Research

Publication:

“Equilibrium Default Cycles,” 2007, joint with Balázs Szentes,
Journal of Political Economy, Vol. 115, No. 3 (June): pp. 403-446. 

Abstract. This paper analyzes Markov equilibria in a model of strategic lending in which (i) agents cannot commit to long-term contracts, (ii) contracts are incomplete, and (iii) incumbent lenders can coordinate their actions. Default cycles occur endogenously over time along every equilibrium path. After a sequence of bad shocks, the borrower in a competitive market accumulates debt so large that the incumbent lenders exercise monopoly power. Even though the incumbents could maintain this power forever, they find it profitable to let the borrower regain access to the competitive market after a sequence of good shocks. Equilibria are computed numerically, and their attributes are qualitatively consistent with numerous known empirical facts on sovereign lending. In addition, the model predicts that a borrower who accumulates debt overhang will regain access to the competitive credit market only after good shocks. This prediction is shown to be consistent with data on emerging market economies.

Working Papers:

“Debt Contracts with Partial Commitment,” August 2009, submitted

Abstract. This paper analyzes a model of a dynamic lending relationship where the borrower cannot be forced to make repayments and the lender can only commit to short-term contracts. The main result of the paper is that in such an environment, social welfare can be lower than in an environment where the lender cannot commit at all. Providing more incentives for the borrower to stay in the relationship by lowering her outside option can also decrease social welfare. The model predicts that investment size is positively related to liquidity, consistent with empirical findings. Long-run equilibrium dynamics exhibit fluctuations between underinvestment and overinvestment. Extension to long-term contracts is also analyzed.

“Specialization under Uncertainty,” May 2004, joint with Andrei Kovrijnykh

Abstract. We analyze a general equilibrium model with two sectors, sector-specific skills, and stochastic sector-specific productivity shocks. The main focus of this paper is the choice of specialization by the workers. That is: How much sector-specific human capital should a worker acquire? We identify three reasons for less than perfect specialization: 1) risk-aversion, 2) decreasing returns in human capital accumulation, and 3) substitutability/complementarity between outputs. For a simple distribution of shocks, where the realization of the shocks can take one of two values and the shocks are perfectly negatively correlated, we show that there are always some workers who fully specialize in a competitive equilibrium. Furthermore, if the productivity shocks have large enough variance, there will be some workers who acquire both skills. We prove that the competitive equilibrium is generally inefficient, and generates too little specialization compared to the social optimum where the social planner can use transfers among the workers. We also argue that if the planner is not allowed to use these transfers, there will be less specialization in this constrained optimal outcome than in the competitive equilibrium. In order to see how the correlation between the productivity shocks affects the specialization by the workers, we compute the equilibrium skill distribution numerically.

Work in Progress:

“Paying for not Buying: Delegated Search with Private Information Acquisition,” October 2009, joint with Hector Chade

Last modified 10/2009