Pounding the pound.

Photographer: Munshi Ahmed/Bloomberg.

HSBC Currency Traders Got Greedy on Christmas

Matt Levine is a Bloomberg View columnist. He was an editor of Dealbreaker, an investment banker at Goldman Sachs, a mergers and acquisitions lawyer at Wachtell, Lipton, Rosen & Katz and a clerk for the U.S. Court of Appeals for the Third Circuit.
Read More.
a | A

Oh man:

Federal agents surprised an HSBC Holdings Plc executive as he prepared to fly out of New York’s Kennedy airport around 7:30 p.m. Tuesday, arresting him for an alleged front-running scheme involving a $3.5 billion currency transaction in 2011.

Mark Johnson, HSBC’s global head of foreign exchange cash trading in London, was held in a Brooklyn jail overnight and will appear in court Wednesday, according to prosecutors.

Here is the Justice Department's announcement of charges against Johnson and Stuart Scott, the former head of FX cash trading for Europe, the Middle East and Africa. Here is the complaint, charging them with front-running the purchase of $3.5 billion worth of pounds sterling by a corporate client, apparently Cairn Energy.

As is so often the case, the alleged front-running happened in two ways: gradually, and then suddenly. In 2010, Cairn signed an agreement to sell part of its ownership interest in an Indian subsidiary for about $3.5 billion, subject to regulatory approvals. "If the sale was approved, the Victim Company planned to convert approximately $3.5 billion in sale proceeds into Sterling, which it then intended to distribute to its shareholders," and it sent around a request for proposal seeking a bank to do the currency conversion. In October 2011, Cairn picked HSBC to manage the trade, and HSBC agreed to keep Cairn's information confidential and not to trade on it. Johnson and Scott were brought in on the trade. From the complaint:

In or about October 2011, the defendant MARK JOHNSON became an "insider" to the Victim Company FX Transaction. In or about November 2011, the defendant STUART SCOTT also became an "insider" to the Victim Company FX Transaction. As "insiders," JOHNSON and SCOTT knew they had an obligation not to misuse the Confidential Information, including by front-running.

But, says the Justice Department, they did. On Nov. 28, 2011, they "were notified that the Victim Company FX Transaction might occur soon"; two days later, Johnson bought some sterling in his proprietary trading book. A week later -- Dec. 5 -- they "received additional information" about the trade; "specifically, a news article was circulated to them reporting that the underlying sale by Victim Company of its Indian subsidiary had received regulatory approval." That day and the next, Johnson and Scott bought some more pounds. Cairn actually traded on Dec. 7.

The Justice Department says, or at least implies, that these proprietary purchases were front-running. It's an odd claim. The FX market is big. The U.S. dollar/British pound cross trades something like $472 billion a day. These guys ran FX trading at a big U.K. bank. They bought pounds in the week before a $3.5 billion client transaction. It would be a little weird if they hadn't bought pounds -- their home currency! -- for a week. And while of course foreign-exchange rates are affected by big corporate cross-border flows, it would be a little weird if knowing about one $3.5 billion corporate transaction -- in the trillions of dollars of dollar/pound transactions in any week -- would be enough to predict the future price of the pound over that week.  

More generally, there is a reason that "insider trading" isn't really a thing in the FX world. These guys trade pounds all day. They know that their clients are buying and selling pounds. If they were restricted from trading any time they knew that there might be a big client trade in the next week, they could never trade. Front-running a live order is bad, but just buying pounds when you know a client might later want to buy pounds seems ... inevitable? 

But this is the less interesting claim. More important, and more damaging, is what Johnson and Scott allegedly did in the hour before Cairn's trade, when they did have a live order. Actually the problem started when they were finalizing the order. Cairn wanted to buy at the fix, the hourly benchmark FX price. But which fix?

Supervisor 1 arranged for a call at approximately 1:35 PM London time between the Victim Company, the Advisor, Supervisor 1 and the defendants MARK JOHNSON and STUART SCOTT to discuss whether the Victim Company FX Transaction should be executed at the 3 PM fix or the 4 PM fix. Both SCOTT and JOHNSON knew that because there was less liquidity at the 3 PM fix, currency prices at that earlier time were easier to manipulate than prices at the 4 PM fix, so it was advantageous to them and HSBC, and disadvantageous to the Victim Company, to execute the Victim Company FX Transaction at the 3 PM fix.

Initially during the December 7, 2011 call, the defendant STUART SCOTT falsely and fraudulently suggested to the Victim Company that the 3 PM fix had more liquidity than the 4 PM fix. When confronted by the Advisor about that assertion, SCOTT falsely and fraudulently stated that the fixes were the same in terms of liquidity. SCOTT then stated there was more volatility at the 4 PM fix and the defendant MARK JOHNSON stated that he "personally would recommend" the 3 PM fix "so there's an element of surprise." SCOTT further stated that: "That's an excellent point, actually, yeah. Because people do look for that, for the significant flows to happen at 4 o'clock and once they get a smell of that or a smell of significant flow going through, they will try to jump in front and start to muck around in the markets."

That is: The 4 p.m. fix is the most liquid trading window, so it's the "right" time to exchange dollars for pounds, so that's just what they'll expect you to do. So you should do your trading at the "wrong" time, when no one else is trading, so no one will notice. 

Look, when you read it in a criminal complaint, obviously it sounds bad.  

In any case, the HSBC guys talked Cairn and its adviser into using the 3 p.m. fix, and got the order. They were pleased:

For example, during a consensually recorded phone call at approximately 2:28 PM London time, JOHNSON commented to SCOTT, "Seems that they're starting to bite," in reference to the Victim Company's orders. In response, SCOTT stated, "full amount" (indicating that the Victim Company had authorized the full order of 2.25 billion Sterling). JOHNSON then responded "No, you're kidding?" SCOTT then re-confirmed that Victim Company had indeed authorized the full purchase, to which JOHNSON replied, "Ohhhh, f***ing Christmas."

Indeed. Knowing that a client might buy 2.25 billion pounds in the next week is an interesting data point. Knowing at 2:28 p.m. that the client will buy 2.25 billion pounds at exactly 3 p.m. is, you know, Christmas.

We have talked about the FX fixes before. Dan Davies has written the essential explanation of the FX-fixing problem, which if you have not read you should go read now and come back. The idea is that HSBC was on the hook to sell Cairn approximately 2.25 billion pounds at exactly the market price at 3 p.m. This was not something that HSBC could do on an agency basis: Johnson and Scott couldn't just go to the FX store at 3 p.m. on the dot and buy 2.25 billion pounds to turn around and sell to Cairn. Instead, HSBC was at risk, as a principal, for that 3 p.m. price. So it had to buy 2.25 billion pounds before or after 3 p.m., for its own account, and hope that the prices that it paid for those pounds would be lower than the price it got from Cairn.

And so it was perfectly legitimate for HSBC to trade ahead of Cairn's order before the 3 p.m. fixing. I guess I should say that again in italics. It was perfectly legitimate for HSBC to trade ahead of Cairn's order before the 3 p.m. fixing. Don't take my word for it. Here's the U.K. Financial Conduct Authority, in an order fining HSBC for currency manipulation:

A firm legitimately managing the risk arising from its net client orders at the fix rate may make a profit or a loss from its associated trading in the market. Such trading can potentially influence the fix rate. For example, a firm buying a large volume of currency in the market just before or during the fix may cause the fix rate to move higher. This gives rise to a potential conflict of interest between a firm and its clients.

There is a conflict of interest, but it is a legitimate and inevitable conflict of interest. HSBC was selling pounds, as principal. Cairn was buying them. HSBC had to get them from somewhere. If it got them cheaper than it sold them, it made a profit; if it didn't, it had a loss. Everyone involved expected HSBC to try to make a profit. How it tried to make a profit was its business.

In this case, it was pretty easy for Johnson and Scott to figure out how to make a profit. They knew they had a giant buy order for 3 p.m. The 3 p.m. fix wasn't all that liquid. They could be pretty sure that, if they bought a lot during the half-hour before the fix, that would push the price up. The math is simple: If you start buying at 2:30 p.m. at $1.569, and you push the price up smoothly and finish buying at 3 p.m. at $1.571, then you will probably have paid an average of about $1.57 for your purchases. And then, at 3 p.m., you sell all of your pounds for $1.571, for a tenth of a cent of profit. (These numbers are made up, though of the right order of magnitude: "HSBC gained approximately $5,000,000 from its execution of the Victim Company FX Transactions," or about 0.14 cents per dollar of the $3.5 billion transaction. )

The fact that some form of this conflict is inevitable and legitimate doesn't mean that it was handled well here, or that Johnson and Scott didn't front-run. There is bad stuff in the complaint! The push for the 3 p.m. fix instead of the 4 p.m., and the gloating about Christmas, don't look great. Even worse is what they told the client as the price went up:

When the Victim Company and the Advisor observed upward movement in the price of Sterling between 2:00 PM and 3:00 PM London time, they questioned Supervisor 1 about these price movements. At approximately 2:45 PM London time, Supervisor 1 told the defendants MARK JOHNSON and STUART SCOTT that the Victim Company was calling at "every uptick" in reference to the price of Sterling. Finally, just after 3:00 PM London time, Supervisor 1 told JOHNSON and SCOTT that he had told the Victim Company that "a Russian name" was buying at the same time as the Victim Company.

Always a great idea to blame anonymous Russians. Later, Scott allegedly "represented to the Victim Company and the Advisor that the Victim Company FX Transaction went 'okay' despite an 'initial jump' in the price which he falsely and fraudulently attributed to trading by a Russian bank," and told them "that HSBC began 'taking action' in the FX market approximately five minutes prior to the 3 PM fix," instead of the much longer period during which HSBC was actually trading. 

This stuff is not good. Lying to the customer isn't good service. It might even be fraud. More than that, Johnson and Scott's execution itself -- their trading in advance of the fix in a way that pushed up the price -- might also be fraud, or front-running. Legitimate risk management to try to make a reasonable profit on a principal trade with a customer is one thing; intentionally "ramping" up the price to worsen the execution and take advantage of the client is another.

I mean, I couldn't tell you offhand what the difference is between them! They'd look similar, trading-wise. Either way, HSBC was trying to buy pounds for its account before the fix, to sell to Cairn at a profit. But I can imagine that things like saying "Ohhhh, f***ing Christmas" before the trade, and lying to the client about what happened afterwards, might be reasonable indications that you've moved beyond legitimate risk management into shady "ramping" of the price. I suspect that a jury would take them that way.

But all of this exists on a continuum. The essential allegations here are that the HSBC traders convinced a client to give them an inefficient order, and then executed it inefficiently, in a way that made a lot of money for HSBC and cost the client a similar amount of money. But that is the nature of trading: HSBC was dealing with Cairn as a principal, and of course it preferred to do a trade that made it more money rather than one that made it less money. If you showed traders this complaint, some would tell you that Johnson and Scott did a masterful job of convincing the customer to do a good trade for the bank, and others would tell you that they committed criminal fraud. And many would tell you that Johnson and Scott were piggish and shady, but that their piggishness was within the bounds of typical trader behavior, the sort of ordinary bluffing and advantage-seeking that happens on trading desks all the time. And they would be surprised to learn that Johnson and Scott might end up in prison for it.

But that is the deal, these days. Different financial markets have, over long periods of time, evolved standards of behavior that are not, perhaps, entirely honest in the most traditional sense of the word.  To outsiders -- and to some less experienced participants in those markets -- those standards can seem shocking, even criminal. To regular participants, they can seem normal, even admirable. Johnson and Scott weren't paid by HSBC to mindlessly exchange dollars for pounds. They were paid to make money for HSBC on its trades with customers. That's a tricky job, and they brought the traditional tricks of traders to bear on it. But times have changed: Prosecutors are looking at a lot of trading tricks -- like bond traders being squirrelly about the price they paid for bonds -- and they don't like what they see. And the argument that this was standard behavior, or near-standard behavior, or that many traders saw nothing wrong with it, doesn't seem to win much sympathy from prosecutors.

  1. In fact, while I don't know exactly what prices they paid, per Bloomberg data the closing price on Nov. 30, 2011, when Johnson first bought with his alleged inside information, was $1.5704 per pound. The pound then closed below $1.5704 each day until Dec. 7, when it closed at $1.5710. The next day it closed at $1.5629. So it worked out for Johnson -- the pound was just a bit higher on the day he sold than on the day he bought -- but only barely.

  2. Is the government's theory here that any time a bank recommends trading at the 3 p.m. fix, that's fraud? Surely some people voluntarily trade at the 3 p.m. fix, no?

  3. Strictly, the fix is calculated based on a trading window around the fix time, but we can ignore that here. See the methodology here.

  4. Or some before and some after, or some before and some after and some at 3 p.m. on the dot, or sell from inventory, or go short, etc.

  5. It also made $3 million on Johnson's and Scott's earlier prop trades in sterling, for a total of $8 million.

  6. Trading over a half hour, instead of the five minutes that they allegedly told Cairn, might also have been intended to move the price more -- though of course sometimes it's the other way, and jamming a lot of transactions in right before the fix is more manipulative.

  7. I mean, $8 million on a $3.5 billion at-risk transaction, which is a lot of money for a trade in a liquid currency pair.

  8. I am not sure that they are all that different from the standards of behavior in other, non-financial businesses. When the appliance salesman pushes that extended warranty, is that really so different from an FX trader pushing the 3 p.m. fix? The general nature of sales is that you pretend to have your customer's best interests at heart, but you are always thinking at least a little about your own.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

To contact the author of this story:
Matt Levine at mlevine51@bloomberg.net

To contact the editor responsible for this story:
James Greiff at jgreiff@bloomberg.net