A prolific researcher, Rob Arnott edits the Financial Analysts Journal and his work frequently appears in it. Recently, he created some turmoil with the publication of his paper (co-authored with Jason Hsu and Phillip Moore) on fundamental indexation, essentially a treatise on a radical way to improve the indexing industry.

The paper posits that there are better ways to construct a benchmark than market capitalization, the way in which most indices are constructed. The drawback to using market-cap weighting in creating funds is that companies that are often dearly valued are more heavily weighted in an index because their high prices result in larger market caps. On the flip side, firms that rarely sell at rich premiums are underweighted in an index, meaning index investors miss out on potential outperformance.

This also means major indices are usually tilted toward large-cap stocks. Recently, the 50 largest stocks accounted for more than 50% of the Standard & Poor’s 500 composite index. A single colossus such as General Electric Co. carries more weight than legions of smaller stocks in that index.

Looking to solve this problem, some index providers include only companies that meet predetermined balance-sheet ratios. For example, the Barra large value index includes the bottom half of the S&P 500, as sorted by price/book ratio.

But this kind of surgery on the equity universe can yield under-diversified portfolios, as William Bernstein warns in his Ef-ficient Frontier newsletter. And within the universe of remaining stocks, you are still weighting by market cap — overweighting overvalued stocks and shouldering significant transactional costs that occur when firms jump in and out of the universe.

Arnott and his colleagues suggest an innovative fix. Rather than splitting the cap-weighted universe, they contour it by including the 1,000 largest companies, weighting them according to their economic footprint as reflected in measures such as earnings, sales, book value, dividends and number of employees.

Arnott et al first created an alternative model fund for each measurement. They then weighted companies by that measurement, re-weighting each alternative fund once a year.

Using data from 1962 to 2004, the researchers found that a portfolio of companies weighted by these measures outperformed a portfolio weighted only by their market cap.

Each of the model funds outperformed the S&P 500 by about 2% a year, they report. And each alternative model fund outperformed in both bull and bear periods, and in periods of moving interest rates.

These findings imply investors have been systematically undervaluing large companies relative to small ones. To offset this, Arnott’s company, Research Affiliates, has established some exchange-traded funds to trade on this fundamental approach.

But Bernstein is skeptical: “While it is clear that Bob Arnott, with hindsight, has discovered a theoretically profitable anomaly, what basis is there for assuming it will continue in the future, now it has been publicized?”

The theory may work, he adds, providing you believe whoever has made the poor valuation choices that produced the study’s superior returns will continue making them, and other investors will not fritter them away bidding up the price of larger firms’ stocks. Plus, investors need to consider not only the methodology used, but who is executing it.

John C. Bogle, founder of The Van-guard Group and the godfather of index strategies, agrees. Despite Arnott’s findings, he maintains there are several points in favour of funds using cap-weighted indices.

First, these new fundamental-weighted index funds have higher costs than passively managed index funds because of higher management fees, turnover and operating expenses. And, because of the forced turnover in the entire fund each time a company changes its dividend, fundamental-weighted index funds will be less tax-efficient than existing funds.

All known methods of fundamental indexation overweight value and small-cap stocks, Bogle says — the sectors that outperformed in the measurement periods cited in Arnott’s research, and the real reason for the resulting outperformance. If large-cap or growth sectors begin to dominate the markets, Bogle predicts fundamental index-ation approaches will underperform. IE