Funds that don't survive are an important piece of the puzzle when you study the predictive power of expense ratios. In the U.S., where robust costs data stretching back a decade or more is available, I set out to look at survivorship rates of low-cost and high-cost funds so that I could get a true feel for your odds of success when picking funds of different expense ratios. Putting the pieces together that way led to a pretty stark picture.
Survivorship Bias,
or Where High-Cost Funds Go
True, it's also possible for a low-cost fund to perform poorly and get merged away--it's just much less likely to happen than at a high-cost fund. Costs, after all, reflect not just fee structure but asset size. A small fund with high costs is more likely to be unprofitable for the fund company (and therefore a better candidate for the scrap heap) than a large fund with low costs. In addition, higher-cost funds sometimes take on greater risk than they otherwise would in order to compensate for their cost disadvantage--in turn, they put up ugly numbers that can lead to them being merged away or liquidated.
Consider, for example, that, using data for the U.S. market, all the technology funds in the cheapest quintile of their category in 1998 still existed five years later when the bear market in tech was ending and tech funds were rebounding sharply. However, 47% of funds in the priciest quintile were merged away or liquidated before the rebound. The disparity is a little extreme, but the trend is borne out across the broad asset classes and within individual Morningstar US fund categories. I looked at rolling five-and 10-year periods from 1995 on to see whether high costs were a good predictor of whether a fund would be merged away or liquidated. Within each category, I separated funds into quintiles based on their costs, and I excluded categories where there were fewer than five funds per quintile.
It turns out there's a high correlation between costs and survivorship as high-cost funds had a striking attrition rate. For U.S. stock funds, the quintile with the lowest expense ratios had an attrition rate of 13% over five-year periods and 25% over 10-year periods. However, the attrition rate for the most expensive quintile was double that: Over five-year periods, 29% of the high-cost funds had merged or liquidated, and nearly half (49%) had merged or liquidated over 10-year rolling periods.
The pattern was similar for international stocks and taxable-bond funds, though both groups had higher attrition rates across the board. For international-stock funds, the cheapest quintile had an attrition rate of 19% for five-year periods and 33% for 10-year periods, while the attrition rates for the priciest quintile were 37% and 57%, respectively, for the five- and 10-year periods. Taxable-bond funds in the cheapest quintile saw a five-year attrition rate of 21% and 35% for 10 years, while the priciest quintile had a five-year attrition rate of 37% and a 10-year attrition rate of 57%. The really striking part here is that not even half of those high-cost funds survived over a 10-year period.
All things being equal, funds with high costs are much more likely to produce poor performance because of their cost disadvantage. However, all things aren't equal and many high-cost funds in the U.S. market also have weaker management, higher risk strategies, and fewer resources. Put that all together and you have a recipe for failure. When a fund comes out of the gate with a few years of bad performance, the fund company, realising that it will have an awfully hard time attracting assets, kills it off.
Interestingly, municipal-bond funds (muni funds), a tax-advantaged asset class in the States, broke the pattern, as attrition rates of low-cost and high-cost funds were pretty close. The five-year attrition rate for cheap muni funds was 29% and the 10-year rate was 40%, while the priciest quintile shed 25% in five years and 39% in 10 years. I'm not sure why that is, but it might be that the small firms launching trendy junk are less likely to venture into munis, where there's no chance of producing the sort of huge returns that quickly attract money. In addition, the gap in costs between pricey and cheap funds is smaller and therefore the higher-cost funds have less of a disadvantage than high-cost stock funds.
How Costs Predict Success
Now that we've added the missing piece of attrition back in, we can see just how powerful expense ratios are as a predictor of performance. I again separated funds within each category into quintiles based on their costs, then checked on what percentage within each group managed to beat the category average. The category average overstates returns a bit because it reflects only funds still in existence today. That's why even the best-performing quintile in our survivorship-free database saw slightly fewer than half its funds beat the official category average.
The results show just what a bad bet high-cost funds are in the U.S. market. Funds in the cheapest quintile were more than twice as likely to succeed--that is, beat the average for their categories--than those in the most expensive quintile. Success declines rapidly as you move up in price: Of domestic-stock funds, 47% in the cheapest quintile succeeded over a 10-year period, 33% of the next cheapest quintile succeeded, 30% of the middle quintile succeeded, 27% of the second priciest quintile succeeded, and just 19% of the most expensive quintile beat the category average. The pattern was similar in other asset classes. In foreign stocks, funds in the cheapest quintile were twice as likely to succeed as those in the priciest (40% to 18%). For bond funds they were five or six times as likely to succeed (48% to 7% in taxable bonds and 49% and 9% in munis).
Your Chances Against a Cheap Index Fund
Those statistics are pretty compelling, but I also decided to look at how a fund's costs affect its chance of success versus a more formidable benchmark: the cheapest index fund in a category.
In particular, I looked at what percentage of funds within each cost quintile survived and beat the cheapest index fund in the category. It's a harder test than the previous one, of course, because the cheapest index fund in a category is usually the cheapest of any kind and index funds usually incur lower trading costs as well.
| Cheapest Quintile in: | vs. Priciest Quintile |
vs. Second Priciest Quintile |
vs. Mid Quintile |
vs. Second Cheapest Quintile |
|---|---|---|---|---|
| U.S. Equity | 2.5x | 1.7x | 1.6x | 1.4x |
| Intnl. Equity | 2.2x | 2.2x | 1.7x | 1.3x |
| Taxable Bond | 6.9x | 3.2x | 2.3x | 1.5x |
| Muni Bond | 5.4x | 3.1x | 1.5x | 1.1x |
| This table displays how investing in the cheapest quintile of funds within a given asset class would multiply your odds of success versus other cost quintiles. | ||||
Once more we see that low-cost funds fared much better. The cheapest quintile of equity funds was nearly twice as likely to beat the cheap index fund as the highest-cost quintile. The cheapest quintile of bond funds was six times more likely to beat the index fund than the high-cost quintile.
The cheapest quintile of domestic-stock funds survived and beat the cheapest index fund 29% of the time compared with just 17% for the most expensive quintile. For international-stock funds, we see a similar picture: 32% of foreign funds in the cheapest quintile survived and beat the cheapest index fund while just 18% of the most expensive funds were successful. For taxable-bond funds, 19% of the cheapest quintile were successful over 10 years and a mere 3% of expensive bond funds succeeded. There are no municipal-bond index funds to provide a benchmark test for tax-free funds.
Does this mean that all index funds are superior to all actively managed funds? No. There are high-cost index funds, too. If we asked how the average index fund fared against the lowest-cost actively managed fund, you'd likely see that most index funds lag the cheapest actively managed fund.
I'd be wary of drawing too broad a conclusion, such as that indexing is always better than active management, because the above compared the cheapest index fund against all active funds. Of course you can improve your odds by focusing on low-cost funds, whether choosing actively managed funds or index funds. Both types have higher-cost examples with lower chances of success.
Low-Cost Funds Double Your Chances of Success
Factoring in survivorship to the expense equation shows just how powerful expenses are as a predictor. Sure, there will always be exceptions. Winning lottery tickets are great investments--the catch is that you have no way of knowing if yours will be a winner, and they are bad investments for 99.9% of people. From a fund company's perspective, high-cost funds are like free lottery tickets because it has convinced investors to pay the bill. If the fund company is lucky, it will produce big enough returns to attract large sums of money. If it isn't, it'll simply fold up the fund and hide its mistake under the rug. Of course, that doesn't make the fund's investors' losses any less real.
Rather than making a bad gamble, the savvy investor should look for low-cost funds with sound fundamentals. Our research in the U.S. market shows that doing that will double or triple your chances of success and greatly reduce your chances of dramatic underperformance. In fact, our U.S. Analyst Picks (our qualitative research team's favourite funds in that market) apply those fundamental criteria and they have succeeded about two thirds of the time. With the field narrowed, consider which funds, active or passive, have the best combination of low costs, good-quality management, sound strategies, diversification, and stewardship you want. With patience, those funds should have a better chance of getting you to your goals.
