Join our community of smart investors
OPINION

Half-efficient markets

Half-efficient markets
January 29, 2015
Half-efficient markets

There are two answers to this (actually more, but I'll keep it short).

One is that volatility isn't proof of market inefficiency. As Cardiff University Business School's David Meenagh has shown, such price swings might be due to reasonable investors making small but reasonable changes to the probabilities they attach to different possible futures. For example, in October and December they thought there was a chance of a sharp global economic slowdown, but they have since reduced the probability they attach to this scenario. This isn't necessarily irrational.

Let's however, put this aside. Let's suppose that the aggregate market is indeed too volatile.

You don't need to believe that individual investors are irrational to believe this. As Stanford University's Mordecai Kurz has shown, prices can fall because rational investors worry that others will worry, and they can rise because they worry less that others will worry. (Or they can fall because some worry that others will worry that others will worry - and so on). Maynard Keynes famously likened stock market investing to a newspaper beauty contest in which people tried to guess what average opinion would consider to be the prettiest face. Pierre Monnin, an economist at the Council on Economic Policies in Zurich has shown that such beauty contest-type guessing about others' opinions can generate the volatility we see in share prices. And Makoto Nirei at Hitotsubashi University in Tokyo shows that it can also explain why crashes are more common than a normal distribution of returns would predict.

Let's concede all this. Does it follow that markets are so inefficient that my preference for tracker funds is wrong?

Not at all. As the late Nobel laureate Paul Samuelson said, stock markets are "macro inefficient" but "micro efficient". The aggregate market can be mispriced even though individual shares are fairly priced relative to each other.

To see how this can happen, imagine you think BP is underpriced relative to other oil stocks. You have a simple way of exploiting this. You can buy BP and short-sell other oil stocks. This protects you from the danger that sentiment towards oil stocks generally will deteriorate and so make BP even cheaper.

But what if you think global stock markets generally are under-priced? You can't so easily protect yourself from the possibility that sentiment will deteriorate even further. All you can do is buy and hope. And this is dangerous: as Keynes famously said, "markets can stay irrational longer than you can stay solvent." Sensible investors might reasonably not want to take this risk. For this reason, Insead's Bernard Dumas has shown that there is little that investors can do about excessively volatile markets.

This means that whereas mispricings of individual stocks will often quickly be corrected, mispricings of aggregate markets won't be. So the market will indeed be micro efficient but macro inefficient.

Back in 2002 Yale University's Robert Shiller provided evidence for this. He showed that, for individual shares, dividend yields predicted dividend growth; "cheap" shares (those on high yields) were often cheap for a good reason - they had poor growth prospects. However, for the aggregate market, this was not the case. The dividend yield did not predict dividend growth. Instead, a high yield predicted rising prices and a low yield falling prices. That's consistent with the aggregate market rising and falling too far.

This is perhaps the best example of Simpson's paradox; what's true of the whole is not necessarily true of the parts.

Now, this does not mean that the efficient market hypothesis is wholly correct for individual shares. The tendency for defensives, momentum, and quality stocks to out-perform suggests it might not be. However, this evidence against efficient markets is wholly different from the (alleged) fact of excess volatility. And you might reasonably think that the hassle and risk of investing in these anomalies is so great that tracker funds are a reasonable investment. Ideas don't have to be wholly true to be useful.

My point here is simple. There are few useful universal truths in the social sciences. The efficient market hypothesis might (only might) be wrong for the aggregate market. But it doesn't follow that it is grossly wrong for individual stocks. There is, therefore, no inconsistency between believing that aggregate markets over-react and believing that tracker funds are a reasonable way of investing in shares.