Research Areas
  • Development economics
  • Health economics
  • Education economics
  • Labor economics

Contact information



Blog Index
The journal that this archive was targeting has been deleted. Please update your configuration.
Disasters can have long lasting effects, but understanding the breadth, variety and longevity of their effects can be challenging. This paper examines the long term effects and subsequent intergenerational transmission of exposure in childhood to the natural disasters that have occurred in Latin America in the last 100 years. The identification strategy exploits the exogenous variation in geographic location, timing and exposure of different birth cohorts to natural disasters. This study measures individuals' exposure to each disaster based on their geographic location at birth to avoid any bias in the estimates due to possible selective migration caused by each disaster. The main results indicate that children in utero and young children are the most vulnerable to natural disasters and suffer the most long-lasting negative effects. These effects include less human capital accumulation, worse health and fewer assets when they are adults. Effects are found to have a non-linear relationship with the level of development of each country. Furthermore, the results provide evidence of the intergenerational transmission of shocks, indicating that children born to mothers who had been exposed to natural disasters also have less education and increased child labor.
Over the past years, a greater number of studies has been conducted to evaluate and measure the impacts of shocks in early childhood. Additionally, there has been a growing concern among economists and policy makers about how negative conditions experienced early in life may have persistent effects or long lasting effects. In this context, the present investigation estimates the long-run effects of shocks in early childhood and their intergenerational transmission.  In particular, we estimate the long-run effects of the 1970 Ancash earthquake on human capital accumulation for the affected and subsequent generations 37 years after the shock. We exploit the localized nature of the earthquake and the exogenous timing of this event to capture a child’s exposure and identify the effect of the earthquake on welfare. The main finding of this paper shows that in utero males exposed to the earthquake completed on average 0.5 years less schooling than their unaffected cohorts, while exposed females completed 0.8 years less schooling. Surprisingly, children whose mothers were affected at birth by the earthquake have 0.4 less years of education while those whose fathers were affected by the earthquake at birth suffer no effects on their educational achievement. The evaluation of other outcomes also suggests that the level of welfare of affected individuals was negatively impacted in the long run.

Evidence documenting the linkages between migration and climate at regional scale is limited. Knowledge on the matter is particularly important for Central America and the Caribbean, a region of the world characterized by exceptionally high (internal and international) migration rates and substantial exposure to disasters. We link individual-level information from multiple censuses for seven countries with georeferenced climate data at the province level to measure the impact of heat exposure on internal mobility. Our results imply that a 1-standard deviation increase in heat would affect the lives of 7,314 and 1,578 unskilled, young (15-25) women and men, respectively. The total effect is slightly smaller than observed in our previous work which focuses on displacement from droughts and hurricanes, but could increase with climate change. Of notable importance is youth facing heat waves are more likely to respond by moving to urban centers than when exposed to disasters endemic to the region. Additional research is warranted over the welfare implications of these choices in the long term and the interventions available to minimize distress migration.

While evidence on the linkages between migration and climate is starting to emerge, the subject remains largely under-researched at regional scale. Knowledge on the matter is particularly important for Northern Latin America and the Caribbean, a region of the world characterized by exceptionally high migration rates and substantial exposure to natural hazards. We link individual-level information from multiple censuses for eight countries in the region with natural disaster indicators constructed from georeferenced climate data at the province level to measure the impact of droughts and hurricanes on internal mobility. We find that younger individuals are more likely to migrate in response to these disasters, especially when confronted with droughts. Youth exhibit a stronger inclination towards relocating to rural and small town settings, motivated possibly by opportunities for nearby off-farm employment and financing limitations for urban transport and living expenses. Migration decisions are mediated by national institutional arrangements. These findings highlight the importance of social protection and regional planning policies to reduce the vulnerability of youth to droughts in the future and secure their economic integration.

This paper addresses an important source of variation within democracies — the degree of institutionalization. The concept of institutionalization describes the extent to which politics takes place, and is believed to take place, via formal political institutions. Countries vary in their degree of institutionalization, hence, in the degree to which political actors pursue their goals via conventional politics or via “alternative political technologies”. This paper postulates that if politics is conducted largely outside of formal channels, the structure of the formal channels should not matter much as a determinant of policy outcomes. To address this issue this paper proposes a new index of institutionalization and with it revisits seminal work regarding the impact of constitutions on public spending. The findings show that the effect of constitutional rules on policy outcomes is conditional on the degree of institutionalization.

We approach the problems of measuring the dimensionality of welfare and that of identifying the multidimensionally poor, by first finding the poor using the original space of attributes, and then reducing the welfare space. The starting point is the notion that the `poor' constitutes a group of individuals that are essentially different from the `non-poor' in a truly multidimensional framework. Once this group has been identified through a clustering procedure, we propose reducing the dimension of the original welfare space using recent blinding methods for variable selection. We implement our approach to the case of Latin America based on the Gallup World Poll, which contains ample information on many dimensions of welfare.

This chapter reviews recent methods to quantify the dimensionality of welfare and its relation to deprivation. We discuss two alternative strategies based on factor analytic methods and on variable selection after cluster analysis. Unlike latent variable methods, variable selection strategies are immediate to interpret and resample, since they choose variables originally in the data set. The advantages and disadvantages of both strategies are discussed as well as some recent empirical applications of these methods. The methods discussed in this chapter are shown to be able to summarize an initially large list of variables into a few new variables (as in factor analytic methods) or a subset of the original ones (as in feature selection / cluster methods), that can serve the purpose of characterizing the poor. These methods can assist the conceptual search for relevant dimensions of welfare, or provide confirmatory analysis of alternative, likely multidisciplinary studies aimed at isolating relevant factors for poverty analysis.

Working Papers

When shocks such as natural disasters occur in early childhood, they can have lasting health and economic effects on the lives of the affected kids that can be transmitted to the next generation. This paper uses household survey data from Tanzania to estimate the short and long term effects of exposure to the Tanzania Flood of 1993 on the health of young victims and to show their intergenerational transmission. The identification strategy exploits exogenous variation in the disaster's geographic extent, timing, and the exposure of different birth cohorts to the disaster. The children exposed to the flood have lower height-for-age Z-scores three years after the shock, with bigger effects for girls than for boys. Women who were less than 18 years old during the flood experienced negative impacts even 12 years after the flood. The kids of women affected before their 18th year have lower height-for-age Z-scores, while the kids of the men affected before their 18th year experience no effect on their height-for-age Z-scores. One of the main mechanisms identified for the intergenerational transmission of health effects is the poor performance of the affected females in the marriage market. The effects are robust to code those individuals who migrate as residents of the affected region at the time of the flood.

This is the first paper using household survey data from two countries (Eritrea and Ethiopia) involved in an international war to measure a conflict’s impact on children’s health in both nations. The identification strategy uses event data to exploit exogenous variation in the conflict’s geographic extent and timing and the exposure to the fighting of different birth cohorts while in utero or early childhood. War-exposed children in both countries have lower height-for-age Z-scores, with the children in the war-instigating and losing country (Eritrea) suffering more than the winning nation (Ethiopia). The paper uniquely incorporates GPS information on the distance between survey villages and conflict sites to more accurately measure a child’s war exposure; results indicate negative impacts are 35-75% larger than if exposure is measured at the imprecise region level. Effects are robust to including region-specific time trends, alternative conflict exposure measures, and addressing potential bias due to selective migration.

  • Do social programs reduce poverty duration?  An application for the Supplemental Security Income Program. Under Review.
This paper evaluates the impact of the Supplemental Security Income (SSI) program on poverty duration.  We use a duration model to estimate how the program affects the hazard rate of leaving poverty.  We deal with the endogeneity problem in poverty duration models by applying an instrumental variable approach.  For a more robust causality analysis,  an instrument variable approach is used to determine the effect of the program on the probability of being poor and also on income not related to the subsidy.  The results indicate that the SSI program reduces poverty duration by 25.5%, increases income not related to the subsidy and reduces the probability of being poor.  Our theoretical framework and our empirical results suggest that cash transfer programs  are  an  efficient  poverty  alleviation  policy  for  breaking  the  persistence  of poverty traps.
Work in Progress
  • Life-Long Effects of Gestation during Ramadan: Evidence from Nigeria (with Seyed Karimi), in progress
We use a different type of natural experiment, namely exposure to the Islamic fasting month of Ramadan, in order to measure the scope of the effects of malnutrition in utero. To uncover the effects in the life-long range, we measure a wide array of adverse effects from birth and childhood to school age and adulthood and examine health, education, and labor market outcomes. Testing the wide range of outcomes allows us to investigate both cognitive and physical channels of adverse labor market effects of malnutrition in utero. A unique feature of this paper is using a uniform socioeconomic context, Nigeria, to measure the effects that are specific to the different stages of the life. This property, which is missing in the related literature, enables us to provide a complete and proportionate picture of the consequences of Ramadan-induced malnutrition in utero.
  • Medium term natural disasters effects: Household income, individual health and child schooling after the 2005 Atlantic Hurricane Season in Mexico (with Maria Cucagna), in progress
This paper analyzes the medium term effects (5 years after the shock) of the hurricanes on income poverty, health and schooling. In this paper, the identification strategy uses household GPS information to compare those households within the hurricanes’ trajectories with those households outside of the trajectories. The fact that the hurricanes passed through different states allows us to obtain a treatment group that it is not necessarily related to one specific region, thus separating region specifics from the effect of the hurricanes. I find that five years after the hurricanes, exposed households are poorer, while exposed individuals have a greater probability of disability or mental health problems. Moreover, I find that school age children have lower rates of school enrollment.
  • Effects in the Long Run of the Salvadoran Civil War (with Pablo Acosta), in progress
This paper studies how exposure to the Salvadoran armed conflict subsequently affected the education of individuals exposed in utero. The identification strategy of this paper compares the educational performance of individuals born in the war-affected regions with that of those that born in non-affected regions. We exploit the exogenous variation in the conflict’s geographic extent and timing to identify exposure to the fighting of different birth cohorts while in utero or early childhood.