REMITTANCES AND INEQUALITY IN RURAL MEXICOPINAR KESKIN AND GUSTAV RANISABSTRACT. In this paper, we investigate the impact of internal and external remittanceson rural income inequality in the context of Mexico, and try to understand how that im-pact evolved over time. Using income decomposition techniques on a Mexican householdpanel carried out every two years, the paper finds that both internal and external remit-tances have a positive effect on rural equality. Although the external remittances becomemore and more equalizing over time, we find that internal remittances are always moreequalizing than external ones. However, after imputations for the potential home lost inthe earnings of migrants, our findings indicate that external remittances actually had beenincreasingly unequalizing for rural Mexican households.Department of Economics, Yale UniversityDate: May 4, 2009.2 REMITTANCES AND INEQUALITY IN RURAL MEXICO1. INTRODUCTIONThe importance of migrant remittances as an external flow to developing countries hasgrown rapidly in the past decade, and the growth of remittances has exceeded that ofprivate capital flows and development assistance in this period. Once overshadowed bydevelopment assistance or foreign direct investment to these countries, as of 2001 remit-tance income in developing countries was 260% of total official development assistanceand counted for more than 40% of the total foreign direct investment (Ratha, 2005).Given the increasing magnitude and the potential growth-enhancing impacts of mi-grant remittances, many economists examined the impact of remittances on the origincountries in recent years. Indeed, these studies investigated many important aspects ofthe remittances-origin country relationship: relaxation of credit constraints, increasing in-vestment in micro enterprises, improved educational attainment of children, etc.1One ofthe key messages of these studies is the fact that migrant remittances can help the receiv-ing households increase income and consumption. Moreover, it’s been well-documentedthat remittances flows can help families to cope with economic shocks.However, little is still known about the impact of remittances on income inequalityin these regions. In fact, many studies attempted to study the remittance-inequality re-lationship in a developing country context. However, most of the time, these studiesoffered conflicting results about the impact of remittances on inequality.2There may be several reasons for these conflicting outcomes. First, each of these stud-ies focuses on different countries, i.e. different economic and social environments. Sec-ond, these researchers have examined the distributional effects of migrant remittancesusing different techniques: with and without endogenous migration costs, by usingincome-source decompositions of inequality measures, by comparing migration outcomeswith no-migration counterfactuals, etc. Finally, although there is a general consensus in1See, for example, Lucas (1987); Rozelle et al. (1999); Hanson and Woodruff (2003); Dustmann andKirchkamp (2002); Woodruff and Zenteno (2007).2See, among others, Lipton (1980); Stark, Taylor and Yitzhaki (1986) ; Adams (1989, 1991); Adams and Al-derman (1992); Taylor and Wyatt (1996); Barham and Boucher (1998); Rodriguez (1998); Taylor et al. (1999);Taylor et al. (2005); Acosta et al. (2008).REMITTANCES AND INEQUALITY IN RURAL MEXICO 3the literature that the impact of remittances on regional inequality varies over time de-pending on the particular stage of migration, most of these studies are static. Thus, theyfocus on a certain point in the migration history of a country.In the present study, we examine the impact of internal and external remittances onrural income inequality in the context of rural Mexico.3Our main contribution to theliterature will be the dynamic setting we use by utilizing a Mexican household surveycarried out by INEGI every two years. We will be able to follow Mexican states overtime, starting from 1992 to 2005 by using this nationally representative dataset to estimatethe marginal effects of both international and internal migrant remittances on incomeinequality among Mexican rural households and how these effects changed over time.Later, we test the robustness of our finding by using an alternative dataset, Progresa.Our study is closest in spirit to the Stark, Taylor and Yitzhaki (1986) paper on remit-tances and inequality. In their paper, the authors argue that migration out of a ruralcommunity has some risks which cannot be totally foreseen by the migrants. Moreover,migration potentially brings a large cost to the budget of a rural household in terms ofinformation gathering, transportation expenses, etc. Obviously, these costs and uncer-tainty will be a lot higher in the case of international migration. Such reasoning makesthe authors believe that earlier migrants tend to come from richer households, and re-mittances sent by these migrants are more likely to increase income inequality in theremittance-receiving communities. However, as time passes, information about the des-tination region is spread in the migrant-sending regions by the help of migrant networks.This information spread dissolves some uncertainty, reduces costs, and potentially re-verses the initial unequalizing impacts of remittances by increasing the rate of migrationamong members of the poorer households. In our work, we will have a chance to testthe hypotheses proposed by Stark, Taylor and Yitzhaki (1986) and see how the impactsof internal and external remittances on rural inequality changed over time.3In the last couple of decades, international migration, in particular migration to the United States, gainedincreasing importance in Mexico. According to Mishra (2007), the share of Mexican population working inthe United States, as a percentage of the Mexican labor force, increased fivefold, from 3% in 1970 to 16% in2000. In the same period, remittance transfers to Mexico also rose rapidly. In 2004, remittances are estimatedto exceed the foreign direct investment inflow of the country and reach 16.6 billion dollars, which is equal to2.5% of GDP. This number is approximately as high as 80% of the crude oil exports of Mexico.4 REMITTANCES AND INEQUALITY IN RURAL MEXICOThe remainder of the paper is organized as follows. Section 2 summarizes the analyti-cal framework we use in our study. Section 3 describes the data used in the analysis andpresents our empirical