The patent doctrine of inequitable conduct—which allows a patent to be held unenforceable on the basis of misbehavior by the applicant during patent prosecution—has been the subject of intense criticism from the bench and bar alike. And yet to date there has been no systematic attempt to determine whether the doctrine is or is not working as theorized. This study fills that gap. We evaluate the performance of the inequitable conduct doctrine with a novel methodological approach: by empirically characterizing the differences between patents found unenforceable and several other types of patents (unlitigated, litigated, invalid, obvious, and underdisclosed), we use those differences to reveal the real-world impact of the inequitable conduct doctrine. We find that pat ents held unenforceable have clear hallmarks of risky prosecution behavior, such as longer pendency and fewer disclosures of prior art, as compared to all other types we studied. These results indicate that the doctrine is likely operating better than conven tional wisdom would suggest.