This blog post was originally published on the MoneyGeek.ca blog by Jin Choi. The website no longer exists, but Jin has graciously allowed us to re-publish his research for the benefit of future investors forever.
Some months ago, a reader sent me the performances of two funds, and wanted to know what I thought about them.
On surface, both funds had eye catching numbers. The most popular measure of risk adjusted returns is called the Sharpe ratio, which is roughly the ratio of past returns to risk. One fund had a Sharpe ratio of 1.4, while the other had a ratio of 2.3. Both numbers were way above the Sharpe ratios of the U.S. stock markets during their respective time periods. In other words, both funds had outperformed the U.S. stock market when you accounted for their different risk levels.
However, instead of endorsing them, I expressed my skepticism regarding these funds. Not because I doubted these numbers, but because they didn't have long enough track records.
At the time the reader sent me their performances, the fund with the Sharpe ratio of 1.4 had exactly 3 years long track record. On the other hand, the fund with the Sharpe ratio of 2.3 only had 1 year and 3 months long track record.
Now, you might understand why I think 1 year and 3 months might be too short. But what about the fund with 3 years? Isn't 3 years a long enough time to judge if the fund manager has skill or not?
I don't think it is.
Partly, this is because the average market cycle lasts longer. The average bull market (period of time in which the stock market keeps marching up) has historically lasted about5 years and 7 months. Unfortunately, we often don't know how much risk a fund is actually taking until the bull market ends and the markets go down.
But that's not all. With young funds, you also have to beware of something called survivorship bias.
Let me explain survivorship bias by explaining how you can take advantage of it yourself. However, let me just say that I'm not at all encouraging you to do what I'm going to describe.
How To Appear Like A Stock Market Genius
Let's say that there's a guy named Gecko. Gecko is no one special. He has more or less average intelligence, and he's not especially hard working. But he is interested in the stock market, and he wants to get rich.
Gecko devises a plan to make money by selling an expensive stock tips newsletter. But since he's just starting out, he doesn't have any credibility, so no one seems willing to subscribe to the newsletter. So he devises a cunning plan.
First, he purchases a list of 10,000 emails from an unscrupulous website. Then, he picks a stock; let's say he picks Twitter. To half the people, he emails them and tells them that Twitter's stock is going to soar next month. To the other half, he tells them the exact opposit; he tells them that Twitter's stock is going to tank.
Let's say that a month later, Twitter stock goes up. This means half the people would have received a prediction that came true, while the other half would have received a prediction that didn't.
So then, Gecko picks another stock - let's say he picks GM. He takes the list of 5,000 for whom the Twitter prediction came true, and he divides the list again in half. To one half, he says that GM stock will soar next month. To the other, he says that GM stock will tank.
Let's say that another month later, GM stock tanks. Half of the 5,000 would have received a prediction that came true, while the other half wouldn't have.
So then, he discards the list of people to whom he sent the wrong prediction, picks another stock, and emails them again. In total, he does this 10 times.
By the end, he will have been left with just 10 people for whom his predictions became true each and every single time.
Now, imagine being one of these people. Every month, you would have received an email from Gecko, saying such and such stock will either soar or tank. To your surprise, Gecko would have nailed every one of his predictions.
So if Gecko says you'll now have to pay for his predictions, ask yourself - would you do it? I think many people would,.
This is survivorship bias in action. If we only see the good outcomes, we can be fooled into thinking that there's skill when it's really just dumb luck.
So what does this have to do with funds? Because survivorship bias comes into play as well in the mutual fund world - both intentionally and unintentionally.
Are Surviving Mutual Funds Just Lucky?
Mutual fund companies often 'incubate' new funds. That is, they start new funds and let them experiment with different investment styles and managers. If the fund doesn't do well in the first few months, they fold it quietly. If the fund does well, they start to promote it to outside investors.
As investors therefore, we would only ever hear about the funds with good performances. Unfortunately, that means we wouldn't be able to tell whether that's due to luck or skill.
But even with funds that are past the initial incubation stage, we can't rule out survivorship bias. When a fund underperforms and loses money for its investors, those investors tend to flee. When the fund's total asset under management drops below a certain threshold, they fold the fund. Again, this means that we only ever get to see the funds that have performed well, whether by skill or by dumb luck. This is why many people feel that even 3 years is too short to judge a fund.
Minimum Of 5 Years?
So how long of a track record should a fund have before we can feel comfortable that survivorship bias no longer plays as great a role?
There are no hard and fast rules, but we can take a clue from theGlobal Investment Performance Standards (GIPS). To be GIPS compliant, a fund must initially show the last 5 years of performance. This suggests to me that many knowledgeable people believe that we need at least 5 years of performance history to discern skill.
Given this, if you're a MoneyGeek member, you might wonder about Moneygeek's own model portfolios. Both the regular and premium portfolios are just over a year old. Does this mean you should wait and see how the portfolios do before they invest?
If you're very conservative, yes you can wait until we get a long enough history. However, I personally don't think it's necessary (though I'm obviously biased).
First, the regular portfolios mostly consist of ETFs. These ETFs track indices and these indices have very long track records.
Secondly, the premium portfolios are modeled after what I personally invest. Although the portfolios themselves have a limited history, I myself have a longer track record(though I should note that the performances are unaudited for cost reasons, and also that I hold some penny stocks as well as options that are not included in the premium portfolios).
That's what I have to say in MoneyGeek's defense. As for the funds I mentioned in the beginning, the funds haven't lived up to expectations since and have significantly underpformed the U.S. stock market over the past year.
Was survivorship bias at play? It sure looks like it.
This blog post was originally published on the MoneyGeek.ca blog by Jin Choi. The website no longer exists, but Jin has graciously allowed us to re-publish his research for the benefit of future investors forever.
