A popular research design identifies the effects of corporate governance by (changes in) state laws, clustering standard errors by state of incorporation. Using Monte-Carlo simulations, this paper shows that conventional statistical tests based on these standard errors dramatically overreject:...