Abstract: Motivated by mixed evidence related to the pricing of measures of risk, we investigate the information content of innovations in implied idiosyncratic volatility. Using both cross-sectional and time-series methodologies, we find that innovations in implied idiosyncratic volatility explain future returns for a sample of 2,864 optionable firms examined during the 1999-2010 sample period. We find that long-short portfolios formed using innovations in implied idiosyncratic volatility produce much larger abnormal returns than long-short portfolios formed using the level of implied idiosyncratic volatility. We also find that significant (both statistically and economically) abnormal returns continue to be generated for portfolios formed using innovations in implied idiosyncratic volatility when the sample is partitioned by: firm size, book-to-market equity, implied idiosyncratic volatility persistence, volatility risk, jump risk and option skewness. Our cross-sectional results lead to similar conclusions and are robust to the inclusion of traditional risk factors. Therefore, our results provide evidence of a significant and, to this point, undetected ex ante measure of risk.