Photographer: Spencer Platt/Getty Images

UBS Overcharged Itself for Some Derivatives

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

We've talked a lot about how the job of a dealer in financial markets is to find people who will buy stuff from you at a higher price than you paid for the stuff. Most of the people who are buying the stuff from you, meanwhile, have as part of their job description thwarting you in those efforts. It's not their whole job description. They have other things to worry about, like the value of the stuff, plus they might actually want to reward you for your services and be willing to pay you a healthy markup. But it is usually part of their job to avoid paying you a comical markup, and when I write about bond traders who charged their customers comical markups, one reaction that I get is to the effect of "those customers should be fired." 

One great trick, though, is to charge comical markups to customers whose job description doesn't even include getting a good deal. I don't ... I don't exactly want to say that that describes UBS's structured notes business, but hoo boy:

UBS, one of the largest issuers of structured notes in the world, agreed to settle the SEC’s charges that it misled U.S. investors in structured notes tied to the V10 Currency Index with Volatility Cap by falsely stating that the investment relied on a “transparent” and “systematic” currency trading strategy using “market prices” to calculate the financial instruments underlying the index, when undisclosed hedging trades by UBS reduced the index price by about five percent.

That is from Tuesday's $19.5 million Securities and Exchange Commission settlement with UBS over structured notes. Structured notes are a way to sell derivatives to retail investors. Here, UBS took a strategy for trading foreign exchange forwards, wrapped it up in a structured note and sold it to investors.

The strategy was what UBS called the "V10 Currency Index with Volatility Cap." You can read a description on pages PS-18 through PS-23 of this prospectus; be warned that it includes both upper- and lower-case sigmas. Basically, though, it is a currency trading strategy that invests in carry trades (sell low-yielding currencies to buy high-yielding currencies) when volatility is low, and invests in reverse carry trades when volatility is high.  It does this by trading in six-month forward contracts on major currencies. You buy the note, UBS invests your money in the strategy for three years, and at the end, you get back whatever the results are.

UBS calls this strategy an "index," but that is mostly just marketing. It turns out that you can't just issue a structured note that is like "you give us $1,000, we'll invest it however we want, and we'll give you back whatever's left in three years." Instead, UBS specified the investing strategy very precisely, so that if you invested your money in the structured note, you'd know what UBS was up to and could follow along at home. From the SEC order:

In the wake of the financial crisis, UBS perceived that investors interested in diversifying their stock and bond portfolios were attracted to these types of structured products so long as the underlying trading strategy was transparent. In an article discussing FX-based structured products entitled “Pushing frontiers in FX” that was published in The Banker on April 1, 2010, and which referenced the V10, a senior UBS employee was quoted as saying, “[i]n the current market, investors are looking for uncorrelated and highly transparent products and since the algorithmic model is pre-defined and the rules are made available, it is a glass box rather than a black box.”

Almost, though! The "index" was very clearly and exhaustively specified, in that UBS spelled out exactly how much of which currency forwards it would buy in which market conditions. The trading strategy was indeed a glass box. But the actual payoff of the notes depended on the level of the index: The algorithm told you which forwards to buy and sell, but then the ultimate payout depended on whether the prices of those forwards went up or down.

And the prices of those forwards were not so clearly specified. Currency forward contracts don't trade on an exchange with an official reference price; they trade over-the-counter between dealers. And for the purposes of these notes, UBS got to decide what the prices were. After five formula-dense pages describing the index, here is all the prospectus had to say about those prices:

The prices of the foreign exchange forward contracts used to calculate gains and losses from the notional settlement of hypothetical positions underlying the UBS V10 Currency Index shall be based on the prevailing prices available to UBS at 3:00 p.m., London time (or shortly thereafter taking into consideration available prices for the notional foreign exchange forward contracts that would need to be hypothetically settled as a result of such determination), on the relevant valuation date.

In theory, this just means that UBS looks at the market, decides the fair price for the forward contracts and plugs that into the index. That is why UBS says "notional settlement of hypothetical positions": The note is linked to an "index," a more or less observable quantity in the world that is in theory independent of UBS's own trading. The investor is just making a bet on the fair prices of some currencies,  and trusting UBS to compute those fair prices accurately.

But UBS is not making a bet on the fair price of those currencies. You buy this note because you think, roughly, that carry trades will work in the future the way they have in the past. But UBS doesn't sell this note because it thinks that carry trades won't work in the future the way they have in the past. UBS doesn't care what carry trades will do in the future. The point of this note for UBS is to hedge it, so that UBS doesn't lose money no matter what happens to carry trades in the future. UBS is betting that it can hedge this note profitably, which means: It is betting that the actual price of carrying out the transactions described in the "index" will be lower than the price at which it sold the note. The difference is its profit.

Some of the sources of profit are obvious. For one thing, if you paid $10 for one of these notes, you got back only $9.75 worth of stuff; the rest was an underwriting fee. For another thing, if you bought a note, you effectively loaned UBS money for three years, but "UBS paid no interest on the Notes," which is a good deal for UBS.

But a less obvious source of profit comes from the fact that, when UBS calculated the index price, it didn't just use hypothetical fair prices for the underlying currency forwards. It used the actual prices at which it actually bought the actual forwards, both at the beginning and each time the index switched from one set of forwards to another. This made total sense for UBS: The investor was buying a note linked to hypothetical derivatives trades, but UBS was selling a note linked to its own actual derivatives trades, and it wanted its obligation to the investors (the note) to match up with its hedge (the derivatives trades). And this was perfectly legal and covered by UBS's disclosure, which said that the index would be calculated based on "prevailing prices available to UBS," not abstract "fair market" prices.

The problem is that UBS bought those forward contracts from itself: The UBS's London office hedged the structured notes, and it sent hedging orders that "were routed through a Swiss intermediary and then executed by the FX spot desk in Switzerland." So the "prices available to UBS" turned out to be the prices that UBS's FX spot desk offered to its London hedging desk by way of its Swiss intermediary. Guess how that worked out! UBS allegedly bought those forwards from itself at huge markups:

First, between May and October 2010, a UBS employee acting as an intermediary added markups to hedge transactions executed on switch days, which led to prices being used to calculate the Index that were not consistent with market prices. Given the intermediary’s largely administrative function, there was little, if any, legitimate business justification for the amount of these markups. Their impact on the Index was to reduce it by approximately one percent.

Second, between May 2010 and December 2011, the FX spot desk at UBS added spreads, which were not adequately disclosed, to internal transactions undertaken to hedge UBS’s exposure to V10 instruments on switch days. These spreads were not added systematically, but rather were determined largely at the discretion of the FX spot desk. The prices of the hedge trades (including the spreads) were used as inputs to calculate the Index. The impact of the spreads on the Index was to reduce it by approximately four percent. Some UBS employees on the FX spot desk used colorful and sometimes obscene language in electronic communications via chat rooms while executing the V10 hedge transactions.

There's a third alleged problem, involving arguable front-running, but I want to stop here and linger on "colorful and sometimes obscene language in electronic communications via chat rooms," because the SEC does not. It just moves blithely along, not even bothering to mention what the obscene language was. I was under the impression that cases like this stand or fall on the level of obscenity and the frequency of misspellings in the chat rooms. It shows impressive, but disappointing, restraint that the SEC chose not to excerpt any of the colorful conversations here. Were they in German? Was this restraint part of the settlement? Was UBS offered a choice of, like, settle for $17.5 million with disclosure of the chats, or $19.5 million without? 

In any case, a 4 percent markup on simple foreign exchange forwards is obscene enough! I said that there is no single objective reference price for these forwards, but they're not, like, super exotic instruments. You can find them quoted online, or on your Bloomberg. The bid/ask spreads tend to be a fraction of one percent of the notional amount. Four percent is bonkers. Four percent is like airport-currency-kiosk levels. Four percent is the sort of markup that the FX desk charges the structured notes desk because the structured notes desk doesn't care. The structured note just silently and seamlessly passes the cost on to the clients, and doesn't even call it a cost. It's just an input into the "index." And how could you argue with an index?

Structured notes are sort of a notorious place for banks to plunder their clients, and the SEC's announcement of this case links to a recent Investor Bulletin warning investors about their dangers. Some of those dangers are simple and obvious: If you want to plunder your clients, and those clients aren't paying attention, you'll probably be able to plunder them.

But the deeper danger is this "glass box" danger, the disconnect between what the customer thinks the note is and what the bank thinks it is. UBS clearly thought of these notes as being pass-through, hedged transactions: UBS would trade forwards, and would pass the economics of that trading on to the investors in the notes. But the investors thought of the notes as being objective transactions: An index existed, and the investors would get the economics of that index, regardless of what UBS did. And that little gap was disastrous: UBS could charge whatever markup it wanted on its derivatives trades, because the customer didn't even realize that UBS's derivatives markup had anything to do with the notes. As a way to sell the derivatives at a higher price than UBS paid for them, that's pretty effective.

  1. As UBS charmingly describes it:

    Generally, the Index is based on the view that foreign exchange forward rates are biased estimators of future foreign exchange spot rates, and that currencies that trade at a forward discount (i.e., currencies whose forward exchange rate is lower than the current spot exchange rate) tend to outperform, on average and over time, currencies that trade at a forward premium (i.e., currencies whose forward exchange rate is higher than the current spot exchange rate). This is commonly referred to as “forward rate bias.” The Index is also based on a view that this tendency may be reversed during periods of high volatility in foreign exchange markets. While historically there has been some empirical evidence to support these views, the basis for these views is not entirely clear.

  2. Though he's also making a bet that UBS will be around over the next three years to pay off. Much of the controversy about structured notes is about this: They are cheap funding for the bank, and buyers may not be aware that they're taking the bank's credit risk, because the notes say things like "principal protection" or "capital guarantee," and the buyer doesn't understand that that guarantee is only as good as the bank's credit. From an SEC warning about structured notes:

    The retail market for structured notes with principal protection has been growing in recent years. While these products often have reassuring names that include some variant of “principal protection,” “capital guarantee,” “absolute return,” “minimum return” or similar terms, they are not risk-free. Any promise to repay some or all of the money you invest will depend on the creditworthiness of the issuer of the note—meaning you could lose all of your money if the issuer of your note goes bankrupt. Also, some of these products have conditions to the protection or offer only partial protection, so you could lose principal even if the issuer does not go bankrupt. And you typically will receive principal protection from the issuer only if you hold your note until maturity. 

    I will ignore this in the text, both because it's not what's at issue in the UBS case, and because, come on, obviously a bank's bonds are subject to the bank's creditworthiness. But people do get exercised about the credit stuff.

  3. From the SEC order:

    Between December 2009 and November 2010, UBS issued approximately $190 million of Notes linked to the V10. The public offering price was approximately $10 and the underwriting discount was approximately $0.25. Most of the Notes had a three year term. UBS paid no interest on the Notes. Investors were entitled to a cash payment at maturity dependent upon the performance of the V10.

  4. UBS settled "without admitting or denying the SEC’s findings."

  5. That is:

    Third, between May 2010 and August 2011, the FX spot desk at UBS, including the head of the desk, engaged in over two dozen trades in management trading books shortly before executing potentially market-moving, internal V10 hedging transactions, which were directionally consistent with those hedging transactions. UBS did not have in place meaningful controls or restrictions during the Relevant Period over trading ahead of internal V10 hedging transactions.

  6. Like, on the Bloomberg page AUDJPY Curncy FRD, I see the "outright" bid/ask prices for a 6-month Australian dollar/Japanese yen forward at about 85.3856 / 85.4185, for a spread of about 0.04 percent of notional, 1/100 of what UBS was allegedly charging.

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:
Zara Kessler at zkessler@bloomberg.net