The Persistence of Fee Dispersion among Mutual Funds
Michael J Cooper, Michael Halling, Wenhao Yang
Review of Finance, Volume 25, Issue 2, March 2021, Pages 365–402, https://doi.org/10.1093/rof/rfaa023
The neoclassical model of mutual funds (Berk and van Binsbergen, 2015) argues that percentage fees are irrelevant, as fund size will adjust in equilibrium such that net alphas are equal to zero. However, we show that from an investor’s standpoint, in the cross-section of net-of-fee returns, fees are important. We regress net alphas on lagged fees and controls in our sample of all US and international equity funds from 1980 to 2017. For both CAPM and BvB alphas (estimated following Berk and van Binsbergen, 2015), we find statistically significant negative coefficients on fees.
Given that fees strongly matter from an investor perspective, we investigate the pricing of mutual funds. Methodologically, we make the fees charged by different funds comparable by estimating residuals expenses from yearly, cross-sectional regressions of total annual expenses on lagged fund characteristics, such as risk and performance characteristics, measures of fund manager skill from Berk and van Binsbergen (2015), the extent of active management, service levels, fund size and age. The residual approach allows us to compare prices across “similar” funds.
We explore the pricing of various groups of funds, including S&P 500 index funds, all index funds, all active US and international mutual funds investing in equities, and largest (top quintile) active funds. In yearly regressions of fees on fund characteristics, we explain between 32% and 42% of the variation in fees in active funds. We find that the average annual spread in residual expenses between the 25th and the 75th percentile (between the 10th and 90th percentile) across the all funds sample is economically large with 47 bps (98 bps). We find similar results in the large fund group. Surprisingly, given the commodity-like nature of these funds, we also document large and economically significant residuals in the sample of S&P 500 index and all index funds. We show that the residual spreads are relatively constant over time. Thus, despite the enormous growth in the mutual fund industry, the dispersion in residual expenses has stayed about the same.
Our findings of economically large fee dispersion across arguably similar funds carries important investor implications. For example, based on residual expenses of the All funds sample, an investor purchasing the lowest expense funds (i.e., the bottom quintile) would have earned compounded CAPM, Carhart (1997) four-factor model, and BvB net-of-fee alphas of 30%, 58%, and 24% higher, respectively, than the investor purchasing the most expensive (i.e., the top quintile) funds over our study period.
Finally, we show important cross-sectional variation for net value added across fee levels. Low fee funds tend to have positive net value added while high-fee funds tend to have negative net value added, suggesting a systematic relation between net value added and percentage fees. Finally, we provide a quantification of misallocated capital (Zhu, 2018) and an assessment of its relation to percentage fees in the paper. This analysis shows that 70% of funds are overinvested and, thus, should shrink.