Journal of Empirical Legal Studies (forthcoming, 2007)
John J. Donohue III & Daniel E. Ho
We use differences-in-differences (DID) to assess the impact of damage caps on medical malpractice claims for states adopting caps between 1991-2004. We find that conventional DID estimators exhibit acute model sensitivity. As a solution, we offer (nonparametric) covariance-adjusted randomization inference, which incorporates information about cap adoption more directly and reduces model sensitivity. We find no evidence that caps affect the number of malpractice claims against physicians.