Most investors, especially in their retirement portfolios, face the task of choosing a mutual fund, often with inertia. According to research by Finance Professor Feng Zhang of SMU Cox, the typical investor tends to lose money investing in mutual funds.  Zhang, with co-authors Bessembinder and Cooper, find that over two-thirds of U.S. equity mutual funds underperform a benchmark SPY ETF (S&P 500), after fees, in terms of compound returns during their 1991-2020 study period. “Though some funds perform very well, the net economic impact of this long-term underperformance is large,” he and his colleagues say.

Generally, mutual funds create wealth, but factoring in investment-related fees changes the equation. While fund manager skill creates $1.23 trillion of wealth to mutual fund holders, it is not sufficient to overcome fund expenses totaling $2.26 trillion. “It does not mean mutual fund managers are not skilled,” says Zhang. “They are.” But when fees are included, the value of investor wealth becomes eroded over the long term. “The mutual fund companies are charging $2.26 trillion in fees over time,” he notes. When using the returns of SPY, the research shows that the overall wealth loss to mutual fund investors is $1.02 trillion.

Considered from another angle, the average mutual fund’s return, before fees, exceeds the returns of the market benchmark at long horizons. Many, if not most, investors tend to buy and hold funds in their portfolio. The sample mutuals funds showed a 394% total return over the thirty-year period; the average SPY buy-and-hold return over matched periods is 298%, after fees. “That works out to about 0.8% per month,” he notes. Nevertheless, less than half, or 45.2%, of individual funds outperform the SPY in the long run —even in a comparison of mutual fund returns, before fees, to SPY returns that are net of fees. In the study, the authors use SPY as an alternative investment opportunity, or the benchmark, when computing wealth creation and loss.

Skewed returns, geometric means

Zhang further clarifies a key point of the research. “You may think that investing in a mutual fund is a good deal, but the numbers are misleading,” he says. The 394% return before fees appears decent on the surface, but “the returns are skewed over the long term: the median fund’s buy-and-hold returns before fees are only 115%,” Zhang explains. He discussed how the skewness should be near zero if the distribution is normal, but it is not normal. That 394% average over the long term is driven by a small fraction of funds, which is what is meant by “the skewness.” Zhang surmises: It means that the 394% average of buy-and-hold returns is largely driven by outliers. And can you correctly pick those outliers?”

Over long horizons, the compounding of average returns is a focal point in financial planning, particularly at pension funds, where average returns rule.  This research indicates that “potentially a large majority of possible future realizations are less than the mean [average] outcome.” A certain proportion of well-performing funds are driving up the averages. 

Further, Zhang highlights how the maths work out in a practical way relative to simple arithmetic versus geometric averages:

“Think of two funds over two years. The first fund A offers a 20% return in year 1, and then -10% the next year. Fund two, B, offers a 5% return in year 1, and in year 2, also a 5% return. It’s kind of boring. The arithmetic mean would be the same, 5% per year. But the total return over the two years in fund A is only 8%; in the second year, the more boring fund B, the total return is more than 10%. This is exactly the point about skewness, your return variance, it’s not zero. It means the lower your variance, the higher your total return will be.”

  A B
 Year 1 20.00% 5.00%
 Year 2 -10.00% 5.00%
 Arithmetic mean 5.00% 5.00%
 Compound returns 8.00% 10.25%
 Geometric mean 3.92% 5.00%

Zhang illustrates the point that if you’re investing in funds with greater volatility, compared to a fund with a more linear return trajectory, you’ll do better. That is, selecting a more value-oriented type investment over one with momentum, or volatility, tends to have better long-term performance. “Precious metals funds don’t perform well over 30 years,” Zhang notes.  “Specific industries with volatility perform badly.”

Measuring Up

The study acknowledges the “stellar overall equity market returns during the 1991-2020 period.” However, over 20% of U.S. equity mutual funds decreased their aggregate investors’ wealth as compared to the wealth they would have attained if they had instead earned one-month Treasury-bill returns, the authors write. “Management fees and trading costs contribute to this striking outcome, [but] the timing of investor flows and the skewness of compound fund returns are also important.” Zhang says, with his co-authors, that they were shocked about the finding that 20% underperformed T-bills. It highlights the effect of the outliers and time horizons with respect to when a fund launches and its financial flows.

In another related paper, “What You See May Not Be What You Get: Return Horizon and Investment Alpha,” Zhang and his co-authors upend a forty-year standard of calculating above-market returns, or alpha, from a monthly point of view. That standard was set in the 1960s and 1970s owing to researchers’ focus on alpha measured at the monthly horizon. They show that the industry practice of monthly benchmarking returns causes inaccurate measures of fund performance over different time horizons, especially over the long term. Zhang says, “It’s time to reconsider how to measure fund performance.”

The research by Zhang and his colleagues offers new thinking about how investment performance can be practically measured and how conventional wisdom needs to be challenged.

The paper “Mutual Fund Performance at Long Horizons” is authored by Feng Zhang of Southern Methodist University’s Cox School of Business, Hendrik Bessembinder of Arizona State University, and Michael Cooper of University of Utah.

Written by Jennifer Warren.