Famed pricing models come under fire

The Capital Asset Pricing Model is broken. Now, a new group of academics is trying to fix it, looking at the role of professional money managers in setting stock prices.

The biggest problem the researchers have identified: Money managers' incentives don't always align with those of their clients, leading the managers to pile into riskier stocks, driving their prices up and, perversely, leading to lower expected returns.

Managers have a huge incentive to retain the business of sophisticated institutional clients, which generates an annuity-like payment of fees. Several years of sustained underperformance can lead to that business walking out the door, observed Brad Cornell and Richard Roll, two finance professors at the University of California at Los Angeles' Anderson Graduate School of Management.

Thus, managers try to pick stocks that outperform a given benchmark and limit volatility of tracking error to that benchmark, instead of picking the optimal portfolio, according to a paper by Messrs. Cornell and Roll that was published in the January/February 2005 issue of the Financial Analysts Journal.

Jason Karceski, an academic at the University of Florida's Warrington College of Business, Gainesville, takes a different tack. He argues that active managers "tilt their portfolios toward high-beta stocks. In equilibrium, this bump in demand for high-beta stocks pushes their prices higher and expected returns lower," he wrote in a December 2002 paper published in the Journal of Financial and Quantitative Analysis.

Together, the papers establish a framework for modifying CAPM to make it a closer approximation to reality. However, much more research needs to be conducted to turn this new Capital Asset Pricing Model into a workable tool for investors.

The Cornell/Roll approach is "interesting, but it doesn't add that much clarity to the Capital Asset Pricing Model in my mind," said William Ricks, chief executive officer and chief investment officer for AXA Rosenberg Investment Management (North America), Orinda, Calif.

For more than 30 years, academics have been trying to figure out why some stocks appear to be vastly overpriced and others grossly undervalued under CAPM, the breakthrough theory that attempts to explain how securities prices are set. That mid-1960s theory says a market portfolio consisting of all assets is the most efficient portfolio, leading to use of beta as a risk measure and adoption of indexed funds. The theory helped win a Nobel Prize in Economics for William F. Sharpe Jr., now chairman of Financial Engines Inc., Palo Alto, Calif.

By the early 1970s, academics were questioning why CAPM was not working in practice, leading some to question whether beta was dead. Efforts to revamp CAPM have used three major approaches, according to the Cornell and Roll paper.

Some academics noted that value stocks and small-cap stocks enjoyed returns greater than those predicted by Capital Asset Pricing Model, leading to the three-factor model put forward by Eugene F. Fama of the University of Chicago's Graduate School of Business and Kenneth French, now at Dartmouth College's Tuck School of Business.

A second approach developed a multiple-period CAPM, instead of the fixed time period used in traditional CAPM. This intertemporal CAPM recognizes that market conditions change over time.

A third approach has come from behavioral finance experts, who argue the market prices can be irrational because of erroneous mental shortcuts made by investors. However, behavioral finance experts lack an overarching theory to explain how securities prices are set.

That's why some academics are seeking a new way forward. "People seem to be getting disenchanted with the beta model, with the three-factor model, and they sense that something is just not working," said Meir Statman, chairman of the finance department at Santa Clara University, Santa Clara, Calif.

"Maybe we should work with more open thinking," he added. Mr. Statman suggested developing a model that includes traditional CAPM-type measures of risk, behavioral observations and authority delegated to professional managers.

The new wave of thinkers is looking at the role of professional money managers in setting stock prices.

What Messrs. Cornell and Roll tried to do was marry CAPM with principal/agent theory, whose main observation is that the incentives of the two parties — the ultimate owners and those hired to manage entities or assets on their behalf — are often in conflict. For example, chief executives have frequently been criticized for placing their own well-being ahead of shareholders'.

The authors expressed surprise that while many academic papers have been devoted to principal/agent theory, the role of delegated investing has largely been ignored.

A possible explanation, they write, is that professional money management skyrocketed in the second half of the 20th century. Institutional ownership of common stocks shot to more than 50% in 1994 from less than 10% in 1950, according to previous research. What's more, institutions accounted for 90% of the volume of New York Stock Exchange trading in 1998, up from negligible levels in 1950.

In the case of money management, the problem is that managers focus most on generating their annual fees, not on building the best possible portfolio. Thus, "managers strive to maintain a precautionary strategy that makes underperforming their benchmark for several years in a row unlikely," the authors wrote.

Managers not only have a strong incentive to outperform the benchmark, but also to keep the difference between their performance and the benchmark at a low level. "In fact, many clients believe that the ideal manager outperforms the benchmark every period by the same margin (i.e., with zero tracking-error volatility)," the paper says.

Managers will typically choose a portfolio that is riskier than the benchmark in the hope of beating the bogey and thus generating more annual fees.

To incorporate the effect of the benchmark and managers' efforts to minimize tracking error, Messrs. Cornell and Roll add a third term to the traditional CAPM formula.

The bottom line: Securities prices are influenced heavily by managers' need to outperform the benchmark, not by the need to generate the optimal portfolio for their clients.

Mr. Karceski also looks at the role of professional money management on CAPM, but from a different angle. He argues that mutual fund stock portfolios have a higher beta than the market itself. Using Morningstar Inc. data, Mr. Karceski found equity stock funds have a weighted-average beta of 1.05 when their cash is not counted. (When cash is included, the beta falls to 0.95.) Logically, the collective beta should be 1, the same as for the market.

Why do stock funds have higher-than-market betas? Because investors chase returns, pouring money into hot funds. This ability to attract assets is far more important during bull markets. when stock-market takes over, he wrote.

In 1995, when the market returned 28.7%, net new cash flows into stock mutual funds grew 11.2% that year and the next. But in 1990, when the market fell 12.9%, net new cash flows were -15% during that year and the following year.

So did the need to compete for investor assets lead managers to create more aggressive portfolios than they otherwise might choose? Mr. Karceski asked.

"Toward the end of the 1990s, stellar U.S. market returns and mutual fund flows intensified the pressure to outperform category peers, and many fund managers seemed to respond accordingly," he wrote.

At the same time, some managers cut their personal stock holdings because of fears of an overinflated market. Also, "style creep"— such as where managers added riskier technology stocks to their portfolio — became a big issue as managers labored to boost performance.

"So there is anecdotal evidence that at least some fund managers' portfolio compositions are influenced by the need to compete for new cash flows," Mr. Karceski wrote.

The irony is that as stock prices went up, equilibrium expected returns went down.

The sharp growth in stock mutual fund assets — from $70 billion at the end of 1983 to more than $4 trillion at the end of 1999 — might account for the death of beta, he added.

"My model suggests that the recent explosion in equity mutual funds may be related to beta's death in the early 1980s," he concluded.