Author
Interview conducted November 2, 2007.
Few economists have had greater influence on financial theory, and practice, than Eugene Fama. His 1964 doctoral dissertation, “The Behavior of Stock Market Prices,” suggested that stock markets are efficient. Because of competition among investors, company share prices respond swiftly to past events, current information and future expectations. Actual prices are good estimates of intrinsic value, therefore, and no analyst can find consistent, profitable anomalies.
And price trends are random, said Fama. Past patterns don’t predict future directions. Published in a nontechnical article a year later, his research popularized the “efficient market hypothesis” and “random walk theory.”
Fama’s work soon transformed Wall Street, and later Main Street, by giving rise to a proliferation of low-cost index funds, as many questioned the value of paying for active portfolio management. “If one takes into account the higher initial loading charges of the [mutual] funds,” he observed over 40 years ago, “the random investment policy outperforms the funds.”
More recently, and often in collaboration with Dartmouth’s Kenneth French, Fama reexamined the capital asset pricing model, a classic model for determining fair cost for equity capital, and declared it an empirical failure unless two other factors, market capitalization and book value to market value, are also included. This work, too, has transformed Wall Street by providing academic support for “small-cap” and “value” funds.
Fama has done significant research in other areas, consistently sustaining connections between financial economics and macroeconomic theory. But his illumination of financial markets will undoubtedly be recognized as the University of Chicago scholar’s paramount contribution to both research and reality.
“Eugene Fama has had pathbreaking insights into the functioning of markets, asset-pricing theory, and corporate finance,” Nobel laureate Myron Scholes noted recently, “[insights] that have benefited market participants worldwide.”
Principal-Agent Issues
Region: In the early 1980s, you authored three key pieces regarding principal-agent conflicts [due to differing incentives of an organization’s owners and employees] and how they play out efficiently in various types of organizations. How have your ideas evolved in light of transformations in the corporate world?
Fama: I haven’t spent a lot of time on these issues since then, but they keep popping up. I haven’t seen anything that would cause me to change my opinions generally, but something that has bothered me is the drying up of the takeover market due to the installation of antitakeover provisions by most companies, enabled by state legislatures.
Region: Poison pills and the like?
Fama: Right, and that is very unhealthy, I think, for the corporate world because it takes away the threat of outside takeovers, which is very important for the economy.
Region: A form of market discipline.
Fama: Yes, it’s a unique discipline that corporations have that other forms of organization don’t have. For example, it’s very difficult to attack the University of Chicago in that way. It doesn’t need a takeover defense because there’s no real way to attack it. For a corporation, on the other hand, there was a way. That allowed corporations to have expert boards because the board wasn’t the court of last resort. But the institution of all antitakeover amendments threw a wrench in the process.
CEO COMPENSATION
Region: Another issue those papers touched on was compensation of CEOs, a controversial question in recent years. How do you view the suggestion that some CEOs are overcompensated?
Fama: If the [compensation] process gets captured by the CEO, then it can get corrupted. But if what you’re seeing is a market wage, then I don’t know why you would say it’s too high. If it’s a market wage, it’s a market wage. I don’t know of any solid evidence that the process was corrupted. So my premise would be that you’re just looking at market wages. They may be big numbers; that’s not saying they’re too high. It’s easy to say that people are paid too much, but when you’re on the other side of the fence trying to hire high-level corporate managers, it turns out not to be so easy.
EFFICIENT MARKET THEORY
Region: Can you give us a lay definition of the efficient market hypothesis? How does it differ from random walk [the idea that movements in stock prices are unpredictable]? And what is the genesis of efficient market theory?
Fama: The basic wording of it is very simple. It says prices reflect all available information. The conundrum is how to determine whether prices reflect all available information, and you can’t do that without a model of market equilibrium. What I added to the story was just pointing out that you need a model of market equilibrium in order to carry out the tests of market efficiency.
In the early 1960s, the advent of computers allowed people to do things with data that they couldn’t do before. And the most easily available data was stock market data. So lots of people started working on stock market returns, and the question arose, well, what would we expect if markets were working properly? “Random walk” was the first manifestation of that. But it’s kind of a clumsy statement because it doesn’t recognize that you need a model of market equilibrium to decide what the market’s trying to do in setting prices.
Region: So that is the “joint hypothesis.”
Fama: Right. The joint hypothesis problem says that you can’t test market efficiency without a model of market equilibrium. But the reverse is also true. You can’t test models of market equilibrium without market efficiency because most models of market equilibrium start with the presumption that markets are efficient. They start with a strong version of that hypothesis, that everybody has all relevant information. Tests of market efficiency are tests of some model of market equilibrium and vice versa. The two are joined at the hip.
Once I pointed that out, it was clear that the random walk model was kind of irrelevant. You could have prices not following random walks because the model of market equilibrium could generate expected returns that had some predictable time-varying patterns to them. So the whole nature of the game changed.
Region: The idea that things are unpredictable doesn’t necessarily mean that they’re efficient.
Fama: No, not necessarily. Sure, prices could be random and still be inefficient. Basically, what market efficiency says is that the deviation of the realized price from the equilibrium expected value is unpredictable based on any past information.
Region: Can you tell us about the genesis of this idea? You were at Tufts University at the time, I believe.
Fama: When I was at Tufts, I was working for a professor who had a stock market forecasting service. My job was to devise rules for predicting the market, and I was very good at it. But he was a pretty good statistician. He always told me to set some data aside so I could test [the rules] out of sample. And they never worked out of sample.*
So when I came to the University of Chicago and people were talking about these things, it suddenly dawned on me that maybe that was the nature of the game, that there just wasn’t much predictability of returns because markets were working efficiently. That was the beginning of the story.
There were lots of people at Chicago and at MIT who were very interested in that issue. Merton Miller. Franco Modigliani. Paul Cootner. Paul Samuelson was very interested in it. And Benoît Mandelbrot.
Region: The subject of your first paper, I think.
Fama: Right. Half of my thesis was on the predictability of returns and the other half was on the nature of the return distribution, which was what Mandelbrot was all about, and still is.
IRRATIONAL EXUBERANCE
Region: Some economists—you know them well—say that the stock market crash of 1929 and the more recent climb and decline of the market in the early 2000s suggest that “irrational exuberance” affects the stock market. How do you reconcile this alleged evidence of herding behavior and animal spirits with the notion of market efficiency?
Fama: Well, economists are arrogant people. And because they can’t explain something, it becomes irrational. The way I look at it, there were two crashes in the last century. One turned out to be too small. The ’29 crash was too small; the market went down subsequently. The ’87 crash turned out to be too big; the market went up afterwards. So you have two cases: One was an underreaction; the other was an overreaction. That’s exactly what you’d expect if the market’s efficient.
The word “bubble” drives me nuts. For example, people say “the Internet bubble.” Well, if you go back to that time, most people were saying the Internet was going to revolutionize business, so companies that had a leg up on the Internet were going to become very successful.
I did a calculation. Microsoft was an example of a corporation that came from the previous revolution, the computer revolution. It was hugely profitable and successful. How many Microsofts would it have taken to justify the whole set of Internet valuations? I think I estimated it to be something like 1.4.
Region: About one and a half Bill Gateses.
Fama: That’s right. And Microsoft was a good example because the worse their products were, the more money they made [laughter]. Who didn’t struggle with DOS and then the first versions of Windows?
INTERNATIONAL EQUITY MARKETS
Region: Are all stock markets equally efficient? Is the Hang Seng as efficient as the Nasdaq as the Australian stock exchange? If not, are there money-making opportunities internationally that don’t exist in the United States?
Fama: Anybody who has studied that issue doesn’t come to the conclusion that there are huge opportunities in other markets that don’t exist in the United States. That’s kind of a standard line of international money managers, that the opportunities are better in international markets. That’s certainly not true in developed markets.
In emerging markets, well, I think maybe insiders have more information than they do in domestic markets, but maybe not. In any case, there’s not enough data to know about emerging markets. And the variances are so big it would be impossible to know anyway. When people study money managers in developed markets, they don’t find any evidence that those markets are inefficient … and there’s very little evidence that they’re inefficient in the United States. But I’ve never taken the extreme position that markets are entirely efficient.
HAS FAMA SOFTENED?
Region: I was going to ask you about that. As you know, I’m sure, there was a lengthy Wall Street Journal profile of you and your colleague Richard Thaler in 2004, suggesting that you had softened a bit.
Fama: [Laughter].
Region: And I’m wondering if that was accurate or if you’ve always believed that markets are less than perfectly efficient?
Fama: I start my class every year by saying, “These are models. And the reason we call them models is that they’re not 100 percent true. If they were, we would call them reality, not models. They’re simplifications.” But the acid test is, How good are the simplifications for your purposes? And for almost all purposes, market efficiency is a very good approximation. There is very little evidence that money managers can beat the market.
DIMENSIONAL FUND ADVISORS
Region: Then this is the right time to ask, I guess, about Dimensional Fund Advisors, on whose board you sit. Why should an investor pay a management fee to Dimensional Fund when an index fund might provide efficient returns at lower cost?
Fama: Well, Dimensional is a passive manager. They don’t charge high fees. Vanguard, for example, is another passive manager that charges very low fees. You shouldn’t pay managers very much. The average management fee for an actively managed mutual fund is about 1 percent. There’s no evidence that they generate anything for that 1 percent. So my answer is, I don’t know why anybody buys them.
Region: And yet we keep doing it, don’t we?
Fama: Well, people want to think there’s money left on the table for them.
THREE-FACTOR MODEL
Region: With Kenneth French, you’ve said that the capital asset pricing model (CAPM) developed by John Lintner and William Sharpe has “fatal problems” in explaining stock market returns because of its reliance on beta [the volatility of an individual stock relative to overall market volatility]. And you’ve found that two other factors are crucial for determining prices. Can you tell us about these factors? Are they inefficiencies, or do they represent hidden risk? And is the CAPM truly dead?
Fama: Let me first tell you what the returns evidence says, and then we can talk about how to interpret it. The returns evidence basically says that if you look at the CAPM market beta, it’s not enough to describe the cross section of average returns.
The CAPM says that all you need to know are these market betas, market sensitivities, in order to fully describe the cross section of average returns. What you find is that other variables contribute to the explanation of average returns above and beyond what you get from beta. Indeed, over the last 50 years, you get very little at all from beta.
The two variables that we’ve focused on are market capitalization (the financial profession calls it size, a misnomer because it’s really market capitalization) and the book-to-market ratio, the ratio of the book value of a common equity to its market value. Now, there’s no magic in that ratio. The ratio of almost anything to price will work as well. These are the two variables.
So, small-cap stocks have higher average returns than large-cap stocks, and stocks with higher ratios of book value to market value have higher returns than low book-to-market stocks. Low book-to-market stocks tend to be growth stocks. High book-to-market stocks tend to be relatively more distressed; they’re what people call value stocks. That’s given rise to what the finance profession—academic as well as applied—calls the size premium and the value premium. The value premium tends to be bigger.
So the issue then is, Are these risk factors or market inefficiencies? One group of people says they’re market inefficiencies—particularly the value premium. The behaviorists tend to say the value premium is a market inefficiency. Their story is: The market overreacts to good and bad past times. It doesn’t understand that things tend to mean revert. So growth companies that have done very well tend to be overpriced, and value companies that have done poorly tend to be underpriced, and then the market realizes this and corrects it. And this story says, basically, that people are dumb; they never learn. So every generation of growth stocks and value stocks goes through the same sort of cycle.
That’s not too appealing to an economist—the idea that people never learn about these things—but that is the behavioral story. And initially they said these are arbitrage opportunities because if you go long value stocks and short growth stocks, you get something with a variance close to zero.
But French and I pointed out that if you do that, you get something with a variance very close to the market variance, not zero. It’s quite a risky strategy. And the premium is about the size of the market premium. So it looks and smells like a risk premium. And we developed a three-factor model with a size premium in addition, basically the difference between the returns on small stocks and big stocks.
So, our model has three factors. Every asset pricing model says you need the market in there. Then they differ on how many other things you need. The CAPM says you only need the market. We basically say a minimum of two other factors seem to be necessary. And these two do a pretty good job.
There’s still a third explanation, which is not based on overreaction. It says that people just don’t like small stocks and value stocks. There’s some amount of utility that people get from the nature of the stocks that they hold. So they like big stocks and they like growth stocks, and they’re willing to hold them even though they have lower average returns.
Now you can’t have an arbitrage opportunity there because then there’d be a sure profit. But the fact that they look like risk factors can sustain that story. You can’t tell the difference between that story and a risk story.
MOMENTUM
Region: Let me ask you about momentum. You’ve said that it’s the strongest challenge to the hypothesis of market efficiency. Can you elaborate on that?
Fama: There’s evidence that if you rank stocks every month based on their last year of returns, the very extreme winners tend to win for a few more months and the losers tend to lose for a few more months.
That seems to be true in U.S. data beginning around 1950. We don’t have foreign data going back that far, but it tends to be there in major foreign markets except for Japan. It doesn’t tend to be there in the U.S. data for the ’30s and ’40s. So there’s some chance that it is just a chance result. There are so many people looking for anomalies in the data, that may just be the biggest one that they’ve found. Maybe it won’t be there in the future. We don’t know yet.
Region: Is there an opportunity to make money there?
Fama: Well, there isn’t much of an opportunity to make money, because as I said, you do this every month. And if you rank and trade stocks every month, the turnover of these portfolios is enormous.
Region: The costs will eat up the profits.
Fama: Right. The costs will kill you. So the people who have written these papers have said, basically, “This is interesting, but forget about trading on it.” But it’s still interesting.
HOUSING MARKETS
Region: Your efficient market hypothesis applies to stocks, of course. Recent events have led to scrutiny of housing markets. Are housing markets efficient? Is there greater potential for irrationality to crop up there, either because housing investors are less sophisticated than stock market investors or because housing markets are less liquid?
Fama: I don’t know. Housing markets are less liquid, but people are very careful when they buy houses. It’s typically the biggest investment they’re going to make, so they look around very carefully and they compare prices. The bidding process is very detailed. The bottom line is that real estate is a huge component of wealth and we have no data on it. So the answer to your question is, Who knows?
CREDIT MARKETS
Region: Some observers have suggested that regulators and others have put too much reliance on ratings agencies to determine the risk of mortgage-backed securities and that even financially sophisticated parties “didn’t really know what they were buying.” Is this evidence that credit markets are inefficient?
Fama: That story just doesn’t appeal to me. First of all, it’s well known that rating agencies tend to lag actual changes in credit worthiness. For example, stock prices predict changes in ratings better. The best models of credit quality are basically options pricing models that work off the stock price. So I’m very skeptical of these stories.
The bond market is a simpler market than the stock market. Bonds are simpler to evaluate than stocks, because there’s downside risk, but you don’t have to worry much about the upside: They’re not going to pay you more than they promised. So bonds are much simpler to deal with. Now bond products have become more complicated because of the securitization of that market, but still not that big a deal.
NEW FINANCIAL TECHNOLOGIES
Region: What about hedge funds, collateralized debt obligations and other newer financial technologies—do they serve a useful purpose in mitigating market risk, or do they heighten it?
Fama: I don’t know. People talk on both sides of that issue. The problem is that we don’t have very good hedge fund data and the data we have only goes back about 10 years. That’s just not enough to come to any conclusions on these issues. So I don’t know if it’s going to take another half century before we really know. You’re talking about returns with such high volatility that it really is going to take that long.
This is a standard part of a talk that I give to investment professionals: People like to tell stories about short periods of data, but the reality is that you can’t measure the market premium over periods shorter than an investment lifetime. The 5 percent stock market premium over bills takes about 35 years before it becomes two standard errors from zero.
EQUITY PREMIUM
Region: How do you explain the equity premium puzzle [the idea that stocks should in theory provide only a 1 percent higher annual return than bonds, but have historically returned nearly 7 percent more]?
Fama: In terms of these consumption-based asset pricing model stories? What I say to the consumption people is: You’re telling me the premium should be about 1 percent a year. Well, you wouldn’t be able to tell the difference between that and zero over a 1,000-year period. And for a 1 percent a year premium, who do you know that would hold stocks? It’s this representative investor, but who is that guy anyway? I wouldn’t hold them. I don’t know anybody else who would. So there’s got to be something missing in those models.
ARE BANKS STILL SPECIAL?
Region: In 1985, you wrote a paper titled “What’s Different About Banks?” It’s a question often discussed by the Fed, for obvious reasons. You wrote that special monitoring services and special transactions services, including the checking system, are part of what makes them unique. As other nonbank organizations take over some of the roles, are banks no longer so different, no longer as special?
Fama: Excellent question. Basically, the only companies that can issue debt publicly are very large companies. I mean directly issue debt, commercial paper or marketable bonds. Everybody else has to go to an institution. Now what institutions have done is to securitize these things, put them into bundles and put them on the market. Lots of people have been working on the extent to which the monitoring function as a consequence has been diluted somewhat because the banks aren’t holding 100 percent of the paper that they create. So that’s an ongoing issue. I don’t know what the answer is about whether banks are less relevant now. They’re doing a lot more different things than they ever did, but so are all financial institutions.
SSRN
Region: You and Michael Jenson helped create the Financial Economics Network, which then broadened into the Social Science Research Network. What role do you see it playing in the future in the creation and distribution of economic research?
Fama: I think it’s great for working papers, but I don’t think you can do without the refereeing process. Will all journals end up online? I think that’s a good possibility. But the refereeing process is still critical for quality, improving work and certifying work. So professional journals may change in nature, but that function will remain, and the editorial function will remain as a consequence.
FINANCE/MACRO
Region: It seems to me that macroeconomists are paying more attention these days to the work of financial economists, especially in trying to understand asset pricing. Do financial economists find equal value in the macro theory work being done these days?
Fama: I think those two areas have always been pretty closely joined. I, in fact, wrote a lot of macro-related stuff in the ’80s or even earlier. For example, rational expectations stuff is basically efficient markets; they’re pretty much the same thing. If you’re talking about the macroeconomy, I don’t see how you can avoid financial markets. That’s a big part of the game. Nobody talks about money and bonds anymore the way they did when I was taking macroeconomics. Now people realize it is a lot more complicated. Finance and macro are joined. Our finance faculty has several people who were trained as macroeconomists, especially on the asset pricing side.
TIME ALLOCATION
Region: I understand that you work every day, even holidays. Is that right?
Fama: Right.
Region: That’s an amazing work ethic.
Fama: Not really.
Region: I’ve also heard that you’re a dedicated athlete.
Fama: Right. I work every day, but I never work a full day. I get up at five o’clock in the morning and I work basically all morning until maybe one o’clock, two o’clock, and then I go play golf, I go windsurfing, I play tennis. And that’s it.
Region: We should let you go then. Thank you very much.
—Douglas Clement
Nov. 2, 2007
* When building forecasting models, statisticians often partition their data. The in-sample portion is used to develop the model, and the out-of-sample batch is then used to test the model's predictive ability.
