First person: Expect the unexpected

As a weakening euro and volatility in China have shown, market practitioners need to rediscover the difference between risk and uncertainty, says the Review columnist Anthony Hilton

anthony_hilton_1920 copy
The Swiss franc was pegged to the euro in 2011. For three years this was uneventful, but in January this year the move came under attack. Waves of buyers seeking a safe haven from a weakening euro forced the Swiss authorities to abandon the peg and set the franc free to soar through its ceiling.

This led to eye-watering losses among foreign currency dealers and others who had failed to see the possibility of this happening and who were short of Swiss francs. Some spread-betting companies and forex trading firms that encouraged retail investors to borrow money to bet against currency movements were wiped out – as were their customers. The City is still trying to clear up the mess.

The blameless gameInterestingly, no one seems to think it is their fault. Even Goldman Sachs has a ready excuse. Its Chief Financial Officer, Harvey Schwartz, described the movement that followed the franc’s unpegging as a 20-plus standard deviation occurrence. In effect, what he was saying was that it was the type of thing that its risk-control models predicted could happen only once every billion or so years, and that no reasonable person would factor this possibility into their calculations. No one was to blame.

Goldman has form in this area. Testifying before a Senate Committee in the aftermath of the 2008 financial crash, Schwartz’s predecessor David Viniar similarly described the events of that time as “25 standard deviation events, several days in a row”. He explained that not only did a once-in-a-billion-years event happen; it happened again the following day and indeed several other days thereafter. Again, his defence was that this kind of movement was so unusual it could not possibly be foreseen, and no one could possibly run a business on the basis that it was likely to happen.

Missing a link?Now, there are some out there who might indeed believe the unpegging of the Swiss franc is something that is likely to have happened only once since the dawn of creation. And there are others who might subscribe to an alternative explanation: perhaps the information in the model was incomplete and so it gave the wrong signal.

Most such models are rooted in financial markets theory that assumes that prices revert to the mean over time. This means that they assume a normal or standard distribution – commonly known as a bell curve – in assessing the probability of an event happening. In a normal distribution, deviations of one or two from the mean do not count as extreme events, but when you get to five, let alone 20 standard deviations, you are way off the scale. However, as the Bank of England’s Chief Economist Andrew Haldane has pointed out, there is precious little evidence that the financial markets theory is correct.

Market practitioners need to rediscover the difference between risk and uncertainty. Risk arises when price movements in the future can be calculated or are known. Uncertainty is when movements cannot be calculated or are unknown. Normal distribution models work on the basis that everything meaningful can be measured and the likely range of price movements can be predicted.

From one extreme to the otherIn this world, uncertainty does not exist, and that allows firms to adopt trading strategies that are the equivalent of running on to a motorway to pick up pennies. Such activity makes sense if your model tells you the chance of being hit by a car is one in a billion. Given what the world has experienced these last ten years – most recently the economic slowdowns and ensuing stock crashes seen last month in China and now in Brazil – it is obvious that extreme events happen a lot more often than normal distribution models assume.

Look at the situation in China, where the Shanghai Stock Exchange Composite enjoyed a record high only weeks before it got hammered by a 15% decline.

This is another example of where economic and financial systems have behaved more like chaotic weather patterns than anything predictable. It would make sense to develop financial models that recognise this fact.

That, however, is easier said than done. Models that reflect the world as it is, rather than as we would like it to be, would predict far more extreme events. This would pull the rug out from under a lot of trading strategies and, in turn, probably usher in demands from the authorities for much higher capital requirements for those who do trade. So, the chances are we will continue as we are. That’s why they say ignorance is bliss.

Anthony Hilton is the award-winning former City Editor of The Times and the London Evening Standard.

The original version of this article was published in the September 2015 print edition of the Review.
Published: 12 Oct 2015
  • Wealth Management
  • The Review
  • People
  • Features
  • economy
  • Risk & Decision Making
  • Risk
  • First Person

No Comments

Sign in to leave a comment

Leave a comment