(2018). [Job Market Paper]
Abstract: Climate change has increased the incidence and severity of temperature and precipitation shocks affecting agricultural production. Observed levels of adaptation remain low, suggesting that rural households are constrained or adaptation decisions are sub-optimal. I integrate panel socioeconomic and demographic data from rural Mexico with high temporal and spatial resolution weather data to assess if individuals adapt to the heat-induced crop losses of neighboring households via anticipatory (ex ante) labor responses. I instrument for the proportion of catastrophic crop loss reports in a community with exogenous variation in extreme daily temperatures to obtain estimates of ex ante migration and local labor reallocation for households who have not experienced recent crop shocks but observe the heat-induced crop losses of others. I find evidence of domestic migration in anticipation of crop shocks, particularly among females and households with lower land-labor ratios. I also show evidence of local labor reallocation onto household land (agricultural self-employment), especially among males and households with higher land-labor ratios. This study highlights the substantial influence of the environment-agriculture mechanism, the salience of anticipatory adaptation, and the relevance of learning from others in the context of climate risk. These findings have important implications for the design and targeting of rural climate change mitigation programs, suggesting that adaptation gaps are likely overstated and that rural households have different capacities to mitigate the risks associated with climate shocks.
(2018) with B. L. Barham. [Revise and Resubmit @ Journal of Population Economics]
Abstract: We explore how migration and return migration shape occupation outcomes for rural Mexicans. We extend the Roy Model to account for unobserved migration motives and link it to a novel empirical method – a Mixed Nonlinear Endogenous Switching Regression. This approach exploits non-linear outcomes to generate estimates of the direct impacts of migration and return migration on occupation outcomes while controlling for selection associated with unobserved heterogeneity. This combination is not recoverable using a standard sample selection method or instrumental variables. Occupation outcomes are directly and positively influenced by both migration and return migration, especially for females. We also provide evidence of negative selection driven by unobserved heterogeneity for both genders, which we attribute to non-pecuniary motives such as family reunification or forced returns. These results highlight the positive welfare effects of labor mobility for the rural poor, particularly women, and conversely the losses associated with forced returns.
(2018) with A. Dillon.
Abstract: This study investigates the role that initial household endowments have on future asset growth in northern Nigeria using 20-year panel data of rural, agricultural households. We find no evidence of poverty traps using parametric, non- and semi-parametric specifications for land, capital, and durable asset measures over a relatively long time horizon. Though we find no evidence of poverty traps, asset growth is slow over our study period and is driven primarily by agricultural land extensification. A secondary result is that distinct definitions of tracking rules that underlie panel surveys do lead to different asset accumulation trajectories, but no particular tracking rule is consistent with the poverty trap hypothesis.
(2018) with van Rijn, J., and B. L. Barham. [Under Review]
Abstract: This study uses a dictator game with a charitable organization as the donation recipient to test whether empathic concern explains persistent gender differences in charitable giving. We first explore whether we can evoke empathic concern by varying the content of a real-world charitable appeal video that highlights children’s stories of struggle with access to clean water. Then we examine whether the evoked feelings help explain gender differences in donations. Despite no gender differences in donation behavior in a baseline control group, we find that females donate 63% more than males in treatments that include the personal stories from children. These treatment videos increase self-reported feelings of empathic concern among both males and females relative to the control; however, empathic concern that results from the treatment videos increases average donations among females but not males. Causal mediation methods show that empathic concern explains up to 17% of the observed gender differences in giving. While the treatments also evoke other emotions in addition to empathic concern, none of these helps to explain observed gender differences in donations. In fact, we find no significant effects (positive or negative) of the treatments on male donations. Our study sheds light on the role of children’s personal stories and the empathy evoked by them in explaining the gender-donations gap found in the literature.
The Impact of Conditional Cash Transfer Programs on Indigenous Peoples: A Quantitative Case Study of Mexico
(2018) with S. Roy.
Abstract: Conditional cash transfers (CCTs) are widely used antipoverty measures in Latin America, and many such programs target indigenous beneficiaries. However, concerns have been raised that the indigenous poor, who have historically been marginalized, may not benefit from CCTs as much as the nonindigenous population, owing to cultural as well as geographic factors. Critics have gone so far as to suggest that CCTs may even result in increased inequality. We assess these issues in the context of PROGRESA (Programa de Educación, Salud, y Alimentación), an integrated approach to poverty alleviation in Mexico, in which over one-third of beneficiaries were indigenous at the program’s inception in 1998. A feature of the program’s initial targeting was that indigenous and nonindigenous beneficiaries were drawn from geographically similar areas, minimizing the potential for geographic factors to lead to differential impacts. Despite an extensive literature showing positive average impacts of PROGRESA on health and education outcomes, few studies have disaggregated these effects by indigenous status. Using the randomized assignment of initial program rollout, we estimate PROGRESA’s impacts on a range of health and education indicators, specifically for the indigenous and nonindigenous populations. We find that, as of November 2000, PROGRESA had significant impacts on many health and education indicators among both indigenous and nonindigenous households in our sample; in addition, in aggregate across most indicators, these impacts are extremely similar. Our results indicate that if geographic disadvantage for indigenous households can be minimized (a nontrivial endeavor), cultural factors may not pose an intrinsic barrier to the indigenous population benefiting from CCT programs. As such, CCTs can promote human capital accumulation among both indigenous and nonindigenous households. Furthermore, the findings suggest that fears of CCT induced inequality in health and education due to PROGRESA are unfounded in Mexico and more general criticisms of CCTs on the grounds of widening human capital disparity are likely overstated.
(2011) with Karamba, W. and P. Winters. Food Policy 36(1): 41-53.
(2011) with Zezza, A., P. Winters, B. Davis, G. Carletto, K. Covarrubias, K. Stamoulis, P. Karfakis, L. Tasciotti, S. DiGiuseppe and Genny Bonomi. The European Journal of Economic Development 23(4): 569-597.
(2010) with Davis, B., P. Winters, G. Carletto, K. Covarrubias, A. Zezza, K. Stamoulis, C. Azzarri and S. DiGiuseppe. World Development 38(1): 48-63.
(2009) with Winters, P., B. Davis, G. Carletto, K. Covarrubias, A. Zezza, C. Azzarri and K. Stamoulis. World Development 37(9): 1435-1452.
Works in Progress
Anticipatory Behavior as Adaptation to Climate Change?
(2018) with Nobles, J., F. Riosmena and R. Nawrotzki.
A Comparison of Bargaining Power Measurement Methods
with Klein, M., B.L. Barham, and K. Wu.
Addressing Young Child Mortality in Uganda through Food-Based Micronutrient Interventions
with R. Frattarola Hernández.