Asness and Liew on Efficiency

Source: Institutional Investor

Cliff Asness and John Liew -- Chicago PhD's and now founding principals of AQR -- have a nice piece in Institutional Investor on Fama, Shiller, Nobel Prizes and efficiency.

They do a good job on the joint hypothesis theorem -- maybe a more important part of Fama's 1970 paper than efficiency itself -- and value and momentum strategies.

They point out one big difficulty for the inefficiency view (p.5). If value stocks are just overlooked and growth stocks irrationally overpriced, why do value stocks all subsequently rise or fall together, and growth stocks go the other way? "Cheap stocks would get cheaper across the board at the same time. It didn't matter if the stock was an automaker or an insurance company. When value was losing it was losing everywhere."

A second very important theorem: the average investor must hold the market portfolio, so alpha is a zero sum game. If you're going to profit, it helps a lot to identify just who the morons are whose money you are taking and why they're willing to give it to you. Everyone thinks the other guy is "behavioral." Are you sure it's not you?

Cliff and John have a theory (p.5) "we've seen that a lot of individuals and groups (particularly committees) have a strong tendency to rely on three to five-year performance evaluation horizons. Of course, looking at the data, this is exactly the horizon over which securities most commonly become cheap and expensive... these investors act like momentum traders over a value time horizon."

Be careful though, so far they are assuming a value effect -- "securities become cheap" means expected returns have already risen. How does this behavior than cause a value effect? They go on: When these people buy or sell, "price pressure" (from people slowly and passively rebalancing over a 3-5 year horizon?) "leads to some [additional!] mispricing (inefficiency) in the direction of value." Hmm, it sounds pretty thin guys.

A deeper and more important criticism of the risk-factor interpretation: "many practitioners offer value-tilted products.. But if value works because of risk, there should be a market for people who want the opposite...Some should desire to give up return to lower their exposure to this risk... we know of nobody offering this systematic opposite product (long expensive, short cheap)."

This is one my MBA students have heard for a while. Go back to Fama and French's story for the value premium as risk factor. People whose labor income or nonmarketed business income is correlated with value stocks should shun value stocks on diversification grounds. If there are more such people (steel workers, say) than people whose human capital or business income is correlated with growth stocks (computer programmers), then the former push down the price and up the expected returns of value stocks.

The trouble is, for everyone that an investment firm finds who has not yet thought about this premium, bought value to write insurance, there should be someone else who has not thought about insurance and wants to sell value to buy insurance. Or, the premium will go away.

So where are these customers, who hear about AQR and DFA and say, "Great, I want to short that!" (The problem is even worse for momentum, for which there isn't even a story.)

In fact, Cliff and John are exaggerating here. There are plenty of people offering the opposite product. There are lots of growth funds for sure! There must be shortable value ETFs as well.

And I think that observation paints a partial answer. Lots of people do overweight growth stocks, and thus underweight value stocks, though the value premium says the growth stocks are overpriced. Why? Well, each one of them thinks their growth stocks are the good ones, and they're taking alpha from the other morons who picked the wrong growth stocks. Growth stocks are where all the trading and volume and information and new products and excitement is.  Invest in railroads, steel and banks (value)? That's nuts, you want to be in the wave of the future, Google and Facebook, thinks the average investor. Jim Davis (Table 3 here) actually finds positive alpha among growth managers.  This is not to say it's all "rational." A rational story for information trading is hard to find. But it certainly does paint a different picture of who is buying growth stocks than just "morons" and "dumb committees."

Source: Journal of Political Economy

A little example. Here is a revealing table from Owen Lamont and Dick Thaler's classic 3 com and Palm paper. In the 1990s internet boom, 3com carved out 5% of Palm's shares, keeping 95% which it spun off -- gave to 3com shareholders -- 6 months later. Palm share prices exploded, to the point that you could buy Palm much cheaper by buying 3com, waiting for the spinoff, and throwing away the 3com shares than directly -- a pure arbitrage opportunity. (My response to this fascinating episode, not the point here.)

Notice the huge short interest in Palm, the overpriced security. At the peak, 147% of Palm's shares have been sold short. OK, that's what you expect. But notice that 2.6 percent of 3com's shares are sold short too!

Now, who in the world was short  the long end of the greatest arbitrage opportunity of the century? Well, AQR (Cliff and John's fund), I bet. This was the tech boom, and any long-value short-growth strategy will think 3Com is also "overpriced.'' So long as the correlation matrix estimate does not notice that 3com and Palm are perfectly correated at a 6 month horizon -- they are not in daily and weekly data -- it will short 3com.

So, it's quite sensible that people who invest for other purposes end up looking kind of silly when you condition down to one predictor.  People who buy growth stocks and shun value stocks are looking for something else.

Not an answer, but a very interesting set of questions. And the rest of the essay is good too.

In the end, I find the whole "rational" vs. "behavioral" debate quite empty. Anything people have been fighting about for 40 years really can't make much sense, and debating whether whole classes of models are right or wrong is pretty empty. Gene's joint hypothesis theorem proved that 45 years ago. All methodological debates are pretty empty. We make progress by writing specific models and looking at data.

Behavioral finance is however excellent marketing for active managers -- each of whom tells you all the other guys are behavioral.