The Mind Games Investors Lose

The easiest way for mutual fund investors to gauge the performance of their investments is to pull up trailing- or calendar-year returns from a website like Morningstar.com. With a few clicks, they can see how a given fund did on an absolute basis and when compared with its peers and an appropriate benchmark.

But does that exercise reveal a correct picture of performance?

Not really, which is a key reason why most investors would be smart to analyze a fund's investor returns, too. Investor returns, also known as dollar-weighted returns, measure how the average investor fared over a given time period by incorporating the impact of cash inflows and outflows.

By incorporating purchases and sales into a performance measure, it becomes clear when investors chased performance or sold an offering just as it was about to experience a rebound. Indeed, investor returns can reveal considerable information about investor behavior--for better or worse.

To dig deeper into how and why investors make poor decisions, I sat down with Jason Hsu and Russel Kinnel. Hsu is co-founder of Research Affiliates, where he is vice chairman. He is also chairman and CIO of Rayliant Global Advisors. In addition, he is an adjunct professor of finance at the Anderson School of Management at UCLA and has published more than 40 peer-reviewed articles, including "Timing Poorly: A Guide to Generating Poor Returns While Investing in Successful Strategies."1 Kinnel is director of manager research with Morningstar and editor of Morningstar FundInvestor, a monthly newsletter. He runs Morningstar's annual "Mind the Gap" study, which compares the returns of the average fund investor with the average fund. Our conversation took place in April and has been edited for clarity and length.

John Rekenthaler: Why don't we start with this simple question: What are investor returns?

Russel Kinnel: Investor returns are essentially a way to understand the relationship between fund returns and investor behavior, and the returns that investors actually get. Because in the real world, performance is not simply a reflection of an average of fund returns or anything close to that; it's investors making buy-and-sell decisions all along the way.

Investor returns use dollar-weighted returns versus time-weighted returns. Time-weighted is the total returns we're used to seeing every day--the official returns. But if you instead dollar-weight them, you're saying: How did the average investor do? You're factoring the money that was flowing into this fund at this time and out of that fund at that time.

Take giant funds such as PIMCO Total Returns PTTRX and Growth Fund of America AGTHX. They are owned by many people. When you compare the two, you get an understanding of how investors are timing their investments, both on an individual fund basis and at the asset-class level. They also give us some behavioral insights into what mistakes investors make and some suggestions on how people can do a better job of making the most out of their investments.

Rekenthaler: The key with investor returns is looking at the investor return compared with the paper return or standard total return that we see, right? I don't know quite what do with an investor return of, say, 7.54%. That might be really good or really bad, depending on how the fund did. But what I am really interested in is whether or not the standard total return was better than the dollar return. Isn't that correct?

Jason Hsu: You're absolutely right. These returns only start to be meaningful when you put it in the nomenclature of return gap. Return gap is the dollar-weighted results that the investor actually achieved versus the printed manager return, which assumes buy-and-hold.

Return gap arises because investors don't actually practice buy-and-hold. What they like to do is go in and out of funds, trying to time either an asset class or a manager, because they believe recent performance gives them information about future performance. They'll tend to overextrapolate the recent three-year performance and chase trends. What accounts for a lot of the observed return gap is really this trend-chasing into either a hot manager or a hot asset class, only to be disappointed when the style or the asset-class performance subsequently mean-reverts.

Kinnel: When I'm looking at individual funds, I like to use investor returns as a reality check because the gap is essentially telling you how people time their investments.

If you have a fund that's been consistently awful the past 10 years, you might actually have a positive investor return gap because people were selling along the way and all of those sale decisions were good ones. Conversely, a great performing fund, if it has been great the past 10 years, there's probably going to be a sizable gap. But maybe you still have a very nice investor return.

So, on an individual fund basis, the absolute number can tell you something. But I think in the aggregate there's much more information to be had, and there it's the gap that matters.

Rekenthaler: Can we think about this in much the same way we measure the performance of a fund manager by the benchmark gap?

Hsu: You really want to look at the investor outcome as a joint effort between the fund manager--and here his or her job really is to beat the benchmark--and the investor--who is trying to select managers. If they both do a good job, the investor excess return over the benchmark will be more positive. Now, frequently what we observe is that investors do a poor job timing managers; as a result, they actually erode much of the manager alpha. It is not usual that the manager alpha is positive and the investor return gap versus the manager return is so negative that the combined investor return versus the benchmark becomes negative. So, it's really a combination of two sets of skills--the manager's skill that leads to outperformance for buy-and-hold investors and then the investor's timing skill in choosing managers, which creates a return gap. Ideally, you would like both components to be positive. But as I said before, the evidence suggests that the return gap generated from timing is often substantially negative as to make the conversation on manager alpha moot.

On Their Best Behavior Rekenthaler: You've both done research on the return gap. Russ, why don't you start? Tell us about what you've discovered.

Kinnel: When we look at the return gaps in the aggregate--we roll up all the funds in an asset class or even roll up all the asset classes--what we see is that there is consistently a gap. The average investor lags the market and the average fund.

Now, we also find that end dates matter a lot. One 10-year period can be a little different from another. For example, the numbers I've run through the end of 2015 show a relatively small gap. In 2014, I saw a small gap, as well. Other times there are bigger gaps. I don't think it necessarily has anything to do with skill or investors doing a better job this time versus other times. It's much more about the end dates and the start dates.

The past couple of 10-year periods came after the 2002 bear market. A year or two after markets start to rally, money started to flow back in to the market. So, we have a smaller gap.

For me, the best way to look at the gap is to look at an average of those 10-year numbers. That gets you to something around between 100- and 150-basis-point gap for all funds. When you smooth that out, it really is pretty consistent.

Rekenthaler: Are investor returns best used for a pool of funds rather than for evaluating a single fund?

Kinnel: Occasionally with a single fund you can understand the story because, for example, maybe it's a volatile fund and it has greater risk that throws people off more. But other times, you want to avoid reading too much into the gap figure of a single fund. For example, a new fund might start off with a good gap because there was little money in it to begin with. Just keep in mind that there can be a lot of noise when you move down into individual funds and shorter time periods.

Rekenthaler: I've found that investor returns are one of the more difficult figures to interpret. There are whole categories where the timing of new funds is an issue. For example, take target-date funds. If you look at the investor returns for target-date funds, they're massively positive.

There are two things going on there. One is that investors have used target-date funds well. As we know, 401(k) assets are very sticky, and 401(k) investors are often defaulted into these funds, or they make a selection once and they forget about it. They don't move the money. That's a good thing. They're staying in them through the downturns. You would expect the investor returns to be fairly attractive.

But the other big deal is, because target-date funds surged in popularity after the Pension Protection Act of 2006. So, you have most of the money in target-date funds coming in recent years, after the 2008 market downturn.

It's an accident of timing. If you have a downturn next year, all of a sudden investor returns are going to look a lot worse because there is a record amount of assets in target-date funds. It won't be because investors all the sudden became stupider.

Hsu: The return gap is a noisy measure. There're a lot of things going on. Like Russ mentioned, end point matters. And, of course, whenever end point matters, it just means there's enough noise in the data and you want to be careful to not draw the wrong conclusions.

But the point you make about target-date funds is a particularly interesting one, because there is literature in finance that says that if you dollar-cost-average your investments, which basically means if you put equal parts of money into the market as it rises and falls, you tend to do better than a simple buy-and-hold strategy.

That's why we generally see very sticky and disciplined investment programs have a positive return gap. That's very much the case with target-date funds. When you're in a 401(k) account, you have regular contributions from your paycheck even during downturns. This is essentially a dollar-cost-averaging approach to investing. This has certainly been more successful for investors than trend-chasing in bull markets and panic-selling in bear markets.

Kinnel: Target-date funds are a wonderful lab for people like us to understand that behavior, because they're almost exclusively in 401(k) funds. Jason talked about this being a confluence of manager and investor behaviors. This is really where both are on their best behavior.

Fund managers have created a less volatile fund that is fairly easy to understand. 401(k)s are a good device that get people investing with every paycheck and just about everyone keeps investing with every paycheck, even through bear markets--unlike what they do when they invest outside of their 401(k) account.

A Link to Fees Rekenthaler: Russ, what other things have you noticed when you've gone through the data?

Kinnel: Fees and investor returns have a very strong relationship. At first thought you think, "Of course low-cost funds have better performance." But the gap goes beyond even the fee. It's because lower-cost funds have better performance, so there's a little better reward for the people who got into those funds, whereas the higher-cost funds may be taking a little more risk. There's also the fact that the lower-cost investor may be a little bit more savvy and a little more patient--or at least their advisors are more patient and savvy.

Rekenthaler: Is this the dumb investor theory? Dumb investors self-select for high-cost funds and then make dumb decisions with them?

Kinnel: That's part of it. But also some higher-cost fund managers will take on greater risk to try to make that fee up.

Look at some of Vanguard's nominally active funds, like their bond funds. They essentially say, "You know, we already have a fee advantage. Let's just add value on the very margins and not be too different from the market." That alone is going to bring good results for their investors.

Rekenthaler: You are also suggesting that more volatile funds have a higher return gap?

Kinnel: Highly volatile funds trigger emotions. "That fund gained 50%. I wish I owned it. Everyone else owns it. I need to get in now, because I'm missing out and it is going to keep going up."

Or, "My fund is down 50%. It wasn't my mistake. It was that stupid manager's or the stupid market's. I'm getting out. This game is rigged!"

Take that versus a fund that has a top return of 20% and a worst return of 10%. That's boring as hell. Jack Bogle once praised Dodge & Cox as being really boring, which is high praise in Jack's world. Dullness leads us to leaving investments alone, where no emotions are triggered. Generally, the emotional response leads to bad behavior.

Rekenthaler: Jason, you found the same sort of thing about fees? It's fascinating. I never would have even thought about sorting on fees for behavior.

Hsu: We wanted to test this "naive investor" theory. Naive investors may be more prone to behavioral mistakes such as trendchasing or overreacting to short-term performances. We really attribute the investor gap that Russ and I have been writing about to investor behavioral biases.

What we did was to examine a number of investor attributes that could be interpreted as a proxy for investor naivete. The first thing we thought of--and it's exactly Russ' intuition--is investors who pay lower fees are investors who either work with a financial advisor who understands that high fees result in poor fund performance or are themselves more financially savvy.

When you look at fees, investors who buy lower-fee funds tend to have much lower return gaps, whereas investors who pay high fees could have three times the return gap shortfall of the smarter cohort. You see the same thing when you look at value versus growth. The investor return gap for people who buy value funds versus those who buy growth funds are quite stark. The growth-fund return gap is about twice as large as the value-fund return gap.

Similarly, investors who buy institutional share classes have a lower return gap than those who buy other share classes. Clearly, the advantage of the institutional share classes is they're cheaper, but everything else is roughly the same.

So, there is something to the theory. That investor return gap could be a proxy for investor naivete.

Kinnel: I thought your point on value was really interesting. The negative part being that when it comes to investor returns, value-fund investors pretty much give back that value premium that the academic research suggests is there. The only good news is they still did a lot better than growth-fund investors.

Hsu: That's right. So, there is still some value to that knowledge.

Kinnel: If you step back and think about value stocks, not just value funds, perhaps the reason the value premium exists is that not just because of fund investors but because everyone else tends to go out of those stocks at the wrong time. Go back to 2008 and 2009. Understandably, people were selling a lot of the value sectors because the economy and the financial sector were getting crushed.

But to enjoy that value premium, you needed to be there for the rebound, which started around March 2009. If people really were buy-and-hold, then maybe that value premium would go away.

Hsu: Absolutely. It's something that I've actually argued with Bill Sharpe a number of times. Bill's point is, "Look, in aggregate all investors must combine to earn the market return--alpha is a zero-sum game. So, if there's a value premium that you're earning, who's losing the value premium to you?" And that's an incredibly insightful and really important question to answer. I want to earn the value premium and you want to earn the value premium. So, who's losing?

If you look at the value premium that actually has been extracted--you look at the dollar weight of profits for value investors--what you see is it's not actually positive. It's negative by more than 100 basis points.

Just looking at mutual funds, you have this trading in and out of value funds. Investors go into value funds after value funds already outperformed massively, only to earn their anemic and often very negative value excess return for the next holding period.

I think the tech bubble is a good example of that. The global financial crisis is another great example of that, where people got out at the wrong time and got back in at the wrong time.

When you look at that, it starts to make sense why the value premiums could persist because when it pays the most, when it makes sense to be a value investor, most people are not. The value premium is there but, funny enough, people don't want to earn it when that's available. Most of them only want to get access to it after value stocks have outperformed and are already very expensive and not priced to offer much premium.

Rekenthaler: Are you suggesting the value premium is associated with the emotional difficulty, the psychological difficulty of owning value securities? That it is related to the return gap? Are these things all interconnected?

Hsu: Yes. But it is more than the value premium. It's probably related to many of the premiums we see that are persistent. People tend to not want to hold these exposures, these styles, when they're priced to pay a meaningfully positive premium. This is true for the low-volatility premium and many other factor premiums. A lot of these behavioral anomalies that seem to persist and haven't been "arbitraged out" by investors are probably sustained by the fact that, when the exposure pays the most, most people are unwilling to hold those positions.

Rekenthaler: Is the return gap coming from investors mistiming their moves to or from value and another investment style? Or is it people using value funds poorly, trading one value fund for another at the wrong time? Have you been able to separate out those effects?

Hsu: Anecdotally, we see both effects. Did my value manager land in the top quartile during the recent value style rally? If he didn't, maybe it is time to switch.

When that happens investors are really just switching from a normal, average value manager to someone who's a much deeper value manager. And again that switch tends to happen at exactly the wrong time.

Kinnel: There have to be some people who create those premiums for others. There are some fairly clear lessons for fund managers, investors, and advisors. One way to summarize it would be you should have a high bar for what gets into your portfolio and you should have a high bar for what goes out.

You want to understand the strategy and, therefore, why a fund has performed the way it did. That may well lead you to that more nuanced conclusion that maybe my fund strategy is out of style. It's because of its strategy that those are funds that are going to have some big years and some down years.

If investors understand that and raise the bar so that they don't sell unless, say, fees have gone up or the manager has changed, they're probably going to end up with a better result.

Shun the Winners Rekenthaler: Is there any evidence that shows short holding periods are associated with lower investor returns?

Hsu: I did a recent study2 with Brad Cornell and David Nanigian, and we found evidence that the short evaluation period used by investors contributes to poor investor outcome. Many investors review their managers for replacement every two or three years.

If you fire the manager within your portfolio who's done poorly the past two or three years and then hire the top-quartile manager who's done well during that time, you get a much worse outcome than if you just randomly select the managers out of the Morningstar database. The funny thing is you actually would do better if you did the opposite in your portfolio: Keep the underperforming managers and fire the outperforming ones. Truth be told, that strategy would do the very best historically.

Rekenthaler: Would a five-year or a seven-year period or even a one-year evaluation period be more effective? Hsu: If you push out the review to something closer to a full market cycle, this counterintuitive effect starts to become statistically insignificant. Meaning, the performance over the past five years doesn't become a negative indicator. The advice I would lean toward is, if you lengthen out your review cycle, you're going to do better.

Rekenthaler: The lessons are: 1) Work hard to resist the temptation to make a trade and 2) if you are to make a trade, make the opposite trade of what other people are doing.

Kinnel: There are some obvious behavioral issues there. Recency bias is one that everyone suffers from in the investment world. If a manager's got a 20-year record, people still want to focus on those last three. But the correct way to do it is to use that full 20 years. That's going to be much more instructive than the past three years.

Simply understanding a fund's performance pattern over history can help. For example, a lot of people look at trailing returns. But even the annual calendar returns can at least give you a hint: "Here is this fund that has great trailing returns, but it got killed when high yield sold off." Every few years, it'll lag the S&P 500 by a significant amount. The trailing returns can hide a lot of that volatility and variation with the market.

Rekenthaler: What about investor returns from the perspective of different segments of buyers, owners, institutions, advisor clients versus direct clients, or share class?

Hsu: The institutional data is hard to get your hands on, so studies based on them are not comprehensive. However, research shows a strong pattern where the managers that institutional investors allocate to tend to perform much worse than the ones they fire. That's exactly what we see in the mutual fund data as well. That behavioral bias that Russ is talking about is very much present both in the retail investor aided by their advisors or institutional investors aided by institutional consultants.

I have looked at buyers of index funds versus buyers of nonindex funds. What you see is that the return gaps for buyers of index funds are meaningfully lower than a buyer in an active fund.

Rekenthaler: Is there anything that we should be thinking about when it comes to items that might minimize or maximize changes that would turn you off?

Kinnel: Complexity is an issue.

Rekenthaler: By complexity, you mean the complexity of the fund's investment style? Kinnel: Yes, if the strategy is hard to understand. We've seen, for example, unconstrained bond funds and alternative funds sometimes lead to bad investor results. That's partly because there may be an implication of a free lunch that doesn't exist, and so that leads to investors being disappointed.

Rekenthaler: That backs up what I've seen. Complex investment strategies tend to be owned by people who don't understand all the downsides associated with it or the trade-off that comes from the positive aspects of that strategy. Then, when they underperform, they're disappointed and they bail.

Hsu: In a new research I'm doing with Brad Cornell, we also look at high tracking-error strategies--and you can think of that as high-active-share funds, funds where managers take much larger risks versus the benchmark. If you have high active shares, if you take more tracking-error "risk," then your fund will have a higher probability of coming in bottom quartile.

The standard heuristic is, after every two or three years investors look back and ask, "Who are the managers that really stunk up my portfolio? I'm going to fire them and feel vindicated in doing so." Then, what you see is that the managers who are most likely to be fired and fired at the wrong time over that two-, three-year review cycle are going to be the high-tracking- error, the high-active-share managers. That's very consistent with Russ' intuition that these funds do tend to have a higher negative investor return gap for that reason.

I should note that this trend certainly isn't limited to any particular region or market. As I continue focusing my research and business in Asia, I expect the investor gap to be even stronger in more speculative markets, such as the Chinese stock market. In fact, I expect that investors who take a more disciplined approach in these markets will benefit the most.

Rekenthaler: We tell active managers to earn their keep. Earn your expense ratio. Create alpha by having a high active share. Then, when they do that, investors misuse the funds.

Kinnel: I've never bought into the idea that there should be a link between expenses and active share. To me, it's sort of backward. You always want low cost, but I'm not sure that it's worth paying more for the manager to own fewer stocks. You can add value with more stocks. Active share isn't a predictor of outperformance or alpha, so I think it's a nice idea largely put out there by high-active-share, high-fee managers.

Rekenthaler: How can the investing public become better?

Kinnel: I go back to target-date funds, where you see the best behavior.

For the fund industry, it means designing well-diversified funds that don't take on unnecessary volatility. Even lower-volatility strategies may often be the best solution. It's not going to get you a lot of money in the door quickly, but it can lead to better results. Communication around those strategies is important.

For advisors, I think finding ways to continue to educate investors and build better plans around their needs are keys. Individual investors really need to understand the plan that's being put forth before them. When you have these elements working together, it's far less likely for anyone in that chain to act emotionally and make big shifts at the wrong time.

Hsu: My take here is that we spend way too much time, resources, and educational dollars on figuring out the top-quartile manager versus the average manager. In reality, when you look over a long horizon, the difference between them is not really large.

The biggest return detractor for investors actually comes from this pursuit of the top-quartile or even top-decile managers. In that pursuit, we observe investors reallocating too frequently across managers, styles, and asset classes. To meaningfully improve investor outcome, we should shift our focus to how funds are being sold and being bought, and really educate people on the danger of switching in and out of funds rapidly. That has a much, much larger impact on the investor outcome over the long run than trying to see if you can get lucky and pick a top-quartile manager.

Kinnel: I would suggest, too, that we need to build tools to give the right feedback to investors. People often learn the wrong lesson, and it's a subtle point to explain to them you shifted out of stocks and into bonds at the wrong time.

But it's very doable. Look at the technology out there and all of the firms that have investor portfolios. They generally are also doing analytics on their own managers that would point that very fact out. They just don't show that to the individual investors.

If we had all of those firms providing that kind of response and saying, "Here's what you screwed up," if we shared that, maybe people would start to learn the right lessons.

Endnotes

1 With Brett W. Myer and Ryan J. Whitby, May 1, 2015. 2 "The Harm in Selecting Funds that Have Recently Outperformed," Feb. 25, 2016.