Galen Stops speaks to industry sources to try and present the other side of the story regarding the recent legal action against FXCM.
The FX industry has, by and large, been swift and united in its condemnation of the actions of FXCM, for which the firm was banned from the US and fined $7 million for defrauding FX customers.
But, as they say, there are always two sides to every story and so Profit & Loss has been talking to various market sources that provide different perspectives on this case. This is challenging because as part of the legal agreements between FXCM and the Commodity Futures Trading Commission (CFTC), the firm neither denied nor admitted the allegations against it, and therefore cannot speak to the press about the issue.
But some sources close to the firm, as well as former employees of the firm, have been keen to try and argue that the punishment levied against it was out of proportion to the crime alleged.
Firstly, it is important to highlight some disparities between the complaints issued against FXCM by the CFTC and the National Futures Association (NFA).
The CFTC alleges that between September 4, 2009 until “at least” 2014, FXCM “engaged in false and misleading solicitations” regarding its retail FX business. Specifically, the Commission found that FXCM told customers that when they traded on its No Dealing Desk platform, the brokerage itself would have no conflict of interest about who is on the other side of these trades, because it acts merely as a credit intermediary. In contrast to this model, there are other retail FX brokerages run trading books and therefore actively trade against their clients.
However, the CFTC says that contrary to these representations that FXCM had an “undisclosed interest” in the market maker that was winning the largest share of trades on its No Dealing Desk platform, “and thus was taking up positions opposite FXCM’s retail customers”.
The NFA complaint is largely similar up until this point, but this is where the first major divergence between the two seems to occur. The CFTC complaint then goes on to immediately note that the market maker in question “was an FXCM-backed startup firm”, while the NFA names the market maker as Effex Capital, and says that this firm was “purportedly independent, but which FXCM actually supported and controlled”.
Being “backed” by a firm and “controlled” by it are clearly two very different things, so let’s get into the details.
Define “Dealing Desk”?
According to both complaints, FXCM provided Effex with a $2 million interest free loan, sponsored an account for the firm at its prime broker, allowed it to operate out of FXCM’s main office until it moved to its own in New Jersey in 2011, provided it with systems support and access to FXCM’s corporate network through a virtual private network, and even assigned some FXCM employees to assist in operating Effex’s trading systems.
In addition, the CEO of Effex, named in the NFA complaint (but again, not in the CFTC one) as John Dittami, was a former FXCM employee, leaving the brokerage only to start the market making firm in 2010, having developed an algorithmic trading system designed to trade opposite FXCM clients.
According to the CFTC complaint, it was agreed when Dittami resigned at FXCM to start this new firm that it “would not changethe economic relationship between FXCM and HF Trader [Dittami], including, as stated in his employment contract, HF Trader’s retention of 30 percent of his algorithmic trading profits, with FXCM capturing the residual 70 percent”. All of which doesn’t sound too good in terms of Effex’s independence from FXCM.
That being said, the NFA complaint doesn’t appear to supply hard evidence that FXCM actually “controlled” Effex. Legally, it appears to have been a separate entity. Crucially, Dittami did resign from FXCM as an employee and sign a new services agreement with the firm as Effex. And while the firm might have worked very, very closely with FXCM, and may have even been de facto dependent on its support to begin with, there doesn’t appear to be any concrete evidence provided that senior staff at FXCM had the ability to dictate to or overrule Dittami on any matters relating to Effex.
This is also the argument from sources close to Effex, who claim that the firm has been wholly independent since its inception in 2010, pointing out that it has subsequently opened four different offices and traded with over 30 different counterparts, which included pricing into FXCM competitors.
One source close to the matter adds that, in fact, the funds in the $2 million loan never actually left FXCM because the two firms couldn’t agree on business terms, and the loan was closed before Effex funded its own PB at Citi in early July 2010.
“In short, Effex is not now, nor has ever been, a dealing desk arm of FXCM,” says one source with knowledge of the firm’s business operations.
Or as another source familiar with the matter puts it: “If FXCM truly had a dealing desk then it wouldn’t have lost so much money on SNB day.”
If FXCM did not own or control Effex then it was, in fact, not trading against its own customers on the No Dealing Desk platform. But, because Effex only kept roughly 30% of its trading profits and rebated 70% to FXCM this arguably becomes something of a technicality. The whole point of the “no dealing desk” model was to eliminate any potential conflicts of interest, so that retail clients could be sure that they were being routed to whichever liquidity provider gave the best price without regard to how this might impact the brokers’ P&L.
But this, argue those defending FXCM, is exactly what did happen.
As to the rebate that Effex paid FXCM, payments for order flow was not against NFA rules, and FXCM disclosed in its annual reports between 2010 to 2013 – during the time of the alleged offences – that payments for order flow was one of its revenue sources for its retail business. In the 2014 annual report, FXCM said that this practice was discontinued, effective August 1, 2014.
In addition, one source makes the case that the idea that Effex paid a 70% rebate to FXCM is factually incorrect and claims that the CFTC is in possession of sworn affidavits from audited financials to that effect. They say that although the fee that Effex paid FXCM in 2010 worked out at a roughly 70/30 split in FXCM’s favour, the payment agreed was actually a fixed amount that had to be paid regardless of Effex’s P&L, and that from 2010 onwards the split looked very different.
Regardless of how the payment worked, the section regarding payments for order flow in the FXCM annual report reads: “Our order routing software ensures that payments for order flow do not affect the routing of orders in a manner that is detrimental to our retail customers”.
Sources with knowledge of how FXCM operated during this time claim that the client orders were always routed to the best price, it’s just that Effex was giving the best price more often than the other liquidity providers and therefore was able to become the biggest market maker by volume on the platform. How was it able to do this?
Well, in exchange for this 70% rebate, FXCM allegedly agreed that it would favour Effex over other liquidity providers by enabling it to win all “ties” with other market makers, it would provide it with a real-time view of the price quotations offered by other market makers and add smaller markups to Effex’s prices than to prices provided by other market makers.
Therefore, because it could see the quotes from other market makers, Effex could price just inside them to win a trade. And because they suffered a smaller markup from FXCM they could consistently afford to do this.
One source close to Effex even argues that the benefit to customers went deeper than just price, saying that customers got best execution from this arrangement and had trades honoured, even when the market moved against Effex. The source adds that the markup advantage that Effex enjoyed was exaggerated in the complaints because it refers to the markup being reduced in “pairs” whereas in fact, it is claimed, it was reduced by 1/10th of a pip in one pair during the first few years of the arrangement and then dropped altogether.
Thus, sources familiar with the issue, argue that this arrangement was a win-win that ultimately benefitted the brokerage, because it gained revenue via payment for order flow; Effex because it won more flow; and FXCM customers, because they got tighter prices on which to trade.
This is the crux of the argument from those defending FXCM: how could the brokerage have defrauded customers – the reason why they were ostensibly banned and fined – if these customers actually saved money by receiving tighter prices?
As to Effex being given the advantage on trades where it tied on best price with other market makers, one former FXCM employee comments: “If you prioritise a firm that is giving you a rebate in your aggregator over firms that aren’t, is that wrong? I don’t know, but it’s common market practice in our business and I don’t believe that FXCM did anything wrong because ultimately the client got the best price in the aggregator at the time of the trade. There’s no conflict of interest on behalf of FXCM simply because the best price did always win.”
Likewise, a source close to Effex argues that the firm only won execution requests if it offered the best price, and that it only won ties against other providers upon review of its execution statistics being the best of all the providers. Because of this they say that there were times – albeit it infrequently – when other providers did win ties over Effex.
The CFTC compliant even notes that Effex and FXCM were forced to amend the rebate agreement between the two firms because the market maker’s profits dropped as spreads tightened.
This claim that the arrangement between the two firms benefited FXCM’s retail clients becomes more problematic, however, when compared to comments in the CFTC and NFA complaints.
From the CFTC legal document: “In addition to favoring HFT Co [Effex] over other market makers, HFT Co made use of, and FXCM allowed HFT Co to use, a hold timer that enabled HFT Co to execute a trade at the start or end of a hold timer period, whichever was better for HFT Co. In late 2011 and 2012, HFT Co continued to make use of a hold timer but would accept or reject the trade based on the price at the end of the hold timer period.
“HFT Co also made use of a “previous quote” practice whereby HFT Co submitted a quote to FXCM and FXCM would respond with an execution request based on the trading limits contained in a customer limit order and not the previous quote provided by FXCM.”
If how this worked isn’t clear enough, replace the words “hold timer” with “last look” and you’ll get the idea.
But whereas the CFTC complaint noted that Effex used this strategy to its own benefit, it stops short of explicitly saying that it had a negative impact on client pricing.
The NFA document, though, does not: “In cases where the price at the end of the hold timer had moved against Effex and in favour of the customer, Effex would reject the trade. On the other hand, if the price had moved in favour of Effex and against the customer, Effex would execute the trade and give the customer negative price slippage.”
It continues: “For a time, Effex used a particularly egregious version of the Hold Timer where the firm would hold market orders for a preset period and then execute the order at a price beneficial to Effex, and to the detriment of FXCM’s customers.”
This is where the claim from those sympathetic to FXCM and Effex runs headlong into the arguments from the regulators.
A source with knowledge of Effex’s business practices claims that in the firm’s seven-year history it has a rejection rate of less than 1% on execution requests. The source goes even further, claiming that Effex in fact provided price improvements, reduced execution slippage by adding substantial liquidity at best price, and used price tolerance settings to fill deals biased to the benefit of the end customers, and to Effex’s own detriment.
The source also adds that while Effex co-located servers with FXCM in order to provide the fastest response time possible on execution request, the firm operated behind network security layers to prevent it from having access to customer information, and from seeing or knowing customer resting orders.
Noting the disparity between the CFTC and NFA complaints, this source claims that the latter of these two regulators never contacted Effex during its investigation, never requested any execution data or asked any questions of Effex, its principals or its employees.
In contrast, they say that the CFTC received both execution information and testimony from Dittami, pointing out that the Commission subsequently did not name Dittami or Effex in its complaint, nor did it allege any wrongdoing by them.
Further contrasting the different complaints, the source claims that the “Hold Timer” referred to in the NFA complaint was used in 2010 for five or six trading days with a hold period of 25 milliseconds, whereas the CFTC complaint refers to a time after Effex built a tolerance into its system, which gave the asymmetric advantage to the customer during execution.
“The cost to customer in these five days was likely single digit thousands, versus tens of millions of benefit over time,” the source argues.
Defenders of FXCM also argue that if there was evidence that customers being disadvantaged by the arrangement between the brokerage and Effex, the NFA should require the former to pay restitution to these clients.
They point out that when FXCM was fined in 2011 for engaging in asymmetrical price slippage practices, it was forced to pay a civil monetary penalty and restitution to clients, which in that case jointly came to $16 million.
Granted, pointing to the last time your firm was fined for deliberately disadvantaging its customers as evidence of why this new punishment is wrong in its scale seems counter-intuitive.
Yet when Profit & Loss questioned a spokesperson for the NFA about why FXCM had not been required to pay restitution, the spokesperson simply said that all the public details regarding this issue can be found in the various case documents.
Similarly, when a spokesperson for the CFTC was asked how the Commission had reached the figure of $7 million for the fine against the brokerage, they simply deferred to the online case documents, which do not provide details of this.
As we move on, however, the actions of FXCM become slightly harder to defend. The obvious question thus far is: if the relationship and set-up between FXCM and Effex was an innocent win-win that ultimately benefitted the broker’s clients, then why the appearance of a cover up?
The CFTC document notes that in 2011 FXCM published a diagram showing what percentage of volume each of its market makers provided to the platform. Despite the fact that Effex was providing the most liquidity by volume, it was bracketed under the banner “Citi-Prime Broker (All Others)”, rather than being listed individually like the banks.
An oversight? Perhaps. After all, the individually listed liquidity providers were all banks, and Effex wasn’t, so it could be that’s simply how they divided it up and decided to represent it on the diagram.
The next segment of the CFTC complaint makes this explanation slightly harder to swallow, however.
From the CFTC complaint: “When a list of FXCM’s top liquidity providers for the first half of 2010 was circulated internally, showing HFT Co [Effex] with over 17 percent of the company’s volume, an FXCM employee asked whether that reference should be ‘blacked out’.” Another employee responded: “Good catch, put them as HFT 1….”
And then: “In September 2011, an FXCM executive raised ‘Compliance Concerns’ with members of FXCM’s Compliance department, noting a problem with HFT Co ‘hanging orders’ that arose from ‘the fact that the [HFT Co] adapter is different technology than the rest of the bank adapters’.”
An FXCM compliance officer responded: “Given the sensitivity of the counterparty involved we would prefer the events don’t happen so that we do not draw unwanted attention.” In addition, in response to an online posting, FXCM claimed that non-disclosure agreements meant that it normally did not disclose the name of the 17 liquidity providers that stream prices on its No Dealing Desk feed.
“In fact, there was no non-disclosure agreement between FXCM and HFT Co, and, as noted above, FXCM had disclosed ten of its other market makers to the public,” it says in the CFTC document.
Then to make matters worse, FXCM apparently misled the NFA about its relationship with Effex. Consider these statements from FXCM’s compliance department to its regulator:
One: “FXCM LLC does not have any direct or indirect ownership, interest, or affiliation with entities that provide liquidity to retail clients….”
Two: “To my knowledge, there are no present or past owners, principals, APs, or employees of affiliates of FXCM LLC that have direct or indirect ownership, interest, or affiliation with entities that provide liquidity to retail clients on our No Dealing Desk Model.”
Three: “[HF Trader] served as a consultant for FXCM from October 2009 through April 2010. [HF Trader] worked primarily on software coding.”
In the first statement, the compliance staff might be technically correct in saying that the firm didn’t have ownership or affiliation to any of its liquidity providers, but to say it didn’t have an “interest” in a firm that it was giving clear advantages to in return for a 70% rebate that the CFTC says yielded nearly $80 million in revenue might be stretching things slightly. But giving the benefit of the doubt, “interest” could be interpreted meaning a stake in the company, giving the context in which it is used in between “ownership” and “affiliation”.
The second is clearly false because Dittami was, as the CFTC notes, “a former employee (and, indeed, an executive) of FXCM” and also “the principal and 100% owner of HFT Co, which was the primary provider of liquidity to retail clients on FXCM’s No Dealing Desk platform”. Incidentally, the fact that the CFTC complaint agrees that Dittami was 100% the owner of Effex clearly supports the argument that it was a clear and distinct entity from FXCM.
And the third statement is false because, well, there was apparently an employment contract in which Dittami – aka “HF Trader” – was named not as a “consultant” but as an “executive” of FXCM.
The NFA’s list of infractions by FXCM is more extensive than the CFTC’s, including violations such as “failing to treat all customers equally when giving price adjustments and failure to submit trade data information to NFA through Fortress”.
In contrast, the violations cited by the CFTC centre around the fact that FXCM misled its customers and made false statements to the NFA in order to conceal its history and relationship with Effex.
To briefly recap the arguments made in defense of FXCM: Payments for order flow was allowed under NFA rules at the time. FXCM did not at any point own Effex and therefore did not deal against the clients trading on its No Dealing Desk Platform. By having this close relationship with Effex and giving it certain advantages over other liquidity providers it was able to provide them with better pricing, saving them money.
So again, why the appearance of a cover up? Perhaps because even if you accept the argument that the set-up between FXCM and Effex benefitted the broker’s retail clients, it certainly didn’t benefit the other liquidity providers on the platform.
Speaking to Profit & Loss, a source at one market maker expressed great frustration about conversations that they had participated in with FXCM about why they weren’t winning more volume, in light of the revelations contained in the CFTC and NFA documents.
If FXCM had been transparent about its relationship with Effex and the advantages that it was providing it with, it seems unlikely that these other liquidity providers would have kept pricing into the platform.
Some of the sources that defend FXCM’s actions acknowledge that the regulators were right to discipline the firm for a lack of transparency regarding its relationship with Effex, but argue that claims that it “defrauded” customers are completely wrong and therefore the punishment meted out to the firm is disproportionate to the crime allegedly committed.
If this relationship was truly to the advantage of FXCM’s clients, the brokerage would presumably possess data which would prove this claim. When Profit & Loss pressed one source close to FXCM about why the firm did not take up a legal fight to clear its name, it was claimed that there was an internal debate within the company regarding this, with CEO, Drew Niv, arguing in favour of a legal challenge.
But the source says that the first round of any legal battle would take place before CFTC and NFA committees, and therefore it is unlikely that FXCM would be successful in changing the decision.
The source adds that the next option after that would be to go to a Federal judge, but that it was decided internally at FXCM that this legal process could take years to reach a conclusion and would be financially expensive for the firm.
Was this a case where the cover up was worse than the crime? Was the punishment and subsequent censure heaped upon FXCM unfair?
That presumably depends on how convincing you find the arguments defending the brokerage. But it’s worth noting that when Profit & Loss contacted the NFA, its spokesperson made it clear that the regulator had taken previous infractions into account when making its decision to ban FXCM from the US.
“The NFA’s decision barred FXCM from membership given their long history of engaging in deceptive and abusive practices and they are expected to remain in compliance with all applicable rules and regulations until their withdrawal date,’ they said.
In addition to the $16 million penalty issued against FXCM in 2011, the NFA has authorised three other complaints against the brokerage. While a number of sources have made comments about the NFA’s attitudes towards the retail OTC FX market, ranging from those that claim that the regulator has an active bias towards the futures market to those claiming that this market makes them “uncomfortable”, others in the industry say that they haven’t seen anything to corroborate these claims.
Ultimately, when bringing a legal case against a firm, the regulator has to point to areas where the firm in question violated the rules that it is authorised to enforce, and it is hard to argue that FXCM hasn’t given the NFA ammunition over the years in this regard.
Whether or not the relationship between FXCM and Effex was beneficial to the brokerage’s retail clients has actually not been proven, one way or the other, by any hard data released to the general public.
But what is certainly clear is that what some once viewed as a win-win relationship for everyone has potentially become a lose-lose situation as FXCM loses revenue and suffers further reputational damage, Effex – rightly or wrongly – gets it name dragged through the mud, and retail FX clients in the US have less choice as competition between brokerages reduces even further.