FX: XTX’s ‘zero hold time’ adds to debate over ‘last look’ practices

By:
Paul Golden
Published on:

The latest review of last look prompted by the FX global code has set off further debate over the use of holding windows and latency buffers.

Liquidity provider XTX Markets announced on August 10 it was operating a “zero hold time” model in its application of last look to its FX counterparty businesses.

matt clarke-160x186
Matt Clarke,
XTX Markets
This was part of its support for principle 17 of the global FX code of conduct – that market participants employing last look should be transparent regarding its use and provide appropriate disclosures to counterparties.

According to the company, the increased frequency with which market data is updated weakens the rationale for imposing a holding window since the liquidity provider could receive up to 20 primary market price updates before applying the price check.

This is not XTX’s first attempt to increase the transparency of last look. Earlier this year it introduced a tool it said would reveal the true cost of last look by presenting the buy side with data on the cost of rejected trades.

A lot has changed in the interbank primary market during the past 12 months as market data tick increments have fallen dramatically. This means the major liquidity providers know where the price is at a much more granular level and the last-look check can be done in a fraction of the time it would have taken just 12 or even six months ago.

Previously, a liquidity provider might have required 100 milliseconds to take a final look at the primary market – now that can be done in five milliseconds, explains Matt Clarke, a member of XTX’s distribution team for EMEA.

“In this context, it seems perverse to continue to hold a client trade for a 10th of a second when the price check can be done in a fraction of that time,” he says.

Tim Cartledge-160x186
Tim Cartledge,
NEX Markets

Tim Cartledge, global head of FX and head of product at NEX Markets, suggests that the importance of improved speed on primary markets makes last look less about hold time and more about speed of execution, and refers to a reduction of around 15% in average hold times on EBS Direct between January and July.

As clients see changes in pricing due to faster market data, transparency will improve and pricing should be more consistent across the board, he continues.

“However, as clients get more educated on the topic and increase their desire to control their execution costs and venues, increasing the use of primary markets should be more significant,” he says.

Eliminating the buffer

XTX believes that during the next year or so, the other major liquidity providers are likely to eliminate the hold window, although ParFX COO Roger Rutherford accepts that the market is only at the start of a change in how hold times and latency buffers are amended or removed.

Harpal Sandhu, founder & CEO of Integral, reckons the way last look is implemented has changed notably in the wake of the global FX code.

Harpal_Sandhu-160x186
Harpal Sandhu,
Integral

“Market makers who warehouse risk are now able to differentiate themselves from those without that capability,” he says. “By delivering virtually immediate response times and very high acceptance rates, they are able to significantly improve the quality of execution.”

Zero hold time is not necessarily the best way of increasing last-look transparency, but rather the best solution in absence of more definitive guidance on the future of last look.

That is the view of Henry Wilkes, founder of Institutional FX Advisory Partners, who suggests that if market participants want to have a credible platform product they need to decide whether to adopt a zero hold time policy or drop last look completely.

“There is no consistency in the approach to last look from market participants, which creates confusion and uncertainty as to whether buy-side clients are receiving fair and efficient pricing from their liquidity providers,” he says, referencing comments made last month by Simon Potter, executive vice-president of the markets group of the Federal Reserve Bank of New York.

Wilkes is also critical of the use of latency buffers on an FX platform given advances in the technology behind trading systems, saying it raises questions over the acceptable length of time for a buffer and whether it is used in accordance with principle 17 and purely used by the market maker to verify validity and/or price.

Implementing zero hold time is good for the market since liquidity providers can no longer be accused of inappropriate behaviour and it is compatible with the sentiment of the global FX code, concludes Clarke.

“Last look will never disappear completely because there will always be someone who has to do a credit check, for example,” he says. “However, the hold window can be eliminated.”