We model criminal investigation as a principal-agent-monitor problem in which the agent can bribe the monitor to destroy evidence. Building on insights from Laffont and Martimort’s 1997 paper, we study whether the principal can profitably introduce asymmetric information between agent and monitor by randomizing the monitor’s incentives. We show that it can be the case, but the optimality of random incentives depends on unobserved preexisting patterns of private information. We provide a data-driven framework for policy evaluation requiring only unverified reports. A potential local policy change is an improvement if, everything else equal, it is associated with greater reports of crime.

Cartels participating in procurement auctions frequently use bid rotation or prioritize incumbents to allocate contracts. However, establishing a link between observed allocation patterns and firm conduct has been difficult: there are cost-based competitive explanations for such patterns. We show that by focusing on auctions in which the winning and losing bids are very close, it is possible to distinguish allocation patterns reflecting cost differences across firms from patterns reflecting non-competitive environments. We apply our tests to two datasets: the sample of Ohio milk auctions studied in Porter and Zona (1999, “Ohio School Milk Markets: An Analysis of Bidding”, RAND Journal of Economics, 30, 263–288), and a sample of municipal procurement auctions from Japan.

We study the design of randomized controlled experiments when outcomes are significantly affected by experimental subjects' unobserved effort expenditure. While standard randomized controlled trials (RCTs) are internally consistent, the unobservability of effort compromises external validity. We approach trial design as a principal-agent problem and show that natural extensions of RCTs—which we call selective trials—can help improve external validity. In particular, selective trials can disentangle the effects of treatment, effort, and the interaction of treatment and effort. Moreover, they can help identify when treatment effects are affected by erroneous beliefs and inappropriate effort expenditure.

This paper studies the problem of experiment design by an ambiguity-averse decision-maker who trades off subjective expected performance against robust performance guarantees. This framework accounts for real-world experimenters’ preference for randomization. It also clarifies the circumstances in which randomization is optimal: when the available sample size is large and robustness is an important concern. We apply our model to shed light on the practice of rerandomization, used to improve balance across treatment and control groups. We show that rerandomization creates a trade-off between subjective performance and robust performance guarantees. However, robust performance guarantees diminish very slowly with the number of rerandomizations. This suggests that moderate levels of rerandomization usefully expand the set of acceptable compromises between subjective performance and robustness. Targeting a fixed quantile of balance is safer than targeting an absolute balance objective.