Goldman FX Trader Was Loyal to His Chat Room

(Bloomberg) -- FX badness.

Yesterday Goldman Sachs Group Inc. agreed to pay about $110 million for doing bad things in foreign-exchange trading, paying about $55 million each to the Federal Reserve and the New York State Department of Financial Services. (Disclosure: I used to work at Goldman.) This is the same complex of FX badness that we’ve talked about before — basically, a bunch of traders at all the big banks had chat rooms where they shared confidential information about customer trades, allowing the banks to coordinate their trading to profit at the expense of the customers — and there’s nothing really new here substantively.

On the other hand there are more fun dumb trader chats so I should probably quote generously from those. From the New York DFS consent order:

For example, one Goldman Sachs trader, Trader 1, participated in chat rooms where code names were used to discreetly share confidential customer information. Traders in this chat room, for example, referred to certain customers as the “fiddler,” “hat and coat,” and “dodgy aussie seller.

Okay I feel like “fiddler” and “hat and coat” are good code names, you could see a spy using them, but “dodgy aussie seller” is less a code name and more just an insulting description? It’s like a spy using the code name “annoying espionage agent.” I am not going to allow it, as a code name.

Another example of improper sharing of customer information involved a customer apparently located at another large bank; Trader 1 and others repeatedly referred to this customer as “Satan.”

Now that’s a code name!

In a number of chats, an FX trader at another bank, Trader 2, provided specific information to chat room participants about this customer’s trading. At one point, Trader 1 told Trader 2: I “I remember the old days, your satan info was legendary.”

In another chat from August 2008, traders at other banks goaded Trader 1 to provide current information about the customer code-named “Satan,” since “[Satan] does all his spot with u on yr toy now.” Trader 1 quickly replied by providing confidential information concerning a recent specific trade of Satan: “satan sells 8 euros at 17,” indicating the code-named customer sold $8 million worth of the Euro/US dollar at an identified price. Another trader, Trader 3, responded “good shout,” while Trader 2 similarly replied, “[Trader 1] finally adds value.”

Finally! Don't underestimate this sort of hazing and belittling as a motivation. There you are, in a chat room with your competitors, not providing confidential customer information that would allow your competitors to make a profit at your customer’s expense. What a sensible thing not to do! From the perspective of your employer, your customer, society, the Federal Reserve, the New York Department of Financial Services, etc., of course you shouldn’t disclose that confidential information in that chat. But from the perspective of the chat room, of course you should. And you’re in the chat room. There it is, blinking on your screen, a bunch of dudes who give out nicknames like “Satan,” making fun of you for not adding any value. The demands of your employer, your customer, society, the Fed, they all recede, as the chat room blinks at you insultingly. Anyone who has been on the internet can appreciate this sort of pressure, and where it leads. “Satan sells 8 euros at 17,” you type, and your chat beeps as your buddies praise you, and you feel so good for a little while. And then so bad for a long while, but in the moment, it was the most natural thing in the world to type.

Apparently others at the firm knew about these dumb little chats, but were under the impression that Trader 1 was only getting illicit information from them about other banks’ customers, not giving information about Goldman’s customers. (To be fair Trader 3, the chat-room buddy from another bank, seems to have been under the same impression: Trader 1 really wasn’t adding value that often.) But “a senior member of Goldman Sachs’ Global FX Sales Division” asked Trader 1 to really think about what he was doing:

I would like you guys to give it some thought first please. The question stands: “Why do the people that you talk to seem to give you so much clearly improper information week after week, month after month, and year after year? … Are they stupid? Are they getting something from you by keeping you engaged? Are they very poorly trying to impress you to get hired?” … I am working under the assumption that they are getting something or they would not keep doing it, so I believe it is worthwhile our stepping back and making sure that we are not showing our hands in some way that might not be obvious.

“I do not expect that anyone is doing anything malicious or intentionally wrong from your desk,” continued the senior salesperson, but “I do expect that we are being used in some way that we may not expect.” It seems, though, that they were being used in the most obvious and expected way. They came to the chat, they got information, they formed bonds with the people in the chat, they were peer-pressured into sharing information, and they did.

Eventually lots of kinds of trading will become more electronic; FX trading, in particular, has become more electronic both for efficiency reasons and because of regulatory problems like this. Sometimes the human traders were doing shady and disreputable things, and there is a hope that the computers will put a stop to that. My suspicion is that in many cases the computers will just replicate and formalize the shady things, turning nebulous gray-area “market color” information leakage into explicit algorithms. If humans do shady things, and humans design the algorithms, then the algorithms will just do the shady things faster.

Still there are good things to be said for the computers. They seem unlikely to fall for this particular sort of social engineering, anyway. An algorithm told by its programmers not to share information with rival banks just won’t share information with rival banks. It’s going to follow its instructions; it’s not going to share the information to impress its little algorithm friends in their little algorithm chat room. Some kinds of financial misconduct come from particularly human frailty, and can be eliminated if you get rid of the humans.


Here is a story, from Nate Anderson of Hindenburg Research at ValueWalk, about a weird deal. Inpixon is a wee public company ($5.5 million of equity market capitalization as of yesterday’s close) that does some sort of tech thing. (“Indoor positioning,” apparently, which coincidentally is a good description of how I spend my time.) Like several small public companies that do some sort of tech thing, Inpixon decided to announce that it would start doing its tech thing on the blockchain. It made that announcement (“it will use blockchain to develop a global repository of device reputations to help enterprise security and retailers mitigate rogue-device risk proactively”) on the morning of Jan. 9, and the stock went up 30 percent on extremely heavy volume — over 90 million shares, compared to an average of under 3 million shares traded per day in the previous six months. (The numbers on Bloomberg are lower because they adjust for a subsequent 1-for-30 reverse stock split.) This is not exactly a surprise; the basic rule of early-2018 financial markets was that if a public company slipped the word “blockchain” into a press release, its stock would rise on heavy volume.

So I am sure it did not come as a surprise to the people who bought Inpixon’s stock the day before the announcement. Because the day before Inpixon’s blockchain announcement, it sold almost 18 million shares of stock, for almost $3.2 million, in a “registered direct” offering (in which it registered a public offering of shares but only sold them to a few selected investors) run by its bankers, Roth Capital Partners. Those investors were “wall-crossed” on the blockchain press release: The company gave them the press release before it was public, and they agreed not to disclose or trade on it until it was made public. (The Securities Purchase Agreement that they signed on Jan. 8 notes that they received the press release in advance, and commits the company to make it public by 9 a.m. on Jan. 9.) They got advance notice of Inpixon’s pivot to blockchain, they bought stock from the company — at a 19 percent discount to the previous day’s price — knowing about the pivot, and then the company announced the pivot and the stock went up 30 percent. (The stock has since lost more than 90 percent of its value; the blockchain pivot only works for so long.) 

It’s a nice trade for the investors! Anderson thinks that it is somehow akin to insider trading: “If this kind of deal is permitted, Roth may have succeeded at pioneering a legal form of capital raising that is more egregious than any illegal insider trading scheme we could ever conceive of.” This is confused. There is no insider trading here. At every point when any trading occurred, everyone on both sides of the trade was aware of the blockchain announcement. When Inpixon sold shares to its investors, both Inpixon and the investors knew about it: Inpixon because it was the one issuing the press release, and the investors because Inpixon had wall-crossed them to show them the press release. And when those investors were free to sell those shares publicly, both they and the public knew about it: By the time the investors could trade, the press release was public and everyone knew about the blockchain stuff. That’s why the stock went up! The problem for public investors isn’t that the wall-crossed investors knew about the blockchain stuff and they didn’t. The problem was that the public investors did know about the blockchain stuff, and bought shares from the wall-crossed investors because of it, and those shares soon collapsed. 

It is not insider trading, and yet it feels … bad? So what is it? Well, start from the fact that it is a bit odd, if you are going to announce predictably positive news on a Tuesday morning, to sell a big chunk of your stock at a discount on Monday. If Inpixon had not done a wall-crossed registered direct offering, but instead waited until Tuesday to do a public offering after announcing the blockchain news, then it could have gotten the benefit of that news. After announcing the blockchainery, it could have sold stock at a premium to Monday’s close, rather than a discount. Right?

No, wrong, come on. If Inpixon had announced simultaneously (1) “we are pivoting to blockchain” and (2) “want to buy a massive slug of stock?,” that would have been too cute even in the blockchain-saturated era that was January 2018. (“If you tell public investors simultaneously that (1) you have discovered a cure for cancer or whatever and (2) you are selling a bunch of your stock, they are going to get a little suspicious,” I wrote recently about Longfin Corp., another blockchain-pivoter that sold stock to the public using unusual methods. And so, incidentally, will the Securities and Exchange Commission.) Instead, it sequenced those transactions: It sold the stock to a few selected investors before the announcement, and then gave them the benefit of the announcement so that they could sell their stock. (The investors signed a “Leak-Out Agreement” limiting them to selling a combined 35 percent of Inpixon’s daily volume, but the post-blockchain spike in volume would have helped with that.) The net effect of this transaction is that Inpixon was able to sell a big slug of its stock at a price pretty close (down 19 percent) to its last market trade —which is no mean feat for a micro-cap company. Inpixon got a bunch of money for its questionably-valued stock; the investors got to buy stock at a discount knowing that there would be a tailwind of blockchain hype that would let them sell. One assumes they took advantage, and sold.

Probably you should think of the wall-crossed investors as, essentially, middlemen: Inpixon could announce “we have pivoted to blockchain and have sold a bunch of stock,” which is a much stronger move than "we have pivoted to blockchain and would like to sell a bunch of stock, is anyone interested?” The wall-crossed investors took on the risk of selling the stock for Inpixon, and got a discount for their troubles. But the public investors who bought the stock from the wall-crossed investors, but who would have thought twice about buying the stock directly from the company — who thought the news was “pivot to blockchain,” not “big dilutive stock offering” — might feel a bit aggrieved.

Old-timey CDS orphaning.

I wrote yesterday about a delightful credit-default-swap trade involving hedge fund Chatham Asset Management and newspaper company McClatchy Co. In the course of defending that trade, I described a dumb hypothetical trade in which a hedge fund (1) sells a lot of CDS on a company, (2) buys all the bonds of the company and (3) just pays off the bonds itself. That way it loses all the money it invested in the bonds, but on the other hand it pockets all the CDS premium. It’s a dumb trade because, if the facts are anything like they are in McClatchy, you’d spend a lot more buying the bonds than you’d receive in CDS premium, so you’d have an overall loss on the trade.

But if you could sell way more CDS than there were bonds, then you might have something there. That’s unusual — usually companies with lots of CDS also have lots of bonds — but it is not impossible. A couple of people have recently reminded me of a terrific 2009 trade in which Amherst Holdings sold a lot of CDS on a subprime mortgage bond deal and then paid off all of the underlying mortgages, rendering the CDS worthless. “At one point, at least $130 million of bets had been made on the performance of around $27 million in securities,” with the CDS selling for around 80 or 90 cents on the dollar, so there was plenty of room to profit from this simple strategy. And Amherst did.

“Maybe CDS buyers would be a bit aggrieved,” I wrote about this idea yesterday, “but they’d really have no leg to stand on.” They did complain though:

Firms that suffered losses as well as some that didn't have brought the trade to the attention of two financial industry groups, the Securities Industry and Financial Markets Association, and the American Securitization Forum, which are considering their concerns, say people familiar with the trade groups' thinking.

And here we are, nine years later. It’s possible that some of the predictions that clever CDS trades will destroy the CDS market might be a little overblown.

Oh by the way, if you are a CDS holder in McClatchy who is aggrieved that the proposed refinancing will leave your CDS worthless, you could always go to McClatchy and offer it a better refinancing deal than Chatham is offering, in exchange for it not rendering your CDS worthless. Even better, though, you could go to McClatchy and offer it a better refinancing deal, in exchange for doing a trivial default on its existing debt that triggers your CDS

That is of course the Hovnanian trade, which is the opposite of the Chatham/McClatchy trade. It is also the Codere trade, which is the opposite of the RadioShack trade, all of which we have talked about before. I wrote once:

Now, Codere was a lovely trade, and RadioShack is a lovely trade, but the world still awaits the Hegelian synthesis of the two. I want to see a financially distressed company with a lot of CDS outstanding run an auction to decide whether to default or not. CDS writers could offer favorable financing terms to keep the company afloat without a default. CDS buyers could offer even more favorable terms to keep the company afloat with a quick harmless Codere-style default. And then, you know, they keep bidding. They're just giving the company each other's money. Whoever wants it more will offer the best terms.

Codere was a trade in which a hedge fund that had bought CDS convinced a company to default on its debt to trigger the CDS; RadioShack was a trade in which a hedge fund that had sold CDS loaned the company money to avoid a default so the CDS wouldn’t pay out. Hovnanian and McClatchy are refinements of those trades: Hovnanian won’t just default to trigger CDS, but will also issue weird new bonds to maximize the CDS payout; McClatchy won’t just get financing from a CDS seller to avoid a default, but will move its new bonds to a subsidiary to render default impossible forever. The Hovnanian strategy maximizes a CDS buyer’s ability to take the CDS sellers’ money (and give the company some of it); the McClatchy strategy maximizes a CDS seller’s ability to take the CDS buyers’ money (and give the company some of it). Both sides have powerful tools to get what they want here, and to share some of it with the company. Now they just need to have an auction; whoever gives the company more of the other side’s money should win.

How’s the asset management business?

Hedge fund professionals [Ricky] Sandler, Dawn Fitzpatrick, Dmitry Balyasny and Andrew Feldstein, speaking on a panel at the Milken Institute Global Conference in Beverly Hills, California, tried to be upbeat about the future of industry. They couldn’t manage for long -- quickly turning sour on fees, investor demands, the surfeit of funds and the cost of doing business.

Before the financial crisis,“we used to be able to get high returns with limited volatility," Sandler said. But investors’ obsession with protecting against downside risk has “sowed the seeds of very mediocre returns and high fees.”

It is a neat tactic: Hedge-fund clients complain all the time that hedge funds offer mediocre returns and high fees, so what if the hedge-fund managers turned the complaint around and argued that … hedge funds offer mediocre returns and high fees … because … that’s what the clients demanded? 

AllianceBernstein Holding LP plans to relocate its headquarters, chief executive and most of its New York staff to Nashville, Tenn., in an attempt to cut costs, according to people familiar with the matter. That largely ends a 51-year presence in the nation’s traditional finance capital.

The move by the money-management giant is part of a broad cost-cutting effort within a firm that for years has been under extreme pressure from the rising popularity of index-tracking funds and low-cost investing.

There is a spectrum of expense and activity in which hedge funds are at one end and index funds are at the other, and I guess active equity management in Nashville is closer to indexing than is active equity management in New York. “We’re practically an index fund,” you can say; “we don’t even have a New York office.” Though “AllianceBernstein’s “money managers and private client business will remain in New York.”

How’s Jeff Bezos doing?

Kevin Lin, co-founder of Amazon-owned video game streaming platform Twitch, told CNBC on Tuesday that Jeff Bezos plays games "here and there" but he wouldn't call him a gamer.

"Is Bezos a gamer? I guess if you count flying spaceships into space, maybe that's a game," said Lin, who sold Twitch to Amazon for nearly $1 billion in 2014.

Who needs video games when you can play with the mechanics of the simulation in which we all live?

Elsewhere in newsletters.

Bloomberg View is about to become Bloomberg Opinion, and it will have a new newsletter: Bloomberg Opinion Today, an afternoon roundup of our writing about business, politics, economics, tech and more. Here is the link to sign up.

Things happen.

Legendary Investment Banker Richard Jenrette Left These 24 Rules for Success. Xerox CEO Exits in Settlement With Activists; Fujifilm Deal at Risk. Banks Forced Into Qatar-Saudi Feud. John Griffin at Bloomberg View: “Does the VIX Need Fixing? Sure Looks That Way.” C.E.O.s Should Lead on Addressing Social Issues, Poll Finds. Seven Things Wall Street Wants to Know From Elon Musk. “In a recent paper looking at the removal of the Swiss franc cap, we find that algorithmic trading provided less liquidity than usual, at worse prices, and that its contribution to efficient pricing dropped to near zero.” China’s Biggest Tech Unicorns Stampede to Go Public. “If you actually straight-line rent expenses, even community-based EBITDA is negative.” Florida Man Arrested For Practicing Karate By Kicking Swans In The Head.

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Thanks! 

To contact the author of this story: Matt Levine at

©2018 Bloomberg L.P.