Money Stuff

Swap Traders and Stop Orders

Also Martin Shkreli, private equity vs. activism, coders, Vine and non-GAAP accounting.


I guess the new normal in banking is that every time a trading desk makes a little money, it's going to be news, and they'll get an article congratulating them for it. It's like when the employee lottery pool at the toy factory wins the Powerball, and there's a little profile about them in the local paper. Except it's interest-rate swap trading or whatever:

A team of Citigroup Inc. derivatives traders generated about $300 million of revenue this year, thriving from serving companies and investors trying to anticipate central bank decisions, according to people with direct knowledge of the matter.

The windfall was produced by the bank’s U.S. dollar interest-rate swaps desk led by Geoff Weber in New York, according to the people, who asked not to be identified because the firm doesn’t break out results for such businesses. The group includes Dan Leadbetter, Daniel Gottlander and Mark Zaguskin, one person said.

These profiles are not pure jolly human-interest stories, though, because they are tinged with a bit of doubt about the Volcker Rule. Is this prop trading? How can a bank make $300 million in trading if it's not "prop"? And the answer is always sort of obvious, though not universally satisfying. Clients want to do interest-rate swaps. There's a lot of demand. Swaps aren't quite like stocks; you don't buy 100 swaps from a customer one second and sell them to another customer the next. They are principal trades, often with terms of many years; you enter into a swap and then you just ... have it. You hedge it -- perhaps by entering partially or entirely offsetting swaps with other counterparties, or by trading Treasuries or interest-rate futures or -- but you will likely be left with some residual risk, and that risk could be profitable. (Or: risky.) Also of course you just charge clients for the market-making service, and if you do a lot of trades you'll make a lot of money just on bid/ask spreads. And Citi has seen a lot of customer demand, so it's done a lot of trades:

Weber’s team also has been minting revenue tied to the bank’s handling of corporate bond deals by working with clients issuing the debt, according to the people. In such cases, companies typically sell debt to investors with a fixed interest rate, then enter a contract with the bank to pay a variable rate.

The last time we did this -- when Goldman trader Tom Malafronte was reported to have made $100 million trading high-yield bonds -- I said that just buying some bonds and making money when their price went up wasn't in itself prop trading. It depends how you came by those bonds. "If you go out looking for bonds to buy, that is prop," I said. "If the bonds come to you, that is market making." 

But that was oversimplified, and you can see one reason why in the interest-rate swaps case. Clients didn't just wander into Citigroup's lobby looking to buy swaps. Citi's industry coverage bankers built relationships with those corporate issuers, and its corporate bankers loaned them money, and its mergers-and-acquisitions bankers convinced them to do acquisitions, and its debt capital markets bankers convinced them to sell bonds to fund those acquisitions, and its derivatives salespeople tagged along to those meetings to convince them to swap those bonds to floating. And then, after all that effort, the client entered into a big swap trade with Citi on the other side, taking the risk for its own account. That's how derivative dealing works: You think up a trade you'd like to do, you go out and convince a customer to do that trade, and then you take the risk on the other side of the trade. (And, normally, hedge.) The essential reason that it is permitted "market making" rather than forbidden "proprietary trading" is that a customer wanted to do the trade -- but there is no rule against the bank working really hard to convince the customer that that's what it wants. 

Stop orders!

Still a thing:

Some currency investors, worried that events such as the pound’s recent collapse could become more common, are abandoning an automated trading mechanism that was supposed to protect them from losses after big market swings.

The trading mechanism, known as a stop-loss order, is meant to sell an investor’s position as soon as a currency hits a preset level against another currency. Banks and electronic trading platforms often struggle to execute the orders at exactly the level an investor requested, resulting in trades below the indicated price even when markets are relatively calm.

One plausible claim about our modern electronic marketplaces is that they increase the chances of a "flash crash" relative to a "real crash." (Certainly they have increased the amount of attention paid to flash crashes versus real crashes.) That is, every so often a stock or even a currency will drop rapidly by 10 percent. Sometimes it stays there: Earnings were bad, or the currency was unpegged, or whatever. Sometimes it rebounds within two seconds, and everyone scratches their heads, and two years later the market regulator releases a big report about what happened. The report is rarely all that edifying, though it often mentions stop orders.

In the first case -- the real crash -- if you sold halfway down you are thrilled. The thing was worth $100 before, and it's worth $90 now, and you sold at $95. Good risk management; your $98 stop order saved you. In the second case -- the flash crash --if you sold halfway down you are furious. The thing was worth $100 before, and it's worth $100 now, and you sold at $95 during the five dumb seconds that the market freaked out. Your $98 stop order increased your risk.

I guess if you are super worried about real crashes you should put in stop orders: If your stop gets triggered at $98, but only executes at $95, that's still some protection. If you are super worried about flash crashes you should put in stop-limit orders: If your stop gets triggered at $98 with a $98 limit, then it will only execute if you can actually sell at $98 -- which is fine if the thing rebounds, but bad if it drops to $90 and stays there. As the risk, real and imagined, of flash crashes goes up relative to the risk of real crashes, you'd expect people to want to move away from stop orders that could end up looking foolish in flash crashes.

All of this is pretty much true of market orders too, by the way. Elsewhere: "Does High-Frequency Trading Increase Systemic Risk?" And: "U.S. authorities are investigating whether middlemen in the $5 trillion-a-day foreign exchange market posted fake bids, rigged auctions and gave confidential information to others who then profited from it."

Lunch with Shkreli!

Martin Shkreli had lunch with the Financial Times, and he didn't order wine, which is the one unforgivable sin you can commit with a journalist. ("I tuck into the spaghetti, which is delicious but would be better still with something from the wine list, and I silently bemoan that daytime drinking is not among Shkreli’s flaws," writes the FT's David Crow.) He discussed his pending fraud case, which he thinks he'll win because he's notorious (which is subtly different from thinking he'll win because he's innocent), but he's cool with losing too:

He thinks any prison sentence would be short — a year or so — and that he would be popular with his cellmates. “One person told me that the inmates at the white-collar prison are gigantic, huge Shkreli supporters,” he says.

He wins all the best popularity contests. He also discussed his rationale for jacking up the price of lifesaving drug Daraprim, in terms that are ... I think ... slightly wrong?

He boasts of other attempts to buy old drugs for fatal diseases with the “ingenious plan” of inflating their prices as well, and suggests that executives who eschew such tactics are, in effect, defrauding their investors. “If you have a drug that is $100 for one course of therapy, and you know that you can charge $100,000, what should shareholders think when you say, ‘I’d rather not take the heat’?” he asks.

The idea that a corporate executive's only obligation is to maximize shareholder profit, and that he stops being a human being and a member of society when he takes on his role as an executive, doesn't really have much support in law or tradition. Shareholders can't, as a general rule, sue managers for failing to price gouge. 

But as a matter of market forces, that theory may be right? Like if everyone else is raising prices for proprietary drugs, and you aren't, then your relative performance will suffer. Your competitors will get rich and buy your company, or an activist will launch a proxy fight to try to push you out in favor of a more cutthroat executive. If the shareholders have the ultimate control over the company, and if enough other executives at other companies have convinced the shareholders that they are entitled to ruthless amoral price maximization, then that's what they'll demand from every company. But to get that result you need people like Shkreli -- or, I guess, like Shkreli but less cartoonish -- out on the front lines, pushing the shareholder value theory. “My whole life has been one theme of self-sacrifice for my investors,” he says. “I did it for my shareholders’ benefit because that’s my job. The political risk is being shamed — and shame isn’t dilutive to earnings per share.”

Private equity and activism.

There is mission creep, both ways:

Activists, who traditionally buy up minority stakes then agitate for change, are increasingly attempting total takeovers or acquiring stakes at the request of management teams looking for guidance. Private-equity firms, which typically do friendly buyouts, are increasingly buying noncontrolling stakes in public companies, and they aren’t always letting management in on their plans ahead of time.

My favorite bit of mission creep may be this:

KKR has quietly invested $1 billion in “toehold” stakes in public companies over the past two years in an effort to spark divestitures or gain an edge for an eventual takeover.

Back when "corporate raider" was sort of an undivided category, you'd buy a toehold and then agitate for the company to be sold to you, or be sold to someone else, or just pay you a chunk of "greenmail" to go away. And then we developed a clear distinction between activism, where you'd buy the stake but not the whole company, and private equity, where you'd buy the whole company but not start with a surprise stake accumulated in the market. But in a strictly economic sense, if you are going to pay a premium for 100 percent of a company's stock, it always kind of makes sense to buy 10 percent at the market price first. Bill Ackman revived the toehold business in the Allergan fight a couple of years ago, and it's nice to see it creeping back into the private-equity playbook.


Here's a good article about the chummy, elitist old boys' club on Wall Street:

So last year, Furlong, 30, enrolled in a three-month coding boot camp that uses HackerRank, a web platform that trains and grades people on writing computer code. After earning a top ranking for Java developers globally, Furlong was hired by JPMorgan Chase & Co. in December for its two-year technology training program.

I suppose it is better to hire people based on a transparent meritocratic open-access online test of the actual skills that you'd use in your job than based on, you know, where you played lacrosse. Or I guess you can do both; there is a lot of demand for programmers. "By using algorithms to spot talented coders, HackerRank and competitors with names like Codility claim they’ve essentially increased the world’s supply of developers." It seems sort of unpleasantly efficient though. The old system had some slack built into it. You'd hire someone, and he wouldn't be great, and you'd say "oh well, my bad," and expand your team a little to make up for the dead weight. In the future, if relevant skills are transparently measurable on the internet, then you can employ only useful people, and only exactly as long as they're useful. 

Elsewhere: "How Pricing Bots Could Form Cartels and Make Things More Expensive." And: "Google’s neural networks invent their own encryption."

RIP Vine.

Yesterday Twitter, Inc. announced that it will no longer develop Vine, the app for making weird six-second videos that no one really used anymore. I am old enough to remember when the promise of the internet was the "long tail": If something was cool, and some people liked it, it could exist indefinitely on the internet even if it never captured (or captured and then lost) a mass audience. It seems a little like the rise of Facebook has changed that. Now we all know that a social media platform can become a dominant force in business and technology and politics and culture, and also that a dominant social media platform can be overthrown by an upstart replacement. So there is pressure for the big companies to absorb new social and sharing platforms that might pose a threat to them or give them a competitive advantage: Facebook buying WhatsApp and Instagram, Google buying YouTube, even poor little Twitter buying poor little Vine. And once a sharing platform is part of a big social network, it has to have mass appeal to justify the big company keeping it around. Vine was cool, got acquired, lost its mass appeal, and now it's gone. I guess my main point is that I miss Google Reader. Vine was cool too, here are some good vines.

Elsewhere, Twitter "has sketched out a restructuring plan aimed at becoming profitable by next year," which starts with laying off about 9 percent of its employees. Apple has a new virtual keyboard. Microsoft has a new desktop PC. Honestly I miss MS-DOS with some frequency, I should probably never be allowed to write about technology.

People are worried about covered interest parity.

Big week for covered interest parity! Here's another paper from the Bank for International Settlements, on "The failure of covered interest parity: FX hedging demand and costly balance sheets."

People are worried about non-GAAP accounting.

The Securities and Exchange Commission is worried:

A primary focus of the SEC’s latest inquiries is whether companies have featured customized measures too prominently in earnings releases and other disclosures, the people familiar with the matter said. It isn’t known how many companies the SEC is targeting, or when or whether it might take any action.

“The SEC has made it clear it is serious about trying to put at least some of the non-GAAP genie back in the bottle,” said David Trainer, chief executives of New Constructs, an investment-research firm.

I sometimes make fun of this worry because companies are always required to disclose their results under generally accepted accounting principles too. So the theory of non-GAAP worriers has to be something like:

  1. GAAP accounts are right and non-GAAP accounts are wrong.
  2. Investors are too dumb to know this: If you tell them the non-GAAP numbers, they will be too distracted to look at the GAAP numbers in the same press release.
  3. So at least put the GAAP numbers in a bigger font or whatever.

This theory seems a bit rickety, but it is more understandable coming from the SEC than from most non-GAAP critics. Part of the SEC's job really is to regulate the formatting and font sizes of financial disclosures; it has a readability mandate as well as a substantive one. So it's within its rights to criticize the style of corporate press releases.

People are worried about unicorns.

Welp, "the biggest U.S. initial public offering this year," ZTO Express Inc., fell in its trading debut, closing 15 percent below its IPO price in its first day of trading yesterday. Oh well! Time for the unicorns to flee back into the Enchanted Forest; it's colder out here than they expected. 

Also, we talked yesterday about Yeti Holdings Inc., which has been working on an initial public offering, and this is the point in Money Stuff where traditionally I would show you a reader's rendition of a yeti unicorn. So here's Prentiss Nelson's:

yeti unicorn 2

People are worried about stock buybacks.

Google -- sorry, sorry, Alphabet Inc. -- announced yesterday that it is authorized to buy back up to $1,000,000 x 26^e worth of stock, which is about $7 billion if you don't feel like dealing with exponents this early in the morning. This follows a string of other, um, jokes -- its last buyback authorization was for $1,000,000,000 x sqrt(26), and it once did a secondary offering of 100,000,000 x (pi - 3) shares -- that are funny if and only if you think numbers like 26 and e and pi are funny? I don't know. I hope there is an Alphabet employee whose whole job is thinking up corporate-finance math jokes, and I hope she gets paid Avogadro's number of pennies per billion galactic years. Elsewhere: "A More Money-Conscious Alphabet Emerges."

People are worried about bond market liquidity.

Here is Alexandra Scaggs on the regulation of Treasury traders versus Treasury dealers, a curlicue of worry in the broader worrying about Treasury market liquidity.

Things happen.

"Active large-capitalization funds topped their benchmarks by more than one percentage point from July through September." Merger Deals Set Monthly Record, Even as Election Looms. Barclays Is Trying to Cap U.S. Mortgage Penalty at $2 Billion. UBS Profit Falls as Risk-Averse Clients Curb Revenue. Carlyle Group Turns Focus From Hedge Funds to LendingInflation Fear Fuels Bond Rout. Rising defaults cast cloud over US junk bond surge. Gannett-Tronc Deal in Doubt as Banks Pull Financing. General Electric Pursues Deal With Baker Hughes. Why Wall Street’s Top Cop Criticized by Warren Could Stay Awhile. Could China fork Bitcoin? U.S. charges 61 over India-based impersonation scam. Vikings coach Mike Zimmer refutes reports about slaughtering stuffed animals. Florida man driving home from strip club falls out of truck and runs himself over. Sexy Halloween costumes. Ex-government worker allegedly used county credit card to buy a tuxedo for her dog

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

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

    To contact the editor responsible for this story:
    James Greiff at

    Before it's here, it's on the Bloomberg Terminal.