Why do share prices sometimes overshoot, becoming too high or too low? Why is there sometimes momentum in markets, so that buying after prices rise through a 10-month average can lead to good risk-adjusted returns?
Some recent experiments by Pieran Jiao at Nuffield College Oxford provide a novel explanation for such events. He gave some subjects a long position in shares and others a short position, and then showed them the share price moves, so that some saw profits and others losses. He then asked subjects to predict subsequent price moves. He found that those with long positions who had made a profit predicted higher future prices than others. Conversely, those with short positions who had made profits predicted further price declines.
This is not just a laboratory finding. A study of Dutch investors by Arvid Hoffman and Thomas Post has found a similar thing; investors who have enjoyed profits expect further profits, while experience of losses leads to diminished expectations.
What's going on here is what Dr Jiao calls pay-off-based belief distortion: experience of good returns leads us to expect more good returns, and bad experiences lead us to expect worse ones.
This is not the same as wishful thinking, important as that is. In Dr Jiao's experiments, people who had made a profit on long positions expected further price rises even after their positions had been closed. Nor is it the same as overconfidence - our tendency to overestimate our abilities because we've made money. In Dr Jiao's experiments, positions were allocated to subjects at random and so their profits were not due at all to any skill. Instead, it's a separate mechanism (the social sciences are to a large extent an inventory of mechanisms).
Pay-off-based belief distortion helps explain why bubbles sometimes happen; having made profits, we expect more and so buy even though valuation measures might be telling us to sell.
>Pay-off-based belief distortion says we expect high returns after making a profit, but many investors are prone to the disposition effect - they sell winners too soon”
It also helps explain other behaviour. Ulrike Malmendier and colleagues have shown how experiences in our formative years distort our decisions throughout our lives. For example, people who lived through recessions when they were young are less likely to own shares even decades later, and managers who grew up during the great depression of the 1930s were less likely to borrow than others.
People who have direct experience of bad times are affected by them more than those who only read about them. "Experience is often overweighted compared with descriptive and observational information," says Dr Jiao. This echoes David Hume's famous distinction between impressions (our "more lively" sensations induced by experience) and ideas, which are "less lively".
It might also help explain the most catastrophic economic decision of recent times - RBS's takeover of ABN Amro. Fred Goodwin's successful purchase of other banks led him to overweight the expected gains from buying ABN and to underweight wider evidence at the time that bank mergers were often bad for the bidding company.
It's also consistent with why investors stick with high-charging but underperforming actively managed funds. A profit on these is hard experience of a pay-off, a strong impression, whereas the opportunity cost (the fact that tracker funds did better and charged less) isn't a direct pay-off but only a weak idea. This causes investors to overweight the merits of active funds.
You might think there's a puzzle here. Pay-off-based belief distortion says we expect high returns after making a profit, but many investors are prone to the disposition effect - they sell winners too soon. Isn't this a contradiction?
No. Selling winners is by definition always profitable. So it's tempting to keep doing it. The tendency for the stocks you sell to keep rising is, however, an opportunity cost that makes less of an impact upon us than hard cash profits. We therefore don't respond sufficiently to it. There's a difference between a realised profit and a mere paper one.
None of this is to say that pay-off-based belief distortion is always an expensive mistake; the fact that momentum investing works tells us that it's often right to expect past good returns to lead to future ones. And sometimes it is cancelled out by another error - the gamblers' fallacy, the idea that good times must lead to bad (I suspect that a lot of what looks like rational behaviour is in fact errors offsetting each other).
What it does mean, though, is that we have yet another cognitive bias to guard against. Next time you're feeling bullish about a share, just ask: do I have objective evidence for this belief, or am I instead being unduly influenced by my past profits?
