When do losses hurt us most? This question holds the key to understanding differences in average returns on equities.
For years, the conventional view has been that it is in recessions that losses are nastiest. This is partly because share price falls then come at a time when we can least afford them – when we face heightened risk of losing our jobs or businesses. It’s also because there’s a danger that the recession won’t be a mere cyclical downturn but a permanent loss of wealth, as the 2008 crisis was. And it’s also because in recessions we are bombarded with bad news, which makes us anxious and so less able psychologically to bear risk.
This implies that cyclical shares – those that do especially badly in recessions – should do better than others on average simply to compensate us for the risk of them losing us money when we are least equipped to bear such losses.
But is this true? The evidence is ambiguous. Some do. Housebuilders, for example, have delivered good long-term returns. So too, on average, do value stocks generally – if we define these very simply as those that offer higher than average dividend yields, which is often a reward for taking cyclical risk.
On the other hand, though, small-caps have actually slightly underperformed the FTSE 100 since 1990 despite being more cyclical: they underperformed in the recessions of 1990-92 and 2008-09. And another class of cyclicals have done badly. Harvard University’s John Campbell has shown that shares on the verge of bankruptcy tend to lose money on average.
Cyclical risk, therefore, doesn’t always pay off on average across all assets. This might be because it isn’t as nasty as conventional theory says it is. Christoph Merkle at Kuehne Logistics University has given us evidence of this. He surveyed UK investors before and after the 2008 crash. Before the crash, investors told him they’d feel very unhappy with big losses. When those losses materialised, however, investors were happier than they thought they’d be. This tells us that cyclical losses don’t hurt us as much as we might think.
It’s easy to see why this might be. In recessions, we can take comfort from the fact that everybody else is in the same boat. And we don’t need to blame ourselves for the losses, as they are the fault of other people.
Perhaps, therefore, it is other types of losses that hurt us more. Eric Falkenstein at Pine River Capital Management says what hurts us is falling behind others. This is certainly true for fund managers, who are judged by their relative performance. For them, a 10 per cent gain when everyone else has made 15 per cent is worse than a 10 per cent loss when everyone else has lost 15 per cent. But it might also be true for some retail investors. This isn’t just because we feel bad at falling behind others. I suspect it’s more because losing money when others are not makes us kick ourselves for making bad decisions.
From this perspective, cyclical risk doesn’t matter and so it’s no surprise that small-caps don’t carry a risk premium. And this helps explain what is otherwise puzzling – the good performance of defensive stocks. These expose us to the risk of falling behind others if the market rises strongly. Because some people want to avoid this danger, they steer clear of defensives with the result that such shares offer good returns to those investors willing to take the risk of underperformance.
On this view, the riskiest strategies are all of those that are regarded as unconventional because it is these that we’ll regret the most if we lose money on them. Maynard Keynes might have exaggerated when he said “it is better for reputation to fail conventionally than to succeed unconventionally” but it is certainly better to fail conventionally than to fail unconventionally. This might explain why some algorithmic strategies – most obviously momentum investing – pay off well while judgment-based stockpicking doesn’t. It’s simply because unconventional strategies are riskier and so should carry a risk premium.
What we have here is an ambiguity. On the one hand, it’s clear that cyclical risk does pay off in some assets, as any investor in housebuilding stocks in recent years will tell you – perhaps at great length. The conventional view, therefore, isn’t wholly wrong.
But nor is it wholly right. It predicts good average returns on small-caps and on distressed stocks, which we do not see. And it can’t explain good returns on defensives and momentum. Thinking of risk in terms of relative underperformance, however, can explain all these.
Perhaps the message here illustrates Norwegian academic Jon Elster’s view of the social sciences. We have very few if any simple general laws that explain everything. All we have are a set of mechanisms, some of which explain some facts and others of which explain others. And maybe this is all we’ll ever have.
