The foreign exchange markets’ push towards standardized timestamps—and greater transaction cost analysis (TCA)—has stalled. Since February 2020, when the Investment Association published a best execution guide for timestamps, the industry has made no concerted effort to standardize more granular timestamps.

And while regulations may require venues to record trades to the nearest millisecond or better—Europe’s second Markets in Financial Instruments Directive (Mifid II) among them—some order management systems (OMS) are known even to round timestamps up to the nearest second—undermining such demands. 

So, market participants are unsure whether the FX timestamps they receive are precise—making it hard for them to determine that they’re getting best execution when trading with their liquidity providers.

“How can any organization be expected to make critical decisions with such limited information?” asks Medan Gabby, chief revenue officer at Quod Financial. “That’s why granular timestamping is so important—it’s not possible to match actions to events without microsecond granularity.”

Market participants point the finger of blame at the pandemic and the reallocation of resources for the lack of progress toward more granular timestamp practice this past year—but whether this is the only factor is debatable.

FX market participants seek a counterparty’s timestamp to identify when an order is received, executed, or rejected, so that it can make a thorough TCA on a trade. The more granular the timestamp data, the better the TCA—and the better to meet regulatory requirements around best execution. 

“Being able to granular-timestamp usually requires greater tech resources or your own specialized tech team, so that’s a limitation to solving this issue,” says Stephan von Massenbach, director at Modular FX. “Some of that specialist knowledge and understanding just isn’t available within people’s own organizations—especially if they’re competing for resources or there just isn’t enough bandwidth. People’s focus is elsewhere,” he adds.

“Timestamping is not high up on people’s priority list,” says Quod’s Gabby. “It’s a difficult problem to fix. You can’t just add a couple of extra zeros to the end of your timestamp in your coding and now it’s microsecond[-granular], this is a core engineering issue and can be very expensive to change.”

It’s not possible to match actions to events without microsecond granularity

Medan Gabby, Quod Financial

Ideal Prediction CEO John Crouch says that it’s fairly easy to ensure a new OMS timestamps trades with sufficient granularity—market participants can build such a capability into their OMS from scratch, he points out—adding that it’s legacy OMS systems that are hard to solve. Upgrading an old OMS to timestamp in microseconds requires an expensive replacement of hardware and software—but it can also cause risk and disruption to the overall system, says Crouch.

“A lot of people recognize that their system is old but, because it’s still able to settle trades and get their business done, upgrading it is just not really their priority—and the longer you wait, the more expensive that upgrade becomes.”

Timestamps become even more of an issue if a participant’s OMS clocks aren’t synchronized to that of their trading venue—preventing the corresponding timestamps from being meaningfully compared—thereby further hindering cost analysis.

And while Mifid II requires trading venue operators and participants to synchronize clocks used for reportable events, it’s not uncommon for clocks to naturally experience clock drift over time. So, even if a firm’s internal OMS system can micro- or nanosecond-timestamp, its internal clock could be out of synch to its venue’s—again undermining accurate TCA. 

Hurdles and time trials

“I think increasingly this is about the overall IT architecture and reviewing the set-up of financial and capital markets technology,” says Modular FX’s von Massenbach, who sees value in tightening up practice on timestamps.

But not everyone sees the lack of a granular timestamp in FX as a major problem. Tradefeedr co-founder Alexei Jiltsov says that while a firm’s OMS might not timestamp with the desired granularity, it can still approximate its worst-case scenario TCA on a given trade by using market data to compare prices in the market to the price they actually paid.

“I don’t think TCA is a complex problem to solve if you have good technology, reliable price sources, and record your prices correctly,” he says. “In a nutshell, you just compare your execution price to your benchmark price—which is preferably something that is provided by a third party—and average this across relevant execution scenarios.”

I don’t think TCA is a complex problem to solve if you have good technology, reliable price sources and record your prices correctly

Alexei Jiltsov, Tradefeedr

But Andrew Woolmer, chief executive of New Change FX, disagrees. He says that granular timestamping data is a necessity for conducting proper TCA analysis on liquidity providers—especially in a market environment where spreads are so thin that only truly granular data can show you the difference in execution between LPs.

To improve this decision-making process, the smart order router needs smarter data, he says. “The more granular data used to benchmark liquidity providers becomes, systems can perform better analytics and in turn systematically prioritize the LPs that should receive their flow,” he says.

“The goal of all this analysis is to make sure that people are systematically paying the market a fair amount of money for the services they’re getting and aren’t leaving cash on the table—so better timestamps are important.”

In Jiltsov’s view, the onus for resolving such issues falls on the buy side. “OMS systems might not provide timestamps in as granular detail as buy-side traders would like, but it’s up to [them] to know how the OMS system works and not just blindly use it,” he says. “They can program it and impose more granular rules within it. It’s down to them to solve this.

“I’m not saying that a lack of granular timestamps is OK,” Jiltsov emphasizes. “If your internal OMS doesn’t support granular timestamps, then it’s probably time to change your system, but … you can circumvent a lot of issues by approximating the worst-case scenario. This analysis is not ideal but it can be a starting point for a technology upgrade.”

Whether such upgrades become a priority—or hasten progress toward industry-wide best practice—time will tell.