Speech by SEC Staff:
Remarks Before the 2006 Securities Law Developments Conference Sponsored by the Investment Company Institute Education Foundation
Brian G. Cartwright
U.S. Securities and Exchange Commission
December 4, 2006
Thank you very much for that kind introduction.
I'm pleased to have been invited to participate in this 2006 Securities Law Developments Conference. The sponsor of this Conference, the ICI Education Foundation, has said that it "partners with government agencies and other nonprofit organizations to help promote or enhance investor awareness." A central goal of our securities laws is to ensure that investors have the information they need to make well informed decisions. But even if investors are provided the information they need, they will be unable to make good use of that information if they lack the requisite skills. So I'm pleased to learn that the ICI Education Foundation is a part of the effort to "enhance investor awareness." Investor awareness, as you'll hear in a moment, is one of the themes of my remarks today.
Before I go any further, though, I have to pause to remind you that my remarks today reflect my own views, and not necessarily those of the Commission or any of its Commissioners or, for that matter, my colleagues on the staff of the Commission. It would not be setting a good example were the General Counsel of the SEC to neglect to deliver that important disclaimer!
At the SEC, our core mission is to protect the individual investor. And that means mutual funds are of great interest to me.
Why is that?
First, because very few individual investors should be picking their own stocks; instead, they should be investing through mutual funds, ETFs or similar professionally managed vehicles. And the same goes for picking almost any of the other types of securities out there.
Why shouldn't individual investors pick stocks? Well, take me, for example. I spent more than twenty years assisting issuers and underwriters in satisfying their disclosure obligations. I know my way around a 10-K and an S-1. I'm no accountant, but I think I generally know how to read financial statements and where to look in the notes for the good stuff.
Beyond that, my background as a physicist provided me with the mathematical tools to follow the outlines of modern portfolio theory and financial analysis. Talk of covariance matrices or leptokurtosis doesn't intimidate me.
Nonetheless, it would be just plain nuts for me to pick my own stocks, except maybe just for fun with de minimis amounts of capital, the way some people might spend an evening in Las Vegas.
Because I'm just one guy, and like most investors, I've got a day job which means I'm never going to be able to beat talented pros, who work in teams, not just full-time, but overtime.
And, make no mistake about it, if you pick stocks, you're participating in a competition. In fact, it's a zero-sum game. The mathematics is simple: if you beat the market portfolio by delta, other market participants must in the aggregate underperform the market by the same delta.
Why should I think I could outperform the pros, based on an hour or two stolen every now and then from my other activities? It ain't happenin.'
So, I say to most individual investors: "Don't try this at home!"
Of course, most individual investors certainly don't need me to tell them that. They figured it out for themselves long ago. According to ICI figures, the number of Americans investing in mutual funds is nearing the 100 million mark.
And U.S. mutual fund "AUM" assets under management now approaches the nine trillion dollar mark. That's a number so big, I think it's time to start using exponential notation: that's almost 1013 dollars. Big. Even for this former astrophysicist.
If you're my age, you may remember Senator Everett McKinley Dirksen. They named a Senate office building after him. But people remember him most for reputedly having said: "A billion here, a billion there, pretty soon you're talking real money." Well, 1013 is ten thousand billion. I think even Senator Dirksen would be impressed.
And that's the second reason I'm interested in mutual funds.
With almost 100 million investors entrusting almost 1013 dollars to mutual funds, it would be irresponsible not to be. After all, improving the performance on even a small portion of that total by even a small amount would translate to extraordinary aggregate benefit to investors. Especially when you consider that many investors have time horizons measured in decades, so the exponential magic of compounding translates small differences in performance into whopping differences in terminal portfolio value. The ability of many Americans to retire in comfort literally hangs in the balance.
So I would like to ask today: is there room for significant improvement?
Many others, of course, have asked and are asking the same question, and I make no claim here to any novelty or originality whatsoever in these remarks. And I do hope to bring humility and an open mind to the inquiry.
But I do insist on asking the question: is there room for significant improvement?
Now, some perhaps may differ, but I begin from the viewpoint that one should have very good reasons before interfering with a satisfactorily functioning competitive market. So, we should begin by considering whether the mutual fund marketplace has the characteristics of a satisfactorily functioning competitive market.
This certainly isn't the time or place to attempt anything like a thorough analysis of industry concentration, which would require careful consideration of many factors, such as the degree of substitutability of various products, the height of barriers to entry, and so forth. Nor am I particularly qualified to undertake such an analysis.
But, at least at first glance, the mutual fund marketplace, when viewed as a whole, doesn't appear to be highly concentrated. At the anecdotal level, there seems to be lots of competitors and lots of competition. And, more objectively, the Herfindahl-Hirschman index is around 400, according to ICI figures. The antitrust authorities typically don't start worrying until HHI gets over a thousand.
So far, so good.
But monopoly power is not the only potential source of market failure. And, unfortunately, I think there may be other signs the mutual fund marketplace may not be performing in the way one would expect in a satisfactorily functioning competitive market.
It begins with the "law of one price." Econ 101 teaches us that in an efficient, competitive market, nearly identical goods will sell at nearly identical prices. That's obviously because, with full information, no rational seller would accept less than the market-clearing price, and no rational buyer would pay more.
Significant deviation from the law of one price suggests market failure of some kind or maybe regulatory failure, or some combination of both.
How does the mutual fund marketplace stack up? Again, this isn't the place to attempt anything like a thorough analysis of this topic, which is better explored by those who have the time and resources to conduct careful studies and publish them in peer-reviewed journals. Nor do I want to bore you with citations to the literature. But in keeping with the spirit of my remarks today to ask whether further inquiry is called for let's take a look.
Let's keep things simple by considering only index funds. Before those of you in the business leap out of your seats, let me assure you I understand that some index funds have smaller tracking errors than others, some funds invest in futures and are able to deploy cash more quickly, some have less churn, and so forth. So there are, of course, differences among index fund products tracking the same benchmark. But such differences are sufficiently modest in scale it's hard to imagine they could account for large-scale variations in price. For present purposes, then, as Gertrude Stein might have put it: an index fund is an index fund is an index fund.
So, if this were a satisfactorily competitive market, the law of one price would apply, and the price to the consumer the fees and expenses of index funds tracking the same index would vary only a little. But, of course, that's just not the case, as is demonstrated by spending a little time with the data readily available from suppliers of mutual fund statistics. It's not hard to find an index fund charging far more than what another index fund charges. We don't seem to be seeing the degree of price convergence one might have expected in a satisfactorily competitive market.
The situation appears to be no better, and maybe worse, for actively managed funds, though the comparable analysis of actively managed funds requires additional judgments about what products should be considered sufficiently fungible for these purposes.
This suggests at least the possibility that many investors are paying more for the services provided by their mutual funds than they would if the price had been set in a satisfactorily competitive market. And even small differences compounded over time against huge aggregate numbers have big effects.
How could that be: how could the law of one price fail, if indeed it has?
I don't claim to have answers. I'll leave that to those more qualified than I. But with your indulgence, forgive me a brief speculation or two, or three, about factors that could be at work here.
For one thing, the law of one price could fail in part because some investors don't understand what they're buying. An investor may not understand that when investing in a mutual fund he or she is really buying two things bundled together.
First, at the time of purchase, you buy the securities currently in the mutual fund's portfolio and pay the current net asset value for them, or thereabouts. You buy the mutual fund's current portfolio at something close to current cost. And then, after signing the check, or initiating the electronic funds transfer, you have nothing more to do. You just sit back and wait to see whether or not your purchase turns out to be all that you expected and more, or a lemon, much like you would after the purchase of a new car.
This out-of-pocket payment seems to be the price you are paying for your mutual fund investment, just like what you pay if you buy your new car for cash. Unless you stop and think for a moment. If you do stop and think, you realize you would pay the same for any other mutual fund that just happened to have an identical portfolio at the time, even though the management of the two funds might be radically different in type, cost, service and performance.
Which means you should be focusing much more on the second purchase you're making, which is bundled together with your first purchase. It's a separate, more consequential purchase: the purchase of on-going investment management services. Interestingly, you pay the price for those services only in the future, as the charges for those services are, in effect, deducted from your nest egg over time. And those future payments are made without any further decision or involvement by you whatsoever.
To shorthand this, you might say that with a mutual fund you pay for one thing, but buy something else. You pay for the securities in the portfolio at the time you invest. But you buy the future services provided by the mutual fund. In a manner somewhat analogous to the payroll deduction of taxes, you'll be automatically charged for those services later, without any further thought or action by you. Given that these deferred charges are "on autopilot," an investor is perhaps to be forgiven if he or she doesn't end up paying much attention to them. And if enough other investors do the same, then I suppose it wouldn't be so surprising to see the law of one price fail.
Another possible factor that could be at work, if indeed the law of one price is not functioning as it should be, is the stochastic nature of mutual fund performance.
When you buy a mutual fund, you want performance. You want the value of your holdings to rise. The trouble is: returns are volatile. Sharpe ratios are of order unity. Or to use a little less jargon, the fluctuations in return are very roughly of the same magnitude as the return itself.
I note in passing that, given those fluctuations, it's difficult for an individual investor to determine from the usual 1-, 3-, 5-, and 10-year historical return data whether or not there's a statistically reliable basis for believing a fund with a better apparent record than other funds was not just lucky.
Be that as it may, if fees and expenses the price of the services provided are, say, on the order of 10% of the expected return, then that price is quite literally dwarfed by the random fluctuations in return. To put it in terminology an electronics engineer might employ, the "signal" represented by the fees and expenses the price of mutual fund services is typically buried in "noise" that's roughly an order of magnitude greater.
If an investor focuses solely on performance, and so is sensitive to variations in the price of mutual fund services only indirectly through its effect on return, that investor could easily be misled to believe that price is not worth paying much, if any, attention to. The price of mutual fund services could appear like a nearly invisible rounding error. And if enough investors behave like that, the law of one price could fail.
Of course, you and I know that the price of mutual fund services should be compared with the expected return, and that the exponential effect of compounding renders that price a matter of urgent importance to every long-term investor.
Of course, possible reasons for failure of the law of one price do not stop there. Yet another possibility is that investors believe the market to be competitive and so implicitly expect the law of one price to apply, but are sold mutual fund products by intermediaries with insufficient motivation to assure that investors get the best deal. That's well worth further exploration, but not by me today, given the limits of my time and your patience.
The perfectly competitive market of the economists is simple and straightforward, but the mechanisms of market failure can be manifold and complex. Whatever the explanation, the possibility that the law of one price may be failing in the mutual fund marketplace suggests many investors may be paying more than they would in a competitive market displaying greater price convergence. If so, even a partial fix could make a big difference to American individual investors, especially when you consider the effects of compounding over many years or decades. Further inquiry and effort only seems appropriate.
If it's possible we're not seeing the price convergence one would expect, then one has to wonder if there are other shortcomings resulting from related deficiencies that may be depriving American investors of the long-term returns they deserve. And, as you well know, one doesn't have to look too far for critical commentary suggesting such shortcomings exist.
So let me wrap up by considering just one more general area where further inquiry might possibly be worthwhile.
Contrary to what some of you may be expecting at this point, I do not intend to join in the debate over active versus passive management. For one thing, to do so would, I think, be tedious for this audience, as most of you probably can rehearse the arguments in your sleep.
Suffice it to say, after seemingly countless studies over the last 30-plus years, there doesn't seem to be much residual resistance to three basic conclusions:
- Active financial markets are highly, though not perfectly, efficient;
- Attempts to exploit the residual inefficiencies often, but not always, fail to add enough incremental performance to cover the costs of trying; and
- The limitations of small-number statistics applied to inherently stochastic financial market returns make it, at best, difficult for an individual consumer to determine whether an active manager is good, or just lucky, in sufficient time to act on that conclusion.
Given those results, it is difficult not to raise at least an eyebrow at the relatively low percentages invested with mutual funds and ETFs that are indexed. Perhaps most investors committing funds to active management are making a fully informed choice, well understanding their odds of beating the market after expenses. If not, perhaps it is worth asking whether we could be doing a better job of helping investors to make such choices with their eyes wide open.
For our purposes today, at least, let's take the promise of active management at face value and ask: are investors who choose active management getting what they pay for?
We need to start by reviewing the basics of one way to look at active management from the standpoint of modern portfolio theory. Of course, this talk is not the place for a tutorial on portfolio theory, nor am I particularly qualified to offer such a tutorial, so I must of necessity speak here only in the most general of terms.
I'm assuming that many in an expert audience such as this already are familiar with this subject matter. If you're not, and have the requisite mathematical tools, I recommend you spend some time looking into it, for example, by exploring a text like the one by Grinold & Kahn entitled "Active Portfolio Management." I must hasten to add I haven't conducted anything like a survey of the available literature and am not in a position to endorse that, or any other text even if my colleagues at the SEC would permit me to, which they wouldn't. But I can promise you it's a fascinating topic.
In any event, I hope the following is both recognizable to those of you who are expert and understandable to those of you who are not. Here goes.
An actively managed portfolio can be decomposed mathematically into the sum of two portfolios: one identical to a benchmark portfolio and one that represents the active manager's choices to deviate from that benchmark portfolio. For example, an actively managed large-cap fund using the S&P 500 as a benchmark can be decomposed into a portfolio that exactly mirrors the S&P 500 and a second portfolio that represents all the overweightings and underweightings of individual components of the S&P 500 that the active manager has chosen to make. Interestingly, this latter portfolio is a long-short portfolio, in that it includes both long and short positions.
Parenthetically, but for the sake of completeness, I should note that this decomposition technique obviously doesn't work very well when applied to active asset allocation strategies but there's also much in the academic literature that's not too kind to those strategies. But that's a topic for another day.
So, putting active asset allocation strategies to one side, from this mathematical perspective, an active manager is running a combination of a virtual index fund and a second virtual fund that looks a lot like many hedge funds. Just for fun, I'll call this second component fund the "virtual hedge fund."
The presence of this virtual hedge fund is, of course, why the investor has chosen active management. The fees and expenses attributable to the virtual index fund should be consistent with those one would expect to pay for a real index fund tracking the same benchmark. One can, therefore, estimate what those fees and expenses would be. The balance of the fees and expenses incurred by the mutual fund are then attributable to the virtual hedge fund the actively managed long-short portfolio.
Because the virtual hedge fund is the raison d'être of active management, if you were an investor in the mutual fund in question, you would be very interested in the return and volatility attributable to the virtual hedge fund, as well as how much in fees and expenses must be paid to earn that return.
Unfortunately, a precise mathematical decomposition of a mutual fund portfolio into the sum of a virtual index fund and a virtual hedge fund can be accomplished only by someone in possession of all the relevant information, so analysts with access only to publicly available data must make do with statistical techniques. Nonetheless, as you may know, academics have performed such analyses.
It turns out to be quite illuminating to view a mutual fund portfolio this way: as two portfolios bundled together and sold as one: an indexed long-only portfolio and an actively managed long-short portfolio.
I want to highlight two ways this can be illuminating.
First, this form of decomposition focuses an appropriately intense, bright light on the performance of the virtual hedge fund. Unfortunately, indications from the academic literature suggest this performance in some cases can be dreadful. Too often, the expense ratio of the virtual hedge fund portfolio is outsized and the alpha attributable to it I'm sure I can use that jargon with this crowd is big, and negative. In one academic study, for example, the mean alpha deduced by the author from published data covering over 150 large-cap funds was roughly -9%. Not good. Of course, once you dilute the signal by slopping in the performance of the virtual index fund the way performance usually is reported things don't look nearly so bad.
Second, this form of decomposition also focuses an intense, bright light on the relative size of the virtual hedge fund and the virtual index fund. One can at least imagine a fund managed by an old-fashioned stock jock with a portfolio very different from the relevant market portfolio and, therefore, with a relatively small component virtual index fund.
At the other end of the spectrum, indications from the academic literature suggest in many cases the virtual hedge fund is far smaller than the virtual index fund. Which means, once again using Wall Street jargon, investors are paying for alpha, but getting beta. They are paying for the high costs of active management, but getting instead something that looks a lot like an over-priced index fund. Of course, this was understood even before the mathematical decomposition I described above became one of the tools of the trade. As you know, it was called "shadow indexing," and still is. In the case of larger funds with high AUM, crude back-of-the-envelope calculations suggest the advantage of this approach to an investment advisor could reach into nine figures a year. Which is a lot of money for investors to be leaving on the table.
So perhaps it is worth asking whether we could be doing a better job of providing investors who choose active management with the information they need, in a form they can use, to determine whether or not they're getting the desired bang for their buck and at what level of statistical reliability.
In these remarks, I've touched on only a few possible areas of inquiry. Many more questions could be asked, of course. For example, given the central role of asset allocation in determining portfolio performance, could we be doing a better job of providing individual investors with the information they need, in a form they can use, to help ensure their mutual fund investment decisions are appropriately informed in this regard?
But you have an exciting program to look forward to this afternoon, and I do not wish to overstay my welcome. So let me conclude by returning to my initial question: is there room for significant improvement?
As my remarks today suggest, there seems to be ample evidence to suggest complacency is inappropriate. Given the numbers involved almost 100 million investors and almost 1013 dollars allocation of appropriate resources to efforts to find ways to do a better job than we are doing today seems well worthwhile.
Yours is an industry that harbors extraordinary talents that can be brought to bear on these efforts. Over the last several decades, academia has developed robust disciplines in finance and portfolio theory, and we should seek to harvest the best insights for the betterment of investors.
So this is a discussion I look forward to continuing.
I very much appreciate the opportunity to participate in your program today. Thank you very much for having me.