Research

WORK IN PROGRESS

"Risk Sharing, Private Information, and the Use of Fertilizer." (Job Market Paper.)

I study a model of risk sharing in which households' choices of effort and agricultural inputs are private. I relate the level of risk sharing to the households' production decisions. Insurance leads to inefficiently low levels of effort, lowers the use of inputs complementing effort, and increases the use of inputs substituting effort. Moreover, I demonstrate that agricultural input subsidies may improve welfare. Empirically, I provide reduced-form evidence that is consistent with the main predictions of the model using the latest ICRISAT panel from rural India. Then, I structurally estimate the model to quantify the size of the productive inefficiency generated by risk sharing and the welfare gains from a fertilizer subsidy. I show that going from no sharing to full insurance, fertilizer use drops by four times. Finally, I find that cutting the price of fertilizer in half would cause a 8% drop in risk sharing and generate a consumption-equivalent gain in welfare of 47%.

JEL Codes: O12 O13 O33 Q16

"Propensity to Trust and Network Formation." (Joint with Juan Camilo Cardenas, Danisz Okulicz, and Tomás Rodríguez Barraquer.) Paper coming soon.

We study how people's propensity to trust affects the social networks they form relying on an empirical strategy that is immune to reverse causality. We use a combination of survey questions and a standard trust experiment to measure the propensity to trust of 72 members of a cohort of first-year undergraduates before they have a chance to meet. After four months, we elicit five different social networks among the students. We estimate social network formation models for each of the networks elicited to identify how the different measures of trust affect link formation. We control for a large set of observables, including many individual and dyadic traits which are known to play a significant role in network formation. We find that trust poorly explains the formation of the networks we retrieve. In particular, the effect of homophily in socioeconomic background can go so far as being one order of magnitude bigger than the effect of trust.

JEL Codes: C80 D85 D90 D91 Z13

"Better Not to Know: Uncertainty and Coalition Formation." New version of the paper coming soon.

How does uncertainty about the gains from trade affect coalition formation and welfare? Two agents can agree to form a trading coalition. The agents hold a common prior belief about whether the other is a lemon or a peach. Each agent prefers trading with a peach to autarky, but would stay in autarky rather than trading with a lemon. A utilitarian social planner can draw a noisy public signal of whether the agents are lemons before they might agree to trade. Drawing a signal can decrease expected welfare. Moreover, the relationship between the welfare gain of drawing a signal and the noise of the signal can have at most one discontinuity and be non-monotonic. I characterize how the welfare gains of drawing a signal change as a function of noise when the prior probability that an agent is peach is sufficiently high.

JEL Codes: C71 D80 D83 D85

"Uncertainty and Network Formation: Efficiency and Stability." (This is an application of "Better Not to Know: Uncertainty and Coalition Formation" to networks.) New version of the paper coming soon.

I study how uncertainty about the benefits of being connected to others affects the way in which a group of agents form a network through bilateral agreements. Relationships offer benefits that deteriorate with “distance” and maintaining a direct relationship is costly. Each agent can be either a peach or a lemon, and being connected to a peach offers more benefits than being connected to a lemon at the same distance. The agents hold a common prior belief about whether any other agent is a lemon. A utilitarian social planner can draw a publicly observable noisy signal of whether the agents are lemons before they form a network. Drawing a signal can decrease expected welfare. Moreover, the relationship between the welfare gain of drawing a signal and the noise of the signal is discontinuous and possibly non-monotonic.

JEL Codes: C71 D80 D83 D85