Consistent returns and asset flows help explain why fund companies price bond funds cheaper than stock funds.
This article first appeared in the April/May 2011 issue of Morningstar Advisor magazine. Get your free subscription today!
Having spent years doing manager research at Morningstar, specializing in fixed-income managers, I have an appreciation for the level of sophistication that bond managers apply to their analysis.
Their research into credit quality, collateral status, cash flows, optionality embedded in some bonds, and legal provisions that envelope most fixed-income issues makes for a very complex exercise. And there are no exchanges for bonds, so trading is more difficult and involves more effort.
Therefore, I have to roll my eyes when I hear the argument that stock investing is more difficult than bond investing. Most people would agree that the bond departments of PIMCO, BlackRock, Western, and Fidelity are deep into some serious research. How many stock shops do the equivalent of recalculating the durations and option-adjusted spreads of every one of the thousands of bonds in the BarCap Aggregate Index every night?
Yet despite this sophistication, mutual fund firms charge less for fixed-income funds than they do for stock funds. According to Morning- star Direct, the average expense ratios for each Morningstar open-end category in the world's three biggest capital markets (the United States, Europe, and Japan) are remarkably consistent: Firms price equity funds higher than bond funds. Below are the averages for the three markets(1):
We have all heard about the equity premium puzzle, or why stock investors historically have been able to extract such a high return premium over bonds. But what about the equity fees premium puzzle? If the amount of time and effort that goes into managing bonds is the same as managing stocks, why do stock funds charge more?
I present three explanations. Two are straightforward answers, having to do with history and demand-side considerations. Both contain important elements of truth, but it's hard to see them as containing the full story. My third explanation, however, may complete the picture. It is meant to complement the other two, not supplant them, and present a new angle from which to consider the equity-fees puzzle.
A Sleepy Area of Investing
Before the explosion of complicated fixed- income instruments, globalization of capital markets, availability of serious computing power, and reorientation of focus from income to total return, bond management could be a rather sleepy affair. Back then, the thinking goes, bond managers only occasionally bought bonds. They mostly contented themselves with clipping coupons for income. (No doubt there are some firms that still do this.) In other words, there was a good reason why bond managers charged less: They didn't do much.
I would counter that argument by saying that in those good old days many a stock manager did no more research than his bond counterpart- a bond manager would buy IBM, AT&T, and GM bonds and some Treasuries, while the stock manager would simply buy IBM, AT&T, and GM stock. It is probably true that as the industry has become more competitive the effort that goes into stock research has gone up, but the same has happened to bonds.
Still, I will allow that there was a time when bond managers were lazier than their stock-side colleagues, but not by as much as the differences in fees suggest.
It's the Nature of the Asset Class
A second explanation argues that asset managers charge more for equity funds because they can; investors' high demand for equity funds means that managers can jack up prices.
There are two aspects to this thesis. Because equity expected returns are higher than fixed income-the standard assumption in financial theory-an asset manager should charge a higher fee. In effect, investors are willing to pay higher fees if that is the way to becoming wealthier. Also, stock funds are more volatile than bond funds: They have a higher probability of a large, short-term win. Investors, thus, are willing to pay a premium for a small chance to hit the jackpot, much like a lottery.
This answer explains some things, but not all. To begin with, it's hard to come up with an economic rationale for investor behavior that would continually patronize such blatant asset-manager behavior. Also, there are some anomalies. The two most expensive Morningstar U.S. open-end categories are long-short and bear market. You can't have an expectation of a meaningful equity-return premium while positing that bear-market and long-short funds will be performance winners. To some extent, these could be explained away by the two categories' small size. But then we have situations like the U.S. bank- loans category being more expensive than the U.S. large-blend and large-value categories. Again, there are probably few who would posit that, in the long run, bank-loans funds will outperform large-cap equities. Japan's painful experience during the past 20 years should temper the expectations of solid equity outperformance over the long term. Yet equity funds in Japan are, by and large, more expensive than Japanese bond funds.
Of these two explanations, the first is what one might call a supply explanation. The asset-management industry looks at the cost of supplying the services and sets the fees to cover these costs. The second is more of a demand explanation. The asset-management industry looks at demand-how badly people want to buy the asset class-and adjusts the fees accordingly. We have found both of these explanations of the equity-fees puzzle somewhat incomplete. But another supply- side angle might help us fully understand the dynamics of asset-management fees.
To an asset manager, a fund is a stream of income. The larger a fund's assets under management, the more money the fund generates for the manager. Funds have entire sales and marketing staffs charged with the task of attracting assets. Similarly, to an asset manager, a fund that has a stable flow of assets is worth more than one with a wildly gyrating asset base.
This point goes to the heart of standard economic reasoning: Under most plausible utility or risk-aversion assumptions, economic actors will accept a lower stream of payments (yield) if they can have reasonable assurance that the income stream will be stable. For an asset manager, it makes sense to charge lower fees for a fund that's a mixture of Treasuries, mortgages, and high-grade corporate bonds than it does for a fund owning Latin America equities or technology stocks. The former is less likely to lose half of its value in a span of a few months. (Admittedly, we saw bond funds lose 50% in the 2008 financial crisis, but this was a noteworthy event; after the past decade's market vicissitudes, few would find it remarkable that an equity fund could lose 50% of its value.)
Theory, then, stipulates a positive relationship between the volatility of an asset class' returns and the fees charged by funds operating in that asset class: the higher the volatility, the higher the expenses. Data from the United States, Europe, and Japan supports this contention. I ran a regression on the most recent expenses versus 15-year volatilities of realized returns for different asset categories. My analysis shows a coefficient that is both positive and statistically significant. The regression results are displayed in Exhibit 1. I included only categories with all data points for the whole period.
(A linear regression technique finds a trend line that fits the observed data points most closely. "Closely" has a precise meaning of minimizing the squared deviations between the trend line and the observation points-which in our case measure the fees and standard deviations of Morningstar categories. A t-Statistic measures how reliable a given result is. The higher the t-Statistic, the more reliable it is. T-Statistics above 2 typically indicate a statistically significant relationship; 3 and above indicate a very strong significance-in other words, there is a very small probability that the result could have been obtained by pure chance.)
The meaning of the magnitude of the slope number need not detain us here. For our purposes, what matters is that these numbers are positive, meaning that as volatility gets higher the fees trend higher, and that these are statistically significant findings.(2)
You might be thinking, Of course, fund fees are correlated with volatility. Volatility is, after all, the price for higher returns, so all we are seeing is the second explanation (the nature of the asset class) dressed up in a different garb.
I'll take this objection head-on.
First, we are not necessarily trying to displace the nature-of-the-asset-class explanation. We want to bring up another way of looking at the problem that would complement the other explanations and that is grounded in some economic rationale. The relationship between volatility and fees neatly confirms this new way of looking at the puzzle.
Second, the regression results for realized volatility are much more significant than the results of regressing fund fees versus realized returns for the past 15 years.(3) This means that realized volatility explains fund expenses better than realized returns. All of this is suggestive but not conclusive: It is possible that fund fees respond to expected returns, and it just so happened that the higher-volatility asset classes didn't deliver on the expectation (though it still would be hard to explain the high fees of Japan large-cap funds even though they have posted negative returns for the past 15 years). There is no way to disentangle this angle, because there is no way to measure expected returns for most asset classes.
But there is another way to advance the argument that the volatility matters more than returns by looking at a different kind of volatility. Obviously, the volatility of asset-class returns is the most visible risk to the stability of cash flows from the perspective of the asset-management firm, but ultimately, the asset manager should care about the volatility of its asset base.(4) In addition to asset-return volatility, the asset manager is also exposed to the fickleness of investors' cash flows. It is reasonably well documented that investors fall in love with fads, chase performance, and so on. This line of reasoning is promising because, while there is a theoretical reason for why higher volatility requires higher expected returns, it is much less clear why, if funds are priced off expected returns, the volatility of the asset class flows should correspond to higher returns. In fact, the relationship between return volatility and flow volatility is very weak. (We have decided to show pure flow volatility because asset volatility would include both the effects of returns and flows, confusing the picture.)
Exhibit 2 displays the chart of the relationship between fees and flow volatility, and the regression results. (I use U.S. data only, because foreign data is too spotty.) Flow volatility is based on Morningstar Direct monthly data for estimated cash flows and asset levels for all Morningstar categories for periods between January 1995 and October 2010.(5)
The regression results with a positive coefficient, as expected, whose t-Statistic of 3.48 is amply significant even at 1% confidence level. It appears that volatility of cash flows explains the variation in fees better than (realized) returns do.
The focus on asset volatility can also help explain some of the odd facts about fees that we found earlier. The bank-loan category's standard deviation of returns is less than half that of the large-blend category, but during the past 15 years, its asset volatility has been about 50% higher. Another example is the pricing of foreign large-blend funds versus Europe or Japan stock funds (in the United States). If the argument is that more research and effort requires more fees, then one would expect the foreign large-blend category to be more expensive-after all, it should take more effort and cost to scan the whole globe for opportunities. Yet Europe and Japan funds are more expensive than foreign large-blend funds. Longer-term return volatilities for all three categories are roughly similar. (Returns, too, are similar for Europe and foreign large blend; we all know what happened to Japanese equity returns.) The missing key here might be asset volatility. Foreign large-blend's asset volatility is the lowest. Europe's is 25% higher, and Japan's is almost 100% higher.
A Work in Progress
Two standard explanations attempt to explain why asset managers charge more for equity funds than they do for bond funds. It costs more to supply services for equity funds, and demand is strong enough for equity funds that managers can basically charge what they want. It's possible that these two explanations have an effect on fees. But in today's competitive investing environment, bond shops invest as much time and effort running their funds as equity shops do, and investors would shop elsewhere if they thought the fund companies were jacking up fees inappropriately.
A third, new explanation, however, seems to have legs: The returns and asset flows of stock funds are more volatile than those of bond funds. Therefore, managers charge more for stock funds. This makes economic sense. Because the cash flows from fees of volatile funds are uncertain, fund companies charge more for these funds to counter the risk of fluctuating income streams. Regressions also support our thesis. Not only is there a positive relationship between fees and the volatility of a category's returns but the relationship between fees and the volatility of asset flows is even stronger. (We don't have a good theory of why this last point should be true. We can speculate that high-volatility asset classes attract more short-term traders who want to make a quick profit and get out; short-term gains are much more likely to materialize among, for example, precious- metals funds than among municipal- bond funds.)
This new idea is a hypothesis. It is meant to complement, not supplant, the two other explanations, and it deserves further exploration. The evidence we presented, while not conclusive, certainly seems to support our theory. Besides, there is a good rationale for the asset managers to consider the volatility of assets as an important factor in setting fund fees. We may not have solved the equity premium puzzle, but we hope to have advanced our understanding of the equity-fees premium puzzle.
1) The table excludes allocation and alternative funds, as well as muni funds in the United States.
2) Only 20 open-end Japan categories had enough data for the whole 15 years, which affects confidence intervals.
3) The t-Statistics for realized return versus fees are: U.S., 2.72; Europe, 3.85; Japan, 2.39.
4) We measure the asset volatility by calculating the standard deviation of monthly changes in asset levels for each category.
5) Flow volatility is calculated as the standard deviation of each category's monthly flows expressed as a percentage of its beginning-of-the-month assets.
Maciej Kowara, Ph.D., CFA, is a senior research consultant with Ibbotson Associates, a Morningstar company.