By LARRY SWEDROE
Once investors have decided on the asset allocation that will most likely allow them to achieve their goals (with due consideration for their ability, willingness and need to take risk), the challenge is to identify the funds that are most likely the best in class. This can be a challenge, given that there are thousands of funds to choose from, and past performance alone has been found to not be an effective tool. In addition, measures such as active share have also been found wanting.
Fortunately, there is actually a simple methodology that can be used to identify funds that are likely to outperform the vast majority of actively managed funds and give investors the best chance to achieve their goals. For example, studies going back to William Sharpe’s 1965 paper Mutual Fund Performance have found that a fund’s expense ratio has a strong influence on future performance. As one example, the 2016 study Fund Fees Predict Future Success or Failure, by Morningstar’s Russ Kinnel, found that the average performance of funds increased monotonically as one moved from the highest to the lowest cost quartiles.
The question is: Can we improve the odds of selecting future top performers by examining other metrics? For example, since a fund’s expense ratio matters, intuitively it might be the case that a fund’s total expenses (including implementation costs incurred due to turnover) would also provide valuable information.
Supporting this hypothesis, research, including Russ Wermers’ 2000 study Mutual Fund Performance: An Empirical Decomposition into Stock-Picking Talent, Style, Transactions Costs, and Expenses, and Eugene Fama and Ken French’s 2010 study Luck Versus Skill in the Cross Section of Mutual Fund Returns has shown that, while mutual funds demonstrate stock-picking skills on a gross-of-fee basis, they fail to outperform appropriate risk-adjusted benchmarks net of fees. Mark Carhart’s landmark 1997 study On Persistence in Mutual Fund Performance found that expenses reduce returns on a one-for-one basis (explaining much of the persistent underperformance of active funds he found), and turnover reduces pretax returns by almost 1 percent of the value of the trade. In addition to total costs, perhaps there are other metrics that can be used. For investors, this search is the financial equivalent of the Holy Grail.
Burton G. Malkiel and Atanu Saha contribute to the literature with the May 2019 study Mutual Fund Returns and Their Characteristics: A Simple Approach to Selecting Better Performing Actively-Managed Funds. Using a survivorship bias-free dataset of over 4,300 actively managed US equity and international equity funds for the period 2000-2018, they examined whether funds chosen based on various fund characteristics in a given year can yield superior performance the following year. In their introduction, they noted that S&P Dow Jones Indices has published SPIVA studies of net returns generated by actual portfolio managers compared with their index benchmarks and found that the vast majority of active managers underperform their benchmarks (and the longer the horizon, the higher the failure rate tends to be) and that the difference in average returns between the active managers and their benchmarks was well approximated by the difference in fees. Following is a summary of their findings:
- The prior year’s active returns were negatively correlated with the following year’s returns — superior performance in one year is generally followed by an inferior return in the next, suggesting that a fund’s higher recent past return has no power to predict future returns.
- While a fund’s expense ratio is an important predictor of a fund’s future returns, it is not the only important factor.
- Factors other than expense ratios, such as a fund’s turnover and its past Sharpe ratio (or SR, the measure of risk-adjusted returns), at least for domestic funds, are also important predictors of future returns. However, the lack of explanatory power of the SR in international markets (the correlation was actually slightly negative, though statistically insignificant) calls this metric into question (at least as a standalone measure). Pervasiveness (it exists around the globe and even across asset classes) is an important test of the value of a metric, minimising the risk of outcomes that are nothing more than exercises in data mining.
- Neither the size of a fund, nor the recent inflows into a fund, may be as much a drag on fund returns as is commonly believed.
Malkiel and Saha found that the combination of multiple characteristics appears to have a “multiplier effect” on returns. For example, portfolios chosen based on lowest quartile of expenses and lowest quartile of turnover produce an annual average return of 4.53 percent and 5.11 percent for U.S. equity funds. When both metrics are used, the return increases to 5.29 percent, and translates to a 1.2 percent excess to the average returns of all U.S. equity funds; this difference is significant at the 99 percent level of confidence. This multiplier effect is also pronounced for highest quartile SR funds. Funds chosen based solely on their SR contribute 0.5 percent for U.S. equity funds relative to the average return of all funds in this category. When highest quartile SR is combined with lowest quartile expense ratio and turnover, the incremental effect on return is three times higher at 1.5 percent, and this excess-to-average return is statistically significant at the 99 percent level of confidence. The positive effect of the combination of multiple characteristics exists internationally, though it is less pronounced.
When considering the results, it is important to note that the impact on expenses (and thus returns) of a fund’s turnover is not limited to only brokerage commissions and bid-offer spreads but also includes market impact costs, which are typically greater for active funds, which tend to be demanders of liquidity. The cost of market impact is greatest in small stocks. Malkiel and Saha concluded: “It is possible to create a portfolio of funds based on a combination of factors (such as expense ratio, turnover and Sharpe ratio) that will have considerably higher average returns and higher risk-adjusted returns than the average active fund.”
One contrary and thus interesting finding was that the expense ratio for international equity funds (12 basis points higher than for domestic funds) was positively correlated with future returns. Malkiel and Saha hypothesised: “Equity funds specialising in developed and emerging foreign markets incur considerably higher expenses. Portfolio managers often need to exert greater effort to understand foreign assets and the competitive, accounting and regulatory challenges facing foreign companies. Greater expenses are also involved if the manager travels to visit the companies in the portfolio. The positive relationship between performance and expenses for international equity funds appears to suggest that they may utilise their research investments more efficiently than domestic funds to generate better returns.” Unfortunately, the SPIVA data shows that active international funds persistently perform just as poorly as domestic funds.
For example, the 2018 report found that, over the prior 15 years, 83 percent of global funds underperformed, 90 percent of international funds underperformed, 76 percent of international small funds underperformed and 96 percent of those supposedly inefficient emerging market funds underperformed. And these are all based on pretax returns. Since taxes are typically the greatest expense for active funds, the odds would be even worse for taxable investors. That said, Malkiel and Saha provide us with a way to at least improve those miserable odds.
Implications for investors
In their conclusion, Malkiel and Saha note that, unfortunately, a strategy of investing in portfolios of the selected funds each year is unlikely to achieve the excess returns shown because only about 25 percent of the chosen funds repeat from year to year. Therefore, considerable turnover would be required to continue to be invested in the portfolio of the chosen actively managed mutual funds. The transactions, and especially the tax costs, to execute such a strategy would reduce the net returns achieved.
There’s another implication. It’s easy to identify funds with low cost and low turnover, as they are likely to be index funds or other passively managed structured portfolios (such as those of Dimensional). Of course, index funds provide benefits relative to active management; they have a low expense ratio and low turnover (resulting in relatively high tax efficiency and relatively low transaction costs), avoid the risk of style drift, have minimal cash drag and are totally transparent. But they also come with some negatives. The negatives result from the fact that the sole goal of index funds is to replicate the indexes they are tracking. What can be referred to as “structured investing” (no individual security selection or market timing) incorporates the benefits of indexing while at the same time minimising its negatives. My May 27, 2014, Advisor Perspectives article covered nine such negatives.
Sensitivity to risk factors varies over time. Because indexes typically reconstitute annually, they lose exposure to their asset class (or factors, such as beta, size, value, momentum and profitability) over time as stocks migrate across asset classes during the course of a year.
Forced transactions as stocks enter and leave an index result in higher trading costs as index funds demand liquidity. Instead, a fund can engage in patient trading, using algorithmic programs that allow them to provide liquidity instead of demanding it.
Risk of exploitation through front-running.
Inclusion of all stocks in the index. Research has found that very low-priced (“penny”) stocks (note that the Russell indexes do exclude all stocks priced under $1), stocks in bankruptcy and IPOs have poor risk-adjusted returns. A structured portfolio could exclude such stocks using a simple filter.
Limited ability to pursue tax-saving strategies, including avoiding intentionally taking any short‐term gains and offsetting capital gains with capital losses.
Ability to preserve qualified dividends. A fund must own the stock that earns the dividends for more than 60 days of a prescribed 121-day period. That period begins 60 days prior to the ex-dividend date.
Limit securities lending revenue to the expense ratio. When lending securities, otherwise qualified dividends become non-qualified, losing their preferential tax treatment. However, from a tax perspective, securities lending revenue can be first used to offset the expense ratio of the fund.
Screen for the other factors that have expected premiums. Companies such as Dimensional use momentum screens, incorporating them into their fund construction strategies, allowing them to avoid buying stocks that fall into their buy range but are exhibiting negative momentum. They will wait until the negative momentum ceases before purchasing a stock. At the same time, hold ranges allow the funds to benefit from positive momentum. Patient trading strategies (using algorithmic programs) allows them, when selling shares, to give priority to stocks with the least favourable momentum. Just as the momentum factor can be incorporated, so can the profitability factor (and other factors).
Avoiding forced trades due to cash flows. Accepting the risk of tracking error against a benchmark allows structured funds to engage in patient trading activity that provides opportunities to “sell liquidity” and earn a premium by purchasing stock at a discount. The opportunities arise (specifically in small-cap and especially micro-cap stocks) from the desire of active investors to quickly sell more stock than the market can absorb at the current bid. This can be a large benefit during periods of crisis, as long as the fund itself is not subject to investors fleeing the fund in a panic.
Another important benefit of a structured portfolio is that it can provide greater exposure than an index fund to the factors you are seeking exposure to (such as value, size, momentum, quality/profitability and so on). Note that an active fund could engage in many of these same behaviours — behaviours which have expected benefits.
The bottom line is that if you are going to use an active fund, it should be not only one that has a low expense ratio and low turnover (Vanguard’s active funds fit those two criteria) but one that might also use at least some of the above strategies. Doing so will give you the greatest chance of achieving your goals. In other words, active funds are not bad simply because they are active: they are bad because of high total costs (and they create the risk of style drift). Remember, the efficiency of the market not only prevents skilled professionals from outperforming, it also protects dumb money from being exploited (as long as it invests in low-cost, low-turnover funds).
The bottom line is that, while expenses are an important consideration when choosing a fund, other metrics and fund construction utilised should be considered, including how much exposure to a factor a fund is providing (one should consider the cost per unit of expected return, not just cost).
(Full disclosure: My firm, Buckingham Strategic Wealth, recommends Dimensional funds in constructing client portfolios.)
Larry’s other recent articles for TEBI:
Active managers no better able to manage risks than passive indices
Four reasons why the SEC’s Best Interest rule doesn’t cut it
Active investing is becoming harder, not easier, as passive grows
Be careful how you frame the problem