Pricing Sin Stocks: Ethical Preference vs. Risk Aversion

Topic and Objectives

Several empirical studies find that sin stocks (i.e., stocks of companies that are believed to make money from exploiting human weaknesses, such as alcohol, tobacco, and gambling companies) yield (on average) higher returns than those of non-sin comparable stocks. This so-called “sin premium” is often rationalized by a “boycott” risk factor, namely, the risk that socially responsible investors refuse to hold stocks of sin companies. As a result of the boycott strategy, sin companies are underpriced relative to non-sin companies and must promise higher returns to attract investors. This approach leaves some questions open. 

First, according to the boycott approach, socially responsible investors always refuse to hold sin stocks, independently of their expected cash-flows. Therefore, the boycott-based approach leaves no room for a trade-off between ethicalness and cash-flows, even when the latter become more appealing. Second, the behavior of individual and institutional investors towards sin stocks may significantly differ. Institutional investors, such as mutual funds, pension funds, and foundations, may be subject to social pressures rising from their public exposure, and, accordingly, tend to be reluctant to hold stocks that conflict with their customers’ ethical principles. Individual investors, on the contrary, are generally free from transparency and accountability concerns and, consequently, may be more open to investing in any kind of stocks. Motivated by these considerations, we build a theoretical model intending to relax the boycott assumption. We assume that investors have preferences for gain opportunities (dividends) but weigh them according to their perception of firms’ responsible behavior (ethicalness). Investors do not necessarily boycott sin companies and are willing to receive dividends from both sin and non-sin stocks. 

Key Findings

• We show that sin stocks have higher average returns and volatility than non-sin stocks in two cases: (i) when dividends and ethicalness are substitute goods and investors have low-risk aversion (i.e., smaller than log utility), and (ii) when dividends and ethicalness are complementary goods and investors have high risk aversion (i.e., higher than log utility). In both cases, the desired marginal rate of substitution between dividend payments and ethicalness is positive, which implies that investors would like to receive more dividends from non-sin stocks than from sin stocks. However, since dividend payments are beyond investors’ control, the expected returns must adjust to offset the “ethical” cost of holding less desirable stocks (sin stocks). In this way, our model produces the average positive return and volatility spreads between sin and non-sin stocks. The above cases suggest that two mutually exclusive preference specifications can explain the average return and volatility spreads between sin and non-sin stocks. However, these two settings generate opposite patterns of moments differentials, conditional on the sin dividend share. Precisely, the return and volatility spreads exhibit a negative relation with the dividend share in case (i), and positive in case (ii). 
• To understand which of the two preference specifications is consistent with the data, we investigate the empirical relation between conditional moments and dividend payments. Using U.S. data, we provide evidence supporting case (ii), that is, we find a positive relationship between the dividend share of sin stocks and the return/volatility spread between sin and non-sin stocks.  We conclude that risk-averse investors are more likely to treat monetary gains and ethical behaviors as complementary goods.
• Using data on the portfolio choices of retail investors, we show that investors prefer to receive dividends from ethical companies and, as a consequence, re-balance their portfolio away from sin companies after those companies pay dividends, in line with our model.
• Since institutional investors are more prone to boycott, one would expect that the predictability results illustrated above are stronger for companies with low institutional ownership. Our empirical analysis confirms that this is indeed the case. All together, these empirical results suggest that our model provides a realistic description of the economic forces behind the sin premium observed in the data.