Abstract: Negative screening (of ”sin” stocks) is the most common strategy used by socially responsible investors. There is no consensus in the literature whether these exclusions result in higher cost of capital (and hence higher expected returns) for targeted firms. The existing literature identifies sin companies using industry classification codes (IC). We propose an alternative measure of firms’ exposure to sin activities (sinfulness) based on textual analysis (TA). Sinfulness captures both cross-sectional and time-series variation in firms’ exposure to sin activities. The correlation between the IC and TA sin indicators is only 0.69, with twice as many sin stocks in TA than in IC. TA reveals several important false positive and numerous false negative sin stocks in IC. While the number of publicly listed sin-related stocks has declined by 43% between 1997 and 2021, their total market capitalization has increased almost threefold from about $200bn to $600bn during the same period. A sin-weighted portfolio of sin stocks earns an annualized Fama-French 6-factor alpha of 4%. Overall, our study highlights important shortcomings of using IC to identify sinful firms and resurrects the sin premium, that is, more sinful stocks have higher expected returns.