4

In Survivorship Bias, David McRaney writes:

In finance, this is a common pitfall. The economist Mark Klinedinst explained to me that mutual funds, companies that offer stock portfolios, routinely prune out underperforming investments. “When a mutual fund tells you, ‘The last five years we had 10 percent on average return,’ well, the companies that didn’t have high returns folded or were taken over by companies that were more lucky.” The health of the companies they offer isn’t an indication of the mutual fund’s skill at picking stocks, said Klinedinst, because they’ve deleted failures from their offerings. All you ever see are the successes.

Is this true? Do mutual funds edit/censor underperforming investments to make their returns look better, and if so, is there any way one can figure out if they are doing it?

10

Do mutual funds edit/censor underperforming investments to make their returns look better, and if so, is there any way one can figure out if they are doing it?

No, that's not what the quote says.

What the quote says is that the funds routinely drop investments that do not bring the expected return, which is true. That's their job, that is what is called "active management".

Obviously, if you're measuring the fund by their success/failure to beat the market, to beat the market the funds must consistently select over-performers. No-one claims that they only select over-performers, but they select enough of them (or not...) for the average returns to be appealing (or not...) for the investors.

  • 4
    I think the quote is poorly written; I think it's actually about mutual fund companies that didn't get lucky, going out of business or getting acquired by luckier fund companies. Individual companies in which a fund invested, that went out of business, were acquired, etc. by active management, would indeed count against the performance of the mutual fund, which would not demonstrate the survivorship bias which the article in question was about. – stannius Oct 13 '15 at 16:02
4

There is a survivorship bias in the mutual fund industry. It's not about individual stocks in which those funds invest. Rather, it's in which funds and fund companies/families are still around. The underperforming funds get closed or merged into other funds. Thus they are no longer reported, since they no longer exist. This makes a single company's mutual funds appear to have a better history, on average, than they actually did.

Similarly, fund companies that underperform, will go out of business. This could make the mutual fund industry's overall history appear to be better than it actually was.

Most companies don't do this to deliberately game the numbers. It's rational on the part of fund companies to close underperforming funds. When a fund has a below average history, investors will likely not invest in it, and will remove their existing money. The fund will shrink while the overhead remains the same, making the fund unprofitable for the company to run.

  • So, you think there is a survivorship bias, just not the one the article describes? I have noticed that mutual funds do tend to report ridiculously good returns, making me wonder why everyone does not use them. But even if a mutual fund has been closed, shouldn't they still report it as part of the period it was open/active? – Faheem Mitha Oct 13 '15 at 16:17
  • 1
    @FaheemMitha Re: "report it as part of..." ... report it, where? If it is for marketing material, they can pick and choose and ignore what came before. – Chris W. Rea Oct 13 '15 at 18:42
  • @ChrisW.Rea Well, I don't exactly know how these systems work. But if some funds existed during part of a period that is being reported, even if those funds were shut down later, shouldn't those funds also be reported? – Faheem Mitha Oct 13 '15 at 21:09
  • @FaheemMitha It all depends on what you mean by "reported". Past regulatory filings would have information about funds that existed at that time. As for what the fund company is advertising today, they tend to mention only funds they currently offer (or a subset), i.e. from the survivors. Usually, fund companies advertise the performance of specific funds over specific time periods, not their all-time average performance. (Else, investors would quickly see that most fund companies tend to have average performance before fees, and worse than average after fees.) – Chris W. Rea Oct 13 '15 at 22:45
  • @ChrisW.Rea I see. Thanks for the clarification. It sounds like potential investors should look into funds that used to exist, as well as funds that do exist, for the reasons you mentioned. Are the regulatory filings publicly available? And "the performance of specific funds over specific time period" could be misleading, because they could/would pick their best performing fund in hindsight. Correct? – Faheem Mitha Oct 13 '15 at 23:00
2

If I invest in individual stocks I will, from time to time, sell stocks that aren't performing well. If the value of my portfolio has gone up by 10%, then the value of my portfolio has gone up by 10%, regardless of whether selling those stocks is labeled as "delete[ing] failures". Same thing for mutual funds: selling underperforming stocks is perfectly ordinary, and calling it "delete[ing] failures" in order to imply some sort of dishonesty is simply dishonest.

  • Thank you for the reply. So, you think that removing underperforming stocks from a mutual fund portfolio is not a valid demonstration of survivorship bias? – Faheem Mitha Oct 13 '15 at 16:07
  • No fund claims to perfectly pick winners. All funds change their specific contents as needed. The .decisions may be made by a human or by a computer, but thus management is part of what you're paying the fund to handle for you. The fun's reported returns are the fund's actual returns with these operations accounted for. Nothing misleading is going on; if the fund says it earned 10% over a given period of time, shareholders earned 10% minus the documented fees. – keshlam Oct 13 '15 at 16:44

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .