Quis custodiet? Put the regulators on trial if they screw up

The Treasury’s amendments to the Banking Reform Bill mean that senior bankers could face up to seven years in jail for ‘reckless misconduct’ which leads to the collapse of a bank. Certainly, the behaviour of prominent individuals in the run up to the crisis left much to be desired. If only we could have put a few of them on show trial in 2009 and given them 20 years in jail, regardless of their objective guilt! But this option was never available, we live under the rule of law.

We need evidence to establish the case for conviction. How are we to judge what constitutes ‘reckless’ behaviour? We can hardly apply the elephant test. Describing an elephant may be difficult, but everyone knows one when they see one. But the concept of ‘reckless’ can be seen from different perspectives.

From the pensioned security of their bureaucracies, regulators exhibit a deep-seated tendency to judge outcomes with the benefit of hindsight. If something went wrong, someone must be responsible. The idea that in many ways the future is inherently unknowable rarely crosses the bureaucrat’s mind. We have seen this in the case of Britannia, where the then-FSA vetted its loan book prior to its absorption by the Co-op in 2009. But we are now told, after the event, that it is these very loans which are at the root of the Co-op Bank’s problems.

Even more importantly, the regulatory bodies themselves bear a heavy responsibility for the crisis. The dominant methodology of assessing risk on assets, Value at Risk, assumed that the statistical distribution of returns followed the bell-shaped curve of the so-called normal distribution. Very large changes are exceptionally rare, indeed beyond a certain size would be unlikely to be observed during the entire lifetime of the universe. Various bits and pieces were crafted on, but this basic assumption remained at the heart of the regulatory approach to risk assessment.

From a scientific perspective, physicists had established by 2000, beyond any conceivable doubt, that this assumption was not warranted. Asset price changes do follow approximately the normal distribution, but they have ‘fat tails’. This means that whilst extreme events remain rare, they are very much more frequent than would be observed if the normal distribution described all the data. The polymath Benoit Mandelbrot established the result decades ago, but by the early 2000s there was a flood of scientific papers demonstrating fat tails in asset markets of all conceivable kinds.

Why did the regulators not insist that banks used these findings in their Value at Risk models? If the current legislative proposals had been in place at the time of the crisis, who would have gone on trial, the bankers or the regulators?  Regulation, too, can be reckless. It can succumb to the fallacy that everything is knowable, especially with hindsight, and through sheer inertia can ignore advances in knowledge. More amendments should be tabled, to enable the regulators themselves to be dealt with if they are shown to have been ‘reckless’.

Paul Ormerod

As published in City Am on Wednesday 9th October 2013

Join our newsletter and get 20% discount
Promotion nulla vitae elit libero a pharetra augue