We examine how the relation between mutual fund performance and fund flows has changed over time by separating our sample into two periods (1978-87 and 1988-97). We document an increase in the flow-performance asymmetry in the second period that exacerbates the adverse incentive for fund managers to increase portfolio risk. We develop a measure ofthe elasticity offund flows with respect to performance, which filters out the confounding influence of greater aggregate fund flows in the second period and allows an examination of whether current investors place more emphasis on prior performance when selecting funds. We conclude that, though top performing funds are rewarded with greater fund flows in the second halfofour sample, the change is due solely to the increase in aggregate fund flows and not to an increased reliance on performance by individual fund investors.JEL classification: G1, G2, L1.

The brokerage commissions paid for portfolio transactions by a large sample of equity mutual funds are investigated. Median brokerage commissions measured as a percentage of net assets are 21 basis points per year with a standard deviation of 27 basis points. The commission levels are negatively correlated with fund size and positively correlated with fund turnover and expense ratio. The average brokerage commission measured as a percentage of assets traded exceeds the typical executiononly commissions for large institutional traders. This finding is consistent with many mutual fund brokerage commissions including payments for research, so-called soft dollar payments. Funds' expense ratios are positively correlated with commissions per trade, inconsistent with the idea that mutual fund managers who pay soft dollars for research have a corresponding reduction in management fees.

We examine portfolio credit quality holding and daily return patterns in a large sample of bond mutual funds and document evidence of window dressing. Using portfolio credit quality holdings data, we find that bond funds on average hold significantly more government bonds during disclosure than nondisclosure, presumably to present a safer portfolio to shareholders. Multiple-index market models estimated with daily returns data corroborate these findings. We detect differences in factor loadings on days surrounding disclosure dates that indicate systematic tilting of the portfolio toward higher quality instruments. 2006 The Southern Finance Association and the Southwestern Finance Association.

Many individuals purchase shares in mutual funds as investments. With a lack of evidence supporting performance persistence in fund returns, investors should consider expenses as a fund-selection tool since fund expenses have a negative effect on fund returns. One of the largest expenses incurred by fund investors is distribution expenses, which include both load charges and annual fees. Close to two-thirds of all equity funds charge investors for fund distribution. The true cost of these distribution fees to investors is hard to measure because a myriad of distribution arrangements have evolved that vary both the timing and magnitude ofdistribution charges. We derive a simple methodology that expresses the present value of distribution costs as a percentage of the original investment in fund shares for any expected holding period. This methodology allows direct comparison of the effect on investors of distribution fees for mutual funds with different types of sales arrangements.

No abstract