1. Extreme returns are more likely than a normal distribution would predict.
Between 2003 and the Deepwater Horizon explosion on 20 April, the standard deviation of weekly changes in BP's share price was 3.39. That implies that the 24.9 per cent fall we've seen in the last five weeks is a 3.28 standard deviation event. If returns were normally distributed, this sort of fall should happen only one month in 160 years. In fact, this is the second such fall in the last two years; a similar drop occurred in October 2008, for very different reasons.
This tells us that even defensive stocks carry tail risk - big falls are more likely than you'd expect from a bell curve. A better description of equity risk is the cubic power law.
2. Shares are state-contingent securities.
Looked at in narrow financial terms, Deepwater Horizon is small beer for a firm of BP's size. Even if clean-up costs reach $3bn, as some claim, this would be less than three months of last year's profits. Instead, the danger is that among the mountain of law-suits BP faces in coming years some Judge John Taylor will clobber it with massive punitive damages, with the result that BP might have to break up. The problem for BP's share price, then, is that a disaster scenario has emerged which was previously not there. BP's price will depend upon the probability investors attach to this scenario. As this will change from day to day, BP might remain volatile.
3. There's a difference between risk and uncertainty.
Risk is a known unknown; it can be (roughly) quantified. Uncertainty is an unknown unknown. And because BP has changed from an oil business to a law-plus-oil business, the amount of uncertainty it carries has increased.
This matters, because the old cliche is true: investors hate uncertainty. The lesson of the Ellsberg paradox is that they hate it more than risk.
You might think this uncertainty means BP's price should rise, as investors need high returns to compensate for the uncertainty. Not necessarily. Economic theory tells us that what matters is not a stock's individual danger, but rather the contribution its danger makes to our overall portfolio. And even for a giant like BP, this is small because...
4. Diversification works.
Since 20 May, BP has dropped by more than 18 per cent. Although this alone has wiped almost two percentage points off the index, the FTSE 350 has risen during this time. So too has our benchmark value portfolio, which had a 5 per cent weight in BP. This reminds us that diversification can spread stock-specific risk well - even within a narrow style of investing. Investors needn't, therefore, lose money from exposure to individual stocks.
You might object here that I've been cheeky with my choice of time period. In the previous three weeks, diversification didn't work, as BP's fall was associated with falls in the broader market.
True, but this shows the point. The falls earlier in May were due to a market-wide danger, prompted by the eurozone's debt crisis. These sort of dangers cannot be diversified away among equities alone. But idiosyncratic stock risk - which is what BP's latest travails are - can be so diversified.
In this sense, conventional portfolio analysis is right: it's important to distinguish between market risk and stock-specific risk.
