William Sharpe

Exclusive Interview: Prof. William Sharpe

William Sharpe

Prof. Sharpe "The key issue is that past performance is a thin reed for how to predict future performance. Expense ratios and turnover are generally better predictors." - William Sharpe

PALO ALTO, CALIF. - Few individuals have advanced theoretical underpinnings of index investing more than Professor William Sharpe, STANCO 25 Professor of Finance at Stanford University and Nobel Laureate in economics. Few individuals are doing more to make the case for prudent, science-based investing understandable to the average investor.

At his office next to Stanford Business School, Sharpe fielded questions on the state of indexing during academic "office hours". As graduate students dropped in to try to squeeze into his classes, he outlined his views on indexing today for IndexFunds

"I am an educator," he said. "The thing I would love most is to educate everyone." He has clearly had an effect on Jack Bogle of Vanguard Group, whose new book Common Sense on Mutual Funds prominently mentions Sharpe in its acknowledgements section.

"My conclusion after worrying a lot about how to help the American investor is that the majority don't want to be educated in any depth," he said, which is rather understandable. People have careers, hobbies and family life, and they can devote only so much time to understanding investments.

So he keeps many of his points simple. "The average dollar in an active fund will net of costs underperform the average dollar in an index fund," he notes. That of course is because active funds have higher fees.


The dominant factors that tend to predict relatively strong future performance in order of importance:

  1. Low expenses
  2. Low turnover
  3. Good past performance


"The growth/return side is very hard to predict, but costs are easy to predict," he said.

"It is important to understand that if you think that index funds will beat Treasury bills by 6% and you pay 1% in fees, then you have given away about 17% of the advantage of taking equity market risk," he said.

"Some people who are really superior could add enough value to offset the costs," he said. The trick is separating the truly superior from the simply lucky. "It's really hard to tell what the benefits will be."

"Don't expect much persistence," he said. "The one area where past performance is helpful is if the fund is really rotten -- usually due to high expenses." Abysmal performance tends to signal poor future performance.

"Increasingly you have to help people do the arithmetic," he said, such as "how much will it cost me if I pay 1%?"

This is one reason he founded Financial Engines, whose Web site at www.financialengines.com lets investors see how basic allocation decisions affect outcomes based on currently available data. In real time the site runs Monte Carlo simulations after putting in risk thresholds, expected retirement age, and other factors. The current service is directed at corporate 401-K plans where employees have a limited number of fund options, but additional services are planned.


Some may find it surprising that he is a partial investor in active funds. "I am not religious about putting all your money in index funds," he said. Ultimately the benefits of index funds are not to be assumed. They must be proven by rigorous study.

Curiously, the theory of efficient markets insists that the more competition there is for top returns the more tenaciously the theory will hold. If everyone fled to index funds no one would be calling the bluff of overheated sectors.

"We have to have active investors or the markets will go crazy," he said.

In an article for the July/August 1998 Financial Analysts Journal, Sharpe examined in depth the various metrics of Morningstar, Inc. for analyzing mutual funds. He found some of the more popular ratings of dubious value.

Sharpe's landmark study Morningstar's Risk-adjusted Ratings calls into question the practical benefit of some, but certainly not all, of the more popular metrics of the world's best known mutual fund research group.

"If you have to use one, use the category ratings because the comparisons are somewhat more homogeneous," Sharpe said. "You have to compare a growth stock manager's performance with a growth index and a value manager with a value index." Morningstar highlights its category ratings as well, but marketers for mutual funds generally ignore them in favor of the imagery of stars.

In principle information is always beneficial to the markets, but when large numbers of investors are basing decisions on their retirement income exclusively on ratings such as Morningstar "stars", there is something wrong with investor education. Although Morningstar has clearly stated that stars are not clear indicators of future performance, investors continue to latch on to the appealing rankings.

Star ratings do not measure how good a fund competes in its asset class, but instead give good marks for any fund that happens to be in an asset class in ascendance. The laws of reversion to the mean cause the star ratings to have little predictive value. The asset class that goes up precipitously often comes down to earth equally fast.

If there is little scientific basis to using Morningstar "stars" as ratings guides, then why are they so popular? "Morningstar ratings came out, took hold, and got into everybody's consciousness," he said. "Everybody looks at the star rating."

People need to understand how hard it is to predict superior performance, he emphasizes. "Past performance gives you a little edge," he said. "There is something there, but there is not a lot."

Generally, well-designed studies show little long-term performance consistency. For example, if growth funds do well over 6 years, a crude study may show that early leaders continue to outperform.

"Many fund families have so many funds they can always find funds with four or five stars after the fact," he said. Fund groups often deliberately seed dozens of small funds in anticipation that many will fail but some will show above average performance by luck if nothing else. "Pruning" by the industry of 3-4% of funds every year has a big effect on performance figures posted prominently in the financial press.

Sharpe takes issue more generally with the traditional method of establishing a target percentage of funding for each asset type and classifying funds according to that asset type.

"There are a number of arguments why this is not an efficient way to approach the problem."

"Many times with an active fund the standard descriptions don't tell you what you are getting in terms of asset exposures" he said. "You ought to know what you are getting."

"The other big problem is that if you are not using index funds, most funds are not a pure play," he said. Ensuring that you get the asset allocation you want can involve considerable extra work.

"What if all growth funds are bad?," he added. "If you are going to look at someone's performance, you ought to do it relative to a passive portfolio with a similar style."                    

Will McClatchy