A better way to index?

(Fortune Magazine) -- The virtues of index investing are well known: low turnover, low cost, and you are sure to keep pace with the market. But there's one potential problem: The S&P 500 and other popular indexes are capitalization-weighted: Companies are represented based on their market value (share price times number of shares outstanding), so as stocks run up, so does their presence in the index. As a result, during times of market exuberance highflying stocks account for a disproportionate share of the index. The prime example came in the late 1990s. People who put money into an S&P 500 index fund in those years were getting a huge helping of wildly overvalued tech shares. They paid the price when the bubble burst.

To address that problem, Rob Arnott, chairman of Research Affiliates, an asset-management firm, set out to develop indexes that aren' t linked to stock valuation. After testing various formulas using decades of U.S. and foreign stock market data, he found that indexes that weight companies based on a combination of sales, cash flow, book value, and dividends would almost always outperform traditional ones. His RAFI U.S. 1000, for example, would have earned an annualized average of 12.35% from 1962 through 2005, vs. 10.25% for the S&P 500; that means a hypothetical $1,000 invested in the RAFI index would have grown to $186,000, compared with $73,000 if invested in the S&P 500.

Armed with his results, Arnott has been peddling his "fundamental" indexes to institutional investors since 2005, and PowerShares Capital Management has introduced 11 exchange-traded funds based on them. The most popular vehicle available to individual investors is the PowerShares FTSE RAFI U.S. 1000 (Charts).

The secret seems to be simply that these fundamental indexes reduce the presence of overvalued stocks - the kind most susceptible to steep declines. Overall, the average P/E of stocks in the Russell 1000 is 17, compared with 15 for Arnott's RAFI U.S. 1000. That sounds like a good idea, but adherents of traditional indexing claim that cap-weighting offers the most accurate representation of the stock market. They say fundamental indexing is a gimmick that creates a value-stock bias and entails more turnover and higher expenses.

Let's look at those objections. The PowerShares RAFI 1000 fund, for example, has annual expenses of 0.6%, vs. 0.2% for the cap-weighted iShares Russell 1000 Value Index (Charts) , and annual turnover of 10%, vs. 7% for the iShares fund. Those differences are small, and to Arnott they're justified. "If price goes up and the fundamental scale of a company hasn't, we' re going to be trimming it," he says. "And to my way of thinking that's good turnover. Why should we own twice as much [of a stock] just because the price has doubled?"

And without wading into the theoretical dispute over indexing, we don' t see anything wrong with a slight value tilt. While it will depress returns in a raging bull market, it offers some protection in a meltdown. In the crash, for example, the Russell 1000 fell 50%, while the RAFI U.S. 1000 would have dropped only 25%. For conservative investors, a fundamental index fund like PRF would seem to be well worth considering.