Stock Selection and the Domini Social Index Securities
In this section of the study we specifically address the issue of stock with
the 400 stocks of the Domini Social Index (DSI) during the June 1990–De-
cember 1994 period. If one applied the seven-factor robust regression-
weighted composite model ranking to the 1,200-stock universe for the
June 1990–December 1994 period and used the simulation conditions dis-
cussed in the preceding section, one would have outperformed the S&P
500 by 297 basis points instead of the 400 points of outperformance previ-
ously found. The portfolios turned over approximately 228 percent annu-
ally during the June 1990–December 1994 period. If one used an equally
weighted portfolio rule in which one sold securities when the expected re-
turn ranking fell into the bottom half of the distribution, one could sub-
stantially slow down turnover and enhance performance. If one used a
selling criterion of selling when the alpha fell below –.7, the seven-factor
model would earn an excess return of 439 basis points with turnover of
only 107 percent. If one used the same –.7 selling criterion and employed
the seven-factor model including only the 400 stocks of the DSI for the
June 1990–December 1994 period, the excess returns would be 299 basis
points and annualized portfolio turnover would be 79.3 percent. One can
effectively pick stocks within the socially responsible DSI universe, and the
excess returns of the 1,300-stock universe and DSI 400 would be virtually
identical. Furthermore, if one wanted to be even more socially responsible
and not invest in DSI stocks in which KLD has noted minor environmental
and product concerns, one could create a portfolio strategy using the –.7
alpha sell rule and outperform the S&P 500 by 323 basis points. The differ-
ence between the 323 basis points of outperformance of the environmental
and products concerns portfolio and the 299-basis-point outperformance of
the DSI portfolios using our seven-factor model represents the additional ex-
cess returns occurring with the implementation of two KLD screens. Clearly,
more research needs to be undertaken with respect to the effective use of so-
cial screens in a socially responsible universe.
Recent Socially Responsible Research
Drhymes (1997) and Guerard (1997b) showed that the use of all KLD
concerns and strengths should not significantly alter unscreened universe
returns during the 1992–1997 period. Guerard found that only the mili-
tary screen cost the investor returns during the period. The military screen
is the only social screen producing a statistically significant cost during the