Cornering the Elephant in the Algo Room

While algo execution is undoubtedly on the rise in FX markets, as has been the case for the past few years, the pace of growth remains steady rather than the predicted spectacular. One of the reasons for this is a basic tenet of execution risk – market impact thanks to signalling.

Another trend of recent times in FX markets is the rise of the dark pools, many of which started as peer-to-peer offerings, but that are now expanding their model to all market participants subject to certain rules. This has made these pools of liquidity more meaningful for algo execution teams – the right model means signalling risk is reduced, potentially eliminated.

The challenge for the industry is providing the right model that offers sufficient liquidity to attract the algos – especially those executing larger parent orders – whilst making it a fair and “safe” environment in which to operate for all parties. Solving this problem, in the words of Mathijs Peeters, head of distribution for Europe at Siege FX, “allows the market to corner the elephant in the algo room”.

The solution, as far as Peeters is concerned, involves providing a way to work part of the inactive balance of the algo amount in a safe environment without signalling or market impact. “This can now be done through our partner banks’ resting algos which have been tweaked such that they can float a large amount in the Siege MidPool whilst they execute the rest as per usual (in small clips) into the lit market,” he explains.

If the amount in the MidPool is matched it will be replenished from the inactive balance and if it is only partially matched the balance will be executed seamlessly in sequence as a “working” amount as per usual in the other venues accessed by the algo (including internalising pool of the algo bank used). All MidPool trades face central PB counterparties, therefore provide the necessary anonymity  and are matched at the New Change FX mid-price at the time the match happens.

“As there is no price formation in the pool there is no market data and thus there is no market impact as a result from trading in the pool. You either match at NCFX mid or there is no one there and you move on,” Peeters explains. “For the user of these algos there is no change apart from the upside that potentially a good part of the order is quietly matched in Siege, shortening the time frame the algo is exposed to the market,” suggests Peeters.

Peeters’ analogy of cornering the elephant in the room is apt, for this represents a step forward in reducing the market impact from algo execution, rather than solving it. Until liquidity in the dark pools can be raised to the levels required to satisfy the bulk of orders, there will still be a balance to be executing in environments that leak information. Thus, the elephant, to continue the analogy, is cornered, but could still break out and cause problems!

As a model, however, the dark pool concept is very likely to work in FX markets, possibly even better than in equities given the OTC nature of the business. The challenge now, for all the firms pushing to offer dark, or mid-market matching, is building liquidity levels further.

There is hope, naturally, thanks to the truism that liquidity begets liquidity on a platform, but it remains early days. Equally, these pools have to face the real challenge of what happens in a trending market, an environment in which pegged algos – the type most likely to operate in these pools – often under-perform.

Perhaps the biggest test of this model, indeed the measure by which its success or otherwise is judged, will be how much volume can it attract from what remains seen as the best way to reduce market impact – internalisation at the major dealers. Achieve that and the elephant disappears.