| The first two chapters of this dissertation examine some interactions between social protection, social security, and employment decisions. The case study is Mexico, in order to exploit the availability of its data, along with policy interventions that allow for exploring different dimensions of the aforementioned interactions. In the first chapter, I estimate that Mexico City's Universal Pension, which was unexpected and is neither means nor asset-tested, has on average decreased by 5 percentage points the labor force participation rate of Mexico City's male elderly. This change represents approximately 60 percent of the decline in this rate that took place between the periods prior to and following its introduction. This result is relevant to understanding the consequences of policies that developing countries have been implementing to deal with faster aging and the increased strain on informal sources of old-age assistance.;The second chapter looks at a federal program in Mexico that provides health care for informal workers. I analyze the wage and employment effects for male household heads of Seguro Popular, a program introduced between 2002 and 2005 that extended health-care benefits to informal workers. Before Seguro Popular, which is free for poorer workers and inexpensive for the rest, informal workers lacked any type of health insurance. I find that Seguro Popular made informal employment more attractive, mainly for younger and less educated workers, increased the wages of formal workers, and decreased the wages of informal workers. The estimated positive effect on the share of informal workers is of 1 percentage point. Seguro Popular had a large effect on the informal wage premium, by decreasing average informal wages by 9 percent and increasing average formal wages by 5 percent. This program, instituted at a time of rising oil revenues, thus subsidizes the informal sector.;The third chapter leaves behind the Mexican setting, but continues with the topic of decision-making by senior citizens. In particular, this chapter, which is joint work with Rosa L. Matzkin, analyzes how donation decisions are made between elderly husbands and wives in the US, a country where charitable giving represents nearly 2 percent of GDP. To do so, we propose the nonparametric identification and estimation of a collective household decision model with nonadditive unobserved heterogeneity. The estimation combines recent developments in nonparametric identification and estimation of nonseparable models with a novel approach to estimate monotone multiple-index models. Using the University of Michigan Health and Retirement Survey, we estimate a couple's unobserved taste heterogeneity, the wife's distribution of power index as a function of observable characteristics of the couple, individual demands for donation-making, and the common marginal utility from other consumption. From this, we learn that wives' distribution of power index decreases with observables that make them more dependent from their husbands and that wives' demands for donations are more responsive than their husbands' to analogous changes in their observed and unobserved heterogeneity. From the perspective of a fictitious fundraiser, this would imply, for instance, that he should target couples with younger and more educated wives who enjoy relatively good health. Our framework could easily be extended to accommodate individual demands for a larger range of goods. The empirical challenge, in any case, consists of finding suitable variation in preferences affecting plausibly only individual demands of each spouse. |
