The geography of intergenerational mobility in Latin America and the Caribbean

Muñoz S., Ercio Working Paper The geography of intergenerational mobility in Latin America and the Caribbean IDB Working Paper Series, No. IDB-WP-01620 Provided in Cooperation with: Inter-American Development Bank (IDB), Washington, DC Suggested Citation: Muñoz S., Ercio (2024) : The geography of intergenerational mobility in Latin America and the Caribbean, IDB Working Paper Series, No. IDB-WP-01620, Inter-American Development Bank (IDB), Washington, DC, https://doi.org/10.18235/0013050 This Version is available at: https://hdl.handle.net/10419/300515 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/3.0/igo/ The Geography of Intergenerational Mobility in Latin America and the Caribbean Ercio Muñoz WORKING PAPER No IDB-WP-01620 InterA merican Development Bank Gender and Diversity Division July 2024 The Geography of Intergenerational Mobility in Latin America and the Caribbean Ercio Muñoz InterA merican Development Bank Gender and Diversity Division J uly 2024 Cataloging-in-Publication data provided by the Inter-American Development Bank Felipe Herrera Library Muñoz, Ercio. The geography of intergenerational mobility in Latin America and the Caribbean / Ercio Muñoz. p. cm. — (IDB Working Paper Series ; 1620) Includes bibliographical references. 1. Education-Economic aspects-Latin America. 2. Education-Social aspectsLatin America. 3. Education-Economic aspects-Caribbean Area. 4. EducationSocial aspects-Caribbean Area. 5. Economic development-Effect of education on-Latin America. 6. Economic development-Effect of education onCaribbean Area. I. Inter-American Development Bank. Gender and Diversity Division. II. Title. III. Series. IDB-WP-1620 JEL codes: D63, I24, J62. Keywords: Socioeconomic mobility, Education, Latin America and the Caribbean. http://www.iadb.org Copyright © 2024 Inter-American Development Bank ("IDB"). This work is subject to a Creative Commons license CC BY 3.0 IGO (https://creativecommons.org/licenses/by/3.0/igo/legalcode). The terms and conditions indicated in the URL link must be met and the respective recognition must be granted to the IDB. Further to section 8 of the above license, any mediation relating to disputes arising under such license shall be conducted in accordance with the WIPO Mediation Rules. Any dispute related to the use of the works of the IDB that cannot be settled amicably shall be submitted to arbitration pursuant to the United Nations Commission on International Trade Law (UNCITRAL) rules. The use of the IDB's name for any purpose other than for attribution, and the use of IDB's logo shall be subject to a separate written license agreement between the IDB and the user and is not authorized as part of this license. Note that the URL link includes terms and conditions that are an integral part of this license. The opinions expressed in this work are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent. The Geography of Intergenerational Mobility in Latin America and the Caribbean∗ Ercio Mu˜noz† Abstract This paper estimates intergenerational mobility in education using data from 91 censuses in 24 countries in Latin America and the Caribbean spanning over half a century. It measures upward mobility as the likelihood that individuals will complete one educational stage more than their parents (primary education for those whose parents did not finish primary school, or secondary education for those whose parents did not complete secondary school). It measures downward mobility as the likelihood that an individual will fail to complete a level of education (primary or secondary) that their parents did attain. In addition, the paper explores the geography of educational intergenerational mobility using nearly 400 “provinces” and more than 6,000 “districts,” finding s ubstantial c ross-country a nd w ithin-country h eterogeneity. I t d ocuments a decline in the mobility gap between urban and rural populations and small differences by gender. It also finds t hat u pward m obility i s i ncreasing a nd d ownward mobility is decreasing over time. Within countries, the level of mobility correlates closely to the share of the preceding generation that completed primary school. In addition, upward mobility is negatively correlated with distance to the capital and the share of the workforce employed in agriculture, but is positively correlated with the share of the workforce employed in industry. The opposite is true of downward mobility. JEL-Codes: D63, I24, J62. Keywords:Socioeconomic mobility, Education, Latin America and the Caribbean. ∗I thank Bennett Callaghan, Miles Corak, Christos Giannikos, Marco Ranaldi, Nuria Rodriguez-Planas, Roy van der Weide, and Wim Vijverberg, and several seminar/conference participants for their helpful comments, as well as Pablo Vargas and Joaquin Prieto for their outstanding research assistance. I am deeply grateful for the financial support of the Center for Latin American, Caribbean and Latino Studies at CUNY and the Mario Capelloni dissertation fellowship from CUNY. The views expressed in this paper are those of the author and should not be attributed to the Inter-American Development Bank. †Email: [email protected]rg. Inter-American Development Bank, 1300 New York Avenue NW, Washington, D.C. 20577, United States. 1 I Introduction Measuring intergenerational mobility (IGM) at a geographically disaggregated level can shed light on localized patterns and drivers of IGM, as argued in Narayan et al. (2018) and shown in the seminal work of Chetty, Hendren, Kline, and Saez (2014) for the United States. This type of analysis has not yet been conducted in the countries of Latin America and the Caribbean (LAC) due to the inadequacy of most survey data for this purpose. This work aims to fill that gap in the literature by generating estimates of IGM in education at smaller geographical levels. In this paper, I estimate intergenerational mobility in education for countries in LAC at the national and subnational level using data from 91 censuses. The analysis covers 24 countries and spans over half a century (from 1960 to 2012). I follow the empirical approach of Alesina, Hohmann, Michalopoulos, and Papaioannou (2021), which relies on samples of coresidents (i.e., children living with their parents or older relatives) and allows me to create indicators that are highly comparable to those recently estimated for 27 countries in Africa, a region that, like LAC, has high income inequality levels (see Alvaredo & Gasparini,2015), despite its lower levels of income and higher poverty rates.1This approach focuses on the most disadvantaged population in terms of educational attainment (a large share of parents in the sample attained less than primary education) and minimizes the potential impact of coresidence bias, since low levels of educational completion can be measured with a high degree of confidence between ages 14 and 18 (see Munoz & Siravegna,2023). Estimates of upward mobility, measured as a person’s likelihood of finishing primary education when their parents failed to finish primary school, show wide cross-country heterogeneity. The same is true of estimates of downward mobility, measured as the likelihood of a person failing to complete primary school when their parents did attain that level of education. In LAC, the distance between the most and least upwardly mobile countries is similar to what has been recently documented in Africa, although the least mobile countries in Africa are less mobile than the least mobile country in LAC. I find only small differences in mobility by gender, but I document a decline in the mobility gap between urban and rural populations. Upward mobility is increasing over time, while downward mobility is decreasing. At the sub-national level, mobility is heterogeneous across districts/provinces. Some countries show lower levels of mobility in the northern regions (e.g., Brazil), whereas the opposite is true for Mexico. However, there is much less variability in countries with fewer 1An important stylized fact in the literature on IGM is its negative association with income inequality (see Corak,2013). Hence, one could expect IGM levels in LAC to be similar to those in Africa. 2 regions and smaller populations. Mobility at the sub-national level is highly positively correlated to the share of the preceding generation that completed primary school, which suggests that the factors affecting educational attainment are persistent. In addition, geographical correlates do not appear to be highly associated with mobility, with the exception of distance to the capital. Similarly, some proxies of economic development, like share of the workforce employed in industry and agriculture at the beginning of the sample period, seem to be associated with mobility at the district level. I.1 Related literature There is a growing body of literature on the association between parents’ and adult children’s socioeconomic outcomes, which can be referred to as intergenerational mobility. At the theoretical level, the workhorse model for thinking about the mechanisms of transmission of advantage between generations was developed by Becker and Tomes (1979,1986). The empirical literature has measured IGM using a variety of indicators (see Deutscher & Mazumder,2023, for a synthesis of approaches) that capture different aspects of the phenomena. Similarly, IGM has been studied using a range of outcomes, including income and education. Research on IGM in income has mainly focused on developed countries (see Black & Devereux,2011, for a survey). Early estimates mostly centered on the United States (e.g., Solon,1992) and, later, on how the United States compares to other developed countries (e.g., Bjorklund & Jantti,1997). Given the challenges associated with finding suitable data sets, estimates of income mobility for developing countries are scarce and have only recently started to increase (see Emran & Shilpi,2021, for recent surveys focused on developing countries). Recent papers on IGM in income have focused on documenting patterns at the sub-national level. In a seminal work, Chetty et al. (2014) shows substantial variation in income mobility across commuting zones in the United States. Several others show geographical patterns for other countries (for example, Corak,2020;Cort´es Orihuela et al., 2023;G.C. Britto, Fonseca, Pinotti, Sampaio, & Warwar,2022;G¨uell, Pellizzari, Pica, & Rodr´ıguez Mora,2018). Research on IGM in education has been more global, in part because of better data availability (see Torche,2021, for recent surveys focused on developing countries). Hertz et al. (2007) documented mobility for 52 countries, including seven in Latin America, and concluded that this region has the highest level of persistence, while Nordic countries have the lowest. More recently, Van der Weide, Lakner, Mahler, Narayan, and Gupta (2024) created a new database with estimates for 153 countries (18 from LAC) using survey data. Apart 3 from these two papers with global coverage that include LAC, other early contributions were Dahan and Gaviria (2001), which estimated sibling correlations in schooling for 16 countries, and Neidh¨ofer, Serrano, and Gasparini (2018), which documented IGM for 18 countries. Moreover, researchers have used estimates from LAC to study the relationship between inequality and IGM (Neidh¨ofer,2019), as well as the impact of IGM on economic development (Neidh¨ofer, Gasparini, Ciaschi, Gasparini, & Serrano,2024). A recent wave of papers have analyzed IGM in education at the sub-national level, mostly with census data. For instance, Card, Domnisoru, and Taylor (2022), Derenoncourt (2022), Hilger (2016), and Feigenbaum (2018), among others, use coresident samples from census data to study different aspects of upward mobility in the U.S.; Asher, Novosad, and Rafkin (2023) study mobility among different marginalized groups and analyze geographic differences in India; Van der Weide, Ferreira de Souza, and Barbosa (2020) study mobility at the sub-national level in Brazil; Neidh¨ofer et al. (2024) compute educational mobility for 52 subnational divisions in 10 countries of Latin America to estimate its impact on regional economic indicators; and Alesina et al. (2021) and Alesina, Hohmann, Michalopoulos, and Papaioannou (2023) study patterns of IGM in Africa using census data, applying methods similar to those used in this paper. This paper contributes to this literature in several ways. First, it complements previous studies by estimating a new indicator that is focused on the most disadvantaged segment of the population in terms of educational attainment. Second, the paper improves the country coverage with at least six additional countries that account for 62% of the population in the Caribbean region, for which, to the best of my knowledge, no estimates of intergenerational mobility were available so far. Third, it exploits census data, which contains large samples and provides high cross-country comparability, to study IGM in LAC. In addition, it uses the same approach as recent estimates for Africa to improve the current cross-regional comparability.2Fourth, the paper provides novel evidence on changes in mobility gaps by gender and urban/rural population. Fifth, this study is the first to map IGM in education at a highly disaggregated regional level for almost the entire population of LAC. Lastly, the paper explores how IGM is associated with a set of correlates at the sub-national level. The paper is organized as follows. Section II describes the data and methodology. Section III reports the main descriptive results at the country level. Section IV explores the geography of mobility and looks at correlates of IGM. Section V concludes with final remarks. 2Hertz et al. (2007) allowed this type of comparison but included only seven countries from Latin America and four from Africa. More recently, Narayan et al. (2018) and Van der Weide et al. (2024) allow regional comparisons but pool together estimates from different types of surveys (e.g., opinion surveys and household surveys) and mix these estimates with retrospective information and coresident samples, which may be problematic (see Munoz & Siravegna,2023). 4 II Data and Methodology In this paper, I use census data obtained from IPUMS International (Integrated Public Use Microdata Series, IPUMS,2019), which is hosted at the University of Minnesota Population Center and reports harmonized representative samples (typically 10%) of full census microdata sets for a large number of countries. In particular, I use 91 samples of population and housing censuses from 24 countries. The censuses are conducted to compute total population and contain an educational attainment question in their questionnaire.3The key advantage of this data set is that it contains detailed information about the location of a large share of the entire population, which allows me to analyze mobility at a very disaggregated geographical level.4Moreover, the information about educational attainment is, for the most part, collected directly from each household member, in contrast to previous research that used retrospective questions (i.e., individuals being asked about the educational attainment of their parents), which may introduce recall bias. However, the main disadvantage of this data set is that it does not link all individuals to their parents because to be linked, both individuals and parents have to be part of the same household. Below, I explain how I addressed this issue and share recent evidence showing that the coresidence bias is likely very small for the indicators used in this paper. II.1 Countries and smaller administrative units The 24 countries analyzed are: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Cuba, Dominican Republic, Ecuador, El Salvador, Guatemala, Haiti, Honduras, Jamaica, Trinidad and Tobago, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay, Saint Lucia, Suriname, and Venezuela (see Table A1 in the appendix for details about the fraction of data available by census). I drew 91 samples from these 24 countries at various points from 1960 to 2012. Importantly, the study includes six countries from the Caribbean region for which no estimates of mobility were available. Regarding geography, IPUMS reports residence at the time of the interview for at most two levels of administrative units in which the households were counted. These variables contain the geographies for every country harmonized spatio-temporally to provide spatially consistent boundaries across samples in each country. This allows me to assign individuals to 3I do not use Chile 1960, Colombia 1964, Costa Rica 1963, Dominican Republic 1960 and 1970, Ecuador 1962, Honduras 1961, or Mexico 1960 because the individuals are not organized into households in those censuses. I also omitted the 1995, 2005, and 2015 interdecennial census counts for Mexico. 4Previous literature on Latin America has used household survey data or public opinion surveys (see for example, Hertz et al.,2007;Narayan et al.,2018;Neidh¨ofer et al.,2018), given that long panel data or administrative/registry data that allow the researcher to link generations are rare. 5 Figure 5: Transition matrix for selected countries 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (a) Jamaica 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (b) Guatemala Notes: The sample is made up of individuals over age 25 that coreside with at least one individual from the preceding generation. The figures display the transition matrix between the educational attainment of individuals in the sample and their parents. The horizontal axis is divided according to the share of parents with each level of educational attainment. The height of each rectangle within each figure is the likelihood of child educational attainment conditional on the attainment of their parents. to 18 (or 14 to 25), for whom the preceding generation (parents or older relatives) have on average less than primary education. Hence, αup cis the parameter of interest and, for each country c, measures the likelihood that children whose “parents” did not complete primary school will themselves complete primary education, net of cohort effects for both children and parents, as well as of census year effects. This empirical approach, similar to the one used in Alesina et al. (2021) with data from Africa, delivers a measure of mobility that is comparable between countries and that captures some long-term patterns over half a century by netting out birth-cohort and census-year effects that are common across countries. Downward mobility at the country level. To estimate downward IGM, I use a similar econometric specification, pooling observations from all the censuses and countries: ydown icoyt =αdown c+γb o+γb y+θt+ϵicoyt (2) where ydown icoyt is a dummy variable that takes a value equal to one when individual idoes not complete primary education and zero otherwise. The parameters γb o,γb y,θtagain refer respectively to fixed effects by decade-cohort of the preceding generation that coresides with individual i, decade-cohort of individual i, and census year. These fixed effects aim to 12 control for differences driven by the birth cohort of children, the birth cohort of their parents, and other potential factors associated with each census. This regression uses a sample of individuals aged 14 to 18 (or 14 to 25), for whom the preceding generation (parents or older relatives) have on average completed at least primary education. Hence, αdown cis the parameter of interest and, for each country c, measures the likelihood that children whose “parents” did not complete primary school will themselves fail to complete primary education, net of cohort effects for both children and parents, as well as of census year effects. Upward and downward mobility at a finer geographical level. To estimate IGM at a more disaggregated level (i.e., provinces or districts), I run the following econometric specifications, country by country: yup icroyt =αup cr +γb o+γb y+θt+ϵicroyt ydown icroyt =αdown cr +γb o+γb y+θt+ϵicroyt (3) The variables and subscripts that appear in equations 1and 2are interpreted in the same way here, and the additional subscript rrefers to the district or province, according to the level of geographical disaggregation used in the analysis (provinces for baseline estimates and districts in the additional exercise in the appendix). Why is primary education a suitable variable for measuring IGM? There are three reasons for focusing on primary education. First, a substantial share of the population of Latin America and the Caribbean in the period spanned by this data set has attained less than a primary education, as shown in the previous subsection. Second, this focus makes the analysis directly comparable to the recent work of Alesina et al. (2021) in Africa and minimizes the potential bias that comes from using samples of coresidents. Third, the focus on the lowest level of education can also be justified from a conceptual point of view. Development policy discussions often claim that the poorest should not be left behind, and this focus is related to the school of moral philosophy exemplified by the principle of justice proposed by Rawls (1971).15 Robustness. As a robustness check, I compute upward and downward mobility using some alternative options for data construction. First, I use the maximum attainment of the preceding generation instead of average attainment. This change produces estimates with negligible differences (for example, the Pearson correlation coefficients between measurements using average versus maximum at the country, province, and district level are approximately 1). Second, I estimate mobility using a sample of individuals linked to (probable) parents, as done by IPUMS (2019). This change produces estimates that are also 15See Ravallion (2016) as an example of the focus on the poorest in the context of poverty measurement. 13 highly correlated (for example, the Pearson correlation coefficients between measurements using older relatives versus (probable) parents at the country, province, and district level are 0.98, 0.97, and 0.93, respectively). Alternative measures of IGM. I estimate a set of additional measures of intergenerational mobility that are less focused on the bottom of the educational attainment distribution. In contrast to the estimates focused on primary education, these measures are computed using individuals aged 19 to 25. First, I estimate upward and downward mobility considering secondary education instead of primary. Second, I estimate upward mobility as the likelihood of an individual finishing at least secondary education when the preceding generation was not able to complete primary school. These indicators are more prone to coresidence bias, but they still provide valuable information. For example, Munoz and Siravegna (2023) show that the rank correlation between indicators of upward mobility using secondary level computed with all children versus coresidents is approximately 0.86.16 III Intergenerational Mobility in LAC III.1 Country-level estimates Table 1summarizes the estimates of mobility at the country level. On average, close to 50 percent of children with parents who did not finish primary education (from now on, illiterate parents) are able to complete primary school. On the other hand, downward mobility is close to 10 percent, since one out of 10 children with parents who finished primary education (from now on, literate parents) do not complete primary school.17 There is substantial heterogeneity among LAC countries. The probability that children of illiterate parents will complete primary school ranges from 18% in Guatemala to 87% in Jamaica. In the case of downward mobility, the estimated probability that children of literate parents will not complete primary school ranges from null in Jamaica to 23% in Haiti. The heterogeneity found in upward mobility in Latin America (e.g., the gap of 69 percentage points between Jamaica and Guatemala) is relatively similar to the heterogeneity that Alesina et al. (2021) documented for African countries (e.g., the gap of 75 percentage points between South Africa and South Sudan), although minimum and maximum values were higher in Africa. Furthermore, the level of upward mobility among countries in LAC substantially overlaps with that of Africa. Countries with the lowest levels of upward mobility in LAC, such as Haiti, Guatemala, and Nicaragua, are more upwardly mobile than the five 16Using 72 countryand five-year birth cohorts that span 18 countries in Latin America. 17Note that these estimates are computed net of cohort and census year effects according to equations 1 and 2, which may result in estimates outside the 0–1 range, as seen in Jamaica. 14 lowest of the 27 countries for which Alesina et al. (2021) provide estimates (Malawi, Ethiopia, Sudan, Mozambique, and South Sudan). In contrast, the overlap in downward mobility is much less pronounced (see Figure A15 in the appendix.). Table 1: Country-level estimates of educational intergenerational mobility (1) (2) (3) (4) (5) (6) mobility / N census years upward upward downward downward N N age range 14–18 14–25 14–18 14–25 14–18 14–25 Jamaica 1982,1991,2001 .868 .864 -.004 .003 43,404 77,227 Trinidad and Tobago 1970,1980,1990,2000,2011 .839 .833 .023 .023 41,253 81,100 Argentina 1970,1980,1991,2001,2010 .762 .789 .035 .034 1,068,471 2,017,618 Chile 1970,1982,1992,2002 .682 .709 .05 .044 344,149 651,737 Uruguay 1963,1975,1985,1996,2006,2011 .668 .685 .064 .052 108,528 199,653 Cuba 2002,2012 .662 .688 .027 .024 101,268 214,486 Panama 1960,1970,1980,1990,2000,2010 .635 .665 .049 .04 86,527 157,906 Costa Rica 1973,1984,2000,2011 .634 .643 .086 .068 107,088 197,018 Bolivia 1976,1992,2001,2012 .609 .634 .068 .057 206,745 358,013 Mexico 1970,1990,2000,2010 .602 .622 .048 .042 2,811,581 4,961,471 Ecuador 1974,1982,1990,2001,2010 .543 .572 .089 .074 373,130 667,055 Suriname 2012 .535 .563 .042 .031 2,999 6,141 Venezuela 1971,1981,1990,2001 .533 .587 .096 .08 517,834 940,766 Saint Lucia 1980,1991 .523 .492 .126 .142 2,089 3,679 Peru 1993,2007 .48 .524 .115 .088 357,472 668,806 Paraguay 1962,1972,1982,1992,2002 .432 .463 .116 .096 118,082 207,766 Colombia 1973,1985,1993,2005 .402 .437 .142 .114 886,765 1,605,718 Honduras 1974,1988,2001 .398 .433 .151 .133 109,458 182,786 Dominican Republic 1981,2002,2010 .376 .442 .15 .124 173,340 312,654 Brazil 1960,1970,1980,1991,2000,2010 .367 .422 .171 .128 10,755,296 18,713,402 El Salvador 1992,2007 .342 .374 .164 .138 85,402 150,582 Haiti 1971,1982,2003 .212 .266 .226 .178 104,465 183,588 Nicaragua 1971,1995,2005 .194 .238 .223 .18 93,635 167,740 Guatemala 1964,1973,1981,1994,2002 .181 .212 .159 .129 238,047 402,133 mean / total .52 .548 .101 .084 18,737,028 33,129,045 Notes: Columns (1) and (2) give upward IGM estimates. They reflect the likelihood that individuals aged 14–18 or 14–25 whose parents have not completed primary schooling will manage to complete at least primary education. Columns (3) and (4) give downward IGM estimates. They reflect the likelihood that individuals aged 14–18 or 14–25 whose parents have completed primary schooling or higher will not manage to complete primary education. Columns (5) and (6) give the number of observations used to estimate the countryspecific IGM statistics (children with parents whose education is reported in the censuses). Countries are sorted from the highest to the lowest level of upward IGM based on the 14–18 sample (column (1)). “Mean” gives the unweighted average of the 24 country estimates. Figure 6maps the country-level estimates of upward and downward mobility in education. It highlights the heterogeneity found across the continent, showing that patterns of upward mobility are inversely related to downward mobility and that there are combinations of low and high mobility countries in South America, as well as in Central America and the Caribbean. The estimates of upward and downward mobility are significantly negatively correlated at the country level (see Figure A16 in the appendix). Country-level estimates of intergenerational mobility focused on secondary education can be found in Table A8 in the appendix. The samples are smaller, and the level of upward 15 Figure 6: Intergenerational educational mobility in LAC Argentina Bolivia Brazil Chile Colombia Costa Rica Cuba Dominican Republic Ecuador El Salvador Guatemala Haiti Honduras Jamaica Mexico Suriname Nicaragua Paraguay Peru Panama St. Lucia Trinidad and Tobago Uruguay Venezuela Argentina Bolivia Brazil Chile Colombia Costa Rica Cuba Dominican Republic Ecuador El Salvador Guatemala Haiti Honduras Jamaica Mexico Suriname Nicaragua Paraguay Peru Panama St. Lucia Trinidad and Tobago Uruguay Venezuela Upward Mobility 0.66 to 0.87 0.59 to 0.66 0.49 to 0.59 0.37 to 0.49 0.18 to 0.37 (a) Upward Mobility Argentina Bolivia Brazil Chile Colombia Costa Rica Cuba Dominican Republic Ecuador El Salvador Guatemala Haiti Honduras Jamaica Mexico Suriname Nicaragua Paraguay Peru Panama St. Lucia Trinidad and Tobago Uruguay Venezuela Argentina Bolivia Brazil Chile Colombia Costa Rica Cuba Dominican Republic Ecuador El Salvador Guatemala Haiti Honduras Jamaica Mexico Suriname Nicaragua Paraguay Peru Panama St. Lucia Trinidad and Tobago Uruguay Venezuela Downward Mobility −0.004 to 0.046 0.046 to 0.072 0.072 to 0.115 0.115 to 0.154 0.154 to 0.226 (b) Downward mobility Notes: Upward mobility reflects the likelihood that children aged 14–18 whose parents have not completed primary schooling will manage to complete at least primary education. Downward mobility reflects the likelihood that children aged 14–18 whose parents have completed primary schooling or higher will not manage to complete primary education. Both estimates are net of cohort and census year effects. mobility is considerably lower, while the level of downward mobility is considerably higher. As with the estimates using primary education, there is significant variation across countries. In the case of upward mobility measured as the likelihood that children complete at least secondary education when their parents were not able to complete primary, we see lower levels of mobility at the country level, as could be expected (see Table A11 in the appendix). III.1.a Urban-rural Given that an important feature of most developing countries is the gap in living standards between rural and urban residents (see Lagakos,2020), I explore the heterogeneity in IGM between these populations and document how they have evolved across birth cohorts. I do so by estimating upward and downward mobility by country, birth decade of the “children,” and urban/rural status of their residence.18 Figure 7shows the gap in the upward/downward mobility between urban and rural areas over the various birth cohorts. I find a gap in upward mobility that favors urban areas and that has been declining from 36 percentage points to 18Equation 1and 2in this case becomes yup/down icys =αup/down cys +ϵicys, where αup/down cys is a fixed effect by country, decade-cohort of individual i, and rural/urban residence s. 16 Figure 7: Intergenerational educational mobility in LAC - urban/rural BOL BOL BOL BOL BRA BRA BRA BRA CHL CHL CHL CHL COL COL COL COL CRI CRI CRI CRI GTM GTM GTM GTM HND HND HND HND PRY PRY PRY PRY ARG ARG ARG BOL BRA BRA COL CRI DOM DOM ECU ECU ECU SLV SLV GTM HTI HTI JAM MEX MEX MEX MEX NIC NIC PAN PAN PAN PAN PAN PRY PER PER PER LCA URY URY URY URY VEN VEN VEN gap = 10.692 + -.0053 cohort (3.1274) (.0016) 0 .1 .2 .3 .4 .5 urban upward IGM - rural upward IGM 1940 1950 1960 1970 1980 1990 Birth decade countries with data for 1950s, 1960s, 1970s, 1980s other countries mean among countries with data 1950s-1980s (a) Upward mobility BOL BOL BOL BOL BRA BRA BRA BRA CHL CHL CHL CHL COL COL COL COL CRI CRI CRI CRI GTM GTM GTM GTM HND HND HND HND PRY PRY PRY PRY ARG ARG ARG BOL BRA BRA COL CRI DOM DOM ECU ECU ECU SLV SLV GTM HTI HTI JAM MEX MEX MEX MEX NIC NIC PAN PAN PAN PAN PAN PRY PER PER PER LCA URY URY URY URY VEN VEN VEN gap = -9.263 + .0046 cohort (2.2662) (.0012) -.5 -.4 -.3 -.2 -.1 0 urban downward IGM - rural downward IGM 1940 1950 1960 1970 1980 1990 Birth decade countries with data for 1950s, 1960s, 1970s, 1980s other countries mean among countries with data 1950s-1980s (b) Downward mobility Notes: These estimates correspond to the probability that individuals whose parents did not finish primary school will manage to complete at least primary education, in the case of upward mobility, and the probability that individuals whose parents completed primary school will not manage to complete primary education, in the case of downward mobility. The estimates are for individuals aged 14–18 by country (pooling all available waves), birth decade of the “children,” and urban/rural status of the household. 20 percentage points in more recent birth cohorts. In other words, upward mobility is on average 36 percentage points higher in urban areas compared to rural areas for the cohort born in 1950–1959, and this gap declines to 20 percentage points for those born between 1980 and 1989. The gap in downward mobility is also closing from below, moving from 29 percentage points for the 1950 birth decade to 15 percentage points for the 1980 birth decade. Figure A17 and Figure A18 in the appendix show estimates by sub-population rather than the gap between them for countries with data for at least four decades. These estimates suggest that the gap has been shrinking because of an increase in upward mobility and a decrease in downward mobility in rural areas.19 III.1.b Gender As discussed in a recent survey on IGM in developing countries (see Torche,2021), gender gaps in education have been disappearing or even shifting in favor of women. I examine whether these patterns hold in this census data set by estimating IGM for males and females 19The urban-rural gap can be affected by migration from rural to urban areas. However, the gap is unlikely driven by this migration, as it would require negative selection into migration, which contrasts with the evidence of positive selection shown in Munoz (2022). 17 Figure 8: Intergenerational educational mobility in LAC - gender ARG ARG ARG ARG BOL BOL BOL BOL BRA BRA BRA BRA CHL CHL CHL CHL CRI CRI CRI CRI HND HND HND HND NIC NIC NICNIC PAN PAN PAN PAN PRY PRY PRY PRY 780 780 780 780 VEN VEN VEN VEN ARG BOL BRA BRA COL COL COLCOL COL COL CRI 192 192 DOM DOM DOM ECU ECUECU ECU ECU ECU SLV SLV SLV GTM GTMGTM GTM GTM GTMGTM HTI HTI HTI JAM JAM JAM MEX MEX MEX MEX NIC PAN PAN PRY PER PER PER LCA LCA 740 780 URY URY URYURY URYURY URYURY URYURY gap = 3.535 + -.0018 cohort (1.854) (.0009) -.2 -.1 0 .1 .2 male upward IGM - female upward IGM 1940 1950 1960 1970 1980 1990 Birth decade countries with data for 1950s, 1960s, 1970s, 1980s other countries mean among countries with data 1950s-1980s (a) Upward mobility ARG ARG ARG ARG BOL BOL BOL BOL BRA BRA BRA BRA CHL CHL CHL CHL CRI CRI CRI CRI HND HND HND HND NIC NIC NICNIC PAN PAN PAN PAN PRY PRY PRY PRY 780 780 780 780 VEN VEN VEN VEN ARG BOL BRA BRA COL COL COLCOL COL COL CRI 192 192 DOM DOM DOM ECU ECUECU ECU ECU ECU SLV SLV SLV GTM GTMGTM GTM GTM GTMGTM HTI HTI HTI JAM JAM JAM MEX MEX MEX MEX NIC PAN PAN PRY PER PER PER LCA LCA 740 780 URY URY URYURY URYURY URYURY URYURY gap = -.8180 + .0004 cohort (.6631) (.0003) -.05 0 .05 .1 .15 .2 male downward IGM - female downward IGM 1940 1950 1960 1970 1980 1990 Birth decade countries with data for 1950s, 1960s, 1970s, 1980s other countries mean among countries with data 1950s-1980s (b) Downward mobility These estimates reflect the probability that individuals whose parents did not finish primary school will complete at least primary education, in the case of upward mobility, and the probability that individuals whose parents completed primary school will not manage to complete primary education, in the case of downward mobility. The estimates are for individuals aged 14–18 by country (pooling all available waves), birth decade of the “children,” and gender. separately and documenting how the gap between these populations has evolved across birth cohorts. I estimate upward and downward mobility by country, birth decade of the “children,” and gender.20 I do not find systematic differences by gender for older birth cohorts, but there appears to be a trend towards higher upward mobility for women because their upward mobility is 3 percentage points higher in the 1980s birth cohort (see Figure 8). For downward mobility, there is a similar gap in favor of women of approximately 3 percentage points for the 1980s birth cohort, with a flatter trend. Figure A19 and Figure A20 in the appendix show estimates by sub-population rather than the gender gap for countries with data for at least four decades. These estimates suggests that the gap has been increasing in favor of women because of a more-than-proportional increase for them in upward mobility and a less than proportional decrease in downward mobility. III.1.c Shifts over time As mentioned in the data section, data coverage over time is unbalanced, with data for some countries spanning more years than others. This limits analysis of trends over time 20Equation 1and 2in this case becomes yup/down icyg =αup/down cyg +ϵicyg, where αup/down cyg is a fixed effect by country, decade-cohort of individual i, and gender g. 18 Figure 9: Intergenerational educational mobility in LAC across cohorts 0 .2 .4 .6 .8 1 IGM UP Haiti Guatemala Nicaragua Honduras El Salvador Venezuela Brazil Saint Lucia Paraguay Dominican Republic Colombia Panama Argentina Ecuador Peru Costa Rica Suriname Chile Uruguay Mexico Cuba Trinidad and Tobago Bolivia Jamaica 1940 1950 1960 1970 1980 1990 (a) Upward mobility 0 .1 .2 .3 .4 IGM Down Haiti Nicaragua Guatemala Honduras Brazil Saint Lucia El Salvador Paraguay Venezuela Dominican Republic Ecuador Argentina Peru Colombia Trinidad and Tobago Costa Rica Chile Uruguay Panama Bolivia Mexico Jamaica Suriname Cuba 1940 1950 1960 1970 1980 1990 (b) Downward mobility These estimates correspond to the probability that individuals whose parents did not finish primary school will manage to complete at least primary education, in the case of upward mobility, and the probability that individuals whose parents completed primary school will not manage to complete primary education, in the case of downward mobility. The estimates are for individuals aged 14–18 by country (pooling all available waves) and birth decade of the “children.” and the conclusions that can be derived from cross-country comparisons at given points in time or for a given cohort. Nevertheless, I document estimates of mobility by country for the different birth cohorts that are available.21 Figure 9reports these estimates. The level of upward mobility has clearly been increasing, while downward mobility has been falling. This is unsurprising, given that educational attainment has increased in the region over the last decades. For example, in Brazil, Panama, and Uruguay, the countries with the most censuses in the data set, upward mobility increased from 0.05, 0.36, and 0.48, respectively, for those born in the 1940s to 0.66, 0.74, and 0.85 for those born in the 1990s. Despite this shift and the improvement in schooling in the region, in several countries the probability that those born in the 1990s to poorly educated parents will complete primary education is still less than 80 percent. 21Equation 1and 2in this case becomes yup/down icys =αup/down cy +ϵicy, where αup/down cy is a fixed effect by country and decade-cohort of individual i. 19 III.1.d Discussion of the results The estimates of intergenerational mobility at the country level reveal the following patterns. First, there is substantial heterogeneity in upward and downward mobility across countries in LAC. Caribbean countries, such as Jamaica and Trinidad and Tobago, show the highest levels of upward mobility in LAC, while Haiti has among the lowest. As mentioned earlier, these country-level estimates complement the existing literature by using an indicator focused on the most disadvantaged population. Hence, the associated rankings based on IGM do not necessarily align with previous studies that use other indicators.22 The range of estimates for LAC is similar to the range found for Africa, although with higher average upward mobility and lower downward mobility. Second, upward mobility is higher in urban areas, and downward mobility is lower. However, the urban-rural gap is decreasing over birth cohorts because IGM in rural areas is catching up. Third, upward mobility is slightly higher and downward mobility slightly lower for females, and the gender gap shifts in favor of females over birth cohorts. This trend in upward mobility favoring females is consistent with the findings of Neidh¨ofer et al. (2018),23 and this finding adds to the scarce literature on gender differences in IGM (see Torche,2021). Fourth, upward mobility has increased significantly across birth cohorts, while downward mobility has decreased substantially. This is consistent with the trends shown in Neidh¨ofer et al. (2018) and Van der Weide et al. (2024) for absolute mobility. IV Spatial Variation and Correlates of IGM in LAC In this section, I map intergenerational mobility in education across regions in LAC and then explore whether the observed patterns are associated with a set of correlates. IV.1 Spatial variation of intergenerational mobility in LAC Table 2summarizes the estimates of mobility at the province level. These results show that in some countries, mobility levels vary substantially across provinces.24 This is the case in Paraguay, Mexico, Guatemala, and Peru, where the difference between the most upwardly mobile and least upwardly mobile provinces is more than half the range between countries 22This is also recognized in Neidh¨ofer et al. (2018) when comparing ranks based on indicators of relative mobility to those of absolute mobility. 23These results are documented in the online supplemental material of the paper and compute upward mobility as the probability that children with parents who achieved less than secondary education will complete secondary school for nine countries. 24Figure A6 in the appendix visually represents variability within countries. 20 in Latin America. But there are also cases with either high or low upward mobility at the country level and very little variation within country, such as Jamaica and Haiti, although this is somewhat expected since these countries have few administrative units and a small population. For downward mobility, variability (measured by standard deviation or the difference between the province with the highest and lowest level) is less pronounced. However, the full range of downward mobility in some countries is larger than the difference between the country with the highest mobility (Jamaica) and the country with the lowest mobility (Haiti). Nicaragua and Honduras stand out as cases where the range between the provinces with the minimum and the maximum downward mobility is relatively wide (see Figure A6 in the appendix). Figures 10 and 11 map out the estimates summarized in Table 2. They reveal interesting patterns in some countries. For example, Mexico has a somewhat lower level of upward mobility in the south, and a lighter spot in the middle of the country where the capital is. In contrast, Brazil has a much lower level of mobility in the northern regions and higher mobility on the east coast near the states of S˜ao Paulo and Rio de Janeiro. Overall, the region shows higher levels in the south, especially on the Pacific coast, and some heterogeneity in the Caribbean, with major contrasts between Cuba and Haiti. In the appendix, I share similar estimates (see Table A7) and maps (see Figure A7 and A8) at the district level, which is the finest administrative unit available in the data set. The patterns are qualitatively similar, but the level of disaggregation means that estimates for some districts with few observations end up outside the [0,1] range. The appendix also contains summary statistics of alternative estimates of intergenerational mobility that consider secondary education at the province and district levels (see Table A9,A10,A12, and A13). These estimates are consistent with the country-level ones, in the sense that relative to my baseline estimates using primary education, they show lower levels of upward mobility, higher levels of downward mobility, smaller samples, and significant within-country variation. IV.2 Correlates of intergenerational mobility Given the substantial heterogeneity observed across regions, I explore a set of correlates of regional IGM to uncover stylized facts that help characterize its geography. The set is relatively small given the difficulty of collecting data that is comparable across all adminis21 et al.,2022). Additionally, this finding suggests that patterns of IGM in income may also vary significantly in the region, although this hypothesis is not verifiable at the moment. For example, Card et al. (2022) use a similar measure of intergenerational mobility in education (fraction of children who completed at least ninth grade among those whose parents completed 5–8 years of education) to map IGM at the county level in the U.S. using the 1940 census. They find important variability and geographical patterns that are similar to the map with the estimates of income mobility for the cohort born in years 1980–83 from Chetty et al. (2014). Second, IGM is strongly correlated to the average educational attainment of parents. Furthermore, beyond the effect of initial conditions, IGM appears to be correlated to some proxies of development, as well as distance to the capital, suggesting that educational opportunities are centralized. These results are qualitatively similar to the findings in Africa (Alesina et al.,2021) and complement previous studies that have shown a positive association between economic development and measures of intergenerational mobility in the region (Neidh¨ofer et al.,2024,2018) and globally (Van der Weide et al.,2024). V Final Remarks In this paper, I examine intergenerational educational mobility for countries in Latin America and the Caribbean at a disaggregated regional level using census data spanning more than half a century. I investigate mobility in education at the bottom of the educational attainment distribution by focusing on the likelihood that children whose parents did not complete primary education will themselves manage to complete that level, which can be measured with a high degree of confidence between ages 14 and 18. Similarly, I measure downward mobility as the probability that children whose parents completed at least primary education will not themselves manage to complete that level. I find wide cross-country and within-country heterogeneity. In LAC, the distance between the most and least upwardly mobile countries is relatively close to what has been recently documented in Africa, although the least mobile countries in Africa are less mobile than any country in LAC. Similarly, the median country in LAC shows higher upward mobility than the median country in Africa. There is significant overlap in the levels of upward mobility between these two regions, but much less overlap in levels of downward mobility. I do not find significant differences by gender, but I do document a decline in the mobility gap between urban and rural populations. I also find that upward mobility increases and downward mobility decreases over time. Within-country mobility shows a variety of patterns. For example, some countries have higher mobility in the northern regions (e.g., Mexico), whereas others show higher mobility 28 in the southern regions (e.g., Brazil). The level of within-country heterogeneity also varies by country, with the lowest levels found in the smallest and least populated nations. Moreover, level of mobility closely correlates to the share of the preceding generation that completed primary school. In addition, there appears to be a weak positive correlation between upward mobility and share of the workforce employed in industry or distance to the capital, whereas downward mobility is significantly correlated to the share of the workforce employed in industry and the share of the workforce employed in agriculture, and is only weakly correlated to distance to the capital. Given the unbalanced nature of the data set in terms of coverage over time and across countries, further research could shed more light on potential determinants of mobility in Latin America by analyzing countries with relatively high data coverage, such as Chile, Mexico, or Brazil, where it is easier to collect correlates by administrative unit. This paper contributes to this goal by creating estimates of mobility at a disaggregated geographical level and making them available in an online data appendix for future research. 29 References Alesina, A., Hohmann, S., Michalopoulos, S., & Papaioannou, E. (2021). Intergenerational Mobility in Africa. Econometrica,89(1), 1–35. Alesina, A., Hohmann, S., Michalopoulos, S., & Papaioannou, E. (2023). Religion and Educational Mobility in Africa. Nature,618(7963), 134–143. Alvaredo, F., & Gasparini, L. (2015). Recent Trends in Inequality and Poverty in Developing Countries (1st ed., Vol. 2). Elsevier B.V. Retrieved from http://dx.doi.org/10.1016/ B978-0-444-59428-0.00010-2 Asher, S., Novosad, P., & Rafkin, C. (2023). Intergenerational Mobility in India: New Measures and Estimates Across Time and Social Groups. American Economic Journal: Applied Economics (forthcoming). Becker, G. S., & Tomes, N. (1979). An Equilibrium Theory of the Distribution of Income and Intergenerational Mobility. Journal of Political Economy,87(6), 1153–1189. Becker, G. S., & Tomes, N. (1986). Human Capital and the Rise and Fall of Families. Journal of Labor Economics,4(3), S1–S39. Bjorklund, A., & Jantti, M. (1997). Intergenerational Income Mobility in Sweden Compared to the United States. The American Economic Review,87(5), 1009–1018. Black, S. E., & Devereux, P. J. (2011). Recent Developments in Intergenerational Mobility (Vol. 4b). Card, D., Domnisoru, C., & Taylor, L. (2022). The Intergenerational Transmission of Human Capital: Evidence from the Golden Age of Upward Mobility. Journal of Labor Economics, 40(S1), S39–S95. Chetty, R., Hendren, N., Kline, P., & Saez, E. (2014). Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States. The Quarterly Journal of Economics,129(4), 1553–1623. Corak, M. (2013). Income Inequality, Equality of Opportunity, and Intergenerational Mobility. Journal of Economic Perspectives,27(3), 79–102. Corak, M. (2020). The Canadian Geography of Intergenerational Income Mobility. The Economic Journal,130(631), 2134–2174. Cort´es Orihuela, J., D´ıaz, J. D., Guti´errez Cubillos, P., Montecinos, A., Troncoso, P. A., & Villarroel, G. I. (2023). Decentralizing the Chilean Miracle: Regional Intergenerational Mobility in a Developing Country. Regional Studies,57(5), 785–799. Dahan, M., & Gaviria, A. (2001). Sibling correlations and intergenerational mobility in Latin America. Economic Development and Cultural Change,49(3), 537–554. Derenoncourt, E. (2022). Can You Move to Opportunity? Evidence from the Great Migration. American Economic Review,112(2), 369–408. Deutscher, N., & Mazumder, B. (2023). Measuring Intergenerational Income Mobility: A Synthesis of Approaches. Journal of Economic Literature (forthcoming),61(3), 988–1036. Emran, M. S., Greene, W., & Shilpi, F. (2018). When Measure Matters: Coresidency, Truncation Bias, and Intergenerational Mobility in Developing Countries. Journal of Human Resources,53(3), 579–607. Emran, M. S., & Shilpi, F. (2021). Economic Approach to Intergenerational Mobility: Measures, Methods, and Challenges in Developing Countries. In V. Iversen, A. Krishna, & 30 K. Sen (Eds.), Social mobility in developing countries: Concepts, methods, and determinants. Oxford University Press. Feigenbaum, J. J. (2018). Multiple Measures of Historical Intergenerational Mobility: Iowa 1915 to 1940. Economic Journal,128(612), F446–F481. Francesconi, M., & Nicoletti, C. (2006). Intergenerational Mobility and Sample Election in Short Panels. Journal of Applied Econometrics,21, 1265–1293. Gallup, J. L., Gaviria, A., & Lora, E. (2003). Is Geography Destiny? Lessons from Latin America. Stanford University Press and the World Bank. G.C. Britto, D., Fonseca, A., Pinotti, P., Sampaio, B., & Warwar, L. (2022). Intergenerational Mobility in the Land of Inequality. IZA Discussion Paper(15611). Goldin, C., & Katz, L. F. (1999). Human Capital and Social Capital: The Rise of Secondary Schooling in America, 1910-1940. The Journal of Interdisciplinary History,29(4), 683– 723. Gomez-Lobo, A., & Oviedo, D. (2023). Spatial Inequalities in Latin America: Mapping Aggregate to Micro-level disparities. LSE Research Online Documents on Economics, 120691. G¨uell, M., Pellizzari, M., Pica, G., & Rodr´ıguez Mora, J. V. (2018). Correlating Social Mobility and Economic Outcomes. Economic Journal,128(612), F353–F403. Haider, S., & Solon, G. (2006). Life-Cycle Variation in the Association between Current and Lifetime Earnings. American Economic Review,96(4), 1308–1320. Hertz, T., Jayasundera, T., Piraino, P., Selcuk, S., Smith, N., & Verashchagina, A. (2007). The Inheritance of Educational Inequality: International Comparisons and Fifty-Year Trends. The B.E. Journal of Economic Analysis Policy,7(2). Hilger, N. (2016). The Great Escape: Intergenerational Mobility in the United States Since 1940. NBER Working Paper(21217). IPUMS. (2019). Integrated Public Use Microdata Series, International: Version 7.2 [dataset]. Minneapolis, MN: IPUMS: Minnesota Population Center. Lagakos, D. (2020). Urban-Rural Gaps in the Developing World: Does Internal Migration Offer Opportunities? Journal of Economic Perspectives,34(3), 174–192. Maloney, W. F., & Valencia Caicedo, F. (2016). The Persistence of (Subnational) Fortune. Economic Journal,126(598), 2363–2401. Munoz, E. (2021). Intergenerational Educational Mobility within Chile. Available at SSRN: https://ssrn.com/abstract=3969270 or http://dx.doi.org/10.2139/ssrn.3969270. Munoz, E. (2022). Does it Matter Where You Grow up? Childhood Exposure Effects in Latin America and the Caribbean. Policy Research working paper, World Bank Group, 10037. Munoz, E., & Siravegna, M. (2023). When Measure Matters: Coresidence Bias and Intergenerational Mobility Revisited. IDB Working Paper,01469. Narayan, A., Van der Weide, R., Cojocaru, A., Lakner, C., Redaelli, S., Gerszon Mahler, D., . . . Thewissen, S. (2018). Fair Progress?: Economic Mobility Across Generations Around the World. The World Bank. Neidh¨ofer, G. (2019). Intergenerational Mobility and the Rise and Fall of Inequality: Lessons from Latin America. Journal of Economic Inequality,17(4), 499–520. 31 Neidh¨ofer, G., Gasparini, L., Ciaschi, M., Gasparini, L., & Serrano, J. (2024). Social Mobility and Economic Development. Journal of Economic Growth,29, 327-359. Neidh¨ofer, G., Serrano, J., & Gasparini, L. (2018). Educational Inequality and Intergenerational Mobility in Latin America: A New Database. Journal of Development Economics, 134, 329–349. Ravallion, M. (2016). Are the World’s Poorest being Left Behind? Journal of Economic Growth,21, 139–164. Rawls, J. (1971). A Theory of Justice. Cambridge, MA: Harvard University Press. Solon, G. (1992). Intergenerational Income Mobility in the United States. The American Economic Review,82(3), 393–408. Torche, F. (2021). Educational Mobility in the Developing World. In V. Iversen, A. Krishna, & K. Sen (Eds.), Social mobility in developing countries: Concepts, methods, and determinants. Oxford University Press. Van der Weide, R., Ferreira de Souza, P., & Barbosa, R. (2020). Intergenerational Mobility in Education in Brazil. mimeo. Van der Weide, R., Lakner, C., Mahler, D. G., Narayan, A., & Gupta, R. (2024). Intergenerational mobility around the world: A new database. Journal of Development Economics, 166. 32 Appendices This appendix provides details on how the sample was constructed, as well as some additional tables and graphs. Table A1 lists the census samples obtained from IPUMS-International and the size of the data set extracted. Table A2 reports sample size, from raw data to samples restricted by age and by availability of information on education. Table A3 reports coresidence rates by country for different ages. Table A4 details how the information on relationship to head of household is used to identify different generations. Table A5 reports coresidence rates by country-sample for different ages. Figure A1 displays coresidence rates by urban/rural population and by gender. Figure A2 compares estimates of upward mobility with all children versus coresident children for the same country-cohort. The source of these estimates is Munoz and Siravegna (2023). Figure A3 displays the educational attainment transition matrix for individuals 14–25 years old. Figure A4 displays the educational attainment transition matrix for individuals 14–25 years old in selected countries. Table A6 summarizes education level by cohort using data on individuals at least 25 years old. Figure A5 reports the CDF of the sample size by province. Figure A6 reports the variability of province-level estimates of intergenerational mobility within countries. Table A7 reports district-level estimates of intergenerational mobility. Figure A7 and A8 map out mobility at the district level for LAC. Figure A9 and A10 map out mobility at the district level for LAC using secondary education. Table A8,A9, and A10 report estimates of IGM based on secondary education. Table A11,A12, and A13 report estimates of IGM based on the likelihood of an individual completing secondary education when their parents completed less than primary school. Figure A11,A12,A13, and A14 display mosaic plots of educational attainment by country. Figure A15 compares upward and downward mobility in LAC and Africa. Figure A16 shows the negative relationship between upward and downward mobility. 33 Figure A17 shares estimates of upward mobility by urban/rural status for selected countries. Figure A18 shows estimates of downward mobility by urban/rural status for selected countries. Figure A19 displays estimates of upward mobility by gender for selected countries. Figure A20 shares estimates of downward mobility by gender for selected countries. Figure A21 shows scatter plots between IGM and share of the preceding generation that completed at least primary education, by district. Figure A22 shows the association between IGM estimates using secondary education and correlates. 34 A Sample coverage and construction Table A1: Census samples N Country Year Fraction Households Persons N Country Year Fraction Households Persons (%) (%) 1 Argentina 1970 2 129,728 466,892 47 Haiti 2003 10 219,633 838,045 2 Argentina 1980 10 672,062 2,667,714 48 Honduras 1974 10 49,064 278,348 3 Argentina 1991 10 1,199,651 4,286,447 49 Honduras 1988 10 77,406 423,971 4 Argentina 2001 10 1,040,852 3,626,103 50 Honduras 2001 10 123,584 608,620 5 Argentina 2010 10 1,217,166 3,966,245 51 Jamaica 1982 10 54,526 223,667 6 Bolivia 1976 10 121,378 461,699 52 Jamaica 1991 10 62,291 232,625 7 Bolivia 1992 10 177,926 642,368 53 Jamaica 2001 10 64,317 205,179 8 Bolivia 2001 10 239,475 827,692 54 Mexico 1970 1 82,856 483,405 9 Bolivia 2012 10 292,117 1,003,516 55 Mexico 1990 10 1,648,280 8,118,242 10 Brazil 1960 20 3,066,365 14,983,769 56 Mexico 2000 10.6 2,312,035 10,099,182 11 Brazil 1970 25 5,111,039 24,789,716 57 Mexico 2010 10 2,903,640 11,938,402 12 Brazil 1980 25 6,716,885 29,378,753 58 Nicaragua 1971 10 36,063 189,469 13 Brazil 1991 10 4,024,553 17,045,712 59 Nicaragua 1995 10 82,815 435,728 14 Brazil 2000 10 5,304,711 20,274,412 60 Nicaragua 2005 10 119,339 515,485 15 Brazil 2010 10 6,192,502 20,635,472 61 Panama 1960 5 11,869 53,553 16 Chile 1970 10 199,041 890,481 62 Panama 1970 10 31,755 150,473 17 Chile 1982 10 282,356 1,133,062 63 Panama 1980 10 47,726 195,577 18 Chile 1992 10 373,964 1,335,055 64 Panama 1990 10 61,458 232,737 19 Chile 2002 10 486,115 1,513,914 65 Panama 2000 10 84,346 284,081 20 Colombia 1973 10 349,853 1,988,831 66 Panama 2010 10 95,579 341,118 21 Colombia 1985 10 571,046 2,643,125 67 Paraguay 1962 5 18,307 90,236 22 Colombia 1993 10 774,321 3,213,657 68 Paraguay 1972 10 43,883 233,669 23 Colombia 2005 10 1,054,812 4,006,168 69 Paraguay 1982 10 60,465 301,582 24 Costa Rica 1973 10 36,323 186,762 70 Paraguay 1992 10 100,704 415,401 25 Costa Rica 1984 10 56,186 241,220 71 Paraguay 2002 10 113,039 516,083 26 Costa Rica 2000 10 106,973 381,500 72 Peru 1993 10 564,765 2,206,424 27 Costa Rica 2011 10 124,693 430,082 73 Peru 2007 10 821,675 2,745,895 28 Cuba 2002 10 371,878 1,118,767 74 Saint Lucia 1980 10 2,674 11,451 29 Cuba 2012 10 416,577 1,115,643 75 Saint Lucia 1991 10 3,394 13,382 30 Dominican Rep 1981 8.5 103,904 475,829 76 Suriname 2012 10 14,037 53,636 31 Dominican Rep 2002 10 247,375 857,606 77 Trinidad and Tobago 1970 10 15,871 69,349 32 Dominican Rep 2010 10 309,624 943,784 78 Trinidad and Tobago 1980 10 23,870 105,464 33 Ecuador 1974 10 145,902 648,678 79 Trinidad and Tobago 1990 10 27,561 113,104 34 Ecuador 1982 10 195,401 806,834 80 Trinidad and Tobago 2000 10 35,715 111,833 35 Ecuador 1990 10 243,898 966,234 81 Trinidad and Tobago 2011 8.8 41,606 116,917 36 Ecuador 2001 10 354,222 1,213,725 82 Uruguay 1963 10 79,403 256,171 37 Ecuador 2010 10 386,944 1,448,233 83 Uruguay 1975 10 95,935 279,994 38 El Salvador 1992 10 125,695 510,760 84 Uruguay 1985 10 105,761 295,915 39 El Salvador 2007 10 172,012 574,364 85 Uruguay 1996 10 118,067 315,920 40 Guatemala 1964 5 40,220 210,411 86 Uruguay 2006 8.4 85,316 256,866 41 Guatemala 1973 5.5 59,622 289,458 87 Uruguay 2011 10 118,498 328,425 42 Guatemala 1981 5 65,555 302,106 88 Venezuela 1971 2 284,336 1,158,527 43 Guatemala 1994 10 160,603 833,139 89 Venezuela 1981 10 323,321 1,441,266 44 Guatemala 2002 10 222,770 1,121,946 90 Venezuela 1990 10 468,808 1,803,953 45 Haiti 1971 10 95,145 434,869 91 Venezuela 2001 10 646,080 2,306,489 46 Haiti 1982 2.5 28,698 128,770 35 Table A2: Sample sizes All observations Obs. with education All observations Obs. with education Country Year age: All age: 14-18 age: 14-25 age: 14-18 age: 14-25 Country Year age: All age: 14-18 age: 14-25 age: 14-18 age: 14-25 Argentina 1970 466,892 42,317 96,744 31,411 59,124 Haiti 2003 838,045 103,088 218,016 72,705 130,436 Argentina 1980 2,700,000 241,353 532,289 193,448 348,232 Honduras 1974 278,348 32,262 64,660 24,018 37,966 Argentina 1991 4,300,000 392,977 844,871 347,074 611,881 Honduras 1988 423,971 47,258 95,944 37,642 62,769 Argentina 2001 3,600,000 321,380 764,630 295,621 596,468 Honduras 2001 608,620 73,272 154,339 62,008 105,745 Argentina 2010 4,000,000 354,910 813,073 323,256 621,385 Jamaica 1982 223,668 27,612 58,456 17,270 28,729 Bolivia 1976 461,699 51,674 109,380 35,230 57,307 Jamaica 1991 232,625 25,145 56,810 17,326 32,498 Bolivia 1992 642,368 69,992 147,085 46,235 75,965 Jamaica 2001 205,179 21,357 47,770 14,349 25,241 Bolivia 2001 827,692 90,786 199,275 63,080 111,001 Mexico 1970 483,405 54,069 111,210 41,915 64,605 Brazil 1960 15,000,000 1,600,000 3,500,000 1,300,000 2,200,000 Mexico 1990 8,100,000 1,000,000 2,100,000 900,739 1,500,000 Brazil 1970 25,000,000 2,800,000 6,000,000 2,300,000 3,700,000 Mexico 2000 10,000,000 1,100,000 2,400,000 963,638 1,700,000 Brazil 1980 29,000,000 3,300,000 7,400,000 2,700,000 4,600,000 Mexico 2010 12,000,000 1,300,000 2,700,000 1,200,000 2,200,000 Brazil 1991 17,000,000 1,800,000 4,000,000 1,600,000 2,800,000 Nicaragua 1971 189,469 22,601 44,957 16,771 26,368 Brazil 2000 20,000,000 2,200,000 4,800,000 1,900,000 3,400,000 Nicaragua 1995 435,728 51,956 107,402 42,619 74,447 Brazil 2010 21,000,000 1,900,000 4,500,000 1,700,000 3,200,000 Nicaragua 2005 515,485 60,691 136,084 50,811 95,961 Chile 1970 890,481 96,432 203,625 73,392 123,911 Panama 1960 53,553 5,481 11,869 3,368 5,498 Chile 1982 1,100,000 130,958 293,439 106,794 197,946 Panama 1970 150,473 15,817 34,219 11,310 18,797 Chile 1992 1,300,000 121,069 290,349 100,838 199,734 Panama 1980 195,577 22,673 47,420 17,725 30,333 Chile 2002 1,500,000 130,506 297,907 110,343 214,019 Panama 1990 232,737 25,536 57,471 19,537 36,604 Colombia 1973 2,000,000 245,355 493,144 172,222 281,047 Panama 2000 284,081 27,438 62,585 21,924 41,171 Colombia 1985 2,600,000 312,063 705,404 245,920 466,142 Panama 2010 341,118 30,266 70,017 26,170 49,837 Colombia 1993 3,200,000 336,233 758,037 263,014 485,909 Paraguay 1962 90,236 10,003 20,431 6,011 10,224 Colombia 2005 4,000,000 399,870 860,151 325,438 579,432 Paraguay 1972 233,669 27,630 54,005 18,806 31,105 Costa Rica 1973 186,762 23,539 46,832 18,809 30,070 Paraguay 1982 301,582 34,248 74,515 25,177 45,971 Costa Rica 1984 241,220 28,005 64,067 23,982 44,198 Paraguay 1992 415,401 41,705 89,839 30,061 52,473 Costa Rica 2000 381,500 40,582 88,091 36,085 63,624 Paraguay 2002 516,083 59,365 125,811 48,042 85,609 Costa Rica 2011 430,082 40,703 98,328 36,805 74,880 Peru 1993 2,200,000 245,196 539,320 183,244 335,766 Cuba 2002 1,100,000 82,556 180,787 69,378 132,152 Peru 2007 2,700,000 280,035 636,955 222,254 419,885 Dominican Republic 1981 475,829 62,387 126,838 49,358 84,310 Saint Lucia 1980 11,451 1,516 2,985 1,076 1,754 Dominican Republic 2002 857,606 85,616 194,479 69,843 128,140 Saint Lucia 1991 13,382 1,455 3,406 1,138 2,154 Dominican Republic 2010 943,784 98,661 221,932 78,426 142,857 Trinidad and Tobago 1970 69,349 8,259 16,684 6,398 10,873 Ecuador 1974 648,678 72,812 162,826 49,142 82,561 Trinidad and Tobago 1980 105,464 13,096 28,713 11,078 20,578 Ecuador 1982 806,834 89,627 194,868 64,889 112,394 Trinidad and Tobago 1990 113,104 10,646 24,520 9,232 18,279 Ecuador 1990 966,234 108,806 237,150 83,171 146,856 Trinidad and Tobago 2000 111,833 12,444 26,458 10,890 20,515 Ecuador 2001 1,200,000 126,354 287,034 100,955 186,327 Trinidad and Tobago 2011 116,917 8,325 22,630 7,288 17,595 Ecuador 2010 1,400,000 145,454 326,549 117,218 212,597 Uruguay 1963 256,171 20,618 47,079 15,749 28,722 El Salvador 1992 510,760 62,794 129,373 44,508 74,325 Uruguay 1975 279,994 24,213 53,152 18,704 33,222 El Salvador 2007 574,364 62,912 131,762 55,338 100,318 Uruguay 1985 295,915 23,728 55,355 18,881 35,368 Guatemala 1964 210,079 22,674 46,804 17,177 27,249 Uruguay 1996 315,920 26,188 60,440 21,870 41,399 Guatemala 1973 289,446 33,148 71,814 24,569 39,263 Uruguay 2006 256,866 21,943 45,451 20,277 36,604 Guatemala 1981 302,106 33,771 72,879 26,958 45,277 Uruguay 2011 328,425 26,825 60,496 23,925 43,382 Guatemala 1994 833,137 97,480 196,310 82,505 135,877 Venezuela 1971 1,200,000 133,044 282,119 87,971 144,465 Guatemala 2002 1,100,000 127,311 269,696 114,181 200,981 Venezuela 1981 1,400,000 166,729 367,032 133,566 238,340 Haiti 1971 434,869 51,096 101,984 35,014 58,427 Venezuela 1990 1,800,000 199,055 445,482 149,752 269,185 Haiti 1982 128,770 15,471 36,494 8,349 15,840 Venezuela 2001 2,300,000 234,403 534,204 204,784 394,511 Notes: This table reports the total sample size by census year and country, as well as the sample population restricted by age and by the presence of information on education for children and parents. 36 B Rates of coresidence This table shows the coresidence rate by country for different ages. The coresidence rate is the total number of individuals who coreside with at least one member of an immediately preceding generation in the household divided by the total number of individuals in the age group. The sample only includes individuals for whom educational attainment and relationship to head of household are observed. Table A3: Coresidence rates Rate Observations (thousands) 14–18 18–25 21–25 20–23 14–18 18–25 21–25 20–23 Argentina 95.7 72.1 63.1 72.2 1246 1746 1067 870 Bolivia 86.7 57.6 48.8 56.4 263 358 218 180 Brazil 93.7 63.0 51.7 62.0 12292 16695 10015 8312 Chile 95.4 72.7 63.8 73.3 410 570 351 285 Colombia 93.4 68.4 59.6 68.2 1086 1451 888 717 Costa Rica 94.5 68.3 58.8 68.0 122 173 105 87 Cuba 91.6 74.6 68.7 74.8 141 217 136 107 Dominican Republic 89.0 63.4 54.1 62.7 222 307 182 153 Ecuador 92.8 64.8 55.2 64.2 451 621 378 311 El Salvador 90.8 66.8 57.8 66.1 110 138 82 68 Guatemala 92.8 63.4 52.8 62.6 286 363 214 180 Haiti 94.4 71.6 60.3 71.1 123 158 88 76 Honduras 91.1 62.3 52.1 60.9 136 168 98 83 Jamaica 90.5 65.2 55.2 64.7 58 76 45 37 Mexico 93.8 69.1 59.4 68.5 3363 4318 2536 2112 Nicaragua 92.4 67.7 59.1 67.2 120 156 93 78 Panama 92.5 66.8 57.7 66.3 108 150 91 74 Paraguay 94.7 67.4 57.4 67.2 136 177 107 89 Peru 93.3 69.8 61.8 69.4 436 604 371 301 Saint Lucia 94.7 66.3 55.7 65.2 2 3 2 2 Suriname 95.7 81.2 75.6 82.2 4 5 3 3 Trinidad and Tobago 96.1 78.1 70.4 78.8 47 66 40 32 Uruguay 95.4 68.9 59.2 68.6 125 175 107 87 Venezuela 92.6 67.7 59.0 67.1 630 858 518 428 37 Table A5 – continued from previous page Year 14-18 18-25 21-25 20-23 14-18 18-25 21-25 20-23 Guatemala 1994 93.5 64.5 53.5 63.5 88 104 61 51 Guatemala 2002 93.5 67.8 57.8 67.2 122 159 93 80 Haiti 1971 94.9 66.9 52.8 66.8 37 45 25 21 Haiti 1982 93.8 67.7 56.3 67.9 9 14 8 7 Haiti 2003 94.3 74.3 64.3 73.6 77 99 55 47 Honduras 1974 92.0 59.4 48.2 58.2 26 31 18 15 Honduras 1988 92.6 64.6 54.7 63.5 41 48 29 24 Honduras 2001 89.9 62.0 52.0 60.5 69 89 51 45 Jamaica 1982 90.9 65.0 53.7 64.2 20 25 14 12 Jamaica 1991 91.7 67.5 57.8 67.6 20 28 17 14 Jamaica 2001 88.5 62.5 53.6 61.6 17 23 14 11 Mexico 1970 94.7 58.2 44.0 56.8 44 51 29 24 Mexico 1990 94.0 66.5 55.3 65.8 958 1191 689 579 Mexico 2000 93.0 66.8 57.4 66.4 1079 1442 869 708 Mexico 2010 94.2 73.3 64.7 72.8 1282 1634 949 801 Nicaragua 1971 93.1 61.8 49.6 60.8 18 20 12 10 Nicaragua 1995 93.5 69.4 60.4 68.8 46 56 33 28 Nicaragua 2005 91.2 68.1 60.5 67.6 56 80 49 40 Panama 1960 91.3 52.8 40.6 52.3 4 5 3 3 Panama 1970 91.7 57.8 46.4 56.5 12 16 10 8 Panama 1980 92.7 65.9 55.1 65.1 19 24 14 12 Panama 1990 93.2 69.9 61.0 69.6 21 30 18 15 Panama 2000 93.3 68.8 60.4 68.4 24 33 21 16 Panama 2010 91.8 68.6 61.0 68.5 29 41 25 20 Paraguay 1962 95.7 63.1 51.9 64.3 6 8 5 4 Paraguay 1972 96.0 66.5 55.3 67.1 20 23 14 11 Paraguay 1982 94.7 67.6 57.9 67.8 27 37 23 19 Paraguay 1992 93.2 61.8 52.0 60.9 32 44 27 22 Paraguay 2002 95.1 72.0 62.5 71.5 51 64 37 32 Peru 1993 94.0 69.5 60.9 69.1 196 267 165 135 Peru 2007 92.7 70.0 62.5 69.6 240 337 206 166 Saint Lucia 1980 95.3 64.5 51.5 63.2 1 1 1 1 Saint Lucia 1991 94.0 67.7 58.6 66.9 1 2 1 1 Suriname 2012 95.7 81.2 75.6 82.2 4 5 3 3 Trinidad and Tobago 1970 97.0 72.1 59.5 72.0 7 8 4 4 Trinidad and Tobago 1980 95.2 73.2 63.2 73.6 12 16 9 8 Trinidad and Tobago 1990 95.8 76.9 69.4 78.0 10 14 9 7 Trinidad and Tobago 2000 96.4 81.4 74.2 81.2 12 15 8 7 Trinidad and Tobago 2011 96.5 84.4 80.1 86.2 8 14 9 7 Uruguay 1963 97.1 70.6 60.0 70.3 16 23 14 11 Uruguay 1975 96.5 67.6 56.0 66.6 19 27 16 13 Uruguay 1985 96.9 67.1 57.5 67.6 19 29 19 15 Uruguay 1996 94.0 69.3 60.5 69.3 23 34 21 17 Uruguay 2006 95.0 74.9 65.7 74.6 21 27 16 13 Uruguay 2011 94.0 65.6 56.5 64.8 25 36 22 18 Continued on next page 44 Table A5 – continued from previous page Year 14-18 18-25 21-25 20-23 14-18 18-25 21-25 20-23 Venezuela 1971 93.7 60.5 48.4 58.8 97 121 71 60 Venezuela 1981 92.8 66.7 57.3 66.1 144 192 115 96 Venezuela 1990 91.7 66.6 57.9 65.8 168 227 137 112 Venezuela 2001 92.7 71.7 64.6 71.7 221 318 195 160 45 C Schooling by cohort In this section, I summarize the education level by country and cohort using data on individuals at least 25 years old. Table A6: Education by cohort cohort mean years less primary primary secondary tertiary Argentina 1950 9.1 17.8 50.6 24.1 7.5 Argentina 1960 10.0 11.3 48.9 31.4 8.4 Argentina 1970 10.8 8.1 45.8 36.1 9.9 Argentina 1980 11.4 7.4 46.3 35.9 10.4 Bolivia 1950 6.3 46.1 30.7 17.4 5.9 Bolivia 1960 7.8 31.4 38.5 22.8 7.3 Bolivia 1970 9.2 22.7 37.7 28.7 10.9 Bolivia 1980 10.7 13.7 34.7 35.6 16.1 Brazil 1950 5.7 58.2 18.0 15.7 8.2 Brazil 1960 6.7 44.6 25.4 21.5 8.5 Brazil 1970 7.2 33.8 28.1 27.9 10.2 Brazil 1980 18.9 28.9 39.0 13.2 Chile 1950 9.2 19.0 47.2 28.7 5.1 Chile 1960 10.1 12.0 45.9 37.3 4.8 Chile 1970 11.3 6.8 39.4 46.2 7.6 Chile 1980 Colombia 1950 6.5 34.9 39.7 18.9 6.5 Colombia 1960 7.5 24.5 42.2 26.6 6.7 Colombia 1970 8.8 18.2 34.6 34.6 12.6 Colombia 1980 9.4 14.1 30.0 43.0 12.9 Costa Rica 1950 7.9 23.1 46.9 18.3 11.7 Costa Rica 1960 8.6 14.7 51.8 19.2 14.3 Costa Rica 1970 8.7 15.1 50.3 17.2 17.4 Costa Rica 1980 9.7 11.4 44.1 20.1 24.3 Cuba 1950 10.4 7.6 46.1 32.6 13.7 Cuba 1960 11.4 2.8 39.7 43.4 14.2 Cuba 1970 11.7 1.9 37.6 46.9 13.6 Cuba 1980 12.3 1.5 24.3 52.1 22.1 Dominican Republic 1950 6.3 50.5 29.0 12.3 8.1 Dominican Republic 1960 8.0 33.5 37.1 17.8 11.7 Dominican Republic 1970 8.6 27.4 39.4 22.3 10.9 Dominican Republic 1980 9.7 19.8 33.6 34.0 12.6 Ecuador 1950 7.4 34.3 39.8 17.7 8.2 Ecuador 1960 8.8 22.4 41.7 26.1 9.8 Ecuador 1970 9.4 16.4 42.7 30.9 10.0 Continued on next page 46 Table A6 – continued from previous page cohort mean years less primary primary secondary tertiary Ecuador 1980 10.2 11.1 39.5 37.2 12.1 El Salvador 1950 5.2 55.6 27.3 12.9 4.2 El Salvador 1960 6.5 45.0 31.7 18.3 5.0 El Salvador 1970 7.5 37.3 33.3 23.0 6.4 El Salvador 1980 8.1 31.5 37.2 25.7 5.6 Guatemala 1950 3.5 71.9 18.5 6.4 3.3 Guatemala 1960 4.5 62.3 24.6 9.1 3.9 Guatemala 1970 5.2 55.2 29.0 11.3 4.5 Guatemala 1980 Haiti 1950 3.0 71.9 21.3 6.0 0.8 Haiti 1960 3.4 67.7 18.5 12.6 1.2 Haiti 1970 5.2 52.6 28.1 18.2 1.2 Haiti 1980 Honduras 1950 4.5 61.8 25.8 9.7 2.7 Honduras 1960 5.4 50.7 33.2 13.0 3.0 Honduras 1970 6.0 41.5 42.2 13.9 2.5 Honduras 1980 Jamaica 1950 9.7 7.4 60.5 29.4 2.7 Jamaica 1960 11.2 2.7 44.7 50.2 2.4 Jamaica 1970 12.4 2.1 20.9 74.5 2.6 Jamaica 1980 Mexico 1950 6.8 37.4 41.8 11.4 9.5 Mexico 1960 8.3 23.6 47.8 17.3 11.4 Mexico 1970 9.2 13.8 54.0 19.9 12.3 Mexico 1980 10.1 9.8 50.3 24.1 15.8 Nicaragua 1950 4.9 59.7 24.4 9.3 6.5 Nicaragua 1960 6.0 48.1 31.6 13.9 6.5 Nicaragua 1970 6.4 42.9 33.2 16.5 7.4 Nicaragua 1980 6.8 39.3 32.7 20.3 7.7 Panama 1950 8.6 21.2 45.5 21.4 11.8 Panama 1960 9.7 12.4 45.0 29.0 13.6 Panama 1970 10.2 11.0 40.8 31.1 17.2 Panama 1980 10.7 8.8 36.7 36.3 18.2 Paraguay 1950 6.2 46.8 39.2 9.9 4.2 Paraguay 1960 7.3 34.1 43.7 16.9 5.3 Paraguay 1970 8.1 26.3 46.0 21.3 6.5 Paraguay 1980 Peru 1950 7.5 38.8 16.8 32.9 11.6 Peru 1960 8.4 28.2 19.2 41.4 11.2 Peru 1970 9.3 16.9 20.8 48.3 14.0 Peru 1980 9.7 11.9 21.2 55.2 11.6 Saint Lucia 1950 9.4 72.3 3.9 20.8 3.0 Continued on next page 47 Table A6 – continued from previous page cohort mean years less primary primary secondary tertiary Saint Lucia 1960 52.7 8.8 35.6 2.8 Saint Lucia 1970 Saint Lucia 1980 Suriname 1950 11.3 69.8 16.1 2.9 Suriname 1960 7.1 70.9 18.5 3.5 Suriname 1970 6.4 66.3 22.7 4.6 Suriname 1980 4.9 57.9 30.8 6.4 Trinidad and Tobago 1950 9.0 15.8 44.6 36.4 3.1 Trinidad and Tobago 1960 10.1 12.1 31.6 52.9 3.4 Trinidad and Tobago 1970 11.5 6.7 20.6 67.8 4.9 Trinidad and Tobago 1980 12.1 5.4 15.9 72.0 6.8 Uruguay 1950 8.9 17.7 53.0 23.3 5.9 Uruguay 1960 9.2 12.1 57.5 22.6 7.8 Uruguay 1970 9.7 11.8 53.0 26.9 8.3 Uruguay 1980 10.2 6.6 54.0 31.9 7.4 Venezuela 1950 7.4 26.0 46.2 25.5 2.2 Venezuela 1960 8.1 18.7 46.3 34.0 1.1 Venezuela 1970 8.6 14.6 43.0 42.1 0.2 Venezuela 1980 48 D Sample size in province-level estimates Figure A5: CDF of the sample sizes used in estimating province-level intergenerational mobility (a) Upward mobility (b) Downward mobility 49 E Variability in province-level estimates Figure A6: Intergenerational educational mobility in LAC: within-country variability 0 .2 .4 .6 .8 LAC Paraguay Peru Nicaragua Panama Brazil Bolivia Suriname Honduras Colombia Argentina Guatemala Venezuela Mexico Dominican Republic Chile El Salvador Ecuador Cuba Costa Rica Saint Lucia Uruguay Trinidad and Tobago Haiti Jamaica Difference between highest and lowest Standard deviation (a) Upward mobility 0 .1 .2 .3 .4 LAC Honduras Nicaragua Haiti Brazil Panama Paraguay Guatemala El Salvador Colombia Ecuador Peru Venezuela Bolivia Dominican Republic Costa Rica Argentina Chile Mexico Uruguay Saint Lucia Jamaica Trinidad and Tobago Suriname Cuba Difference between highest and lowest Standard deviation (b) Downward mobility Notes: The figure reports the extent to which estimates of intergenerational mobility (upward and downward) vary within countries and for all LAC using the difference between the province with the highest level and the province with the lowest level, as well as the standard deviation of mobility at the province level, by country. The underlying estimates of mobility reflect the probability that those whose parents did not finish primary school will complete at least primary education, in the case of upward mobility, and the probability that those whose parents completed primary school will not complete primary education, in the case of downward mobility. LAC displays this statistic after pooling all provinces in the sample. The blue line marks the (unweighted) average mobility at the province level. The red line marks the difference between the highest and lowest mobility at the country level, as reported in Table 1. 50 F District-level estimates Table A7: Summary statistics: district-level estimates of educational IGM upward downward country districts mean median stdev min max Nmin Nmean mean median stdev min max Nmin Nmean Cuba 137 .845 .872 .112 .722 .94 50 58 .012 .01 .007 0 .043 178 726 Uruguay 67 .798 .793 .056 .684 .94 50 151 .046 .043 .022 .003 .098 238 737 Chile 179 .758 .752 .079 .534 .969 68 378 .069 .065 .026 .014 .157 140 1181 Costa Rica 55 .714 .719 .07 .498 .878 110 627 .075 .072 .027 .033 .156 313 1320 Argentina 312 .713 .732 .123 .407 .986 56 756 .066 .054 .035 .013 .194 276 2674 Peru 168 .702 .688 .127 .339 .935 111 857 .097 .081 .053 .016 .342 64 1275 Bolivia 80 .627 .642 .13 .345 .948 179 1114 .111 .104 .059 .027 .317 80 1471 Mexico 2,331 .615 .612 .132 .192 1.133 50 551 .083 .071 .055 -.052 .504 50 702 Ecuador 78 .591 .599 .115 .306 .847 180 1930 .109 .095 .047 .054 .291 244 2915 Panama 35 .588 .593 .153 .253 .803 184 766 .095 .08 .052 .031 .241 152 1706 El Salvador 103 .553 .549 .091 .327 .754 92 459 .177 .168 .068 .043 .383 50 381 Venezuela 157 .52 .513 .103 .255 .746 194 1412 .158 .151 .05 .068 .334 135 1886 Colombia 434 .509 .498 .127 -.043 .88 123 967 .151 .145 .065 .037 .371 133 1076 Paraguay 63 .474 .477 .119 .116 .781 208 1146 .152 .143 .051 .039 .259 96 788 Dominican Republic 66 .462 .463 .082 .301 .667 73 770 .154 .147 .036 .082 .273 94 953 Brazil 2,040 .386 .387 .15 .019 .827 366 2514 .203 .184 .087 .046 .602 65 1089 Nicaragua 68 .361 .373 .11 .138 .582 264 882 .214 .2 .069 .103 .423 51 501 Honduras 96 .355 .346 .109 .112 .576 211 805 .24 .224 .08 .109 .44 52 359 Guatemala 191 .243 .237 .11 .03 .613 286 961 .268 .252 .095 .088 .649 50 329 Haiti 23 .196 .191 .063 .087 .373 845 3559 .412 .426 .087 .221 .569 91 982 total 6,683 .523 .539 .187 -.043 1.133 50 1296 .136 .115 .093 -.052 .649 50 1027 Notes: This table shows summary statistics for district-level estimates of IGM. “Upward” reflects the likelihood that children ages 14–18 whose parents have not completed primary schooling will manage to complete at least primary education. “Downward” reflects the likelihood that children ages 14–18, whose parents have completed primary schooling or higher will not manage to complete primary education. “Total” shows the unweighted summary statistics across all districts. The columns “Nmin” and “Nmean” report the smallest and average sample size, respectively, across districts. Countries are sorted from highest to lowest average level of upward IGM across districts (column “mean”). Districts with less than 50 observations are omitted. 51 G District-level maps of mobility Figure A7: Upward mobility in LAC Notes: Upward mobility reflects the likelihood that children ages 14–18 whose parents have not completed primary schooling will manage to complete at least primary education. This graph uses provinces for St. Lucia, Jamaica, Trinidad and Tobago, and Suriname because these countries do not have a finer administrative units in the data set. 52 Figure A8: Downward mobility in LAC Notes: Downward mobility reflects the likelihood that children ages 14–18 whose parents completed at least primary schooling will not manage to complete primary education. This graph uses provinces for St. Lucia, Jamaica, Trinidad and Tobago, and Suriname because these countries do not have a finer administrative units in the data set. 53 J Estimates of upward IGM using primary-to-secondary education Table A11: Country-level estimates of upward IGM using primary-to-secondary education (1) (2) mobility / N census years upward N age range 19–25 19–25 Trinidad and Tobago 1970,1980,1990,2000,2011 .466 8,506 Peru 1993,2007 .416 131,085 Saint Lucia 1980,1991 .388 1,452 Jamaica 1982,1991,2001 .315 4,304 Bolivia 1976,1992,2001,2012 .237 66,410 Chile 1970,1982,1992,2002 .19 97,017 Brazil 1960,1970,1980,1991,2000,2010 .187 6,142,101 Cuba 2002,2012 .187 4,037 Uruguay 1963,1975,1985,1996,2006,2011 .178 25,192 Argentina 1970,1980,1991,2001,2010 .177 226,100 Dominican Republic 1981,2002,2010 .161 64,387 Panama 1960,1970,1980,1990,2000,2010 .161 23,221 Venezuela 1971,1981,1990,2001 .148 185,993 Costa Rica 1973,1984,2000,2011 .133 28,829 Ecuador 1974,1982,1990,2001,2010 .128 121,410 Colombia 1973,1985,1993,2005 .121 354,007 Mexico 1970,1990,2000,2010 .107 1,008,707 El Salvador 1992,2007 .092 37,462 Paraguay 1962,1972,1982,1992,2002 .085 54,934 Haiti 1971,1982,2003 .073 62,660 Guatemala 1964,1973,1981,1994,2002 .042 125,087 Honduras 1974,1988,2001 .036 52,754 Nicaragua 1971,1995,2005 -.004 47,560 Suriname 2012 -.094 200 mean / total .164 8,873,415 Notes: Column (1) gives upward IGM estimates. It reflects the likelihood that individuals ages 19–25 whose parents have not completed primary schooling will manage to complete at least secondary education. Column (2) gives the number of observations used to estimate the country-specific IGM statistics (individuals with parents whose education is reported in the censuses). Countries are sorted from highest to lowest level of upward IGM (column (1)). “Mean” gives the unweighted average of the 24 country estimates. 60 Table A12: Province-level estimates of upward IGM using primary-to-secondary education upward country provinces mean median stdev min max Nmin Nmean Peru 25 .481 .442 .165 .246 .748 250 5243 Cuba 14 .323 .342 .049 .231 .384 155 309 Bolivia 9 .251 .254 .08 .154 .384 348 7379 Chile 44 .203 .205 .062 .093 .331 114 1694 Dominican Republic 23 .194 .201 .046 .067 .275 588 1795 Costa Rica 7 .166 .167 .042 .12 .244 2051 4118 Argentina 24 .155 .15 .048 .089 .322 219 9421 Mexico 32 .155 .149 .039 .096 .244 2143 31522 Colombia 22 .149 .136 .047 .092 .254 141 16091 El Salvador 14 .148 .145 .044 .097 .27 1209 2676 Venezuela 22 .147 .147 .029 .082 .214 643 8454 Uruguay 19 .144 .139 .027 .094 .191 264 1326 Brazil 25 .14 .128 .05 .074 .249 4716 245684 Ecuador 14 .136 .134 .031 .098 .204 840 8672 Panama 7 .131 .127 .054 .055 .224 457 3317 Nicaragua 12 .107 .114 .051 .04 .19 807 3963 Paraguay 14 .076 .067 .048 .032 .211 1312 4225 Guatemala 22 .049 .047 .021 .012 .106 1614 5686 Haiti 4 .046 .041 .011 .039 .063 4211 15665 Honduras 18 .045 .036 .024 .01 .105 152 2931 total 371 .173 .149 .118 .01 .748 114 23884 Notes: This table shows summary statistics for province-level estimates of upward IGM. “Upward” reflects the likelihood that individuals ages 19–25 whose parents have not completed primary schooling will manage to complete at least secondary education. “Total” shows the unweighted summary statistics across all provinces. The columns “Nmin” and “Nmean” report the smallest and average sample size, respectively, across provinces. Provinces with less than 50 observations are omitted. 61 Table A13: District-level estimates of upward IGM using primary-to-secondary education upward country districts mean median stdev min max Nmin Nmean Peru 168 .395 .365 .175 .084 .807 105 785 Cuba 137 .338 .354 .091 .149 .479 52 81 Bolivia 80 .202 .194 .095 .064 .473 150 830 Chile 179 .2 .192 .082 .04 .478 81 421 Dominican Republic 66 .193 .187 .056 .067 .433 65 635 Costa Rica 55 .179 .181 .06 .057 .304 117 524 Brazil 2,040 .174 .168 .079 -.018 .471 278 1990 El Salvador 103 .155 .139 .076 .023 .381 87 364 Uruguay 67 .15 .144 .065 .024 .355 51 156 Argentina 312 .139 .14 .053 .012 .325 54 727 Colombia 434 .131 .115 .066 -.097 .321 82 816 Venezuela 157 .128 .123 .046 .028 .257 190 1185 Ecuador 78 .126 .115 .049 .026 .241 142 1577 Panama 35 .123 .098 .075 .011 .276 176 663 Nicaragua 68 .11 .101 .054 .017 .232 192 699 Mexico 2,331 .109 .098 .068 -.041 .635 50 452 Paraguay 63 .077 .07 .051 -.067 .211 153 900 Guatemala 191 .043 .039 .03 -.005 .156 172 655 Honduras 96 .037 .032 .028 -.006 .141 152 550 Haiti 23 .034 .033 .024 -.003 .103 664 2724 total 6,683 .144 .127 .093 -.097 .807 50 1042 Notes: This table shows summary statistics for district-level estimates of upward IGM. “Upward” reflects the likelihood that individuals ages 19–25 whose parents have not completed primary schooling will manage to complete at least secondary education. “Total” shows the unweighted summary statistics across all districts. The columns “Nmin” and “Nmean” report the smallest and average sample size, respectively, across districts. Districts with less than 50 observations are omitted. 62 K Transition matrix by country Figure A11: Transition matrix by country 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (a) Argentina 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (b) Bolivia 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (c) Brazil 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (d) Chile 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (e) Colombia 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (f) Costa Rica 63 Figure A12: Transition matrix by country 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (a) Cuba 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (b) Dominican Republic 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (c) Ecuador 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (d) El Salvador 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (e) Guatemala 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (f) Haiti 64 Figure A13: Transition matrix by country 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (a) Honduras 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (b) Jamaica 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (c) Mexico 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (d) Nicaragua 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (e) Panama 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (f) Paraguay 65 Figure A14: Transition matrix by country 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (a) Peru 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (b) Saint Lucia 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (c) Trinidad and Tobago 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (d) Uruguay 0 .2 .4 .6 .8 1 likelihood of child attainment 0 .2 .4 .6 .8 1 fraction by parental attainment Less than primary Primary completed Secondary completed University completed Less than primary Primary completed Secondary completed University completed (e) Panama 66 Figure A15: Upward and downward mobility in LAC compared to Africa South Sudan Mozambique Sudan Ethiopia Malawi Guatemala Burkina Faso Guinea Nicaragua Mali Haiti Liberia Sierra Leona Senegal Rwanda El Salvador Uganda Brazil Dominican Republic Benin Honduras Colombia Morocco Paraguay Lesotho Kenya Peru Zambia Cameroon Togo Saint Lucia Venezuela Suriname Ecuador Ghana Tanzania Mexico Bolivia Nigeria Costa Rica Panama Egypt Cuba Zimbabwe Uruguay Chile Botswana Argentina South Africa Trinidad and Tobago Jamaica 0 .2 .4 .6 .8 Upward mobility (ages 14-18) Africa LAC Jamaica Trinidad and Tobago Cuba Argentina Suriname Mexico Panama Chile Uruguay South Africa Bolivia Botswana Egypt Nigeria Costa Rica Ecuador Venezuela Morocco Peru Paraguay Cameroon Saint Lucia Colombia Zimbabwe Dominican Republic Honduras Ghana Guatemala El Salvador Brazil Tanzania Togo Zambia Kenya Nicaragua Haiti Benin Senegal Mali Burkina Faso Lesotho Ethiopia Uganda Sierra Leona Sudan Guinea Rwanda Malawi Mozambique Liberia South Sudan 0 .2 .4 .6 .8 Downward mobility (ages 14-18) Africa LAC Notes: Estimates for African countries come from Alesina et al. (2021). 67 Figure A16: Highly negative correlation between upward and downward mobility Argentina Bolivia Brazil Chile Colombia Costa Rica Cuba Dominican Republic Ecuador El Salvador Guatemala Haiti Honduras Jamaica Mexico Nicaragua Panama Paraguay Peru Saint Lucia Suriname Trinidad and Tobago Uruguay Venezuela IGM UP = .8199 + -2.8106 IGM down (.0261) (.2211) .2 .4 .6 .8 1 Upward IGM 0 .05 .1 .15 .2 .25 Downward IGM Figure A17: Upward mobility by urban/rural status 0 .5 10 .5 10 .5 1 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 BOL BRA CHL COL CRI GTM HND MEX PAN PRY URY Urban Rural Upward mobility Birth decade Graphs by Country 68 Figure A18: Downward mobility by urban/rural status 0 .2 .4 .6 .80 .2 .4 .6 .80 .2 .4 .6 .8 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 BOL BRA CHL COL CRI GTM HND MEX PAN PRY URY Urban Rural Downward mobility Birth decade Graphs by Country Figure A19: Upward mobility by gender 0 .5 10 .5 10 .5 10 .5 1 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 1940 1950 1960 1970 1980 1990 ARG BOL BRA CHL COL CRI ECU GTM HND MEX NIC PAN PRY 780 URY VEN Male Female Upward mobility Birth decade Graphs by Country 69