Survivorship Bias

Includes: SPY
by: Morningstar

By Michael Rawson

Many funds that were around 20 years ago no longer exist. Not surprisingly, the funds that closed tended to have relatively poor performance. Fund companies have a habit of merging losers with better-performing funds, which allows them to keep their clients' assets and mask poor performance.

The average return for all mutual funds currently in the large-blend category was 8.9% annualized over the trailing 20 years through March 2014. While this statement is accurate, it overstates the performance of the average large-blend fund that existed 20 years ago because it excludes the performance of funds that did not survive the period; this phenomenon is known as survivorship bias. Additionally, some of the funds currently in the large-blend category may have been in a different category 20 years ago. This is called look-ahead bias. Look-ahead bias can occur when a fund's category is changed but those changes were not known at the beginning of the time period.

Since an investor 20 years ago faced a different list of funds, we took a look at all large-blend funds that were categorized as such in 1994. The average return of those funds in the category in 1994 that survived to today, regardless of the category they ended up in, was 8.7% annualized. That is only 0.2 percentage points less than the 8.9% return for those funds that survived the period and are currently in the large-blend category. But if we also include funds that did not survive the full 20 years, the average return drops to 8.1%. In other words, the survivorship bias in 20-year returns for large-blend funds overstates the category average return by 0.8 percentage points per year.

After adjusting for survivorship bias, the relative performance of existing funds looks better. For instance, Vanguard Total Stock Market Index's (VTSMX) 9.5% annualized return placed it in the top 19% of all large-blend funds before correcting for survivorship bias. After this correction, its ranking jumped to the top 9th percentile of the category. That's not bad for a fund that offers passive exposure to the market. Likewise, the ETF SPDR S&P 500's (NYSEARCA:SPY) 9.4% return originally placed it in the top 24% of all large-blend mutual funds, but correcting for survivorship bias would place it in the top 11th percentile.

Low-cost, broad market-cap-weighted index funds, like Vanguard Total Stock Market Index and SPY, have a better chance of surviving than their actively managed counterparts. Over the 20-year period, only 34% of active large-blend share classes survived, while 55% of index fund share classes survived. Consequently, their relative performance is better than many investors realize.

Surviving funds tended to have greater assets, lower costs, and better performance. While this was true for both active and index funds, the predictive ability of these variables was much greater for index funds. High-cost and poor-performing index funds were much less likely to survive than index funds with lower costs and good performance. Investors are not likely to stick with high-cost, underperforming passive mutual funds, particularly in light of competition from ETFs.

While these variables also helped predict active fund closures, the difference in the odds of closing was not as strong as it was for passive funds. For active funds, other factors not examined here might have greater influence over the likelihood that an active fund survives, such as a manager change. And investors may be more likely to stick with an underperforming active fund if the underperformance was explainable or if they expected its performance to rebound in the future.

Taking a look at other major categories over the last 10 years, it is clear that survivorship bias has a bigger impact on growth funds. For example, looking at just the surviving small-growth funds suggests that they returned 8.7%, but including those that died shows a return closer to 7.7%. Index funds were more likely to survive than active funds. For example, just 42% of active large-growth share classes survived compared with 74% of large-growth index fund share classes. Growth funds were much less likely to survive than blend and value funds. High-yield bond funds suffered more from survivorship bias than intermediate-term bond funds, which are better diversified. However, high-yield bond funds were more likely to survive than intermediate-bond funds. This suggests that the high-yield funds that closed had worse relative performance than the funds that closed in the intermediate-bond category.

Because index funds are more likely to survive, their performance tends to improve relative to their active peers after correcting for survivorship bias. The average return for only two of the 12 categories in the table below outpaced the relevant benchmark after removing survivorship bias.

While the survivorship bias-free category average returns may provide additional context, the unadjusted category average can still provide useful information about the performance of an investor's current opportunity set.

This exercise serves as a helpful context for investors when presented with fund company marketing material. Naturally, fund companies are going to pitch the funds with the best performance, hide the funds with poor performance, and not even mention the fact that some of their funds no longer exist.

Disclosure: Morningstar, Inc. licenses its indexes to institutions for a variety of reasons, including the creation of investment products and the benchmarking of existing products. When licensing indexes for the creation or benchmarking of investment products, Morningstar receives fees that are mainly based on fund assets under management. As of Sept. 30, 2012, AlphaPro Management, BlackRock Asset Management, First Asset, First Trust, Invesco, Merrill Lynch, Northern Trust, Nuveen, and Van Eck license one or more Morningstar indexes for this purpose. These investment products are not sponsored, issued, marketed, or sold by Morningstar. Morningstar does not make any representation regarding the advisability of investing in any investment product based on or benchmarked against a Morningstar index.