Stop me if you’ve heard this one before: “Suppose that if you invest £70,000, I’ll offer you a bet on a coin toss. If it comes up heads, you win £14,000 pounds, if tails you lose £7,000.” Should you take the bet?
Early in 2016, the Bank of Japan (BoJ) asked this question of 25,000 people as part of a survey on financial literacy. Almost 80% said that they would not take the bet. For the BoJ this meant that 80% of people displayed “strong loss aversion.”
Loss aversion – the tendency to feel the pain of losses more than the pleasure of gains – is a well-documented behavioural “bias.” The question above is commonly used to measure this; in theory one should take the bet because the “expected return” is positive (5%), the implication being that to not take the bet is “irrational.”
Loss aversion has been used to partially explain the valuation gap between equities and bonds that exists today (as illustrated by the chart below). Despite apparently higher expected returns from equities, many investors prefer bonds because the possible pain of losses is far less (a phenomenon that is perhaps more evident in Japan than anywhere else).
How reasonable is this? Even aside from ambiguities in the BoJ’s question (is it my own money or house money? What is my starting wealth?) the concept of expected return is a challenging one for individuals. You are not going to toss the coin repeatedly; it is a one shot deal. In this sense the odds are: I have a one in three chance of feeling disappointed (I can play and lose, I can play and win, or I can not play at all).
There is something similar going on when investors turn away from equity markets. Yes, history suggests that you should win on average or over the long term, but we do not live in an “on average” world and the order of return delivery matters. Many will not have the time to wait to make their money back after a crash, while others suspect that there is a not insignificant risk of being wiped out entirely before you can experience the long term. To ignore this fact is to fall victim to Nassim Nicholas Taleb’s (and Bertrand Russell’s) turkey problem:
On average the Thanksgiving Turkey will be fed on a daily basis, but that is not much use coming up to Thanksgiving itself. For many, the fear is that equity investing displays the same properties: the system is “non-ergodic,” the path of returns matter, and when there are high levels of leverage in the system, “turkey risk” is likely to be greater.
It is therefore overly simplistic to describe many instances of loss aversion, including a preference of bonds over equities today, as irrational. As is often the case, human intuition can be more useful than cold rationality.
What is irrational is the tendency to fear equity crashes more after having just experienced a crash, and, less when returns have been strong. Equally irrational is to underplay the risks of bonds. As recent weeks have shown, holding bonds is not the same as “not playing.” The bear case of many areas of the government bond market are not predicated on the idea that the market is displaying “irrational” loss aversion, but the fact that prevailing yields offer very little compensation for those risks that are being taken on.