Algo crackdown is an awful idea

HIGH frequency traders are getting a bad press at the moment. They have been widely blamed for the so-called “flash crash” on 6 May, and the latest noises coming out of the European Commission suggest that when its review of the Markets in Financial Instruments Directive (MiFID) is finalised next year it may well look at curbing the activities of these traders, who trade using computer algorithms, or algos, largely without human intervention. The EC would be wrong to do so.

High frequency trading is a big deal. Statistics released by the New York Stock Exchange show that in one week in late May this year up to 60 per cent of trades on the exchange were down to high frequency and algorithmic traders.

Many in the markets have been calling to curb the influence of these traders by slowing down speeds, limiting the numbers of orders and restricting co-location (the practice of locating a trader’s computers alongside exchange servers) as well as increasing the use of circuit breakers such as automatic pauses and introducing limit up/limit down technology.

It is fashionable to look at ways of curbing the high-frequency traders’ activities, but in fact they are simply a natural progression from the move to screen-based electronic trading in the mid 1980s. They steal a head start on their competitors by using technology and leveraging off the inefficiencies inherent in European markets. When Luddism and a sneaking feeling that there is something somehow wrong in algorithmic trading is left to one side, it’s hard to see what is wrong with them.

Algos are not even as bad as their opponents say. Apocryphal stories immediately after the US flash crash blaming it all on the algorithmic and high frequency traders for “turning off their algos” at the crucial time and causing or exacerbating the free-fall do not appear to have been true.

In fact, the CFTC/SEC joint report into the flash crash, which was issued on 30 September 2010, not only points to a large order placed by a broker via an algorithmic programme as a probable tipping point in the crash, it also says that early sell pressure was absorbed by algorithmic and high frequency traders. The evidence suggests that high-frequency trading adds significant liquidity to the markets and is therefore beneficial.

And anyway, the EC is in no position to grumble about these traders. Xavier Rolet, CEO of the LSE, has observed in numerous speeches that algorithmic and high frequency traders were created by the regulators. MiFID swept away the monopolies previously enjoyed by incumbent exchanges leading to today’s proliferation of alternative trading venues which in turn spurred a technological arms race and price war between competing venues. This opened the doors to algorithmic and high frequency trading strategies while (largely) de-regulating proprietary trading.

The result of increased competition and the changes to the execution landscape has doubtless surprised both legislators and regulators. But they might spend their time more profitably worrying about how to maintain liquidity and improve price discovery in the European markets, which have been damaged by its fragmentation, and exacerbated by the decision not to introduce a US style consolidated tape that shows the best price across all trading venues.

Some of the ideas around high frequency trading are good – circuit-breakers, for example, would prevent automated trades from running away with themselves. But a wide-ranging clamp-down would be both hypocritical and anticompetitive.

No doubt the creators of MiFID are surprised by some of the consequences of the market’s deregulation, but that is the nature of a free market. Any moves to curb high frequency trading in the review would be a backward step.

Vincent Mercer and William Garner are financial services partners at Speechly Bircham