Although I have a PhD in economics, I have never taken a course in finance. (Finance courses were simply unavailable when I was a student, first at Presidency College, Calcutta, and then at Stony Brook University.) Nevertheless, I will try to describe my understanding of the work recognized by the awards.
We'd all like to be able to predict asset prices. If we knew which asset would rise in price, we could buy it now, and sell it later at a profit. If we knew which asset would fall in price, we could "short" it -- that is, borrow the asset (from someone who currently has it), sell it right away for, say, $100, buy it later for the lower price of, say, $80, return the asset (to whoever we borrowed it from), and keep the $20 profit (less a small interest payment to the lender, perhaps).
Unfortunately, a theory called the efficient markets hypothesis -- with a long pedigree that stretches back a hundred years to the doctoral dissertation of French math student named Louis Bachelier -- says that if people are rational and if asset markets function well, asset prices would be unpredictable.
Here's how I teach the theory to my (undergrad) students: I say to them, "Imagine you are listening to your car radio on the way to class and you hear that McDonald's shares are rising sharply. What could be the reason for this?" After some prodding, the students provide explanations such as "McDonald's just announced a new burger or salad", "McDonald's just announced unexpectedly high profits in the previous quarter," "There's been an E-coli or salmonella outbreak in Burger King restaurants," etc. I then point out to the students that all the explanations they had offered were news items. Finally, I wrap things up by saying that if you accept the idea that the only reasonable explanation for an asset price change is a news event, and if you agree that news events are, by definition, unpredictable, then it follows that asset price changes are unpredictable! My students generally seem to buy the logic.
Eugene Fama, one of today's winners, looked at asset prices and showed that,over short time periods, the efficient markets hypothesis is true: past asset prices are unrelated to current asset prices, and therefore it is indeed impossible to use past asset prices to predict future asset prices.
Fama also showed that news tends to rapidly get embedded in asset prices. By the time you hear that there's been a salmonella outbreak in Burger King restaurants and call your broker with instructions to buy McDonald's stocks, it is already too late. A frenzy of buyers have already bid up McDonald's shares to such expensive heights that it would no longer be a good idea to buy it. Every bit of news leaves its mark on asset prices just as every tremor in our tectonic plates leaves a mark on a seismograph's printout. And, news being random, asset price fluctuations are a "random walk."
(Click to enlarge)
In measuring fluctuations in an asset's price, Fama had to be able to measure both the riskiness of the asset and the fluctuations in its price that are due to its riskiness. The efficient markets hypothesis understands that the riskier an asset, the faster its price must rise, for otherwise people wouldn't buy it. Consequently, Fama had to measure (a) an asset's riskiness, (b) how fast its price needed to rise to compensate for the asset's riskiness, and (c) how fast the asset's price actually rose. The residual -- that is (c) minus (b) -- is what Fama showed to fluctuate like a random walk.
While the short-run unpredictability of asset prices is now well established (for everybody except Wall Street conmen who have managed to build one of the wealthiest industries the world has ever seen, upon the lie that asset prices are indeed predictable and that they know how to predict stock prices), both Fama and Robert Shiller, another of today's winners, have shown that asset prices are predictable under some narrow circumstances.
The basic theory of the efficient markets hypothesis assumes that an asset's price at any given date is equal to the total income the buyer expects to earn from the asset over the future. In the case of a company's shares, asset income consists of dividends paid by the company. Shiller showed that dividend income tends to be stable over time. Therefore, if the efficient markets hypothesis is right that stock prices are equal to total future dividends, they should be roughly equal to total past dividends as well. Unfortunately for the theory, Shiller showed that, while dividend income tends to be stable over time, stock prices fluctuate wildly around dividend income. Yes, the theory is right in the sense that any gap between stock prices and dividend income tends to narrow over time, but the process tends to be slow. And this slowness in the closing of the gap between stock prices and dividend incomes makes stock prices predictable over longer time horizons: if a stock's current price is significantly below (above) dividend income over, say, the past twenty years, then it is likely that the stock price will rise (fall).
(Click to enlarge)
This finding of Shiller led to a lot of theorizing about the reason why the stock-prices-are-unpredictable result of the efficient markets theory fails over longer time horizons. One group of theorists developed a more sophisticated version of the efficient markets hypothesis to address the failure of the basic version. They argued that the excessive fluctuations in stock prices that Shiller observed may be caused by fluctuations in the way people compare incomes earned at different dates. In the previous paragraph, I noted that the basic theory assumes that "an asset's price at any given date is equal to the total income the buyer expects to earn from the asset over the future". I should have said "total discounted income" instead of total income. A dollar earned today is not equivalent to a dollar promised to you in 2020 (even if you are sure that the purchasing power of a 2013 dollar would be the same as the purchasing power of a 2020 dollar). Consequently, when assessing how much to pay for a stock, the buyer must decide how much is expected dividend income in 2015 worth to him and how much is expected dividend income in 2020 worth to him. These comparisons use a concept called the discount rate; if in your view a dollar in 2014 is equivalent to 0.80 cents in 2013, then your discount rate is 0.80. The basic version of the efficient markets hypothesis assumes constant discount rates. The more sophisticated version -- which was built to explain Shiller's finding that stock prices fluctuated a lot -- assumed that people's discount rates could fluctuate.
Theories, of course, need statistical confirmation. And that's where Lars Peter Hansen, today's third winner, comes in. The sophisticated version of the efficient markets hypothesis -- mentioned in the last paragraph -- could not be tested using available statistical tools. Hansen developed a new estimation technique called the generalized method of moments to address the difficulties. And when he applied his method, he ended up rejecting the sophisticated version efficient markets hypothesis as an explanation for Shiller's findings!
This shock had a major effect on how economists thought about asset prices. Models based on rational decision makers and well-functioning asset markets could not explain the big fluctuations observed by Shiller in asset prices. Some economists became convinced that systematic departures from rational behavior held the key to understanding why the efficient markets hypothesis fails for some longer time horizons. This led to a greater interest in behavioral finance and behavioral economics, which combine economics and psychology to understand markets. (Incidentally, I have been teaching a course on behavioral economics for the last few years.)
Like Shiller, Eugene Fama also showed in later work that there are some ways in which the efficient markets hypothesis fails! (This may have come as a surprise to many, considering that, early in his career, Fama had shown -- in support of the efficient markets hypothesis -- that past stock prices do not predict future stock prices and that news events quickly get absorbed in stock prices.) What's the nature of the failures of the efficient markets hypothesis identified by Fama?
The efficient markets hypothesis predicts that stocks whose prices are positively correlated with the average level of stock prices (for the entire stock market) should rise rapidly in price, whereas stocks whose prices are negatively correlated with the average level of stock prices should rise slowly in price. This has to do with the idea of hedging ones risks. If you love the Yankees and hate the Red Sox, you should bet money that the hated Red Sox would win. That way you'd be perfectly hedged: If the Yankees win (lose), you would be happy (sad) that they won (lost) and sad (happy) that you lost (won) a lot of money on your bet. In the same way, stocks that zig when the market as a whole zags are particularly desirable as a hedging tool. This makes stocks that are negatively correlated with the market very expensive to begin with and therefore unlikely to rise rapidly in the future. Conversely, stocks that are positively correlated with the market are useless as a hedge and therefore their prices must rise rapidly, for otherwise nobody would buy them.
All this sounds rather reasonable, does it not? Unfortunately, Fama tested this hypothesis and found no evidence for it! He found that a stock's other attributes -- such as its "total market value" and "book-to-market ratio," but not its correlation to the market as a whole -- are good predictors of its price (thereby weakening the idea that stock prices are unpredictable).
Economists do not have an explanation for Fama's finding. Again, my guess is that psychology is part of the answer. I also wonder whether the predictability identified by both Shiller and Fama (a) can be reliably exploited as a money-making strategy, and (b) whether they will persist in the data once traders understand them better and try to incorporate them into their trading strategies. After all, the basic argument for the efficient markets hypothesis is that predictability provides a profit opportunity and the rush to exploit the profit opportunity destroys the predictability.
For the average person, the work of today's laureates has led to the introduction of index funds. Economists have long been telling whoever would listen to them that they should not listen to Wall Street charlatans who claim that they are good at picking the right stock (or bond, or whatever). Instead, they should buy index funds that take your money and use it to buy a little bit of everything. The unpredictability of stock prices -- especially over short time horizons -- is now well established. So, by all means, buy stocks, but don't try to pick stocks. Just buy index funds that buy a little of every stock that's out there.
Finally, the public pronouncements of Fama and Shiller may suggest that they are at loggerheads. Fama believes that asset markets are largely efficient and Shiller sees them as occasionally prone to spasms of "irrational exuberance" (the title of one of Shiller's books). But the Bank of Sweden gave Fama and Shiller prizes for their research, not for their punditry in the media. The empirical work of these three prize winners can be kept separate from their ex cathedra pronouncements.
