Could Navinder Singh Sarao really rock the Dow Jones index on his own?

http://www.theguardian.com/commentisfree/2015/apr/25/navinder-sarao-dow-jones-stock-markets

Version 0 of 1.

At first sight, it looked like an April Fools’ story. The US Department of Justice is seeking to extradite a day-trader from Hounslow to stand trial on charges that he brought the US stock market briefly to its knees on 6 May 2010. Navinder Singh Sarao is accused of using a computerised share-trading program to manipulate the market for S&P 500 futures contracts on the Chicago Mercantile Exchange, thereby adding (so the prosecution alleges) to wider selling pressure that caused the Dow Jones industrial average to plunge briefly by 6% before bouncing back.

In that short interval, stocks in huge companies such as Procter & Gamble dropped by 25% and established companies such as General Electric and Accenture briefly traded as penny shares. The British courts, not to mention the rest of us, are invited to believe that this mayhem was caused by a 36-year-old geek in the bedroom of his parents neat semi-detached house under the Heathrow flight path.

There are, it seems to me, only two possible interpretations of this. One is that Mr Sarao is indeed responsible for the chaos. The other is that the US authorities have no real idea who is responsible, but need to make an example of somebody and Mr Sarao will do nicely. Either way, we are left with a really alarming conclusion, namely that we have constructed a world that is totally dependent on systems that are a) astonishingly fragile and unpredictable, and b) incomprehensible not only to the average citizen but to those who are supposed to regulate them.

Complexity, in other words, is the new reality we have to live with. It’s a word we often casually bandy about – I have even heard wine buffs talk about the “complexity” of a premier cru. But there is a science of complex systems and what it tells us about their properties runs counter to our intuitions, which is not surprising, given that those intuitions have been shaped by our experience with the relatively simple systems of the past. It’s not that long ago, for example, that if you visited the London Stock Exchange in Threadneedle Street you saw human beings trading stocks with one another. And at the end of the day clerks in offices organised the necessary paperwork to register the resulting sales and purchases. Now, a “stock exchange” is a virtual space in which algorithms exchange messages at lightspeed with no or minimal human involvement.

In a strange way, we haven’t entirely caught up with this new reality. Last Thursday, the Financial Times provided excellent and extensive coverage of the Sarao case. But even there the old analogue mindset could occasionally be glimpsed. “No one should suggest,” thundered the first leader, “that Mr Sarao was solely responsible for triggering the crash… The US investigation into the crash found that the trades most directly responsible probably amounted to several billion dollars’ worth of sales. Mr Sarao’s alleged phantom offers may have comprised a disproportionate chunk of that day’s implicit interest, but the whole point of the judicial complaint is that his orders were phantom. You cannot ‘spoof’ half a trillion dollars off the value of US equities. To accuse Mr Sarao of causing the crash is like blaming a flea on an elephant’s rump for a stampede.”

Well, maybe. Complexity scientists might beg to differ. They know that in complex systems even tiny perturbations can generate wholly disproportionate effects. Many years ago, the meteorologist Edward Lorenz coined the metaphor “the butterfly effect” to capture the idea that the behaviour of a hurricane could be influenced by “minor perturbations such as the flapping of the wings of a distant butterfly several weeks earlier”. It will be interesting to see if the American prosecutors have to rely on complexity science to establish that Mr Sarao was indeed the flea that triggered the stampede.

Overshadowing all this, though, is the scary reality that our societies have become dependent on complex systems that are fragile, unpredictable and poorly understood. Stock markets driven by high-speed algorithmic trading are one obvious example. The just-in-time logistics system that stocks our supermarkets is another: if, for some reason, trucks were off the road for a week, most shops would run out of everything. A third is the international system that provides the food we eat.

And then there is international banking. At a Centre for Science and Policy conference in Cambridge last week, there was an interesting discussion on whether we’ve learned the lessons of that particular complex system. The most profound answer came from Professor Barry Eichengreen of Berkeley. “Well,” he said, “I think we’ve done a good job of putting in place an early-warning system for the last crisis.”

It’s not just generals who are always preparing for the last war.