The Winner’s Curse in Data-Driven Decision-Making

Abstract:
Data-driven decision-making relies on credible policy evaluation: we need to know whether a learned policy truly improves outcomes. This talk examines a key failure mode—the winner’s curse—where policy optimization exploits prediction error and selection, producing optimistic, often spurious performance gains.

First, we show that model-based policy optimization and evaluation can report large, stable improvements even when common “reassurances” from the literature hold: training data come from randomized trials, estimated gains are large, and predictive models are accurate, well-calibrated, and stable. We give theoretical constructions where true improvements are zero yet predicted gains are substantial. We illustrate these pitfalls in a simulation study inspired by refugee matching, where widely-used model-based evaluation projects large employment gains of over 60% even when the ground truth effect is zero.

Second, we argue that avoiding this optimism pushes us toward model-free off-policy evaluation—but its variance can be prohibitive, making naïve “optimize then evaluate” pipelines unreliable. To this end, we introduce inference-aware policy optimization, which anticipates downstream model-free evaluation by optimizing both estimated performance and the probability that the estimated improvement will pass a significance test on held-out data. We characterize the Pareto frontier of this tradeoff and provide an algorithm to estimate it, enabling policies that are not only promising, but also testable.

Joint work with Osbert Bastani and Bryce McLaughlin.

Bio:
Hamsa Bastani is an Associate Professor of Operations, Information, and Decisions at the Wharton School, University of Pennsylvania. Her research focuses on developing novel machine learning algorithms for data-driven decision-making, with applications to healthcare operations, social good, and revenue management. Her work has received several recognitions, including the Wagner Prize for Excellence in Practice (2021), the Pierskalla Award for the best paper in healthcare (2016, 2019, 2021), the Behavioral OM Best Paper Award (2021), as well as first place in the George Nicholson and MSOM student paper competitions (2016). She previously completed her PhD at Stanford University, and spent a year as a Herman Goldstine postdoctoral fellow at IBM Research.