Crystal, Graef. "Knowing How Much, and How, a CEO is Paid Can Improve the Returns On a Portfolio - And By a Significant Margin." The Crystal Report, December 11, 1998.

For the 1994-97 time period, identifies companies with "good" and "bad" executive compensation, using the following criteria: pay level (adjusted for industry and company size), sensitivity of total compensation to share price, level of management share ownership, and willingness to accept pay cuts during periods of adversity. Finds that "bad" compensation companies had significantly (p<.001) below-average stock returns (the "bad" companies had an average beta of 1.2, suggesting their risk-adjusted performance was worse than their nominal returns). "Good" compensation companies performed only marginally better than average, however. This finding was not statistically significant. Crystal did a small test of industry impacts for the six months ended November 30, 1998, and found that industry adjustments reduced the "bad" compensation companies' underperformance by 1/3.