UC Berkeley

This dissertation consists of three chapters that analyze the effects of social development programs on productivity, risk management strategies, and energy consumption among the poorest population in Mexico. Weather shocks have important negative impacts on poor rural households' livelihood as they are not only closer to subsistence and more vulnerable but also depend on the weather for survival. Nonetheless, due to high administrative costs and information problems insurance markets tend to leave this part of the population unprotected. Similarly, poor rural households usually make use of cheap yet inefficient and potentially harmful sources of energy for cooking, lighting, and heating their homes. This situation does not only affect their health and daily activities, but also keeps them trapped in poverty. In the following chapters I discuss several ways in which government action can in fact improve this population's well-being.

The first chapter entitled "Drought and Retribution: Evidence from a large scale Rainfall-Indexed Insurance Program in Mexico" studies the effects of the recently introduced rainfall-indexed insurance on farmers' productivity, risk management strategies, and per capita income and expenditures in Mexico. Weather shocks are a major source of income fluctuation and most of the world's poor lack insurance coverage against them. In addition, the absence of formal insurance contributes to poverty traps as investment decisions are conflicted with risk management decisions: risk-averse farmers tend to under-invest and concentrate in the production of lower yielding yet safer crops. Recently, weather-indexed insurance has gained increased attention as an effective tool providing small-scale farmers coverage against aggregate shocks. However, there is little empirical evidence about its effectiveness. According to the Ministry of Agriculture, 80 percent of agricultural catastrophic risk in Mexico stems from droughts. Therefore, in 2003 it implemented weather-indexed insurance as a pilot in five counties in the Mexican State of Guanajuato, and by 2008 it already covered almost 1.9 million hectares representing 15 percent of rain-fed agricultural land. The main identification strategy takes advantage of the variation across counties and across time in which the insurance was rolled-out. We find that insurance presence in treated counties has significant and positive effects on maize productivity. In fact, we find that insurance presence at the county level increases maize yields by more than 5 percent. Similarly, we find that insurance presence at the county level has had a positive effect on rural households' per capita expenditure and income of a magnitude close to 8 percent. However, we find no significant relation between insurance presence and the number of hectares destined to maize production.

The second chapter entitled "Voters Response to Natural Disasters Aid: Quasi-Experimental Evidence from Drought Relief Payment in Mexico" estimates the effect of a government climatic contingency transfer allocated through the recently introduced rainfall indexed insurance on the 2006 Presidential election returns in Mexico. Using the discontinuity in payment based on rainfall accumulation measured on local weather stations that slightly deviate from a pre-established threshold, we show that voters reward the incumbent presidential party for delivering drought relief compensation. We find that receiving indemnity payments leads to a significant increase in average electoral support for the incumbent party of approximately 7.6 percentage points. Our analysis suggests that the incumbent party is rewarded by disaster aid recipients and punished by non-recipients. This chapter provides evidence that voters evaluate government actions and respond to disaster spending contributing to the literature on retrospective voting.

The third and final chapter entitled "Conditional Cash Transfers schemes and Households' Energy Response in Mexico" analyzes the relationship between income and energy use in poor households in Mexico using household expenditure surveys that were collected to evaluate the poverty alleviation program "Oportunidades". We argue that Oportunidades cash transfers provide an income shock that is exogenous to a household's energy demand, allowing us to estimate short-run and long-run income elasticities for energy use. Short-run estimates hold household's appliance stock constant and long-run estimates model the household's decision to acquire new appliances. As a general estimation strategy households' fixed-effects are included. We also use instrumental variable estimation and a matching difference-in-differences estimator to check for robustness and correct for pre-selection unbalances between treatment and control groups. Results suggest significant differences between long-run and short-run elasticities as households emerging from poverty become first-time purchasers of energy-using appliances. In particular, we find small and not significant effects of cash-transfers on short-run energy consumption expenditure, but find significant and important effects of cumulative conditional cash-transfers on appliance acquisition (i.e. refrigerators and gas stoves). This has important policy implication since poverty alleviation programs like Oportunidades conditional cash transfers program, although not evident in the short run, have significant effects on energy demand.