Research

Publications


Working Papers

"Simple Market Structures are Incomplete" (with Johannes Gierlinger) R&R International Economic Review [pdf]

We study how risk is shared using simple assets, each of which pays on a specific subset of payoff-relevant variables. We show that a market composed of simple assets is incomplete – no matter how many or which assets exist– because it can never address the joint realizations of some distinct risks. We also show that this inability to refine trades generates spill overs between agents with different types of financial constraints. Agents of type i are exposed to uninsurable income risk unless there exists another type j who can condition on all variables affecting i’s income

"Incentive Contracts and Peer Effects in the Workplace" (with Nicolas Oviedo) BSE Working Paper Series 1429. [pdf]

Risk-averse workers in a team exert effort to produce joint output. Workers’ incentives are connected via chains of productivity spillovers, represented by a network of peer-effects. We study the problem of a principal offering wage contracts that simultaneously incentivize and insure agents. We solve for the optimal linear contract for any network and show that optimal incentives are loaded more heavily on workers that are more central in a specific way. We conveniently link firm profits to network structure via the networks spectral properties. When firms can’t personalize contracts, better connected workers ex- tract rents. In this case, a group composition result follows: large within-group differences in centrality can decrease firm’s profits. Finally, we find that modular production has important implications for how peer structures distribute incentives.

“Learning, Sorting, and Turnover in Unstable Environments”  [pdf]

I develop a framework where workers and firms learn continuously about the worker’s productivity type, which itself fluctuates randomly and continuously. Workers receive a competitive spot wage and must decide when to switch between firms, given their belief about their own type. Under supermodularity in production, I show that positive assortative matching (PAM) extends beyond the stable environment of Eeckhout & Weng (2010) to a situation of residual uncertainty that exhibits periods of unlearning, defined as increasing levels of uncertainty along a match. I show that risk-neutrality of workers and skills evolving as a martingale are sufficient to retain PAM. I then extend this setup to allow workers to exert a level of effort subject to classical carreer concerns (a la Holmstrom) and find that PAM can only be sustained under strong parameter restrictions.

“Partial Insurance Networks: Food Sharing in Tsimane' Villages” (with Michael Gurven, Paul Hooper, and Hillard Kaplan )  [pdf]

We empirically analyze the Pareto-optimal risk sharing transfers subject to local information constraints analyzed in Ambrus, Gao, and Milán (2017). Using data from Tsimane’ indigenous communities in rural Bolivia, we find that the observed exchanges across households match the efficient transfers predicted by the theory, and that the theory can account for the deviation from full insurance observed in the data. We propose a reinterpretation of the standard empirical results of risk sharing, predicting household heterogeneity in response to income shocks. We show that this network-based variation in consumption behavior is borne out in the data, and that it can be interpreted economically in terms of consumption volatility.


Work in Progress  

“Local Risk Sharing with Hidden Income”  (with Johannes Gierlinger)

We analyze a model of dynamic insurance with hidden income and local contracts. Households may share risk only with their direct neighbors along a fixed and exogenous network of connections. Therefore, while indirect neighbors cannot directly share a household’s risk, they may nonetheless lower the cost of providing truth-telling incentives in the future, thus improving overall levels of insurance. We argue that the diffusion of information along the network determines the shape of optimal contracts and the type of stylized patterns of dynamic dependence in consumption. We provide testable predictions on how the location of an income shock affects current and future consumption throughout the network, and we test these unique dynamic relationships using data on village economies in rural Bolivia.