A Win for Socially Responsible Investing: Rethinking “Sin Stocks”

I have a love-hate relationship with sin stocks. On one hand, my first big winning stock pick was Swisher Sweet Cigars, which nearly doubled a month after I bought it in 1998, thus proving that I was a stock picking genius and much self-satisfaction (at least until several years later when I lost 80%+ betting on rain forest cafe).

What made me feel less good was how as an undergraduate at Wharton, I spent countless hours assembling a database of mutual fund manager characteristics for the following paper on socially responsible mutual funds. The experience was extremely painful (increased computing power and better optical character recognition software were still a few years out). So my memories are a mixed bag: I’m hot and cold on sin stocks depending on the day of the week.

That said, the collective research seems to suggest that sin stocks enhance portfolio performance. For example, Hong and Kacperczyk (2009) find that “Sin stocks outperform comparables.” HK finds that from 1980 to 2006, a long/short portfolio of US sin stocks yield a statistically significant return of 26 bps per month after adjusting for a four-factor model. They argue that the effects of social norms can help explain the outperformance of sin stocks:

• Lower Institutional Ownership: They find that the institutional ownership ratio of sin stocks is 23% lower than that of their comparables. Sin stocks are less held by pension funds, university endowments and other socially responsible institutional investors.
• Less Analyst Coverage: What’s more, sin stocks receive less media coverage. HK find that the sin stocks on average are followed by 1.3 analysts, while their comparables received 1.7 analysts on average.

Perhaps sin stocks are another anomaly? But as we know, there are over 300 “anomalies” that have been identified over the past decades, but few hold up in robustness checks. This paper by Adamsson and Hoepner (2015) re-examines HK’s results and finds that the “sin stock premium” is actually driven by a small cap performance bias rather than sin stocks characteristics.

The ‘Price of Sin’ Aversion: Ivory Tower Illusion or Real Investable Alpha?

In their pioneering work, Hong & Kacperczyk [HK] (2009) document a significant outperformance of so- called sin stocks: publicly traded companies involved in producing alcohol, tobacco and gaming. Drawing on theories of neglected stocks and segmented markets, the authors suggest this outperformance to be due to a shunning effect caused by norm-constrained institutional investors who effectively pay a price for sin aversion. Motivated by a line of criticism against their research design, this study re-examines whether the sin stocks premium recorded by HK is actually investable for real world investors or derives from an ‘ivory tower style’ research design?

We limit our analysis to only those stocks included in the global equity index benchmarks of institutional investors thereby excluding thousands of stocks usually considered too illiquid in practice. Among these stocks, we still document the sin stock premium in HK’s equal weighted sin-industry portfolios. Equal-weighted sub-industry portfolios are, however, practically barely investable for many institutional investors with their value-weighted benchmarks and are, if analysed in Fama-French models, exposed to a small cap bias in two ways. First, if small cap firms outperform large cap firms, then the equal-weighted portfolio is likely to outperform its value-weighted equivalent and a value-weighted market. Hence, we value-weight our sin-industry portfolios, which leads the premium for gambling firms to disappear. Second, Fama-French’s model controls for return differentials between small cap and large cap firms within all industrial sectors. The model is not designed to capture return differentials between small cap and large cap firms in a single sector such as consumer goods, which could drive the returns of sub-industries in this sector such as alcohol and tobacco. Hence, we add within sector control variables to our model which leads to a complete disappearance of any premium to sin stocks at the global level. When we conduct the same analysis in HK’s purely US universe, the sin stock premiums disappear even more convincingly. This implies that the sin stock premium and the price of sin aversion could be artifacts of an ivory tower research design and are not necessarily applicable to the real world.