Nanigian then examined the data considering only the funds within the bottom quintile of expense ratios. He found that the index funds outperformed again, though the difference was smaller. Index funds in the bottom quintile of expenses had alphas of -0.37% compared to the -0.59% alphas of the active funds. The difference of 0.22 percentage point was not statistically significant (t-stat 0.68). This finding confirms that of prior research that found that expense ratios are strong predictors of performance.
In fact, the authors of a 2015 study by Morningstar concluded: “Fees matter. They are one of the only reliable predictors of success.”
In the April 2005 issue of the Morningstar Fund Investor, they went even further. “For starters, expense ratios are the best predictors of performance—way better than historical returns. It’s tempting to look at strong past performance and assume a fund can repeat its success, but there’s no guarantee it will. In fact, we’ve found that you’d be better off randomly picking a fund with expenses in the cheapest quartile than a fund with returns in the highest quartile and expenses in the highest quartile. Higher expenses don’t get you better management. If it did you’d expect higher-cost funds to outperform their lower-cost peers—when in fact just the opposite has happened.”
Nanigian ran a third test, the most interesting one, in which he paired each passively managed fund with an actively managed “partner.” To identify the “partner” for each passively managed fund, he first restricted the sample of actively managed funds to only those that belong to the same Morningstar category, and then chose the one fund whose expense ratio was closest to that of the passively managed fund. In the event that multiple actively managed funds were identified as eligible partners, the passively managed fund was paired with a hypothetical “fund of funds” that allocates capital equally among each of the eligible partners.
He found that the mean and median differences in the expense ratios across the fund-months were 2 basis points and 1 basis point, respectively. Nanigian’s results should not be surprising—the outcomes were almost identical. The alpha of the passive funds was -0.38% versus the -0.34% alpha of the active funds, both economically and statistically insignificant (t-stat of the difference was ‑0.10).
The finding of virtually no difference in performance of the paired funds should not really be a surprise. While the efficiency of the market prevents active managers from outperforming, it also prevents them from underperforming once expenses are considered.
It also should not have come as a total surprise that when the expenses of active funds were virtually identical to those of the index funds, the active funds had slightly outperformed. The reason is that indexing (with the exception of total market funds) has some negatives that can be avoided or minimized by intelligent design and patient trading. For example:
- The forced turnover of index funds can allow actively managed funds (such as hedge funds) to exploit that turnover by “front-running” the trades of index funds. For example, the lack of opaqueness has historically created problems for index funds that replicated the Russell 2000 Indices. An active fund could also trade patiently, minimizing trading costs, including market impact costs.
- There are many well-documented anomalies in the academic literature that long-only actively managed funds can exploit by excluding from their portfolios the stocks in the short side of the anomaly—while replicating index funds must include all the stocks in an index. This not only avoids the negative side of the anomaly, but allows them to overweight the positive side. Among the anomalies are poor risk/return characteristics of stocks in bankruptcy, very low-priced stocks, IPOs, extreme small growth stocks with low profitability and high investment, and stocks with high accruals.
- Active funds can also provide more exposure to the well-documented factors of size, value, momentum, profitability and quality than do popular indices. This could allow them to earn higher returns, possibly more than offsetting even higher expenses.
These advantages of actively managed funds can help them overcome potentially higher trading costs related to their generally greater trading activity. It’s also important to note that passively managed funds (i.e., no individual stock picking or market timing) that are not indexed can also exploit these advantages and minimize or avoid the negatives facing replicating index funds.