Designing Menus of Contracts Efficiently: The Power of Randomization

We study hidden-action principal-agent problems in which a principal commits to an outcome-dependent payment scheme (called contract) so as to incentivize the agent to take a costly, unobservable action leading to favorable outcomes. In particular, we focus on Bayesian settings where the agent has private information. This is collectively encoded by the agent's type, which is unknown to the principal, but randomly drawn according to a finitely-supported, commonly-known probability distribution. In our model, the agent's type determines both the probability distribution over outcomes and the cost associated with each agent's action. In Bayesian principal-agent problems, the principal may be better off by committing to a menu of contracts specifying a contract for each agent's type, rater than committing to a single contract. This induces a two-stage process that resembles interactions studied in classical mechanism design: after the principal has committed to a menu, the agent first reports a type to the principal, and, then, the latter puts in place the contract in the menu that corresponds to the reported type. Thus, the principal's computational problem boils down to designing a menu of contracts that incentivizes the agent to report their true type and maximizes expected utility. Previous works showed that, in Bayesian principal-agent problems, computing an optimal menu of contracts or an optimal (single) contract is APX-hard, which is in sharp contrast from what happens in non-Bayesian settings, where an optimal contract can be computed efficiently. Crucially, previous works focus on menus of deterministic contracts. Surprisingly, in this paper we show that, if one instead considers menus of randomized contracts defined as probability distributions over payment vectors, then an optimal menu can be computed in polynomial time. Besides this main result, we also close several gaps in the computational complexity analysis of the problem of computing menus of deterministic contracts. In particular, we prove that the problem cannot be approximated up to within any multiplicative factor and it does not admit an additive FPTAS unless P = NP, even in basic instances with a constant number of actions and only four outcomes. This considerably extends previously-known negative results. Then, we show that our hardness result is tight, by providing an additive PTAS that works in instances with a constant number of outcomes. We complete our analysis by showing that an optimal menu of deterministic contracts can be computed in polynomial time when either there are only two outcomes or there is a constant number of types.