Over the 1977–83 period, approximately 1,500 private pension plans filed complete 5500 forms with the Internal Revenue Service; the forms provide hitherto unavailable data on the performance of entire pension portfolios. A CAPM-based analysis of the data reveals that, net of investment fees and turnover expenses, private pension plans underperformed the S&P 500 by approximately 44 basis points per year but outperformed a weighted stock-bond index by approximately 38 basis points. The pension plans underperformed mutual funds by almost 300 basis points. Much of this difference, however, appears to be attributable to the CAPM’s tendency to overvalue portfolios with high betas.

There was no evidence to indicate that pension plan performance was related to excessive fees or total portfolio turnover. Pension plans that incurred relatively high fees and relatively high total turnover did not perform worse (after netting out all expenses) than pension plans that followed essentially passive strategies. Trading in the stock portion of pension portfolios, however, had a significant negative impact on net performance. The results suggest that, for the average plan, stock trading practices observed in 1983 lowered rates of return by 60 basis points relative to a plan that engaged in no stock trading. Pension plan stock trading would appear to be inefficient in the sense that superior performance (gross of expenses) in active portfolios is not high enough to offset the research costs incurred to achieve higher returns.