Foreign exchange platform puts the brakes on high-frequency traders

EBS has changed its rules to discourage algobots.

EBS, a major interbank trading platform in in the foreign exchange market, is considering imposing a major change in the way it runs its market in order to discourage high-frequency trading from taking place.

EBS currently runs on the principle of "first in, first out" trading, where trades are dealt with in the exact order they are made. That is the way most people expect the market to work – but it also gives an advantage to those who can get their trades in quickest. That leads to the arms race that high-frequency trading has seen in the last few years, where traders pay to place their servers close to the exchange, to whittle off those last few microseconds.

Instead of this model, EBS is considering bundling together incoming trades and dealing with them in a random order. That way, every trade that came in in (for example) the tenth of a second between 12:00:00.0 and 12:00:00.1 would be grouped together and dealt in a random order, removing the advantage that the trader who got in at 12:00:00.01 would normally have.

Speaking to the FT, the chief executive of EBS explained why the company has made the decision:

The first twenty years of algorithmic trading have added great transparency and led to the compression of spreads – all great things. But there is a line beyond which marginal speed and smaller trade sizes add no value and actually harm the markets. At some point we, the public markets across asset classes, crossed that line.

The ‘first in, first out’ model sounds fair and plausible, but in modern public markets it implies ‘winner takes all’.

The classic example of how high(er)-frequency trading can have positive effects comes from the fact that the desire to shave seconds off the response time to financial information is the reason why the undersea cables linking London to New York are so high quality. Without that motivation to profit, the cables might not have been laid for decades after, and certainly wouldn't be as fast as they are now. (In fact, the USD/GBP exchange rate is still known as "cable" now, after the first transatlantic cable laid in 1858).

But as the speed of trades has increased ever higher, the side-benefits are shrinking. The difference in liquidity between a market where a tenth of a second and a thousandth of a second matters is minuscule; even if spreads might be a tiny bit tighter than they otherwise would be, no normal trader is helped by that "improvement".

So EBS's speed limit is a welcome step. By dealing with trades in a semi-random order, it removes the incentive to spend millions on shaving off the smallest fractions of time. Ironically, the companies which will benefit most in the short term are the high-frequency traders themselves, who already have the technology to trade speedily, and now don't need to worry about investing more on ever-diminishing returns. But eventually, more and more traders will match that capability, until the market becomes a level playing field again.

The other reason why traders – even high-speed ones – ought to thank EBS is that if the exchanges get HFT under control, then there's one less reason for governments to step in. Discouraging high-frequency trades is one of the strongest reasons for introducing a financial transaction tax. That hits everyone, not just the speedy traders.

A new data-centre in Manhattan. Photograph: Getty Images

Alex Hern is a technology reporter for the Guardian. He was formerly staff writer at the New Statesman. You should follow Alex on Twitter.

Photo: Getty
Show Hide image

Who will win in Manchester Gorton?

Will Labour lose in Manchester Gorton?

The death of Gerald Kaufman will trigger a by-election in his Manchester Gorton seat, which has been Labour-held since 1935.

Coming so soon after the disappointing results in Copeland – where the seat was lost to the Tories – and Stoke – where the party lost vote share – some overly excitable commentators are talking up the possibility of an upset in the Manchester seat.

But Gorton is very different to Stoke-on-Trent and to Copeland. The Labour lead is 56 points, compared to 16.5 points in Stoke-on-Trent and 6.5 points in Copeland. (As I’ve written before and will doubtless write again, it’s much more instructive to talk about vote share rather than vote numbers in British elections. Most of the country tends to vote in the same way even if they vote at different volumes.)

That 47 per cent of the seat's residents come from a non-white background and that the Labour party holds every council seat in the constituency only adds to the party's strong position here. 

But that doesn’t mean that there is no interest to be had in the contest at all. That the seat voted heavily to remain in the European Union – around 65 per cent according to Chris Hanretty’s estimates – will provide a glimmer of hope to the Liberal Democrats that they can finish a strong second, as they did consistently from 1992 to 2010, before slumping to fifth in 2015.

How they do in second place will inform how jittery Labour MPs with smaller majorities and a history of Liberal Democrat activity are about Labour’s embrace of Brexit.

They also have a narrow chance of becoming competitive should Labour’s selection turn acrimonious. The seat has been in special measures since 2004, which means the selection will be run by the party’s national executive committee, though several local candidates are tipped to run, with Afzal Khan,  a local MEP, and Julie Reid, a local councillor, both expected to run for the vacant seats.

It’s highly unlikely but if the selection occurs in a way that irritates the local party or provokes serious local in-fighting, you can just about see how the Liberal Democrats give everyone a surprise. But it’s about as likely as the United States men landing on Mars any time soon – plausible, but far-fetched. 

Stephen Bush is special correspondent at the New Statesman. His daily briefing, Morning Call, provides a quick and essential guide to British politics.