Tag: Financial crisis

Clive Cookson

Financial mathematicians are unlikely to be impressed by Lord Turner’s review of the banking crisis. Their discipline features quite prominently in the first part of the report, looking at what went wrong, but hardly at all in subsequent sections that prescribe solutions.

Lord Turner, chairman of the Financial Services Authority, blames “misplaced reliance on sophisticated maths” for misleading banks’ top management into a false sense of security about the risks they were taking. That is unfair. As Tim Johnson of Heriot-Watt University puts it, the problem was that banks did NOT use sophisticated maths; their mathematical models were far too simple. They had no incentive to employ more complex and realistic models because the simple “single factor” models used for pricing gave an illusion of accuracy and precision – and lulled the market into believing banks had everything under control.

Putting it another way, the models were built to fit market prices. The result was that whacky prices were reinforced by an overlay of scientific respectability.

Lord Turner should have said that we need more and better maths for the future. In particular we require better models of the interdependency of different financial variables and of the risk of extreme events. That means more funding of research in financial mathematics, by the public and private sector.

Another priority, not mentioned by Lord Turner, is to educate bankers about the mathematical basis of their industry. It should no longer be acceptable for them to use models as “black boxes” without making any attempt to understand the underlying assumptions.

The world of research

The science blog is no longer updated but it remains open as an archive.

Clive Cookson, the FT's science editor, picks out the research that everyone should know about, in fields from astronomy to zoology. He also discusses key policy issues, from R&D funding to science education. He'll cover the weird and wonderful, as well as the serious side of science.