One time a very smart guy told me that the only stock he was really willing to make big bets on was Apple Inc. When I asked him why, he chuckled and said “It goes up!” That was in September 2012. The idea that a stock “goes up,” like some law of nature, is an example of a very important behavioral bias: trend-following.
Trend-following, which has always been a puzzling phenomenon, has received some attention in the finance literature lately. Robin Greenwood and Andrei Shleifer looked at a large number of surveys of investors, and found that people’s expectations of future returns are often just extrapolations of the recent past. Of course, those expectations usually turn out to be wrong, but people apparently believe them enough to bet on them -- when stocks go up and expectations soar along with them, money usually flows into equity mutual funds, only to lose out when stocks return to the long-term average, rather than the recent average.
Now, this might be an illusion -- people’s answers to surveys about their expectations might actually just reflect changes in their attitude toward risk, as suggested by University of Chicago economist John Cochrane (disclosure: I will do an experiment to test this hypothesis next month). But whatever the reason, the trend-following behavior is losing investors a lot of money, as YiLi Chien of the Federal Reserve Bank of St. Louis quantified in a recent study.
Trend-following is related to a phenomenon that psychologists have called the hot-hand fallacy, after the supposedly mistaken notion that basketball players have streaks of good shooting. There’s just one problem with this theory: As it turns out, hot hands probably do exist! The “fallacy” isn’t always a fallacy.
Sometimes the underlying structure of the world changes rapidly, and sometimes it changes slowly. If we assume it changes rapidly, then we think that every run of good returns signals an underlying change for the better, and every bad run means that something fundamental is going wrong. But it seems like when it comes to finance, that sort of assumption is more often wrong than right -- with disastrous consequences for trend-following investors.
The most notorious example of inappropriate trend-following is something I call the put-option illusion. Suppose there’s some asset whose payoff is similar to a short position in a deep out-of-the-money put option on some aggregate variable with a negatively skewed distribution (for example, the Standard & Poor's 500 Index, or the U.S. economy). Most of the time, this asset will make money and the put option, which is a bet on a fall in prices, will expire without being exercised. A trend-following investor will look at this asset and think “Wow, it goes up!” Risk will look low as well, if you look only at the recent data -- as trend-followers tend to do. The problem is, the risk is hidden in the left tail of the bell curve, and once in a while, the asset will crash hard.
I can immediately think of three important examples of such assets. The first is housing-backed financial products. As everyone now knows, before 2007 these products were priced based only on data from their extremely recent lifetimes; in fact, they turned out mostly to just be short positions in deep out-of-the-money puts on the U.S. housing market.
The second example is hedge funds in the 1990s and early 2000s. Vikas Agarwal and Narayan Naik found that most of these funds had payoffs similar to short positions in out-of-the-money puts on the equity-market index.
The third example is from Japan, and involves firm-specific human capital. In Japan, most full-time workers are hired and paid based not on their general skills, but on their knowledge of the inner workings of a specific company. If you ask Japanese people why they are content not to own their own human capital, they almost always answer that companies give them a sense of safety. The companies promise lifetime employment. But in severe economic downturns, companies are forced to lay off even these so-called lifetime workers, leaving the laid-off workers with very little human capital at all. Their company-specific human capital is actually just a deep out-of-the-money put option on the Japanese economy. It seems to me that Japan’s legendary risk aversion is really mostly just an example of put-option illusion and trend-following bias.
To contact the author on this story:
Noah Smith at email@example.com