Pay for non-performance is among the most prominent arguments of executive rent extraction, especially Bertrand and Mullainathan’s (2001) pay for luck. We revisit their finding over the last two decades, 1997 through 2016. Pay for luck presents in the first decade but declines in the second decade. This decrease is robust to different measures of luck, various industry subsamples, and the financial crisis of 2008-9. The structural break in pay for luck associates with transparency-based regulations, such as option expensing and new performance pay disclosures. These regimes plausibly enhance shareholder monitoring, which pushes compensation committees to decrease pay for luck.