A team of economists at MIT and Harvard University gave subjects a choice: to invest in a safe asset at 5 per cent interest rate, or in a risky asset with an expected return of 10 per cent and standard deviation of 18 percentage points. On average, subjects put 66.8 per cent of their money into the risky asset. Then they offered a different choice: between a safe asset paying 1 per cent, and a risky one offering a 6 per cent expected return with a standard deviation of 18 percentage points. The weighting in the risky asset then rose to 75.6 per cent, a jump of 8.8 percentage points.
From the point of view of standard economics, this makes no sense. The risk premium didn't change: it was five percentage points in both cases. Nor did risk. Conventional portfolio theory thus says asset allocation shouldn't change. But it did. And the people who changed it weren't fools from off the street: they were MBA students at Harvard Business School, one of the most prestigious courses in the world.
What explains this strange behaviour?
One thing is that people have reference points - targets for returns that they feel comfortable with. If safe returns fall below such reference points (which are set by historical experience) savers will take on more risk in the hope of achieving what feels like a normal return.
In this sense, reaching for yield is the same sort of behaviour we see when losing gamblers bet on outsiders in the last race of the day or when shareholders cling on to poorly-performing shares in the hope they will bounce back. In all cases, people have lost relative to some target level of wealth, and they take risks in the hope of hitting that target.
Whether this is rational or not is moot. On the one hand, it is: reference points are a matter of taste and de gustibus non est disputandum. On the other hand, though, it's irrational to take a bet you'd otherwise reject if the odds don't change.
There's something else going on. It's suggested by the Weber-Fechner law, which tells us that people respond to proportionate differences, not absolute ones.
Although the absolute gap in expected returns is the same five percentage points in both choices, the proportions are very different; in one choice the risky assets has an expected return of twice the same one, but in the other it offers six times as much.
It might seem absurd that this matters. But just remember that speculation about a quarter point cut in Bank rate has had far more attention in recent weeks than it would have done back in the 1990s when Bank Rate was over 5 per cent. This suggests that proportionate changes get more attention than absolute ones. Some people respond to this.
One implication of all this is that equities might well become overpriced as a result of cuts in Bank rate, insofar as investors buy them even though the fundamentals, such as the outlook for profits, haven't changed.
But there's something else. These experiments show that even intelligent people don't behave exactly as economic theory says they should - a fact that is of wider relevance.
