Christoph Merkle at the University of Mannheim asked clients of Barclays how risky they thought their equity portfolios were. Most said they believed them to be no riskier than the general market. However, when he looked at their portfolios, he found them to be riskier than this. Most investors, he says, "grossly underestimate" the volatility of their portfolios.
One reason for this is simple overconfidence. People believe they are better at managing risk than they really are, just as we overestimate our abilities in other fields.
Another reason, say Professor Merkle, is that investors confuse beta and volatility and believe, wrongly, that a low beta means low volatility. Let's take an example. British American Tobacco (BAT) has a lowish beta - of 0.84 in weekly data in the last five years. Does this mean it is safer than the market? No. During this time its annualised volatility has been 18.5 per cent compared with the market's 15 per cent. Even BAT, one of the safest stocks on the market, has idiosyncratic risk. It's this, says Professor Merkle, that investors underestimate.
Now, this might be because they are using a different conception of risk. Financial economists think of risk as the probability of a price fall. But some investors think of it instead as a feeling of discomfort. Someone who works in, say, the building industry might feel that construction stocks are familiar and so less risky whereas he might regard mining stocks as unfamiliar and hence dangerous. Numerically, the two are similar. But they differ in their ambiguity - and investors want to reduce the latter.
Does this underestimation of idiosyncratic volatility matter?
Textbook theory says it does. It says that investors only get good returns as a reward for taking systematic, or unavoidable risks (such as the risk of a general market or economic downturn) but they don't get rewarded for taking on stocks' idiosyncratic risk.
But this isn't wholly true. We know that defensive stocks do better than they should - which means investors get paid to avoid market risk. If your portfolio has a low beta but idiosyncratic risk you might well therefore beat the market on average. A portfolio comprising only BAT is an example of just this; in the last five years it has risen twice as much as the market.
Underestimating idiosyncratic risk is, therefore, a mistake. But if you are holding defensive stocks, it is far from being the worst mistake you can make.
