| This dissertation studies the role of asymmetric information in dynamic contracts. Through two chapters, I examine the relationship between constraints on liquidity and the dynamics of adverse selection. In general, these two features combine to generate inefficiency in the optimal contract, and each chapter explores the magnitude of this inefficiency in different dynamic settings. In the first chapter, I characterize the relationship between persistence in private information and liquidity. By contrast, in the second chapter, I highlight the impact of liquidity on dynamic network effects.;More specifically, in my first chapter I study a dynamic adverse selection model where a long- lived agent contracts with a long-lived principal for the provision of a good in each period. The agent is liquidity constrained and possesses private information about her preferences for the principal's good that evolve over time and are autocorrelated. Employing techniques from Fernandes and Phelan (2000), I characterize the optimal contract as the solution to a concave dynamic programming problem, where the state variables are the previous period's reported type as well as on and off path promised expected lifetime utilities. Through analytical results and simulations, I find that the optimal contract backloads rents and that there exists a threshold state after which the contract calls for the efficient provision of the good. Moreover, the optimal contract exhibits short-run distortions in the provision of the good when the state is below the threshold. Notably, I find that the liquidity constraint causes inefficiency that lingers long after it would have been resolved if there were no constraints on the agent's liquidity.;On the other hand, in the second chapter I study a dynamic-adverse selection setting where a long- lived principal contracts with overlapping generations of agents for the provision of a service in each period. Throughout the duration of the contract, each agent's preferences for the service are affected by both their privately observed, i.i.d. characteristic and the characteristic of any other agent contracting with the principal. I adopt the techniques of Krishna, Lopomo, and Taylor (2013) and characterize the optimal contract as the solution to a dynamic programming problem, where the state is promised lifetime utility. I find that the optimal contract is monotone in state, backloaded, and exhibits the no distortion at the top property. Moreover, the optimal contract is biased in favor of older agents. |
