Jeremy Siegel, Ph.D. The Future for Investors

In my last column I praised indexation -- the strategy of buying a portfolio of stocks whose performance tracks that of an "index," such as the Standard & Poor's 500. Investments linked to indexes have historically outperformed most actively managed portfolios, in part because active managers have not been able to add enough return to negate the impact of the generally higher fees that they charge.

The vast majority of stock indexes today are "capitalization-weighted," which means that the weights that are assigned to each stock in the index are proportional to the total market value of the company's stock. These indexes are good indicators of the average performance of the market since all investors must together hold the market value of each stock.

I believe portfolios linked to capitalization-weighted indexes have been a great way to invest. But increasing evidence suggests that capitalization-weighted indexes may not be the best way to index an investor's portfolio. To understand why, let's go back to the reason why capitalization-weighted indexation became so popular.

The "Efficient" Versus the "Noisy Market Hypothesis"

The "Efficient Market Hypothesis" theory of financial markets dominated the economics profession in the 1970s when index funds were first created. This hypothesis claims that the price of a stock represents the best estimate of the underlying value of the firm that issued the stock. In other words, if the price of a stock changed, it must be because the underlying factors that determined its valuation, such as dividends, sales, etc., had also changed.

But increasing evidence suggests that this hypothesis may not be the best explanation for changes in stock prices. The prices of securities are impacted by far more factors than those only related to valuation. For example, speculators, insiders, momentum traders, and those who must buy and sell because of tax or fiduciary reasons also impact the price of stocks even though their transactions are unrelated to the fundamental value of the underlying company.

I believe the best way to characterize financial markets is to say that the true value of securities is often obscured by "noise," which are transactions that influence the price, but are unrelated to fundamental valuation. I call this way of characterizing how financial market prices are determined the "Noisy Market Hypothesis."

A growing body of research supports this hypothesis. "Value-based" portfolios and fundamentally-weighted indexes generally have produced higher long-term returns than comparable capitalization-weighted indexes over various historical periods tested (all assume the reinvestment of dividends).

Similarly, research that I and others have done supports the conclusion that weighting stocks by some fundamental metric of value, such as dividends, sales, or earnings, instead of by market value, has historically resulted in generally lower portfolio volatility than weighting by market-capitalization. Of course, past performance is no guarantee of future results and there are limits on the inferences we can draw from research results and back-tested index data. However, I believe the data on fundamentally-weighted indexes are compelling.

Dividend-weighted Indexes

Readers of this column know that I have always been fond of dividend-paying stocks. My research, published in my book The Future for Investors, has shown that on an historical basis, the bulk of the real return from stocks has come from dividends and that high dividend-yielding stocks have historically given higher returns to investors than low dividend-yielding stocks.

We can't be certain that this long-term trend will continue, but based on the historical results I believe it is natural to choose dividends as the metric by which to weight individual stocks in an index. (In the interest of full disclosure, I am a Director of, and a Senior Strategy Adviser to, WisdomTree Investments, Inc., a company that develops fundamentally-weighted dividend indexes and products. Some of the research on dividend-weighted indexes that I discuss is based upon my work for this company.)

This is how an index weighted by cash dividends might be constructed. If company A pays $100 million in dividends annually and company B pays $200 million, then company B will have twice the weight in the portfolio as company A, even though company A may have a higher total market value.

My research into dividend-weighted indexes leads me to believe they may be a viable alternative to traditional cap-weighted indexes. The research involves fundamentally weighting an index by the aggregate cash dividends that companies pay. A backtest of the hypothetical, historical performance from 1964 to 2005 of a dividend-weighted index consisting of U.S. companies that paid regular cash dividends showed that the annualized total return of this index exceeded the annualized total return of the S&P 500 Index for the same period by 138 basis points -- or 1.38 percentage points -- per year and did so with lower volatility.

Moreover, my analysis of the hypothetical backtested data for this period supports the proposition that dividend-weighted indexes generally outperformed comparable cap -weighted indexes during bear markets. (The back tests assumed that all dividends were reinvested, but did not assume any transaction costs. The period from 1964-2005 was selected because it was the longest time period that we could measure using the securities pricing data available to us.)

A number of fundamentally-weighted indexes have launched in recent years. For example, the Dow Jones Select Dividend Index (DVY), launched in November of 2003, selects companies based on high dividend yields and weights index components based on dividend per share. The FTSE RAFI US 1000 Index (PRF) selects and weights components based on a combination of fundamental factors. The Morningstar Dividend Leaders Index (FDL) (selects companies with high dividend yields and weights based on a free float adjusted measure of "available dividends."

Each such index uses one or more "fundamental factors" to weight its constituents. I believe we may see additional fundamentally-weighted indexes in the future as the criticisms of the Efficient Market Hypothesis that I have raised become more widely accepted.

Retort by Capitalization-weighted Supporters

Supporters of the traditional capitalization-weighted indexes have criticized the fundamentally-weighted approach. Most notably, John Bogle and Burton Malkiel recently claimed that the five-year period from 2000 to 2005 is responsible for a large part of the difference between the backtested performance of dividend-weighted indexes and comparable cap-weighted indexes. But the performance of these fundamentally-weighted indexes over the last five years simply helped make up for the underperformance of such indexes during the tech boom of the previous five years, when dividend-weighted indexes significantly underperformed capitalization-weighted indexes.

To take out the last five years distorts the data. Using the same logic would allow you to say that tech stocks have a good track record over the past 40 years if we ignore the data since 2000 when tech stocks crashed. By underweighting the speculative sectors of the market compared to their market-cap weighted peers, dividend-weighted indexes did not experience the roller coaster ride that capitalization-weighted indexes suffered during this time period.

Supporters of market-cap weighted indexes have also questioned whether fundamentally-weighted indexes would, like their cap-weighted brethren, exhibit relatively low turnover when compared to more actively managed portfolios. While this may be a criticism of some fundamentally-weighted indexes, I do not believe this criticism is valid with respect to the category as a whole. There is no theoretical reason why properly constructed fundamentally-weighted indexes could not have nearly the same relatively low turnover rates, particularly if such indexes are reconstituted only once a year like many cap-weighted indexes.

Summary

Capitalization-weighted indexes have done a good job for investors. The introduction of cap-weighted indexes in the 1970s helped revolutionize the way we think about investing. However, a growing body of evidence suggests that fundamentally-weighted indexes may offer investors an attractive alternative to traditional cap-weighted indexes. I believe that fundamentally-weighted indexing is the next wave of indexing, and that these indexes have the potential to change the way we think about constructing investors' portfolios.