FX traders sound alarm on tagging ‘abuse’

Tradition has it that attendees at a Venetian masquerade ball wear elaborate masks to conceal their identity before indulging in acts of immorality and intrigue.

The FX market has its own cloak of anonymity in the form of tags, which traders on electronic platforms use to remain incognito. But some in the industry complain that participants are using tags to mask their own unscrupulous motives, such as to front-run rivals or to operate under multiple different guises.

Anonymous trading is a focus of the Global FX Committee’s three-year review of its code of conduct, scheduled for this year. But the review has been delayed for at least six months due to coronavirus disruption. In the meantime, confusion continues to reign over tagging practices: specifically, how platforms use tags, who is tagged, when tags are shown to counterparties, and whether tags are refreshed.

“To be honest, I’m not even sure which venues use tags,” says the head of e-trading at a Nordic bank. “I would assume it’s disclosed in their operating procedures, but I’m not sure which venues use tagging today.”

Widespread uncertainty provides the ideal cover for those seeking to exploit the rules. The “use and abuse” of tagging, as one expert describes it, threatens to undermine the integrity of electronic trading. So it’s no surprise that platforms, or electronic communication networks (ECNs), are keen to have their say in the debate.

“The identification by the Global FX Committee of best tagging practices is really important in order to ensure there are consistent standards in the market around how anonymous FX ECNs go about providing tags,” says David Newns, global head of Currenex, an e-trading platform operator. “Having the industry confirming a position on tags and the concerns around them would be useful.”

Electronic platforms capture the biggest slice of the $6 trillion-a-day FX markets, with a 60% share, according to the latest data from the Bank for International Settlements. Within that, 13% of trading is on multi-dealer platforms that are either fully disclosed or semi-anonymous, and 16% on fully anonymous platforms.

As the name suggests, fully disclosed venues typically identify counterparties before and after trading takes place. Fully anonymous venues do not identify counterparties – leaving settlement in the hands of the platform itself or a prime broker.

Where I think the use of tags has gone awry is when they’re shown before trading. That’s information which a market-maker can use and abuse

Head of FX at a liquidity provider

Semi-anonymous platforms are where tagging occurs. Instead of naming the counterparties to a trade, platforms will use unique alphanumeric codes – ‘tags’ – to identify counterparties before and/or after trading takes place, depending on which tagging model the platform operates.

The arrangement gives participants the benefits of a disclosed venue, as the use of a tag allows liquidity providers (LPs) and consumers to form a kind of relationship. It also functions similarly to a fully anonymous venue, as no-one knows exactly who their counterparty is, so trading activity is hidden behind closed doors.

Typically, semi-anonymous venues will retain data and analytics on the trading behaviour of specific tags, thus helping LPs to tailor the liquidity they provide to specific tags, or to remove them from their liquidity pools altogether. Similarly, liquidity consumers can use the same data to filter out any LPs they don’t want to see quotes from.

“Tags give us some guidance over client trade activity, the type of clients we’re trading with, and the type of flows we’re seeing,” says Stephane Malrait, global head of market structure and innovation at ING. “Therefore, we can decide whether we want to trade with those clients and flows.”

Front running

Debate centres on when tags are shown to counterparties. Some semi-anonymous platforms disclose counterparty tags before trading takes place, while others show tags after trading has commenced.

Proponents of the pre-trade tag model say it provides them with more information on who they’re dealing with before a trade, allowing them to adjust their prices or trading strategy with that counterparty in real time.

Critics counter that if your counterparty knows your tag before trading starts, they could use that information against you. For example, by analysing the typical trading behaviour and market impact associated with your tag, they could front-run your trade – effectively moving the market against you.

Research by Risk.net shows that two semi-anonymous platforms operate a pre-trade tagging model – EBS Select and Refinitiv’s FXall Order Book – while two platforms operate a post-trade tagging model – Cboe’s Hotspot and Currenex’s FXTrades. Two platforms – Euronext FX and 360TGTX – did not respond to a request for comment on which tagging model they operate.

“Where I think the use of tags has gone awry is when they’re shown before trading. That’s information which a market-maker can use and abuse,” says the head of FX at a liquidity provider.

“There are a number of high-frequency firms who would use that tag information to try to pre-hedge their counterparty, which to me is pure front-running – yet real-time tag dissemination allows for it,” the FX head adds.

In another scenario, a liquidity provider might reject an execution attempt by a consumer and then use the pre-trade tag information to trade in the same direction as the consumer, potentially disadvantaging the consumer.

“The use of tags in this way is definitely a concern, is in contravention of the FX global code, and is certainly one of the reasons why pre-trade tagging is looked at with a tremendous degree of sensitivity within the industry,” says Currenex’s Newns.

However, Neill Penney, co-chair of the Global FX Committee and global head of trading at Refinitiv, points out that front running is barred under the global code, regardless of whether pre-trade tags are being used to facilitate it.

“Front running is prohibited under the global code of conduct in all circumstances. This has nothing to do with whether tags are used. It’s important to separate out behaviours which are prohibited from the code from behaviours which are enabled by tagging but are consistent with the code.”

Similarly, Jon Healey, head of global e-commerce at BBVA, says front running could take place at any time and isn’t limited to a semi-anonymous setting – with front running arguably being most likely to occur within a disclosed trading setting, where counterparties know each other’s identity.

“Pre-trade tags aren’t necessarily a bad thing, as it is not the transmission of that information that is a problem, the issue only occurs if that information is misused,” he says.

Tag refreshing

The practice of changing, or “refreshing”, tags also arouses concern. In bona fide cases, it allows market participants who have legitimately changed their trading behaviour to access liquidity pools that may have been cut off from them previously. A second chance, as it were.

Alternatively, a bank may want to split its offering into two separate liquidity streams, requiring two separate tags – despite the fact that the same bank is behind both of them.

But the practice is open to exploitation by parties with more suspect motives. Dealers are wary of so-called toxic flow – trades that result in the dealer losing money. Often it can stem from counterparties who break up larger orders into smaller chunks in a bid for tighter pricing.

Blacklisting the tags associated with these toxic clients is one remedy. But the client may pop up again under a different tag.

“The most annoying thing about tagging is when a platform comes to us – knowing that we don’t like tag 123 – and says ‘Hey, we’ve got this great new tag 456’. So we try it, and lo and behold it’s actually tag 123 again,” says the LP’s head of FX.

“We have seen highly toxic tags where people give up and stop pricing them, then that liquidity taker shows up under a brand new tag. Obviously, as a market-maker we would love for that not to happen,” the FX head adds.

Partly in response to these concerns, some platforms do not refresh tags, instead ensuring members retain the same tag they were assigned when initially joining the platform.

“I’m not sure what the benefit of refreshing tags is as it creates operational inefficiencies for clients, who have to continually remove tags they don’t want to see from their liquidity stream,” says a senior executive at one FX platform.

Given the direction within the GFXC, I think that an anonymous ECN has a clear responsibility to police activity conducted on that ECN

David Newns, Currenex

Research by Risk.net shows that three semi-anonymous platforms – Currenex’s FXTrades, EBS Select, and Refinitiv’s FXall Order Book – do not refresh tags. One platform – Euronext FX – does refresh tags, and two platforms – Cboe’s Hotspot and 360TGTX – did not respond to a request for comment on whether they refresh tags or not.

Some argue that, given the daily data and analytics that firms are able to collect on tag trading behaviours, it wouldn’t take long for a firm to work out that a new tag is an old counterparty they’ve tried to cut ties with.

An e-FX specialist at one large dealer also says it’s difficult to prove how rampant a practice it is within the market considering the semi-anonymous nature of trading.

“It’s one of those things that is almost mythical as you would never really be able to know for sure whether it happened to you or not. For example, if you had somebody randomly trading something esoteric like CAD/NOK, which you used to see offered by tag 111 and then never saw it again until tag 112 started offering it, you might think it looks a little odd but it’d be very difficult to prove that tag 111 was actually behind the trade,” the specialist says.

Platform police

Nevertheless, concerns over the misuse of tags has reopened a debate over the role that platforms themselves should play in the market – and whether they should more actively police their venues for signs of sharp practices.

“I’d be surprised if any anonymous platform providers believed that the marketplace that they are responsible for should be a free-for-all,” says Currenex’s Newns. “Given the direction within the GFXC, I think that an anonymous ECN has a clear responsibility to police activity conducted on that ECN.”

Platforms say they are proactive in working with counterparties to identify who is responsible for dubious practices and potentially remove them from the platform.

Hugh Whelan is global head of liquidity management for cash markets at CME Group, owner of EBS Select. He says: “It’s not always immediately obvious who is behind activity like [front running], so in many cases we’ll filter through all the various non-disclosed tags, disable or enable them, and run some test trades with the LP in order to identify where the LP sees information leakage – for example.”

ING’s Malrait agrees that platforms should be keeping an eye out for tag misuse, but doing so can come at a cost to the platform. The more aggressively platforms police their venues, the less diverse the flow they’re able to offer becomes – something which could ultimately undermine the amount of volume transacted on the platform.

“That’s where the balance has to be right because, for sure, it’s in the platform’s interest to get market participants trading as much as possible but they also have to respect the interests of people who don’t want to give liquidity to aggressive players,” he says.

Meanwhile, some say it shouldn’t be up to the platforms to monitor activity on their venue, but that trading counterparties should be self-policing based on the principles of the global code of conduct.

“The platforms are just a delivery service,” says Mike Wilkins, head of FX trading at INTL FCStone. “If you order something from Amazon and you’re delivered a substandard product, you don’t blame the delivery driver.”

Disclosure, disclosure, disclosure

What market participants do agree on is that more transparency is needed over the tagging models that platforms employ, so traders can make better informed decisions about which platforms to use.

According to the results of the GFXC’s most recent annual survey, over 20% of respondents either received no disclosures from anonymous venues or found that disclosures were “very poor”.

Risk.net was not able to find a single semi-anonymous platform with fully accessible online tagging disclosure. Two platforms – Currenex’s FXTrades and Refinitiv’s FXall Order Book – have an online disclosure available to clients only. One platform – EBS Select – doesn’t have any kind of online tagging disclosure. Meanwhile, three platforms – Cboe’s Hotspot, Euronext FX, and 360TGTX – did not respond to a request for comment on whether they have any form of online tagging disclosure.

“It’s true that liquidity providers have made efforts for their disclosures to be more in line with each other but such efforts haven’t yet been made for trading venues, so at the moment venues have the flexibility to produce whatever type of disclosures they want,” says ING’s Malrait.

While platforms such as Currenex’s FXTrades typically prefer to divulge tagging information to clients in bilateral conversations, Malrait believes that having a written and fully accessible tagging disclosure would make platforms more accountable and provide greater comfort to market participants.

The Global FX Committee says transparency and increased disclosure are a priority in its three-year review of the code. The committee is looking into standardised vocabulary around tags so that market participants can more easily compare disclosures from different venues.

However, some believe the GFXC could go further. Templates for disclosures would help market participants identify and compare disclosures between venues, they say. This would save platform users the task of having to wade through pages of inaccessible FAQ documents to find the information they’re looking for.

Kevin Wolf, head of fixed income, currencies and commodities at Euronext, believes template disclosures are a step too far. But he agrees that it would be beneficial for the industry to agree on what the bare minimum of disclosed information should be.

“The spirit of the code is voluntary self-regulation so I think the industry could agree on what the key pieces of information platforms should disclose is, and then leave it up to each platform to convey that information in a way which is effective and clear,” he says.

Refinitiv’s Penney believes it’s unlikely the GFXC will create template disclosures – or agree on a single tagging model to be used across the market – but he reiterates that increased transparency from platforms on how they use tags is key.

“For example: if you get tagged, when you get tagged, whether that tag ever changes, and how the tag is used. This information should be shared in the spirit of increasing transparency,” he says.

The issues surrounding tags will remain unresolved for the near future, though, as the ongoing coronavirus pandemic has meant the various GFXC working groups – including the anonymous trading working group – have had their work put on hold.

While June’s GFXC meeting is expected to go ahead virtually, the working groups will not be presenting their findings until a later date – likely December.

“Essentially, the plan is that the code review process will be delayed by six months due to coronavirus. However, there is no change in priorities: the topics of focus we announced after the Sydney GFXC meeting remain the topics that are top of mind for the GFXC. Transparency remains a top priority for the both the industry and for the global committee,” Penney says.

Editing by Alex Krohn