Can Efficiency of Returns Be Considered as a Pricing Factor?

Abstract

We add to the investments literature by employing new techniques to estimate asset performance. We estimate a data envelopment analysis based efficiency score that allows for direct comparison between ex-post efficiency rankings and test the ex-ante relevance of such scores by including them into asset pricing models. We find that knowing the fund efficiency score can help explain time-series returns. When efficiency is included in an asset pricing model, the absolute value of the average mispricing error is decreased, which we take as evidence of the explanatory power of efficiency scores. But more importantly, we show that efficacy scores can be used as next period predictors of stock returns. In addition, we further use the efficiency scores to differentiate between the performance of constrained and unconstrained investment assets, as in the case of socially responsible investments. Our findings give robustness to the literature on constrained investments showing significant underperformance of socially and responsible investments.

Appendix

1.1 A Description of Ethical Investing

According to the Social Investment Forum in the U.S. (SIF), over the past 20 years the total dollars invested in SRIs has grown exponentially, as has the number of institutional, professional, and individual investors involved in the field. Between 1995 and 2010, total dollars under professional management in SRI grew from $639 billion to $3.07 trillion, outpacing the overall market. As of 2015, it has further increased to 6.57 trillion dollars.¹³ SRI investing has become a mainstream asset, and as a result, a number of conventional companies now offer SRI products to their clients. An increasing number of investors focus on ethical funds not only because of the ethical aspect, but because returns are comparable to those of more conventional investments.¹⁴

Managers of ethical funds have to follow socially responsible constraints on environmental risk, social risk, and governance risk. For example, they would not buy shares from the weapons, gambling, alcohol, or tobacco industries; nor would they buy shares from companies known for polluting the environment, using “sweatshops,” discriminating against their employees, etc. On the other hand, these managers would increase their number of shares of ethical companies, philanthropic institutions, and firms well known for their social activities.¹⁵ This is the social screening process. Barnett and Salomon (2006) find that when the number of social screens (both positive and negative) increases, the fund’s annual return initially declines, but then rebounds as the number of screens reaches a maximum. Furthermore, SRI’s managers also have to analyze the companies’ financial statements. An example can be seen in Goldreyer and Diltz (1999) who find that the performance of ethical investment also depends on the level of financial screening.

The question still remains whether or not ethical investment reduces profitability, thereby reducing efficiency. Asked another way, are the risk-adjusted returns of SRI portfolios higher than traditional investing? In response to this matter, Hamilton et al. (1993) proposed three alternative hypotheses: (1) the risk-adjusted expected returns of socially responsible portfolios are equal to the risk-adjusted expected returns of conventional portfolios, as the social responsibility feature of stocks is not priced; (2) the risk-adjusted expected returns of socially responsible portfolios are lower than the expected returns of conventional portfolios, as the market prices the social responsibility characteristic by increasing the value of socially responsible companies relative to the value of conventional companies by driving down the expected returns and the cost of capital of socially responsible companies; and (3) (also suggested by Moskowitz 1972), the risk-adjusted expected returns of socially responsible portfolios are higher than the expected returns of conventional portfolios, as the market prices social responsibility (incorrectly) in the case of “doing well while doing good.”

Overall, the literature is divided on the exact consequences of investing in ethical mutual funds. For example, it has been shown that volatility in said market is lower than the volatility of conventional funds (Bollen 2007; Renneboog et al. 2005); but this does not include a discussion on the effects of reducing diversification. Investing in a closed universe is expected to limit the power of diversification and thus it can result in lower risk-adjusted returns. Yet consistent evidence of underperformance of SRIs has been missing.

Looking at the ethical aspect, traditional numerical indexes used to measure the performance of mutual funds do not take such features into consideration. Particularly, the Sharpe ratio (Sharpe 1966), the reward-to-half-variance index (Ang and Chua 1979), and the Treynor index (Treynor 1965) are computed as ratios between the expected excess return and a risk indicator without considering additional features (Basso and Funari 2003). Still, Goldreyer and Diltz (1999), compute Sharpe and Treynor ratios to show that social screening does not affect the investment performance of ethical mutual funds in any systematic way. Furthermore, Statman (2000) shows that the DSI index¹⁶ (which is one of the most well-known SRI indexes) has a higher Sharpe ratio than the S&P 500, which implies that investors seeking to optimize the mean-variance would prefer the ethical investment.

Believers in the efficient market hypothesis argue that it is impossible that SRI funds outperform their conventional peers (Renneboog et al. 2008). At most, prior research has found that socially responsible investing performs at least as good as traditional investment portfolios.

Hamilton and Statman (1993) and Statman (2000) use a sample of SRI funds and non-SRI funds for the periods 1981–1990 and 1990–1998, respectively, to estimate Jensen’s alphas based on CAPM, using the DSI 400 index and the S&P500 index as the benchmark returns for SRI and non-SRI. In the end, the authors find no statistical difference between the Jensen’s alphas, thus favoring the first hypothesis. These results are consistent with Diltz (1995), Guerard (1997) and Sauer (1997).

Later attempts included the use of Fama and French’s three factor model as well as the inclusion of Carhart’s momentum factor. Bauer et al. (2005) find no statistical difference in the performance of SRIs for the overall period 1990–2001, but they do find a “catching up” period in their early life. Their results are consistent with Renneboog, et al. (2008a) who reiterates the study for the period 1991–2003. Renneboog, et al. (2008a) also find said catching up period in the sub-sample 1991–1995. Surprisingly, though, the authors also find that ethical funds outperform their counterparts for the sub-period 2001–2003 (though it is not statistically significant).

Still using Carhart’s four factor model, Gil-Bazo et al. (2010) compare the pre and after fees performance of ethical funds against non-ethical funds for the period 1997–2005. They find that ethical funds have better before and after fees performance when compared to non-ethical funds of the same characteristics.

Most of the literature, however, focuses on the social factor of ethical funds and how SRIs outperform conventional investing due to that factor. Yet, several authors (Kurtz and DiBartolomeo 1996; Guerard 1997; Kurtz 1997 report that the reason for the over performance of the DSI 400 is due to the large investing on growth stocks and not the social factor.

Geczy et al. (2005) show that investing only in ethical funds carried a significant financial cost for the period 1963–2001. Yet on the other hand, Bollen (2007) justifies the profitability of SRI given that investors see beyond the traditional risk-reward optimization problem, as they may possess a multi-attribute utility function which incorporates a set of personal and societal values.