7

25 YEARS OF INDEXING: WHEN ACTIVE MANAGERS WIN, WHO LOSES?

Keynote Speech
The Superbowl of Indexing IV
Phoenix, Arizona
December 5, 1999

JUST A YEAR HENCE, the world's first index mutual fund, proudly named “First Index Investment Trust” at its incorporation on December 31, 1975, will celebrate the 25^th anniversary of its birth. What is now Vanguard 500 Index Fund began with assets of just $11 million; on November 16, 1999, assets crossed the $100 billion milestone for the first time, a remarkable compound growth rate of nearly 50% per year. At the outset, the Fund was both excoriated and denigrated—“un-American” was the least of it—so, while I'm rarely one who thinks asset totals are very important as such, perhaps that date may serve as a useful landmark of the moment that heresy finally turned to dogma.

Index funds, of course, have spread far beyond Vanguard, and now constitute an important subset of the financial services industry. Assets of equity index funds total $300 billion, nearly 10% of the equity mutual fund total. There are 334 index mutual funds, 139 modeled on the Standard & Poor's 500 Index, and 195 in other categories, some of which make considerable sense, and some of which do not. Among those that do not are 69 funds which carry sales loads; the reason for their existence totally escapes me. Among those that do are international index funds and fixed-income index funds. Because of the extra costs of managed international funds—higher fees and expense ratios, higher portfolio turnover, and higher transaction costs—the advantages of low-cost index funds are proportionately even larger than in the domestic arena. Although so far only a single mutual fund firm has made a serious commitment to bond indexing, that arena too makes consummate good sense. Given the smaller margins of potential performance superiority that are possible among fixed-income funds, together with the truly outrageous fees charged by most bond funds, this area holds great potential.

But the new index industry includes far more than scores of index managers (some managing their funds at low costs—at least for a while—and some at costs that are confiscatory). It also includes large numbers of creative marketers engaged in proving that the number of indexes that can be used to create new “products” is virtually limitless; and numerous purveyors of methods to beat the indexes, or reshape them, by one obscure statistical technique or another—from computer-assisted mathematical models to portfolios composed solely of mortgage-backed bonds and S&P 500 futures—and hundreds of providers of legal, financial, technological, and communications services to all of those countless souls who are engaged in the creation, management, and marketing of index funds. Indexing has also spawned a new magazine—a Dow Jones publication called Indexes, of all things—and, to state the obvious, the annual Superbowl of Indexing itself, now meeting for the fourth year, with 97 speakers and 300 participants.

Our initial 500 Fund, a quiet first-born, was an only child for nearly a full decade. It was not until 1984 that the second index fund was born, and it proved to be a pale imitation of its older sibling. Its high costs, predictably, have utterly destroyed the fund's ability to do what its sponsors implicitly promised: To match the return of the S&P 500 Index. While this fund, which is managed by one of the largest and most respected banks in America, has persevered to this day, its shareholders have been ill-served. A $10,000 investment when the fund began, hit at the outset by a 4½% sales charge, and then by operating costs of some 1% each year, would have been worth just $116,000 as 1999 drew to a close. The same investment at the same time in the first index fund, the fund that the copier was emulating, was worth $143,000—a gap of a mere $27,000. (When asked by the press how the fund could justify its lofty costs, a spokesman answered: “It's our cash cow.” And of course it is.)

A Better Standard?

With the passage of time, I have become convinced that the most intelligent index fund strategy is modeled not on the Standard & Poor's 500 Stock Index, but on the Wilshire 5000 Equity Index. (It recently became the “Wilshire 5000 Total Stock Market Index,” from my perspective a change long overdue.) Later on, I'll discuss the merits of the total stock market index as a performance standard against which active managers of funds should be measured, but for now let me say that owning an index fund based on the total U.S. stock market virtually guarantees that investors will capture 98% to 99% of the stock market's annual return. By way of contrast, active investors as a group are virtually guaranteed to capture but 75% to 85%, simply because of the costs of investing. The search to identify, in advance, the few winning funds is like looking for a needle in the total market haystack. Why bother looking for the needle when you can own the haystack?

Owning an all-market index fund, simply put, means buying businesses—in essence, every publicly held business in America—and holding them for Warren Buffett's favorite holding period: Forever. Owning the average actively managed mutual fund, on the other hand, means trading pieces of paper, holding each sheet for this industry's favorite holding period: 406 days. (Fund turnover now averages 90% per year.) Believe me, there is a difference between these two strategies—in the complex and speculative task of mutual fund selection, in the certainty of future relative returns, in management fees and expense ratios, in portfolio turnover costs, and in sales loads too. And, lest we forget, in tax efficiency. (The grotesque tax inefficiency of most active mutual funds, while assuring irreparable harm to taxable fund investors, has provided no demonstrable advantage to tax-deferred fund investors in IRAs and retirement plans.)

Too Many Index Funds?

The selection of the large-cap S&P 500 as the appropriate benchmark at the outset of indexing has earned a lot of money for index fund shareholders. But I don't believe that it should remain the standard in the years ahead. (Though I wouldn't dream of encouraging satisfied investors in the S&P 500 Index funds to make the switch. After all, those 500 stocks represent 75% of the weight of the all-market index.) The consummate simplicity and comprehensiveness of the all-market index fund has me questioning, more than ever before, the wisdom of other, narrower index strategies.¹ Here, to be forthright, I am questioning no judgment other than my own, for we started the first small-cap index fund just over a decade ago, and the first growth index and value index funds in 1992—just as soon as the corresponding Standard & Poor's/BARRA growth and value indexes made their debut. The problem with small-cap index funds—and, lest I paint with too narrow a brush, small-cap funds in general—is that by maintaining what is called “style purity” (i.e., never having the temerity to jump out of the small-cap style box), such a fund must sell its winners and begin all over again. Such a move not only defeats the long-term nature of investment strategy, but has powerfully negative tax consequences. We all ought to be reconsidering this strategy to make it work more effectively.

The growth and value index strategies, too, ought to be reexamined. By definition, each accounts for 50% of the capitalization of the Standard & Poor's 500 Stock Index. But with the soaring market value of the technology stocks, the number of growth stocks comprising that 50% has dropped from 232 companies a decade ago to 128 today. The 1060% appreciation in Microsoft stock alone since 1994—increasing its weight from less than 1% of the Index to nearly 4%—has taken as many as 25 former growth stocks and transmogrified them into 25 new value stocks, making the value index “growthier” than ever before. Furthermore, as companies are elbowed out of the Growth Index into the Value Index, there are potential adverse tax consequences to taxable investors in a growth index fund. Again, as “style” index strategies are tested by time, we owe them some reconsideration. In any event, it is high time for better disclosure of the implications and consequences of “style purity,” in index funds and active funds alike. Today is no time for dogmatism.

The Newest Fad

But if these modest—and well-intentioned—departures from the majesty of broad stock market indexing should be carefully reconsidered, what can one say about the latest fad in indexing? Exchange-traded funds, already acronymically known as ETFs (which, I confess, I still confuse with the completely unrelated “electronic funds transfer”) are the hot item of the day. They've taken indexing far beyond what any of us present at the creation a quarter-century ago could have imagined. Fair enough. But what is wrong is that this new breed of index funds, by accident or design, seems to be frustrating the original purpose of the index strategy—efficient long-term investing in a diversified portfolio of businesses—giving us instead a vehicle for short-term speculation in the stock market.

This is not to knock, in any way, the Spiders (SPDRs, modeled on the Standard & Poor's 500 Index). I've acknowledged for years my respect for the creativity that went into their design, their low operating costs and fine tracking of the index, and their tax efficiency, in which the investor, in substance, is solely responsible for his own taxes. I happen to believe, however, that the services, conveniences, and low costs offered by the best Standard & Poor 500 index funds, along with the absence of brokerage commissions, make them the superior vehicle for long-term investors. Of course, the tax issue (for taxable investors) remains, and it may well be that some long-term investors may prefer to own conventional index funds through a newly created Spider-like series by today's conventional index funds. Such an eventuality should neither surprise nor concern investors. In any event, I expect no interruption in the robust growth of index mutual funds, which now account for nearly 40% of equity fund cash flow.

The real-time pricing of Spiders, on the other hand, makes them by far the superior vehicle for short-term speculators, doubtless well worth the brokerage commissions involved. The marketplace supports that judgment: While our index fund investors turn their holdings over at a rate of just 10%, the annual turnover of Spider shares is running at an annual rate of 1800% in 1999, about 22 times the 80% turnover rate of the average share of stock on the New York Stock Exchange. In the stock market today, the average share is held for about 456 days (itself a period that is hardly a monument to long-term investing), while the average Spider is held for just 20 days. Not even three weeks! And relative to the burgeoning ETF universe, Spider investors are models of decorum. Their counterparts who trade the even-gamier Nasdaq 100, are turning over their “QQQ's” at an annual rate of 7800%, an average holding period of just over four days! Long-term investors might consider extreme caution before investing in a program—however inexpensive, whatever its tracking success, and irrespective of its potential long-term tax efficiency—in which most investors are, well, four-day wonders.

Of course, ETFs have been carried far beyond Spiders and QQQ's. There are now 17 WEBS (with much higher expense ratios—in the plus 1% range) based on global indexes, soon to be joined by 51 new ETFs called iShares (cost not yet disclosed), presenting an even wider selection of domestic and foreign markets. But these new ETFs have been designed with a purpose diametrically opposed to the purpose of that pioneering index mutual fund of 1975. Our purpose was to create an investment that would serve long-term shareholders, to be bought and held, to be the hedgehog who knows one great thing, rather than the fox who knows so many things, bringing simplicity rather than complexity to the world of investing. The purpose of the ETFs seems to be to create a product that will sell, a product that, if all goes well, will serve speculators efficiently, but will serve its sponsors whether things go well or not. Indeed, an executive of a major ETF sponsor recently said, “Brands have always existed in consumer products. We see investors buying (our) brand just like they buy a brand of toothpaste.” That's one point of view. But it's not mine. For believe me: There is a critical difference between designing a product that sells, and creating an investment that serves.

Active Managers and Indexes

Last January, when I offered up “When Active Managers Win, Who Loses” as the title I would use for these remarks, I assumed that this would be a year in which the average mutual fund, after five consecutive years in the doghouse, would finally beat the S&P 500 Index. During the 1994–1998 period, the average fund lagged the 500 by an astonishing 75 percentage points (+119% vs. +194%). What is more, as a dyed-in-the-wool believer in reversion to the mean in the financial markets, I suspected that the time for the smaller stocks to rise again was at hand. I was well aware that while the 500 Index is invested 100% in large-cap stocks, about one-half of the industry's general purpose stock funds are large-cap funds (and with a slightly smaller average cap size at that) and one-half are mid- and small-cap funds. This distribution almost assures superiority for the average fund when non-S&P stocks outperform.

As the year draws to a close, my guess is looking pretty good. The non-S&P stocks are up 22.0%, while the S&P 500 is up 16.0%. And the average fund has produced a gain of +18.0%—two percentage points better. The 500 triumphed in the first quarter, but the average fund led the second quarter lap, as Indexes magazine headlined, “Man Bites Dog! Pros Beat Indexes.” And the average fund led again in the third quarter lap, which The New York Times (somewhat more temperately) headlined, “A Small Victory for Stock Picking.” If the year ends with the S&P Index still lagging, we can expect to see more dramatic headlines as year 2000 begins.

So, at least for the moment the tables have turned ever so slightly toward the active fund managers, and we indexers must be ready with a rational response. So let's begin with the basics: The fact is that indexers always win. That is, in any financial market—and any segment of any financial market—indexers owning all of the securities in that market at low cost must provide better returns than the other investors in the market in the aggregate, simply because the costs incurred by active investors—commissions, fees, taxes—are substantially higher. Costs matter. Costs always matter. And that is why active managers as a group can never win.

So the answer to the question: “When active managers win, who loses?” is: other active managers. Perhaps hyperactive managers, or inactive managers, or managers not included in the database. Nonetheless, simple data being simple data, there will be years when it looks as if active managers win. Mutual fund history, indeed, shows that the average fund beat the S&P 500 in 1991–1993, in 1977–1982 (right after our index fund began), and in 1965–1968 (the Go-Go era). But there is much more to the data than meets the eye—especially over the long run. Let me give you just a few examples:

• First, the S&P 500 is the wrong index to compare with the diverse and ever-changing mutual fund industry. Why would one compare a 100% large-cap index with the average fund in an industry in which large-cap funds as a percentage of all funds have shrunk from 75% 15 years ago to 44% today?
• Second, “selection bias” distorts fund industry returns. Many mutual funds have records that are more than suspect; invalid records that count in the industry data side by side with the valid records. The worst offenders are incubator funds, which, if they fly, fly high, and if they flop, flop right out of the database. But those that fly remain in the record book, which includes many mainstream funds of today that began with tiny assets, providing remarkable returns until size and reversion to the mean took over, and then fading to average and below. But their long-term records are accepted uncritically. Caveat Emptor!
• Third, “survivor bias” plays a major role in industry records. The meek, in short, do not inherit the mutual fund earth, and the dimension of fund failure is astonishing. For example, of 355 equity funds that existed 30 years ago, 186—more than one-half!—have vanished into thin air. Even in the mutual fund boom of the past 15 years, 70 of the original 454 funds have vanished. Analysts from academe, such as Princeton's Burton Malkiel and University of Southern California's Mark Carhart, have estimated survivor bias ranging from as little as 1.4 percentage points annually in 1982–1991 to as much as 4.2 points in 1976–1991. It is now clear that the evanescence of so many funds—surely the weakest performers—sharply overstate long-term industry-wide averages.
• Fourth, front-end sales charges are almost universally ignored in the publicized performance data. As a result, the returns of the average mutual fund investor inevitably lag the reported returns of the average mutual fund. The resulting diminution of returns, amortized over the typical holding period, may cut the reported returns by as much as a full percentage point.
• Fifth, amazingly, fund comparisons—again, almost universally—ignore taxes. Managed mutual funds are shockingly tax inefficient, while market index funds are highly tax efficient, largely subject only to taxes on dividend income. In the ebullient market of the past 15 years, the relative reduction in active fund returns as a result of taxes has been about 1.8 percentage points per year.

So what's to be done in assessing comparisons of industry returns with the stock market? There are lots of complex statistics we can develop (I've surely developed my share!), but there are also some simple and obvious steps we can take. Here are three:

• When comparing the results of the average mutual fund with the market, abandon the use of the S&P 500 Index and use the Wilshire 5000 Total Stock Market Index. By bringing small- and mid-cap stocks into the equation we get a reasonably fair long-term comparison. In fact, the correlation (R²) of the returns of the average value and growth equity fund with this index has been .998 over the past 15 years. During that period, the annual return of the average surviving fund has been 14.4%, compared to 16.9% for the 5000 Index. Adjusting that 2.5% spread to take into account the impact of both sales charges (.5%) and survivor bias (at least 1%), and throwing in another 2.7 percentage points for tax impact gives rise to an average annual fund return of 11.1% for a taxable investor; all-market index fund return (also after costs and taxes), 15.8%. Cumulative fund return: +384%. Cumulative index fund return: +804%.² No further comment is required.
• When using the S&P 500 Index as the comparator, limit the comparison to funds following the large-cap style (or, even better, Morningstar's 933 “Large-Cap Blend” funds, owning both large growth and large value equities). Yes, you'll still find short-term differences that may favor or disfavor active managers. For example, active funds presently hold positions in Microsoft, General Electric, and Coca-Cola that are well below their market weights. But in the long run, the comparison ought to be fair, though it ought to be adjusted for sales charges and taxes; survivor bias is least significant among large-cap funds.
• When using the Russell 2000 or 2500 Indexes as the comparator for small- and mid-cap funds, be absolutely certain to estimate the impact of survivor bias, for here it is highly significant. Morningstar recently placed survivor bias in small-cap funds, over the five years 1992–1996 alone, at 1.1% per year. Result: While small-cap blend funds—with an annual return of 10.3% over the past ten years compared to 10.9% for the Russell 2000—seemed competitive, if we assess a 1.5-point survivor bias, that 0.6-point loss zooms to 2.1 points, not far from what we'd expect based on fund costs. When we adjust for another one point for sales charges and perhaps two points of tax differential, that deficit burgeons.

And when the day comes that active managers seem to win, remember that it is only because the data—proliferating far beyond what even the most farsighted of the index pioneers might have dreamed 25 years ago—fails to capture the results of all active managers is burdened by statistical errors such as disregarding survivor bias, sales charges, or taxes, or ignores data anomalies, of which the most notable is calculating fund returns based on number of funds rather than assets of funds.

A Retrospective

So, fellow indexers, be of stout heart: Active managers as a group never win. As we begin to commemorate the 25^th anniversary year of the birth of the index mutual fund, amid the surfeit of data available today, let's never lose sight of that immutable fact—time-honored to a fault. As Peter Bernstein tells us in his remarkable book Capital Ideas, it was way back in 1908 that the French economist Louis Bachelier spelled out the unequivocal brute fact in his thesis: “The mathematical expectation of the speculator is zero.” However, Bachelier ignored the costs of investing, the proceeds raked off by the numerous croupiers at work in the marketplace. After these costs, as I expressed it most recently in a New York Times Op-Ed piece last summer: “In the stock market casino, it is the croupiers who win.” Paraphrasing the title of a book of the 1940s, I asked “Where are the customers' private jets?” Where, indeed.

Nobel Laureate Paul Samuelson has said that his view of indexing theory, “oscillated from regarding it as trivially obvious (and almost trivially vacuous) and regarding it as remarkably sweeping.” Of course, it is both, giving further testimony to the ancient wisdom of William of Occam. In 1330, he formulated a rule that came to be known as Occam's Razor, essentially that, “when confronted with multiple solutions to a complex problem, choose the simplest one.” Yes, the secret of success in investing is the obvious one: The haystack trumps the needle, almost every time. No matter what the data that compare active managers with market indexes appear to show, we'd best never ignore that immutable fact. As Oliver Wendell Holmes reminded us, “we need education in the obvious more than investigation of the obscure.” I hope my message tonight has reinforced that wisdom.