Why Have Markets Learned Nothing in the Last 50 Years?
An editorial last week about financial market liquidity, by my new colleagues at Bloomberg View, cited this fascinating paper by Jennie Bai, Thomas Philippon and Alexi Savov called "Have Financial Markets Become More Informative"?* The answer is no.
As the View editors put it, the authors "find that prices of stocks and bonds in the U.S. are no better at predicting the profitability of companies than they were 50 years ago, despite the vast increase in trading volumes and sharp decrease in transaction costs over the same period."
This matters, since the main justification for having stock markets to begin with is that they allow society to allocate capital to its highest and best uses. Companies with bright futures have high stock prices and can raise money easily; companies that are doomed have that news broken to them gently by the stock market. This is useful for companies to know: For instance, if you announce an acquisition and your stock price drops, indicating that the market thinks the merger is bad, you can back out of the deal and narrowly avoid a disaster.** But it's more useful for investors to know: If a company raised a billion dollars of stock and is now worth a million dollars, maybe don't give them any more dollars?
So it's sort of fascinating that the market is no better at predicting companies' futures than it was in 1960, despite the intervening invention of, I'm gonna go with, high frequency trading and Bloomberg terminals.*** The authors don't speculate on why this might be, but the mystery dovetails with two of my little obsessions, and I'm going to share them with you.
First: What else has changed since 1960? One possible answer is that companies are significantly more constrained in what information they can leak to their favorite investors and analysts than they were a half-century ago. Relatedly, there's rather less insider trading now than there was then, what with it being rather more illegal, far more heavily punished, and far more effectively policed. In fact, one way that people sometimes try to measure insider trading is by seeing how well stock prices predict things: If a target company's stock jumps just before it announces that it's being acquired, that suggests that someone knew something they shouldn't have. And the evidence is in fact that there's less insider trading than there used to be.
Defenders of insider trading often argue that it makes prices more efficient: By letting insiders profit from their knowledge, we get markets that, while they might be unfair, are better at predicting performance and allocating capital. Maybe! One possible interpretation would be that markets have gotten a bit better, technologically, but those gains have gone toward increasing fairness rather than increasing allocative efficiency.
Or maybe not. The researchers consider this explanation, though they say that their "analysis does not provide a way to control for changes in disclosure." But they note that there's some evidence that disclosure rules are not the cause.****
Another thing that's changed since 1960 is the rise of indexing. In 1960, if you wanted to buy stock, you bought the best stocks you could find. In 2013, if you want to buy stock, you buy all of the stocks in a market-cap-weighted index fund. Or you do whatever else you do, maybe you buy the best stocks you can find, GOOD LUCK TO YOU, I'm in index funds like a rational human. Index funds were invented in 1975 and have become steadily more popular since.
The appeal of index funds is that you're probably bad at picking stocks, and you're probably bad at picking people to pick stocks for you, so you might as well just take it out of everyone's hands and buy all the stocks. But of course, indexing contributes nothing to efficient capital allocation: If I put $1,000 into a total market index fund, it will dumbly go buy shares in the most hated and most loved companies in the market.***** At a tiny margin, I'm cushioning the bad companies against failure by just buying all the companies at once. An index fund isn't predicting earnings. It's not telling companies to increase or shut down their R&D spending. It's just buying all the stocks.
A thing called the Grossman-Stiglitz paradox says that markets can never be totally efficient: Sometimes stock prices need to be wrong, because otherwise no one would want to invest in making them more right. If everyone indexed, then stock prices would all be wrong, and you could make a killing buying good companies and selling bad ones. If no one indexed -- as happened in 1960, more or less -- then prices would be as right as everyone could make them, but you'd have to work really hard to get them that way instead of just free-riding on everyone else's work.
So another possible explanation for Bai, Philippon and Savov's results is that markets have gotten better, but that the gains have gone toward making it easier to index -- and not try to predict the future -- rather than towards better predicting the future.
* Philippon, incidentally, is among the most entertaining of finance academics, because all of his research is basically on the topic "finance: worth it, or what?" "Or what," is largely his answer; this is a famous-ish paper asking "why do financial intermediaries make so much money," with the answer being something like "because you let them."
** No no no. One thing you'll need to get used to is that sometimes I'm kidding. Basically any time an acquisition is announced the acquirer's stock price drops and the acquirer says "no, market, see, you just don't understand the synergies." The market is usually right though.
*** Maybe even more interesting is that Bai et al. find that stock prices do predict R&D spending, but that "the increased predictability of R&D is not related to increased predictability of earnings, so we cannot conclude that informativeness has increased." The intuition is that companies use the capital-allocation signal of stock prices -- they invest in the things that the market tells them they should invest in -- but it doesn't work. The market is wrong.
**** In particular "we see return surprises grow in the years prior to Reg FD," the 2000 regulation that restricted companies from selectively disclosing information to favored analysts, which suggests that whatever made markets worse predictors wasn't driven by disclosure regulation.
***** Of course it will put more money into the most-loved stocks, if it's a market-cap-weighted index, but it's not doing any work to decide. It's relying on previously produced information, not producing new information.