HND USA NLD PRY URY GRCFRA SLE BGD TUR BLZ ITA GEO FIN IRL NPL LKA GMB DOM WBG ZAR TMP AFG PHLCOL NICMEX ECU JAM SVKHRVSVN ESP YEM BEL IDN PER HTI PAN CRI LSO AUT EST BHS EGY PNG TGO MNG MKD JOR KGZ GUY LVARUS UKR MDA LAOCOM Age 25 wages X-axis is average monthly wages in each country for a 25 year old male with a secondary education. Y-axis is the ratio of wages for this person when he is 40 relative to age 25. Blue countries are low-income, gold countries are lower middle income, and pink countries are upper middle income countries. 31

32 Figure 11: Age profile, West Africa (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Burkina Faso (blue), Cote d Ivoire (orange), Gambia (red), Mali (purple), Mauritania (green), Nigeria (brown), Senegal (navy), Sierra Leone (magenta), Togo (yellow). 32

33 Figure 12: Age profile, East/South Africa (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Botswana (yellow), Comoros (purple), Ethiopia (red), Kenya (blue), Lesotho (mint), Madagascar (brown), Malawi (gray), Mauritius (green), Mozambique (navy), Swaziland (gold), Uganda (orange). 33

34 Figure 13: Age profile, Central Africa (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Burundi (yellow), Cameroon (purple), Chad (brown), Congo (navy), Gabon (red), Rwanda (blue). 34

35 Figure 14: Age profile, South Asia (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Afghanistan (yellow), Bangladesh (purple), India (red), Nepal (mint), Sri Lanka (blue). 35

36 Figure 15: Age profile, South America (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Brazil (yellow), Colombia (purple), Ecuador (red), Guyana (blue), Paraguay (brown), Peru (mint), Uruguay (navy). 36

37 Figure 16: Age profile, Central America (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Bahamas (yellow), Belize (purple), Costa Rica (red), Dominican Republic (blue), Honduras (mint), Haiti (brown), Jamaica (navy), Mexico (green), Nicaragua (gray), Panama (gold). 37

38 Figure 17: Age profile, former Soviet countries (a) Primary education (b) Secondary education (c) Post-secondary education Legend: Armenia (yellow), Croatia (mint), Estonia (red), Georgia (blue), Kyrgyz Republic (brown), Latvia (navy), Macedonia (gray), Moldova (green), Russia (magenta), Slovak Republic (pink), Slovenia (teal), Ukraine (maroon). 38

39 Figure 18: Age profile, Western Europe (a) Primary education (b) Secondary education (c) Post-secondary education Legend: Austria (yellow), Belgium (purple), Finland (mint), France (brown), Germany (red), Greece (green), Great Britain (navy), Ireland (gray), Italy (gold), Luxembourg (magenta), Netherlands (orange), Portugal (teal), Spain (blue). 39

40 Figure 19: Age profile, Middle East and North Africa (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Egypt (yellow), Jordan (purple), Morocco (red), Turkey (brown), West Bank and Gaza (navy), Yemen (green). 40

41 Figure 20: Age profile, East Asia and the Pacific (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: China (yellow), Indonesia (purple), Lao PDR (red), Mongolia (mint), Papua New Guinea (navy), Philippines (brown), Solomon Islands (green), Thailand (gray), Timor-Leste (gold). 41

42 Figure 21: Age profile, North America (a) No education (b) Primary education (c) Secondary education (d) Post-secondary education Legend: Canada (yellow), United States (purple). Appendix A: Years used for each country Table 28: Year of data, Sub-Saharan Africa West Africa East/South Africa Central Africa Country Year Country Year Country Year Burkina Faso 2009 Botswana 2009 Burundi 1998 Cote d Ivoire 2008 Comoros 2004 Cameroon 2007 Gambia 2010 Ethiopia 2006 Chad 2003 Mali 1994 Kenya 2005 Congo, Dem. Rep Mauritania 2000 Lesotho 2002 Gabon 2005 Nigeria 2012 Madagascar 2010 Rwanda 2005 Senegal 2011 Malawi 2010 Sierra Leone 2011 Mauritius 2007 Togo 2006 Mozambique 2008 Swaziland 2000 Uganda

43 Table 29: Year of data, Asia South Asia East Asia and the Pacific Country Year Country Year Afghanistan 2007 China 2002 Bangladesh 2005 Indonesia 2007 India 2007 Lao PDR 2008 Nepal 2008 Mongolia 2007 Sri Lanka 2008 Papua New Guinea 2009 Philippines 2007 Solomon Islands 2005 Thailand 2006 Timor-Leste 2007 Table 30: Year of data, Americas South America Central America North America Country Year Country Year Country Year Brazil 2007 Bahamas 2001 Canada 2001 Colombia 2007 Belize 1999 United States 2005 Ecuador 2007 Costa Rica 2007 Guyana 1999 Dominican Republic 2007 Paraguay 2007 Haiti 2012 Peru 2007 Honduras 2007 Uruguay 2007 Jamaica 2002 Mexico 2008 Nicaragua 2009 Panama 2007 Table 31: Year of data, Europe Former Soviet countries Western Europe Country Year Country Year Armenia 1999 Austria 2007 Croatia 2004 Belgium 2007 Estonia 2007 Finland 2007 Georgia 1998 France 2007 Kyrgyz Republic 1997 Germany 2007 Latvia 2007 Greece 2007 Macedonia, FYR 2005 Ireland 2007 Moldova 2002 Italy 2007 Russian Federation 2007 Luxembourg 2007 Slovak Republic 2007 Netherlands 2007 Slovenia 2007 Portugal 2007 Ukraine 2005 Spain

44 Table 32: Year of data, Middle East and North Africa Country Year Egypt, Arab Rep Jordan 2010 Morocco 1998 Turkey 2007 West Bank and Gaza 2007 Yemen, Rep Table 33: Year of data, Countries not being used Country Year Albania 2005 Azerbaijan 2002 Belarus 2005 Bulgaria 2007 Cambodia 2007 Czech Republic 2007 Denmark 2007 El Salvador 2007 Ghana 2005 Guatemala 2006 Hungary 2007 Iceland 2007 Iraq 2006 Lithuania 2007 Malta 2009 Niger 2007 Norway 2007 Pakistan 2007 Poland 2007 Romania 2007 Suriname 1999 Sweden 2007 São Tomé and Principe 2010 Tajikistan 2003 Tanzania 2006 Tunisia 2001 United Kingdom 2007 Vietnam 2010 Zambia

45 Appendix B: Average wages Table 34: Mean monthly wages, Sub-Saharan Africa Education level None Primary Secondary Post-secondary Country Men Women Men Women Men Women Men Women Western Africa Burkina Faso Cote d Ivoire Gambia Mali Mauritania Nigeria Senegal Sierra Leone Togo East/South Asia Botswana Comoros Ethiopia Kenya Lesotho Madagascar Malawi Mauritius Mozambique Swaziland Uganda Central Africa Burundi Cameroon Chad Congo, Dem. Rep Gabon Rwanda

46 Table 35: Mean monthly wages, Asia Education level None Primary Secondary Post-secondary Country Men Women Men Women Men Women Men Women South Asia Afghanistan Bangladesh India Nepal Sri Lanka East Asia and the Pacific China Indonesia Lao PDR Mongolia Papua New Guinea Philippines Solomon Islands Thailand Timor-Leste Table 36: Mean monthly wages, Americas Education level None Primary Secondary Post-secondary Country Men Women Men Women Men Women Men Women South America Brazil Colombia Ecuador Guyana Paraguay Peru Uruguay Central America Bahamas Belize Costa Rica Dominican Republic Haiti Honduras Jamaica Mexico Nicaragua Panama North America Canada United States