July 1, 2000

The Efficient Pioneer

Eugene Fama's contributions to modern finance over

Illustration of Eugene Fama by Lauren Uram

Are investors demanding too much from stocks? Eugene Fama thinks so. The esteemed University of Chicago finance professor claims that the historical premium equities hold over bonds is actually smaller than previously thought. New research conducted with longtime collaborator Ken French of the Massachusetts Institute of Technology indicates that both price-to-earnings ratios and price-to-dividend ratios have been reverting to the mean. From these as-yet unpublished results, Fama and French have determined that the expected return from U.S. stocks, accordingly, will be lower going forward than recent experience shows.

This prediction about future market returns is the latest stunner in Fama's long and notable career of innovative research. His contributions to modern finance over nearly four decades have been consistently pioneering - and controversial - beginning in the mid-1960s with his Ph.D. dissertation, "The Behavior of Stock Market Prices." In this work, Fama argued that in an active marketplace of rational, informed, and intelligent investors, securities prices will reflect all available and expected information. In such an "efficient" market, competitive advantages are virtually non-existent, so it's futile to attempt to outperform a market index. And if anyone does manage to beat the market, it's thanks to luck, not skill.

Fama's salvo scored a direct hit on active management and became a bedrock of the Efficient Market Hypothesis. A less-academic version of Fama's bold assertion was published in in 1965 under the title "Random Walks in Stock Market Prices." The Random Walk Theory contends that stock market activity today does not follow any predictable pattern and has no bearing on future price movements. So don't bother scrutinizing charts and evaluating financial statements - all that technical and fundamental analysis won't give you an edge over Mr. Market. While you might aspire to profit from the occasional market overreaction or underreaction, these events become clear only in hindsight.

Investors can find exploitable differences within an efficient market, Fama reported, if they tilt their portfolios toward stocks with smaller market capitalizations. His research with French in the early 1990s showed that size matters: small stocks tend to have higher returns than large stocks over long time spans. At least, that's what the historical data indicates - recent investors in small-value stocks would question that judgment. Even Fama concedes that the small-stock premium - while noticeable - has grown fainter.

Fama and French, true to form, ignited another debate with the claim that value strategies outperform growth strategies because the market identifies distressed, unloved stocks as riskier investments than popular growth stocks. On the opposite side of the issue are those who say that value wins over growth because investors overreact to negative news and temporarily push down prices too far.

This tug-of-war between risk and mispricing continues unabated. Moreover, skepticism about rational markets is particularly acute nowadays. Stock prices have touched all-time highs and tumbled back down. Such gut-wrenching market swings have become regular events, and investors in these times understandably want to make sense of uncertainty. They also hold out hope that the market gets prices wrong, and sharp-eyed buyers can scoop up bargains. But Fama - always pushing the limits, whether he's windsurfing, playing tennis, or researching stock prices - is rarely pushed to the wall himself. Through all the noise and criticism, he has remained a stout defender of the faith, volleying back hard shots against market efficiency with his own retorts that, like so many tennis balls, tend to land confidently inside the baseline.

Noteworthy new research you've just completed shows that investors should expect less from stocks in the future. What have you found? Ken French and I have an unpublished paper that says expected stock returns are going to be pretty low. Average returns have been high relative to expected returns, but the historical data tends to overestimate. Going forward, you're not going to project that returns would be as high as they've been in the past.

The growth of prices has way outstripped the growth of dividends and earnings. Dividend-price ratios and earnings-price ratios are basically reverting to the mean. In the last 10 years, the dividend-price ratio has gone way down. That can be a reflection of two things: expected future growth of dividends and earnings is real high, or expected returns are real low. We can't find any evidence that expected growth rates are that high. In the historical data, you also see that growth rates are unpredictable. Earnings growth rates have been high for the last 10 years, but the 10 years before that they were low. And if you look at the last 20 years, it doesn't seem much different. We don't see any reason to expect higher earnings.

If earnings growth and dividend growth isn't particularly high going forward, the only way to justify a low dividend-price ratio is with a low expected return.

How far should shareholders ratchet down their expectations? Not lower than the risk-free rate on a short-term bond, but, for the total market, instead of a premium of about 4 percent, now it's about 1 percent.

Another equity premium that you've studied extensively also seems weaker: the size factor. The groundbreaking Three-Factor Model that you developed with Ken French discounted the singular importance of a market beta and added two new variables in market returns: size and book-to-market. What has happened since then to the size premium? The small-size premium is a little more marginal. Small stocks do have somewhat higher returns than large stocks, just not as big as we thought. In the original study, the size effects were done on New York Stock Exchange stocks, where small size tends to be a proxy for value. So the original study was in fact mixing size and value. If you separate them, then the size effect becomes quite smaller. The premium was around 5 percent. Now it looks more like 2 percent.

All of these numbers are subject to a lot of uncertainty. A number as big as 2 percent, which people would kill for if they could get it for sure, in a long-time series of volatile returns doesn't look that big.

Is the small-cap effect dead? I don't think so. These companies really do have a higher cost of capital. Differences in expected return is the same as cost of capital. If you think small stocks have a 2 percent higher expected return, they have a 2 percent higher cost of capital.

How did you and French develop the Three-Factor Model? In the 1992 Journal of Finance article called the "Cross Section of Variation in Expected Stock Returns," we looked at variables that people had found in the past were related to differences in expected returns across stocks. Variables like market capitalization, book-to-market equity, leverage, earnings-price ratio, long-term return reversals - the whole litany of things that people had looked at. Two of them - market capitalization and book-to-market ratio - seemed to provide a pretty good explanation of everything that was captured in the whole set.

The origin of the Three-Factor Model is to basically say that in addition to an overall market factor, if you want to explain volatility of stock returns and want to explain average stock returns, you need risk factors in addition to the market. The second factor seems to be related to size. Small stocks behave differently than big stocks. They do have somewhat higher returns, just not as big as we thought it was when we did the control for value. And the third factor seems related to value vs. growth.

Do these relationships tell a risk-related story, or do they reveal market mispricing? What we've done since is write a sequence of papers that try to address that issue. These papers show that small stocks tend to move together, big stocks tend to move together, high book-to-market stocks - value stocks - tend to move together, and low book-to-market stocks - growth stocks - also tend to move together.

So which is it: risk or mispricing? It's a risk story. What we found is that earnings of these companies move together and their return moves together. They give rise to a risk that can't be diversified away. That can explain why they have a premium.

Indeed, your research with French contends that value stocks outperform growth stocks because they're more risky. In your work, a variable like book-to-market is a proxy for risk. High book-to-market stocks - low price-to-book - deliver better returns than low book-to-market stocks - high price-to-book. You're being paid more for the additional risk. The premium for value stocks is due to the rational investor's dislike for companies that are doing poorly. Value stocks are basically distressed stocks. People don't like them. If they also weren't associated with the common risk factors you can't get rid of, that story would be nice and clean. The fact is that there is a value vs. growth common factor in returns that makes it very difficult to distinguish between that story and one that says this is just rational pricing of a risk factor.

We're talking about a multi-factor view of the world, so you can't summarize risk solely in terms of return volatility. If all you are concerned about is volatility of return and you believe these results, you'd stick to value stocks. That's all you'd buy. But your returns would look a lot different from the market returns. They wouldn't be that highly correlated. The essence of a multi-factor story is that different sources of volatility have different expected returns.

Finance professor Josef Lakonishok, who believes that human behavior and psychology influence markets, finds fault with the notion that high book-to-market stocks are riskier. He notes that an Internet company like Yahoo! has little book value and a large market capitalization. An underloved utility, in contrast, has a lot of book value and a smaller market cap. By your reasoning, he says, a utility would be more risky than Yahoo! because it has a higher book-to-market value. How would you answer that? You may think that's wrong, but the reality is it's all coming out of the price. People are willing to pay more for a dollar of whatever Yahoo! owns than they are for a utility. Which, if you turn it over, says they're willing to hold Yahoo! at a lower expected return.

You no longer think about risk in terms of variance alone. If you do, that takes you right back to the Capital Asset Pricing Model. You want to think about risk in more expansive ways. The cost of capital of a distressed company is higher than the cost of capital of a growth company like Yahoo! Distressed companies pay more through the lower stock price, and that's the way they generate a higher return.

How would you value an Internet stock? I don't know how you value the Internet sector. You're betting on the distant future. But somehow the stuff has to get priced. Maybe the prices are right, maybe they're wrong. We don't really know at this point. Sure there's mispricing, but is it too high or too low? That's the problem, and I don't think it's predictable. The information is already in the price. You're not going to get something for nothing.

This behaviorist idea of an inefficient market, full of anomalies and miscalculated prices, appears quite in vogue nowadays. Robert Shiller, for instance, is topping the best-seller lists with Irrational Exuberance, his discourse on investor behavior. He's been saying the market is overvalued for years now. At some point, you're going to be right. It's kind of a religious idea. You say the market has excessive exuberance, but I don't know how you'd document it. They have all these behavioral arguments. My bottom line always is if the market is so inefficient, why do professional managers have such a hard time beating a passive index? That's a tough one for these guys. If markets are as easy to time as the irrational exuberance story indicates, why don't market timers do better?

So what would you say is the cause of stock market "bubbles"? I don't think that word has any meaning. "Bubble" assumes there's something predictable, that you can tell when the market is going up and when it's going down. We're real good at identifying these things after the fact, not before.

It's true that expected returns are low by historical standards. But I'm not sure why that's due to irrational exuberance. All that says to me as an economist is that a lot of people out there are willing to hold stocks with low expected returns. It's supply and demand. When the price goes up, expected return goes down. I don't need a behavioral story to justify that.

Why do you suppose the behavioral argument has become so popular? There's a great demand among investors and professional managers to want to think the market's inefficient, so there's some money to be made above and beyond normal returns. The evidence indicates that the market is pretty efficient. Predictability is very difficult to find. There are no managers who beat the market with any degree of consistency. We did an experiment where we looked at 20 years of data on mutual funds. We took the funds that did the best the first 10 years and then looked at how they did the second 10 years. The only persistence you could find was on the negative side; funds that did poorly continued to do poorly.

Market historian and author Peter Bernstein was quoted in this magazine recently (see May 2000 "Capital Ideas Revisited," page 112) as observing that the market efficiency story and the argument of whether or not you can beat the market has weakened. "The case for active management is much stronger than the efficient market people would lead us to believe," Bernstein said. How would you respond to that? The evidence that people can beat the market just isn't there. That argument has never weakened. That's what gets all these guys in the end. There are a lot of people who don't want to look at that evidence. They come up with all kinds of arguments about why markets aren't efficient. But if you look at whether people beat the market, you don't see it. There's persistence in performance on the downside. On the upside, there's nothing.

But even an efficient-markets statesman like Burton Malkiel allows for some inefficiencies in the market. He calls himself a "random walker with a crutch." The question is where do you put the crutch? How big is it? The behaviorists' biggest embarrassment is the fact that professionals don't seem to be able to capitalize on this. Active fund investors are paying for something that's not being delivered. The delivery is random.

Do you ever allow for investor overreaction? That surely happens. But all of those are hindsight observations. You can't tell in advance what is overreaction and what is underreaction. All of this looks totally random, just what you'd expect to see out of an efficient market.

Would you say that the stock market is perfectly efficient? No market is ever completely efficient. The question is how efficient is it, for all practical purposes? It's always a model. For investment purposes, the efficient market is an excellent model.

Has anything ever shaken your confidence in efficient markets? It's clear that the market is not efficient as far as corporate insiders are concerned. Insiders make money when they trade their company's stock. What's surprising is how little they make. The average abnormal returns there are between 1 percent and 1.5 percent. If insiders are in the know and that's all they get out of it, how much can you expect others to get?

The Three-Factor Model has limited use for beta as a risk measure. Is beta dead? Beta is not dead. Beta has a role in every multi-factor model. The Capital Asset Pricing Model is dead. Those are two different things. Now it's always possible that the CAPM is true and we just don't have the big market portfolio that it calls for. But at that point, it seems to me like you're turning the CAPM into a religion. We have to deal with the market proxies we have. The CAPM doesn't seem to do that well with them.

In your estimation, how much bearing on performance does beta carry? If you look at beta alone, it doesn't help. If you look at beta in combination with these two factors, it helps. The value/growth premium is negatively correlated with beta. If you have a three-factor model, where you control for the size effect and for the book-to-market factor, then you see the market factor comes in and you do need beta. The Capital Asset Pricing Model is dead from the standpoint that it takes more than one factor to explain. Beta alone doesn't do it. The CAPM is a one-factor model. If you're convinced the one-factor model doesn't work, then CAPM is dead whether or not beta is related to average return.

Critics charge that these "something-over-price" relationships are just data mining without hard fundamental theory supporting it. That's also why after the initial paper in 1992, we've written several additional papers. That's why we're extending it to international markets. There is the problem that something which exists in the past data may be there by chance. I don't think the risk factors are there by chance, but the premiums on those factors could be there by chance. That's why we're extending the data back to 1926. We try to get a sample period that's different from the period in which the phenomenon was initially discovered. The interpretation of the results is controversial. Is it a risk story or a market mispricing story?

What does your follow-up research show? We know you need at least two more factors for bonds. You need a maturity premium and a default-risk premium - long vs. short, and low-grade vs. high-grade. That seems to do it for the bond market. So if you want to combine bonds and stocks, you have a five-factor model.

Doesn't the real reward from small stocks come only if you are fortunate enough to buy the super-winners? That's not the way the data goes. Everybody obviously wants to pick those stocks, but the studies just do a cut on size and see what happens after that. These are companies of a different risk than the big companies and, as a consequence, they're going to have different expected premiums.

If small stocks outperform, then what do you think of the idea that active money managers can exploit inefficiencies in small stocks and are more likely to hurdle their index? That is indeed the line of active managers. If you look at the actual performance of their mutual funds, small-cap funds don't do any better than the big-cap funds once you adjust for the fact that they choose small stocks. They don't seem better able to choose among the small stocks than large-cap fund managers choose among large stocks. They're both pretty bad.

Why do money managers as a group consistently underperform? They underperform because they don't have the ability to beat the market. There are always exceptions. Sometimes those exceptions can be due to chance, but sometimes they can be due to talent as well. You hear about the same guys all the time.

Two names we often hear are Warren Buffett and Peter Lynch. They certainly have shown an ability to beat the market. Warren Buffett is my all-time hero. What he says is that he can pick one or two things every few years and hold them. And that's about it. Lynch, I think, was lucky. Out of any set of managers there's going to be one who does extremely well. Whatever Lynch had wasn't transferable, that's for sure.

Doesn't investment advice have greater value today? Information is continuously available and markets are more sophisticated. If there's better information available, it's more likely that the market is efficient. The more people trying to beat the market, the more efficient the market is going to be.

If you were running an equity mutual fund portfolio, what would it look like and which strategies would you use? I'd do it just like Dimensional Fund Advisors or any other passive manager. You have a market portfolio, a small-stock portfolio, a value portfolio, a bond portfolio, and you let people balance it.

Is an indexed investment a superior investment in every sector? Indexing - I never liked that term. You want to buy the population - the universe of the sector. You can't fiendishly index - especially small stocks and small value stocks because they're so difficult to trade. If you insist on market proportions, you're going to get killed on the trading costs.

What about the value of persistence? Some managers do outperform. Why not pick them and stick with them? That's a good idea. But what you want to do is avoid the underperformers. I don't think you've got much of a shot at picking the ones that will do better than the market. The best way to pick the ones that are not likely to do worse is to look for those who hold diversified portfolios, charge low fees, and have low expenses. I'm a DFA fan. I'm a Vanguard fan.

Academic theories of finance increasingly are moving from the ivory tower to Wall Street. Does this surprise you? They're almost too successful. People grab them now before they're fully cooked. When I started, it was very difficult to get people to pay attention to what was going on in academic research. Now there's intense interest. The academic financial community has established its credibility. People pay attention.

Reprints Discuss this story
This is where the comments go.