Files
Abstract
This thesis examines the extent to which a sin premium exists in the U.S. stock market in the period 2000 to 2023 on AMEX, NASDAQ, and NYSE. That is, whether positive risk-adjusted returns are harvestable by investing in equities that are classified as being sinful, which is defined as stocks within the alcohol, tobacco, gun, and gaming industries. By running OLS regressions on value-weighted portfolios net of the riskfree interest rate and net of a group of comparable stocks, while controlling for wellknown factor returns, there is strong evidence for sin investing yielding a positive alpha. When excluding the gun industry from the sin portfolio, the results remain intact in the former, whereas, in the latter, alpha turns statistically insignificant after controlling for an adequate number of factor exposures. In general, the findings are less robust when implementing an equal-weighted portfolio construction methodology. By running Fama-MacBeth regressions, evidence for the sin anomaly is lacking. Once controlling for sufficiently many firm-specific characteristics, the sin indicator variable is insignificant, both from a stock-by-stock and from an industry portfolio perspective, despite the point estimates being consistently positive. Hence, from the factor model regressions, the thesis finds that sin stocks do outperform non-sin stocks on a riskadjusted basis, but the results from the Fama-MacBeth regressions are less convincing. The robustness checks echo the fact that methodological inconsistencies provide a plausible explanation for the absence of consensus as to whether a sin anomaly exists.