Danger: Computers at Work

This article appeared in The Spectator, issue of 14 October 2011.

In The Fear Index, the latest thriller by Robert Harris, now heading for the Christmas bestseller lists, a brainbox hedge fund manager with little in the way of interpersonal skills discovers that his computer-driven trading system has flown out of control and threatens to send the world’s stock markets into an unprecedented tailspin. Anyone familiar with Mary Shelley’s Dr Frankinstein, or any Bond film, will recognise the genre: an oddball genius consumed by his own creation, conflagration, nemesis, the whole works – populist fiction at its best.

But is it fiction? Not so fast, reader. As the author makes clear in a footnote near the end of his novel, the market meltdown which Dr Alex Hoffmann’s trading system appears to have prompted in The Fear Index is one that actually happened. The “flash crash” that afflicted the US markets on May 6 2010 sent the main US stock market indices tumbling by more than 9% in just fifteen minutes, causing short term panic. One blue chip company, Proctor & Gamble, saw its shares fall by 37% in a matter of minutes before they – and the market – eventually recovered.

What caused the flash crash remains a matter of controversy. As with the infamous stock market crash of October 1987, when the Dow Jones Industrial Average, the US stock market index, fell by more than 22% in a single day, it was not long before the flash crash was being blamed on computerised trading programmes, and not long also before conspiracy theorists started to take issue with the somewhat inconclusive findings of the official reviews into how such alarming episodes could have come about.

In 1987 computerised trading was still in its infancy. There seems little reason now to dispute the official verdict that a relatively new technique, known as portfolio insurance, was a powerful contributor to the market crash. Unsophisticated by modern standards, and untested in a major market move beforehand, portfolio insurance was sold to a number of large institutional investors as a computer-driven risk management system, designed, ironically, to protect their portfolios against large market moves. Using financial futures contracts to provide that insurance, it turned out in practice to have the unfortunate effect of creating an automatic, self-perpetuating spiral of sell orders as the market started to fall, thereby defeating its own objective. Market regulators subsequently introduced a system of automatic trading halts, known as circuit-breakers, in order to prevent a repeat.

In the case of the flash crash, a joint enquiry by the Securities and Exchange Commission and the Chicago Futures Trading Commission last year failed to identify a single cause for the market’s sudden heart attack. It did however draw public attention for the first time to the extraordinary rise of so-called high frequency traders, to whom it attributed at least partial responsibility for what had happened. High frequency traders are trading firms which employ networks of powerful computers to both generate and execute orders at lightning speed. They use complex algorithms based on the historic relationship between a wide range of different variables to identify minute pricing anomalies that may not persist, in some cases, for more than a few seconds at a time.

High frequency traders are an extreme example of a technique which is also employed in more conventional ways by a number of hedge funds. Like his hero’s firm in The Fear Index, firms such as Renaissance Technologies, Winton Capital and Man Group employ scores of highly trained scientists (rarely economists) to develop sophisticated algorithms that can form the basis of profitable trading strategies. The results can be spectacular. Winton Capital, for example, one of the most successful quant funds in the business, made a profit of £288 million in 2008 and £60m in 2009 after its trading strategies made good money in years while the rest of the world was losing its shirt. Winton’s founder, David Harding, recently gave £20m of his estimated £400 million fortune to fund research at the Cavendish Laboratory in Cambridge.

Not everyone thinks that the advent of these fast-moving computer-driven trading funds is good news. Last week the head of the body that regulates the futures market in Chicago called publicly for high frequency trading to be more closely regulated. In the UK the Department of Business, Innovation and Skills has set up a working party to study the implications.  According to the Government’s Chief Scientist (straying a little way, you might think, from his normal brief), high speed traders now account for a third of the daily volume of equity trading on the London stock market and as much as three quarters of the volume in the United States.

It flies in the face of common sense to argue that such a dramatic change in the trading environment can have taken place without having some impact on the way that markets behave. There is no doubt either that financial markets have experienced much more extreme volatility since the credit crisis in 2008. In the last three months, for example, daily moves in leading stock markets, have been way above the historical norm. Daily moves of 2%, 3% and 4% in both directions have become commonplace, against a long term average of just 0.8% a day. Okay, there is a Eurozone crisis going on, but surely computer-driven trading is at least partly to blame for destabilising the markets in this way?

Some bureaucrats, at least, seem inclined to agree. The International Organisation of Securities Commissions found last year that while algorithms and HFT technology are useful risk management tools, “their usage was also clearly a contributing factor in the flash crash event of May 6 2010”. In contrast, a working group of the Bank for International Settlements earlier this year dismissed high frequency trading as the trigger for the “flash crash”, but warned that such trading could add to the systemic risks of the financial system by “accelerating” and “propagating” shocks initiated elsewhere, just as happened with portfolio insurance.

Not so, say industry apologists. They argue that ever faster trading fulfils a useful function, by providing additional liquidity to the market and ensuring keener prices, just as financial theory says it should. As such it may make markets more, not less, stable. Academics commissioned by the UK Government’s Office for Science conceded (sorry, Vince) that research has so far found “no direct evidence that high frequency computer based trading has increased volatility”. However, they went on, “despite all its benefits, computer based trading may lead to a qualitatively different and more obviously nonlinear financial system in which crises and critical events are more likely to occur in the first place”. Robert Harris puts the same common sense point more clearly and more forcefully. It is the novelist’s prerogative to know what others can only guess at. You have been warned.