Tuesday, August 11, 2015

The Aftermath of LIBOR and Penny-Shaving Attacks

Anyone remember the LIBOR scandal from back in spring 2008? A trader for UBS Group and Citigroup named Tom Hayes was just sentenced by a British court to 14 years imprisonment for his role as a ringleader of the scandal. Darrell Duffie and Jeremy C. Stein discuss both the scandal and--perhaps more interesting to those of us who bleed economics--the economic function of financial market benchmarks in  "Reforming LIBOR and Other Financial Market Benchmarks," in the Spring 2015 issue of the Journal of Economic Perspectives. (All JEP articles back to the first issue in 1987 are freely available online courtesy of the American Economic Association. Full disclosure: I've worked as Managing Editor of the JEP since that first issue.)

For those who have blotted the episode from their memories, LIBOR stands for London Interbank Offered Rate. It's the interest rate at which big international banks borrow overnight from each other. A main use of LIBOR in financial markets was as a "benchmark" for adjustable interest rates. For example, if you are a potential borrower or lender worried about the risk that interest rates might shift, you might be able to agree on a loan where the interest rate was, say, the LIBOR rate plus 4%. Duffie and Stein point out that using LIBOR as a benchmark interest rate for international loans dates back to 1969, when "a consortium of London-based banks led by Manufacturers Hanover introduced LIBOR in order to entice international borrowers such as the Shah of Iran to borrow from them." 

Two key details set the state for the LIBOR fraud. The first detail is that after LIBOR became well-established as a basis for interest rates on loans, the finance industry began to use LIBOR as the basis for lots of more complex financial transactions: for example, "exchange-traded eurodollar futures and options available from Chicago Mercantile Exchange Group, and over-the-counter derivatives including caps, floors, and swaptions (that is, an option to engage in a swap contract)." I won't plow through an explanation of those terms here. The key takeaway is that the benchmark LIBOR interest rate wasn't just linked to about $17 trillion in US dollar loans. It was also linked to $106 trillion in interest rate swap agreements, and tens of trillions more in interest rate options and futures, as well as cross-currency swaps. As a result, if you had some information on how LIBOR was likely to change on a day-to-day basis--even if the change was a seemingly tiny amount that didn't much matter to borrowers or lenders--you could make a substantial amount of money in these more complex financial markets. 

The second detail involves how LIBOR was actually calculated. Banks did not actually submit data on the costs of borrowing; indeed, someone at a bank responded to a survey each day with an estimate of what it would cost that bank to borrow--even though on a given day many of these banks weren't actually borrowing from other banks. In addition, during the financial crisis as it erupted in 2007 and 2008, no bank wanted to admit that it would have been charged a higher interest rate if it wanted to borrow, because financial market would be quick to infer that such bank might be in a shaky financial position. 

So on one side, LIBOR is a key financial benchmark that affects literally tens of trillions of dollars of continuously traded and complicated financial instruments.  On the other side, you have this key benchmark being determined by a survey of the opinions of fairly junior bank officers who have some incentive to shade the numbers. The British court found that Tom Hayes led a group of traders who sent messages to the bankers who responded to the LIBOR survey, requesting that the LIBOR rate be jerked a little higher one day, or pushed a little lower another day. Again, those who were just using the LIBOR rate as a benchmark for loans probably wouldn't even notice these fluctuations. But traders who knew in advance how the LIBOR was going to twitch up and down could make big money in the options and futures markets. 

What's the solution here? Duffie and Stein point out that financial benchmarks like LIBOR are extremely useful in financial markets. However, you need to design the benchmark with some care. For example, instead of using a survey of bank officers, it makes a lot more sense to use an actual market-determined interest rate for a benchmark. Moreover, the LIBOR rate is based on banks borrowing from banks, and so it will reflect risk in the banking sector. For certain kinds of lending and borrowing, it's not clear that you would want your interest rate to rise and fall with changes in the riskiness of the banking sector. Thus, they discuss the virtues of benchmark rates that are market-determined and not linked to the banking sector--like the interest rate for short-term borrowing by the US government. (They also discuss the merits of using some other less well-known  benchmark interest rates, like the Treasury general collateral repurchase rate or the  overnight index swap rate, fo those who want such details.)

More broadly, it seems to me that the LIBOR scandal is the actual real-life version of what seems to be an urban legend plot: the story of how a fraudster finds a way to program the computers of a bank or financial institution so that a tiny amount of certain transaction is siphoned off into a different account (for examples, see the 1983 movie Superman III, or the 1999 movie Office Space). The problem with these "penny-shaving" or "salami-slicing" attacks in real life is that if you steal a little bit from a large number of transactions, it's quite possible that no individual party will notice. But if you take a few million dollars out of a financial institution, the accountants are going to notice! 

In the LIBOR scandal, however, the fraud happened by knowing about tiny little changes in LIBOR a day in advance. Those who lost out from not knowing these changes in advance had no way of knowing that they were being cheated. In a similar scandal from earlier this year, Citicorp, JPMorgan, Barclays, Royal Bank of Scotland and UBS pled guilty to felony charges for their actions in foreign exchange markets. Again, these are very large markets, and so small acts of dishonesty can add up to large amounts. As the US. Department of Justice described it:
"According to plea agreements to be filed in the District of Connecticut, between December 2007 and January 2013, euro-dollar traders at Citicorp, JPMorgan, Barclays and RBS – self-described members of “The Cartel” – used an exclusive electronic chat room and coded language to manipulate benchmark exchange rates. Those rates are set through, among other ways, two major daily “fixes,” the 1:15 p.m. European Central Bank fix and the 4:00 p.m. World Markets/Reuters fix. Third parties collect trading data at these times to calculate and publish a daily “fix rate,” which in turn is used to price orders for many large customers. “The Cartel” traders coordinated their trading of U.S. dollars and euros to manipulate the benchmark rates set at the 1:15 p.m. and 4:00 p.m. fixes in an effort to increase their profits.

As detailed in the plea agreements, these traders also used their exclusive electronic chats to manipulate the euro-dollar exchange rate in other ways. Members of “The Cartel” manipulated the euro-dollar exchange rate by agreeing to withhold bids or offers for euros or dollars to avoid moving the exchange rate in a direction adverse to open positions held by co-conspirators. By agreeing not to buy or sell at certain times, the traders protected each other’s trading positions by withholding supply of or demand for currency and suppressing competition in the FX market."
A trader at Barclay's reportedly wrote in the group's electronic chat room: “If you aint cheating, you aint trying,” Clearly, situations where relatively small groups of people can cause relatively small and almost imperceptible  tweaks in values that affect a very large market are ripe for manipulation.