This dissertation consists of two lines of research. The first line focuses on decision making under uncertainty. Specifically, I analyze the nature of risk preferences in life-insurance choices, and I examine adverse selection and moral hazard in private-healthcare markets in a context where the government provides universal healthcare. The second line focuses on the political economy of institutions, and I study the impact of a multinational company on the economic development of its host country during and after its tenure. In Chapter 1, I estimate a structural model of risk preferences from the choice of insured amount in life-insurance contracts. The literature investigating life-insurance purchases has mostly focused on survey data to estimate how demand for insurance depends on demographic and socioeconomic factors, leaving aside the task of estimating an underlying model of decision making that generates this demand. Using proprietary data on life-insurance choices, I estimate a model of risky choice to explain households' decisions. My results indicate that, in addition to standard risk aversion (decreasing marginal utility for wealth), including decision weights that might differ from the actual probabilities improves the fit of the model. These weights can be interpreted as a combination of state-dependent utility and probability distortions. Moreover, I find support for the existence of heterogeneity in preferences. Women are more risk averse than men, and the decision weights vary depending on gender and age. In Chapter 2, we investigate adverse selection and moral hazard in private healthcare in markets where the government provides universal healthcare. We use proprietary data on health-insurance choices and medical expenditures to examine asymmetric information in this institutional setting. We disentangle adverse selection and moral hazard by leveraging variation in copays and deductibles implemented by the insurance company, and the fact that, for a sample of customers, the insured amount was exogenously assigned, shutting down the adverse selection channel. We find evidence for the presence of asymmetric information, both in the form of adverse selection and moral hazard. Moreover, we develop a model that incorporates sources of heterogeneity that could potentially explain selection in this institutional setting, namely risk aversion, risk type, concern for health, and taste for convenience services offered in the private healthcare sector. In Chapter 3, we analyze the impact of large-scale FDI on economic development by studying the case of the United Fruit Company (UFCo) in Costa Rica from 1889 to 1984. We implement a geographic regression discontinuity design that exploits a quasi-random assignment of land, and census data geo-referenced at the census-block level for 1973, 1984, 2000 and 2011. These allow us to identify the company's effect during its tenure, and after it stopped production. We find that former "UFCo lands" have had higher living standards and better economic outcomes than counterfactual areas without the company's presence, and that convergence is slow. These findings are validated using nighttime lights data. Detailed historical data suggest that the mechanisms behind our results are.