Experimentally Identifying Constraints to Risk Sharing: Separating Limited Commitment, Moral Hazard and Hidden Income

National Bureau Of Economic Research Inc, Cambridge MA

Investigators

Abstract

Households in developing countries face a lot of income risk, yet formal credit and insurance markets often do not exist. To deal with this risk, households participate in informal risk sharing, making transfers between households. This "informal risk sharing" has been one of the most widely studied topics in development economics. While early work focused on the surprising effectiveness of village informal risk sharing, more recent literature has shifted to understanding specific constraints that may be barriers to informal insurance. However, there is little evidence of the relative importance of such constraints. In order to effectively design policies to address household income risk it is important to understand which constraints to informal insurance are important, and also, how such constraints may interact with one another. Our project will design an experiment that lets us look at constraints individually as well as together. Specifically, research has considered three main frictions: (1) hidden income, (2) moral hazard and (3) lack of commitment. Our project studies the joint interaction of each type of friction. First, we will develop a risk-sharing model with all three frictions and characterize the first order equations together with the various multipliers. This gives us a set of modified (inverse) Euler equations that we can take to the. Second, we will conduct a field experiment wherein individuals face job opportunities but the job opportunities differ in terms of whether the wage is known to one's risk sharing partners, whether there is a bonus, and whether the bonus feature is known to one's risk sharing partner. The theory allows us to characterize the relationship between today and tomorrow's consumption in terms of (inverse) Euler equations with Lagrange multipliers. The experiment, by exogenously varying the degree to which a given friction can matter, gives us the power to measure the extent to which various constraints are binding in this framework.