Consider four recent events:
1. On 1 May, Capital and Regional's share price fell 13 per cent as investors feared it would have to raise more cash. Its price has since fallen a further 36 per cent.
2. When Bradford & Bingley announced its rights issue on 14 May, its price fell 9.1 per cent. Since then, it's lost another 54 per cent.
3. Yell issued a profit warning on 7 February, causing its price to fall 15 per cent. It has since fallen another 68 per cent.
4. Taylor Wimpey issued a gloomy trading statement on 17 April, and its share price lost 3.8 per cent on the day. Since then, it's lost another two-thirds.
These - and we could add HBOS, Royal Bank of Scotland or almost any housebuilder - are all examples or variants of post-earnings announcement drift. This is the tendency for share prices to continue falling after bad news - in contradiction to the efficient market hypothesis, which says news should be immediately embodied into share prices.
The queer thing is, post-earnings announcement drift is not a new thing. It was first identified in the academic literature in 1968. It's captured by Warren Buffett's famous saying, "there's never only one cockroach in the kitchen". And Norman Strong of the University of Manchester has found "significant evidence" for it in the UK. He calculates that during the 1990s the 10 per cent of stocks announcing the biggest good earnings surprises outperformed the 10 per cent with the nastiest surprises by an average of 10.8 per cent a year in the 12 months after the announcements. "Returns continue to drift upwards for stocks with extreme positive earnings surprises and they continue to drift downwards for stocks with extreme negative earnings surprises," he concluded.
This happens because of a widespread human tendency. It was perhaps best illustrated by Gordon Parsons, a Leicestershire seam bowler in the 1980s. In one match, he was hit for 24 in a single over. He explained later: "They were all good balls. The batsman just hit them."
Quite simply, people hate admitting they were wrong. In the jargon, we are Bayesian conservatives. We cling to our prior beliefs more than we should in the face of contrary evidence. So, for example, holders of housebuilding stocks, after seeing profit warnings, clung to their view that the shares were on low ratings and continued holding them. The result is that the shares initially stayed relatively high, and only gradually fell as even more bad news hit the sector, eventually forcing the optimists to give in.
Here, though, is the nasty thing. It's astonishingly hard to rid ourselves of this tendency. We have our prior beliefs for good reason - there's evidence for them. If we always abandoned them in the face of conflicting evidence, we would be committing the opposite error to conservatism - overreaction. And we would merely stumble from one belief to another*.
In particular, the efficient market hypothesis - which says that post-earnings announcement drift shouldn't exist - has a mountain of evidence in its favour.
And, of course, there are always exceptions to the tendency. For example, HSBC issued a profit warning in February, but has since risen more than 10 per cent. Such instances remind us that it isn't always a good idea to sell after bad news.
There's no easy message here - merely that it' s very hard to be rational, and even harder to be right.
*The psychologist Max Coltheart gives an extreme example of this error. A patient was suffering from the Capgras delusion - the belief that one's relations have been killed and replaced by imposters. One patient whose cognitive skills were otherwise unimpaired was asked: "Don't you think it incredibly unlikely that your family have been replaced by replicas?" "Oh yes" he replied. "I'd never have believed it until I saw it with my own eyes."
