Grossman, Blake and William Sharpe. "Financial Implications of South Africa Divestment." Financial Analysts Journal, July/August 1986.

Widely regarded as the formative work on analysis of screened portfolios, this study compares the returns of a value-weighted South Africa-free portfolio to those of a comparable unscreened NYSE portfolio. After diluting the South Africa-free portfolio with T-bills to make total risk equal to the benchmark, the authors find that it outperformed the benchmark by 1.87% per year from 1960-83. This benefit, however, appears to be entirely attributable to the size effect.

After adjusting for all other factors, doing business in South Africa appeared to be associated with positive returns of 0.77%, although the difference was significant only for the 1960-1975 time period, when the disparity reached 2.05%. The factor turned negative for 1975-83, and although not significant, the authors observe "[it] is nonetheless sufficient magnitude to reject the hypothesis that [South Africa] stocks have substantially outperformed [South Africa-free] stocks with similar characteristics in the recent past." The study also includes a useful commentary on the transaction costs of a divestment strategy.