Our study finds investors can have greater confidence in finding alpha-producing managers in some categories more than others.
If you’ve had your finger on the fund industry pulse recently, you are well aware that the active-versus-passive investing debate is alive and well. Unfortunately, it seems that many adherents on both sides of the aisle tend to advocate an all-or-nothing approach— either all passive or all active. However, a rational investor should take a more nuanced view where fees are weighed against the potential for outperformance in category-bycategory consideration. In what follows, I will provide a decision framework for the trade-off between the potential for outperformance offered by active management with the cost savings offered by passive management. In addition, I will discuss one method for determining which active fund industry segments may be worth considering for an active allocation.
Are the Costs Worth It?
When faced with the choice of investing in an active or passive fund, an investor must weigh the potential for outperformance against the assuredness of incurred costs resulting from the higher fees charged by actively managed funds. Costs will, by construction, negate performance regardless of whether or not we realized any outperformance in the process of investing actively. Given this adversarial nature between costs and performance, therefore, we need to get a handle on the expense ratio premium of active funds versus their passive counterparts. This premium is essentially an investor’s entry fee for the opportunity to outperform the passive fund. Quantifying the expense ratio premium of an active fund over a passive fund is relatively straightforward and can be done by simply comparing expense ratios of the two products.
Once investors have an idea of what cost will be incurred, they next should form a realistic expectation for the outperformance that the active fund can achieve over the passive fund for a given time horizon. And, perhaps more importantly, investors must estimate their confidence in their ability to choose an outperforming active fund.
There are a variety of ways to do this. For example, the majority of research firms have focused on comparing simple returns historically (e.g., the U.S. Large Blend index returned 10% and 80% of large-blend funds returned less than 10%). However, this analysis fails to capture the different risk exposures of these funds. Some funds that may be designed to deliver lower returns with a lower risk profile in the same category and are thus unduly discriminated against in a simple returns-based analysis.
Using a method of comparison that focuses on risk-adjusted returns, however, avoids this bias and concentrates on the manager’s skill at trading risk for reward. One of the preferred measures for doing risk-adjusted return comparisons is alpha, which provides a fund’s estimated return after controlling for its risk profile. By definition, an index fund’s alpha will be zero whereas an active fund’s alpha can be either positive or negative.
Unlike the expense ratio premium of active funds over passive funds, which is known with certainty, any estimate of an active fund’s future alpha will be prone to error. Therefore, we need to account for this lack of certainty when we compare the expense ratio premium to the expected alpha of the active fund. In order to do so, we estimate the probability distribution of alpha by Morningstar category and see where 0% lies in this distribution. If the bulk of the distribution lies above 0%, that should signify that not only are active funds performing well as a whole but that investors should have higher confidence in their ability to identify those top tier funds. Conversely, if the bulk of distribution falls below the 0% threshold, then managers are not adding value to their investors and investors should be more wary of their ability to identify an outperforming active fund.
Using the probability distribution of alpha has several distinct advantages. First, as a risk-adjusted return measure, the distribution of alpha reflects the expectations for a manager’s skill. We should hope that in the long run, alpha exceeds the manager fees. To estimate manager’s value proposition, therefore, it is wise to use alpha estimated net of fees and judge it relative to 0%.
Second, as opposed to simply estimating averages, estimating a probability distribution will give us the entire range of outcomes for alpha by category historically. These historical ranges or spreads in alpha can be used to indicate which categories possess the largest differences in the skill of active managers. For example, a category with a range of alpha from –1% to 1% would indicate that there is little difference in skill between the best and worst managers compared to a category where alpha ranged from –5% to 5%. Using probability distributions in this way allow for comment on not only the potential for alpha but also enable us to identify the categories where manager skill is feasible and rewarded.