Jensen’s alpha

Jensen’s alpha is computed per country and investment scope for all mutual fund portfolios. The theoretical expected returns are required in order to calculate the mutual fund portfolios’ under or over-performance, which is known as the Jensen’s alpha.

The theoretical expected returns were calculated by applying the single factor Capital asset pricing model known as CAPM.

The CAPM was applied by using the global market premium factor collected from Kenneth French Data Library for the portfolios with global holdings or the excess return on the local MSCI Indices i.e. the index returns minus the local risk-free rates. The certain index and risk-free rate for each portfolio was chosen based on the mutual fund portfolio’s domicile and investment focus. The different indexes and risk-free rates are further examined in Chapter 6.

The intercept obtained from the CAPM calculation represents Jensen’s alpha, which is used to evaluate fund performance.

The following table 7.4 presents the alphas and market factor betas of the ethical mutual fund portfolios and their matched sample portfolio of conventional funds per country and investment focus. Also, the alpha and beta of the “difference” portfolio is included to improve comparability.

Table 7.4: Jensen’s alpha. Alphas are annualized and presented in percentages.

Mutual funds with a global investment universe

All portfolio alphas are negative and all alphas, except the alpha for the Danish ethical portfolio, are significant and indicate that both the ethical and conventional mutual funds underperform in relation to the market. All betas are also significant and less than 1, which imply that the mutual funds are less volatile than the market.

The betas of the difference portfolios indicate that the Swedish and Norwegian ethical mutual funds with a global investment scope marginally tend to be more sensitive to fluctuations on the market than their conventional equivalents. Conversely, the Danish ethical mutual funds are less sensitive to the market than the conventional funds with the same characteristics.

Moreover, the alphas for the ethical mutual funds are less negative than for the conventional mutual funds ones. On a yearly basis generate the Norwegian conventional funds approximately 18 % lower returns than the ethical ones.

R² is lower for the globally investing mutual funds in comparison to R² for the mutual funds with a regional investment focus. This indicates that the Scandinavian factors measures a higher proportion of variation in the regression model and are a better fit. The reason behind this could be that the global factors collected from Kenneth French Data Library take all global markets into consideration. The mutual funds, which are investing globally, do invest in various markets around the globe, but different funds are focusing on different markets.

A higher R² i.e. a better fit might be attained if the country origin of the holdings in every mutual fund had been examined and new factors were constructed for every mutual fund on basis of the holdings’ domicile. However, the authors have had limited access to resources which publish the holdings of every mutual fund and it would be beyond the scope of this paper to investigate all mutual funds’ holdings. The mutual funds, which contain more than 50 % of global stocks have been classified as global mutual funds and have been regressed against global factors in order to get more accurate regression results. The final section of the analysis elaborates more upon the difference in using global versus regional factor proxies.

Mutual funds with a regional investment universe

The Swedish portfolio alphas are not significantly different from zero, which indicates that both the conventional and the ethical mutual funds neither over nor underperform relative to the market. The US mutual funds overall underperform the market with approximately 2%, a result which is statistically significant.

In similarity to the mutual funds with a global investment scope, the betas for the market proxy are all statistically significant at the 1 % level. However, the ethical mutual funds with a regional investment scope show different results depending on the funds’ domicile. The Swedish ethical

mutual funds have higher betas than their conventional counterparts, and are therefore more sensitive to the fluctuations on the market. Contrariwise, the Norwegian ethical funds are less sensitive. There is only a statically significant difference in the exposure to the market factor for the Swedish and Norwegian ethical fund portfolios. The ethical funds based in the US are marginally less sensitive in comparison to the conventional ones, which is not statistically proven.

The most remarkable finding in this part of the study is the alpha of the Norwegian ethical fund portfolio, which is significantly positive in comparison to the conventional fund portfolio alpha.

The findings show that the annualized return is approximately 20 % higher for the ethical funds in comparison to their conventional counterparts.

R² is considered to be high for all regressions. Hence, it measures a high proportion of the variation in the regression model and should be considered to be a good fit.

Partial conclusion

Overall, the results show no evidence of statically proven differences in performance of SRI funds and their conventional peers. This inference is in line with findings of the majority of previous scholars e.g. Hamilton et al. (1993), Bauer et al. (2005), Kreander et al. (2005), Renneboog et al.

(2008b), Leite, Cortez (2014). However, the Norwegian SRI funds are an exception, a significantly proven over-performance relative to their conventional peers is shown. Statman (2000) presents parallel findings indicating that the SRI performance is better than the performance of unscreened funds. However, the difference is not statistically significant.

The superior performance of the Norwegian ethical funds is contradictive to the modern portfolio theory, where a constrained investment universe would negatively affect the possibility of a fully diversified portfolio. However, the traditional “rule of thumb” in finance literature states that a portfolio of 20 – 30 randomly selected stocks is basically considered as well diversified (Fisher, Lorie 1970) . More recent findings indicate that the stock market has become more volatile and the minimum number of stocks to closely approximate a well-diversified portfolio is 50. However, there are still some indications that some specific risk remains. The standard deviation of a random portfolio of 50 stocks was 5 % higher relative to the market portfolio (Campbell et al. 2001).

The findings further imply an underperformance relative to the market among all mutual funds, both ethical and conventional. Renneboog et al. (2008) presented the same evidence. The

under-performance of mutual funds in relation to the market confirms the theory of efficient financial markets.

The Danish and globally investing Norwegian ethical funds are significantly proven to be less sensitive to market fluctuations than their conventional counterparts. Mallin et al. (1995), Gregory et al. (1997) & Kreander et al. (2005) reported similar findings, denoting that ethical funds are less risky than the conventional funds. Also, Bauer et al. (2005) find that ethical investing funds are less exposed to market volatility. Leite & Cortez (2014) found insignificant results that European SRI funds only are marginally less risk-averse than their conventional counterparts. Contrariwise, the globally operating Swedish SRI funds tend to have a higher risk than the market.