The Pre-Tax Costs Of Portfolio Turnover
[Editor's Note: The tables accompanying Figures 2-4 were inadvertently not included in the print version of this article. Please consider this version the full and final version of the article.]
The average turnover in mutual funds has increased with time. According to John Bogle (Common Sense on Mutual Funds ), turnover has increased from 30 percent twenty-five years ago to nearly 90 percent today. Wermers (2000) has similar numbers.
Unlike publicly disclosed expense ratios, the costs associated with turnover are generally unknown, and are invisible to investors apart from their impact on overall performance. These expenses are nonetheless real, however, with both explicit (e.g., commissions) and implicit (market impact) costs that should be considered when selecting an investment.
A variety of research has been conducted seeking to quantify the impact of portfolio turnover (most notably Carhart ). However, with market developments like decimalization and the elimination of fixed minimum commissions, the timeliness of this research is questionable; some of it relies on data dating back as far back as the 1960s. The purpose of this paper is to update that analysis and quantify the pre-tax costs (both performance and risk) associated with portfolio turnover using actual mutual fund returns over a more recent sample period.
Carhart's "On Persistence in Mutual Fund Performance" is perhaps the most-often-cited article discussing the implications of portfolio turnover, as well as many other issues surrounding active management. Carhart tested the performance of three mutual fund strategies (aggressive growth, long-term growth, and growth-and-income) using a survivorship-free database covering January 1962 through December 1993. He found a turnover slope coefficient of -0.95, suggesting that for every 100-point increase in turnover, the annual return drops by 95 basis points (which he interprets as the net cost of trading).
Additional researchers have also noted the negative impacts of turnover. In his book, Bogle on Mutual Funds (1994), John Bogle estimated the cost of turnover to be approximately 1.2 percent for each 100 percent of turnover (page 204). Day, Wang and Xu (2001) noted that a 1 percent relative increase in yearly turnover is associated with a 0.075 percent decrease in risk-adjusted performance. Dowen and Thomas (2004) noted that equity managers who trade less tend to produce greater returns (interestingly, fixed-income managers who trade more tended to produce greater returns). Sharkansky (2001) also noted the impact of turnover across a variety of equity (and fixed-income) categories, ranging from 124 bps per 100 percent of turnover for large-cap equities to 255 bps for small-cap equities.
Two published studies, which are in the composite minority, have noted potential benefits from portfolio turnover. Bauman, Miller and Veit (2005) found that many investment advisors working for 13F institutions (i.e., pension portfolios) have the skill to identify and purchase stocks that generate a higher return per unit of risk than both the market and the stocks they sell. Wermers (2000) also found that high-turnover funds, although incurring substantially higher transaction costs and charging higher expenses, tend to hold stocks with much higher average returns than low-turnover funds. However, Day, Wang and Xu (2001) noted a negative correlation between portfolio turnover and pre-expense performance for actively managed portfolios, suggesting that turnover rates for these funds are not driven by superior information.
The impact of turnover has also been noted for individual investors. Barber and Odean (2000) found an average annual portfolio turnover of 75 percent for U.S. investors using discount brokers, while Shu, Chiu, Chen and Yeh (2004) uncovered an annual turnover of more than 10 times that for Taiwanese investors. Barber and Odean noted that U.S. investors tended to hold small, high-risk securities, and that the risk-adjusted performance of the average investor lagged the market by 3.7 percent per year. Perhaps even more telling was the average 10 percent underperformance of the most active 20 percent of investors on an annual, risk-adjusted basis. Shu, Chiu, Chen and Yeh noted that, while Taiwanese investors had positive abnormal returns from factor models, they would have earned better returns from a buy-and-hold strategy.
Although the tax implications of turnover are beyond the primary scope of this paper, they nonetheless represent explicit costs for taxable investors. Bogle (1994) discusses the impact of turnover in a variety of tax scenarios. Additional research by Peterson, Pietranico, Riepe and Xu (2002) found that equity fund investors from 1981 to 1998 lost an average of approximately 2.2 percent annually to taxes, while more recent research by Longmeier and Wotherspoon (2006) found that lower turnover is significantly correlated with higher tax alpha, leading to higher after-tax investment returns.