The incarceration of a trader convicted of spoofing has heightened awareness of the practice, but how hard is it to spot and how prevalent is it in FX? Colin Lambert investigates.
“You have to be pretty desperate to resort to spoofing markets – especially on exchanges where it’s nigh on impossible to shield your activities,” argues a senior electronic trader in London. “Even in OTC markets it’s not easy to get away with given the MIS capabilities of firms today.”
If that is the case, the question has to be asked – even though it will probably never be answered – what was Panther Energy’s principal Michael Coscia trying to achieve with his trading strategies that led to him being sentenced to three years in jail? He has the notoriety of becoming the first trader to be incarcerated for the practice of spoofing financial markets, but at face value it is hard to see how his controversial strategy could not have been discovered.
In its findings against Coscia, the US Commodity Futures Trading Commission (CFTC) explained how Panther had a strategy that would place a relatively small order on one side of the market, and then place much larger orders on the other to effectively make the first order look attractive. Once the logical trade had taken place, and the first order was filled, the counter-bids/offers were cancelled and the strategy repeated itself, but on the other side of the market.
The UK’s Financial Conduct Authority (FCA) also took action against Panther and Coscia and in its report it provided more details in the form of examples.
According to the FCA, a typical example of the strategy would be to place a buy order in the market at 115.86 for 17 lots (making it best bid) and then placing three sell orders at 115.87, 115.88 and 115.89 for amounts ranging from 50 lots to 125 lots. In its report, the FCA states that the larger orders were on average around 20 times the counter order.
Once the smaller order was executed, the strategy reversed itself. The FCA says that just 0.1% of Coscia’s larger, or spoof, orders were fully filled, compared to 44.5% of the smaller orders.
The fact that such a small percentage of the larger orders was hit highlights the changing character of financial markets brought about by the total transparency on exchanges. As an experienced e-FX trader in Asia points out, “Yes, there is a chance that a trader could get hit on the ‘wrong side’ when they are trying to spoof, but they are largely dealing with artificial intelligence that can only believe what it sees in the form of data.
“A human trader might witness the pattern and smell something, a computer programme can’t,” the e-trader adds.
While Coscia’s conviction was seen as a triumph for the market supervision framework of the exchanges, it has also raised the question, could a spoofer succeed in OTC markets – especially one as fragmented as FX?
There has been a trading strategy referred to as “spoofing” in FX markets that goes back decades. In its first iteration, a trader would deliberately trade “noisily”, generally in public markets through the voice brokers, in the opposite direction to the intended trade, with the hope of shifting the market a few points. If the market did move, then the real order would be executed, but generally on a “dark” bilateral basis using direct calls.
The advent of electronic trading and the increase in transparency saw the practice dwindle, however dealing sources tell Profit & Loss that the dominance of EBS and Reuters Matching in the first decade of the century meant that a different form of the practice emerged in which a trader would place an order against the intended trade on one of the aforementioned primary venues and then place the intended order on a secondary venue.
“Everybody priced off EBS or Matching, depending upon the currency pair, so you could influence the market to a small degree by using those platforms,” explains a senior industry figure. “As high frequency traders entered the market it actually became a little easier because they would typically jump in front of your bid/offer on another platform and reinforce the sense that there was a buyer/seller out there.
“As price engines became more sophisticated it became harder to do but it was – is – still possible,” the source adds. “It’s a smart way to execute because masking a large order has become increasingly important and if you can gain a few seconds using this strategy it can be valuable.”
Of course, using such tactics can also go wrong, especially if the “spoof” trades run into a larger market order, thus leaving the trader with an even larger buy or sell order and the likelihood of more slippage, but as noted earlier, the increased automation of markets means that if a primary market exists, it is still able to influence activity on other venues, because the pricing engines follow the data.
As best execution rules have evolved, the aforementioned practice has come to be frowned upon, even though many execution strategists will argue the resulting outcome is something less than best execution. Overall though, execution algos offered by providers will do no more than pause or increase/decrease market participation in response to different market conditions. Trading against the parent order is generally forbidden.
The Grey Area
But is such practice spoofing in the modern sense? On one hand it seems an attempt to deceive, but as already pointed out, some participants see this as a sensible thing to do with so many “sniffer” algos out there.
On the other hand, however, the aforementioned examples have one important difference to the actions of Coscia. “There is an intention to deal, pure and simple,” argues the head of execution services at a bank in London. “If you have an intention to deal there cannot be spoofing.”
While some may argue that proving intention to deal is a difficult and tricky issue, the execution services head is dismissive. “If you place an order at top of book in a CLOB [central limit order book] with no last look, what else can it be but intention to deal? I accept there is the opportunity to abuse the system by, for example, removing all credit from a platform before placing the order, but that doesn’t work – and it’s easy to spot.”
Inevitably perhaps given the practice’s high profile in FX markets, the key in many eyes is the use of last look. “You could argue that placing counter bids or offers to the real order on a platform with last look could be spoofing,” the senior industry figure accepts. “There is no way of knowing whether a trade will be accepted until it is, so how do you prove intention to deal?
“In those circumstances it comes down to the platform – especially if it is anonymous – monitoring the reject rates of participants. If a trader is rejecting a lot of trades when the market hasn’t moved against them, you have reasonable evidence that they are trying to spoof the market and do not have the intention to deal.”
A challenge in such circumstances, however, is actually proving there was a “real” counter order in the market, for unless the participant in question has placed the counter bid/offer on the same platform there seems no way of linking that participant’s activities. “I am not sure how you deal with such a scenario in FX markets,” says a liquidity manager at a platform provider. “There is no single point of information for trades, let alone interactions, so how are we going to know if a trader is counter-bid/offering on another venue?”
The only way to track such activity would be, it seems, at firm level, but while most major businesses have internal compliance that could spot it, a fair number of smaller firms have no such capabilities, and indeed in many cases have no requirement for them. “For the small, three or four person, tech-heavy, trading firm, there is no compliance – it’s make whatever money you can, however you can,” the senior industry figure observes.
There is, potentially, another source of oversight in the trading firm’s prime broker if it uses one. Prime brokerage sources tell Profit & Loss that while they currently monitor a client’s activities in listed markets, they do not in OTC markets. “It is something we are looking at and it’s not easy,” says a senior source at a major prime broker. “The ECNs have to be part of any solution and we’re talking to them right now to see what can be developed.
“If we are providing sponsored access to certain markets, then we need to ensure our client is behaving responsibly, because there are potentially reputational issues for us as an institution,” the source continues. “At the moment, each individual platform has a responsibility to ensure people behave correctly on their venue, but no-one can check activity across all platforms. If we can collect the order flow data we will be able to provide a more detailed picture that would help spot poor practice.”
While Profit & Loss understands from sources familiar with the matter that the second stage of the Global Code of Conduct will deal with the issue of prime brokers’ responsibilities, not all believe it is necessary. Another prime brokerage source argues, “Getting the data in real time isn’t going to happen so this is more about having the data stored so that if a query is raised we can help with resolving it. I am not sure our surveillance role should extend to monitoring our clients’ behaviour in the market – that’s a job for the venues.”
Regardless of who does monitor the behaviour of market participants there is a growing feeling that something needs to be done to show the industry is on top of what for many remains a potential, rather than real, issue.
While the focus tends to be on smaller, quicker trading firms when discussing spoofing, there are some who argue that banks are also in the firing line, and once again, last look is the key factor. “Banks have compliance teams so out-and-out spoofing is likely to be picked up,” says a senior manager at a non-bank trading firm. “But what about the practice of pricing to multiple venues? A bank will stream a price to 10 venues in the same currency pair, but has no intention of dealing on all of them and will reject multiple trades using last look if they are hit on one, maybe two venues. Could that be construed as spoofing? Personally I don’t think so, because it is a natural protection against machine gunning and other abusive practices, but is a layman, especially in the legal profession, going to understand the nuances? I don’t think so.”
Certainly the different market structure of FX would offer challenges to legal experts trying to decipher a trader’s activity, however even in regulated, fully transparent markets like exchanges, ambiguity still exists.
In a blog earlier this year, Gary DeWaal, special counsel at Katten Muchin Rosenman, discussed the Coscia conviction and noted he was convicted under the provision of law that prohibits spoofing, or the bidding or offering with intent to cancel the order before execution.
“It may be clear what is prohibited by this provision, as the court has written, but by its broad sweep, the provision technically makes illegal relatively ordinary trading conduct that no one – not even the Commodity Futures Trading Commission or any exchange – would likely consider nefarious,” DeWaal wrote. “For example, when a trader places a stop-loss order, he or she does not intend for the order to be executed, because, presumably that would mean the market is trending in a direction opposite his or her expectation. However, he or she will accept a trade execution if the conditions of the stop-loss order are realised.”
While noting that the CFTC tacitly acknowledges the dichotomy, as does CME, DeWaal bluntly states, “Unfortunately, the statute prohibiting spoofing simply has it wrong. There is nothing automatically problematic about all spoofing as now defined under applicable law.
“Deception, to some except, is part of smart trading. No trader knowingly reveals all his or her strategy or intent as part of an order placement (consider, for example, exchange- sanctioned iceberg orders),” DeWaal continues. “As CME Group wrote in a comment letter to the CFTC about what should be deemed illegal spoofing, it is not the intent to cancel orders before execution that is necessarily problematic, it is “the intent to enter non bona fide orders for the purpose of misleading market participants and exploiting that deception for the spoofing entity’s benefit”. Spoofing is simply the big circle in the applicable Venn diagram; what should be prohibited is solely a smaller circle within the larger one – a subset.”
So while Coscia has been convicted and jailed – and UK- based trader Navinder Singh Sarao faces trial in the US for his alleged role in the 2010 Flash Crash – it seems that too many questions remain. “Until the anti-spoofing law is further clarified to reflect what truly is problematic, it will embrace both legitimate and illegitimate activity; potentially scare away bona fide trading and have a deleterious impact on market liquidity; and potentially cause some market participants to run afoul of the law for ordinary order placement activity,” DeWaal wrote. “This is not right or fair, even if, as the court ruled in Mr Coscia’s case, traders have ‘fair notice’.”
Uncovering such activity in FX markets is commensurately harder and requires a degree of luck, according to the head of market surveillance at a major bank in London, who explains, “To uncover something like this you need a break – something that makes you, or the relevant authorities, take a look at a participant’s activity. That’s the hard bit. The good news is that spoofing, if it takes place in today’s market – and I haven’t seen anything – will take place electronically and as such can be tracked.
“To successfully spoof a market you have to use a pre- programmed algo strategy – trying it manually would rarely, if ever, work because it has to be quick,” the surveillance manager, who has previous trading experience, adds. “Once someone is suspected of spoofing, an authority could go through the legal process and gain access to the algo strategies that would tell them everything. If someone has a high frequency arb algo they will deal both sides and will only be passively bidding or offering one side of the market. If they’re spoofing, the algo will be programmed to be on both sides of the market, to skew the balance of the order book, and to only trade one side. It’s easy to spot, once you actually have the ‘in’.”