ETF Special Report

I have championed market-capitalization-weighted index fund investing even before such funds were made available to the general public. In the first edition of A Random Walk Down Wall Street in 1973, I indicated that whenever a strategy of indexing was recommended, the typical answer was, “You can't buy the index.” My response was “It's about time you could.” In 1976, Vanguard introduced the first index funds, and they were based on the S&P 500.

My argument for indexing was based on my belief that our equity markets are remarkably efficient. When information arises about the stock market or individual stocks, that information gets reflected without delay. Stock prices, therefore, incorporate whatever public information is known, and active portfolio management — switching from stock to stock in an attempt to provide superior performance — is unlikely to be effective, especially when you add in trading costs and taxes. Of course, we know that the market occasionally makes egregious errors; trouble is, it is virtually impossible to recognize those errors in advance. While some observers accurately recognized the high-tech Internet bubble during 1999 and early 2000, those same analysts worried about bubble valuations in the mid-1990s — and paid a price for being early. Besides the data strongly suggest that active mangers have not been successful in avoiding the “overvalued” sectors of the market. (See charts.)

Zero-Sum Game and Then Some

Markets need not be efficient, however, to justify a passive indexing strategy based on market capitalization. Even if markets were inefficient, all of the stocks in the market must be held by someone. Thus, (financial) investors as a group must own the capitalization-weighted (float-weighted) index of all stocks. Investing must be a zero-sum game as is shown in Exhibit 1. If some investors happen to be holding only the best-performing stocks in their portfolios, then other investors must be holding the poorer-performing ones. As a group all investors cannot live in Garrison Keillor's Lake Wobegon and be above average.

However, active management must be a negative-sum game. Active managers incur two types of costs: management fees and turnover costs. Mutual fund management fees on average are about 100 basis points per year. In addition, turnover costs (commissions, bid-asked spreads, market impact costs) easily add another 50 basis points to the annual cost of active management. Passively managed index funds and ETFs are available at 10 basis points or less and have minimal turnover. Thus, the average active manager must underperform a low-cost index fund by the difference in fees. Exhibit 2 illustrates the situation with an assumed extra 140 basis points of costs. Indexing must be a winning investment strategy.

The data strongly support the illustrative calculations on the previous page. Over the past 10 and 20 years, a cap-weighted S&P 500 index fund has outperformed more than 75 percent of actively managed, large-capitalization equity funds. As Exhibit 3 shows, the amount of that outperformance after fees has been over 140 basis points annually, entirely consistent with the cost differences hypothesized above. The index fund has also been more tax efficient. While a few active funds have beaten the S&P 500 over the long run, there is no way to predict the winners in advance. There is very little persistence in mutual fund performance. The best-performing funds in one period are usually not the best performing in the next. The capitalization (float)-weighted S&P 500, however, has been a consistent above-average performer after expenses.

Exhibit 4 examines the long run returns for every equity mutual fund that existed in 1970. Note that of the 355 equity funds in existence in 1970, only 117 of those remained in December 2006. All we can do is plot the returns of the survivors, which are the better performing funds. Funds with poorer records tend to get merged into other equity funds in the same complex that enjoy better records. But even though these data are tainted by “survivorship bias,” the actual distribution of returns looks very similar to the theoretical distribution shown in Exhibit 2. Most of the distribution is on the negative side. Very few of the surviving funds beat the market by any substantial amount.

A Weight Problem? Nah

Recently, cap-weighting has been criticized for incorporating an improper weighting mechanism. According to this argument, stocks should be weighted by their economic footprint (the amount of their profits, book values, etc.), rather than by their total market capitalization (or float). These critics argue that a fundamentally weighted index will tend to outperform cap-weighted indexes such as the S&P 500. To be sure, the Fundamental Index has done very well during the early 2000s as markets have adjusted from the high-tech stock bubble that peaked in March 2000. But these are active bets that involve tilts in the weighting toward “value” and smaller-cap stocks. I am far from certain that such biases will prove to be consistent excess performers now that P/E multiples are quite compressed. In any event they are not “indexes” in the sense that all investors can own them. Capitalization-weighted index funds are fully scalable and do not involve tilting the composition of the fund toward value or small-cap stocks in an effort to chase returns. To those who try to “improve” an index, let me say in parting that, sure, the market makes mistakes — it goes crazy some times. But, overall, it does a pretty good job. Such a good job that it's a fool's errand to try to beat the major market indexes — the occasional market madness notwithstanding. That's why cap-weighted indexes will forever be the preeminent choice for many investors, retail and institutional.

Exhibit 3