Grading the Academic Research on Stock Returns
How accurate have been the professors' predictions?
Forty years ago, academic finance had three suggestions for equity investors:
1) The markets were efficient.
In general, stocks were rationally priced. Finding winners wouldn’t be easy, even for well-trained, intelligent buyers. What’s more, investing was a zero-sum game, because for every winner of a stock trade, there was a loser of equal size. Consequently, most professional investors would struggle to beat the market averages.
2) Betas would determine relative stock returns.
Collectively, the relative returns of equities could be predicted. Stocks would behave according to their sensitivity to market movements. That is, the more aggressively that a security responded to changes in the overall stock market, the stronger its expected long-term returns. (This assumes that equities outgain the risk-free rate of return. If not, then the opposite holds.)
3) Investment factors would determine relative stock returns.
This field was in its preliminary stage, with Rolf Banz’s breakthrough article on the performance of small stocks published in 1981, and the studies of value investing just beginning. Nevertheless, the idea that small, cheap (that is, “value”) companies would score the best stock-market returns was already being publicly tested, with the launch of DFA U.S. Micro Cap Fund DFSCX in December 1981.
The finance professors, of course, were entirely correct about the stock market’s efficiency, along with the implications that precept held for professional investment managers. For that insight alone they earn a passing grade. Whether academia excelled, though, depends upon the accuracy of its conjectures about relative stock returns--that is, about how different groups of equities would fare.
As we have seen, they took two separate bites of that apple. The beta hypothesis, advanced by William Sharpe in 1964, is distinct from the analysis of investment factors, which was most famously summarized by Eugene Fama and Ken French in 1993. (That work has since been greatly extended, through the addition of many other factors.) Not only do the two approaches yield competing predictions, but their intellectual roots differ. Sharpe’s concept was theoretical, while the studies of investment factors were predominantly empirical. First came the data, then the theories to explain them.
Let’s see how each forecast performed.
To evaluate the results, I used the investment-style indexes from Russell and Wilshire. (Several companies offer style indexes, but within Morningstar’s database, only Russell’s and Wilshire’s possess sufficiently long histories.) Within each index series, I calculated the returns for the four corners of the Morningstar Style Box.
The period extended for 42 years, from January 1980 through December 2021. The average annualized gain for each Style Box corner, obtained by averaging the results of the two series of indexes, were:
1) Small Value 12.75%
2) Large Growth 12.34%
3) Large Value 11.77%
4) Small Growth 10.26%
The margin between the three top-performing styles was modest, although as Jack Bogle never tired of reminding us, if compounded, a few basis points over a given year becomes a whopping great margin over the long term. (Sure enough, by these figures an investor who put $10,000 into small-value stocks in January 1980 would have held $220,000 more than a large-growth shareholder entering this year, and $474,000 more than somebody who favored large value stocks.) Small-growth stocks, however, performed significantly worse.
Professor Sharpe’s formulation made no prediction about which stocks would post the highest betas. Whereas investment-factor research explicitly demonstrated that small and/or low-cost securities had excelled, beta told no stock-market tales. However, given their volatility, small-growth stocks always figured to record the steepest betas, followed either by small-value or large-growth firms, with sluggish large-value companies at the bottom.
That is indeed what occurred from 1980 through 2021. Small-growth companies registered by far the highest betas, followed by large growth, then small value, and finally (as expected) by large value. How that compares with stocks’ actual results, as computed by the Russell and Wilshire style indexes, appears below.
Beta flunked its exam. The relationship between its predictions and the market’s actual returns was random. Yes, beta correctly foresaw that large-growth stocks would notch the second-highest returns among the four groups, but turn up four cards, and the suit will likely match your guess on one of those four occasions. There’s no glory in that outcome. Critically, and more to the point, beta utterly botched small-growth companies. According to beta, such securities should have easily enjoyed the strongest gains. Instead, they wore style investing’s dunce cap.
Presumably, academia’s second attempt at predicting stock returns would outdo its first. After all, unlike beta, investment factors were based on observation. Their forecasting accuracy therefore did not depend on insight. For investment factors to succeed, all that was required was for the future to resemble the past. The affinity need not be sharp; a casual relationship would suffice.
For the most part, that is what happened.
History repeated for small-value stocks. Their victory margin was narrow, with major perils during the journey--many small-value investors struggled to maintain the faith when the category lagged during the late 1990s, and then again through the most recent bull market--but a win is a win. The investment-factor approach was also correct in suggesting that large-value stocks would outgain small-growth firms, as (at least according to the Fama/French research) the value-stock premium outweighed the small-company premium.
Investment-factor research, however, underestimated large-growth stocks. Possessing two strikes against them, such securities were expected to lag the pack. Instead, they almost led the way. With large-growth companies, the theoretical approach of beta offered more insight than did the empirical work of the early investment-factor researchers.
My grade for the finance professors: B+. (They gave me plenty of grades when I attended business school, so it’s a pleasure to turn the tables.) They were thoroughly right about the biggest item, the stock market’s efficiency. While beta didn’t correctly forecast stock prices, at least it wasn’t directly wrong, either. And the investment-factor research that followed fixed most of beta’s oversights.
Whatever their grade, the academics certainly could have done worse. In 1979, BusinessWeek fussed about “the death of equities,” the U.S. stock market could be bought for a mere 7 times its companies’ earnings, and state public pensions invested an average of less than 25% of their assets in stocks. By those standards, the professors earned an outright “A.”
John Rekenthaler (firstname.lastname@example.org) has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.