How UBS Lost $680 Million
In the fall of 1997, Union Bank of Switzerland entered into a two-part structured transaction with the principals at Long-Term Capital Management that ultimately led to losses of at least 950 million Swiss francs.
The two parties had complementary goals: UBS wanted to buy a significant stake in the fund, and LTCM partners wanted to convert foreign interest income from their off-shore hedge fund into capital gains and defer it for seven years.
In the first part of the transaction, LTCM partners bought a seven-year call option, or warrant, on their own company from UBS. In the second part of the transaction, UBS purchased about $800 million worth of shares in the company to hedge the call option it sold.
The trade was approved at the highest level of management within UBS—the CEO, the head of credit and the head of the derivatives group. The trade also received board approval.
The deal solved a tax problem faced by partners at LTCM and other hedge funds: Because most funds are set up as partnerships or limited liability corporations, income from the fund flows directly to the partners in the form of short-term profits or interest and dividends, which are taxed at 39.6 percent. Long-term capital gains, by contrast, are taxed at 20 percent.
According to sources, the strike price of the option was set at 80 percent of the forward value of LTCM at the time the transaction was struck. “You look at the current net asset value of the partnership, you then forecast what the forward value of that would be, and you compound it at Libor plus a margin of 75 basis points to get the future value for the end of year seven. You then take 80 percent of that number, and that's what the strike price of the option was,” explains one source. That strike price was $575 million, and LTCM partners paid UBS between $175 million and $250 million for the trade. UBS reportedly booked an estimated $30 million to $50 million on the trade up front.
According to one source, UBS needed to purchase $525 million in shares to cover its trade. The bank purchased an additional $275 million in LTCM shares, however. According to this source, the hedge fund shares were booked into UBS's Treasury account. Treasury groups are usually responsible for funding and managing a bank's liquidity, and sources noted that it is extremely unusual for a Swiss bank to hold such a large and relatively illiquid hedge fund investment in its Treasury account.
“Meriwether would not have sold UBS shares in his company,” says one source, “if UBS did not accommodate the partners on a tax trade.” There is some disagreement as to why UBS was so eager to buy so many shares. According to some, UBS intended to repackage its position in the hedge fund to other investors. But during the merger last November, these plans were postponed.
Others, however, doubt that UBS ever intended to repackage the transaction. They argue that UBS had the perfect vehicle to sell the shares but declined to take advantage of it. Last year, the bank was offering a product that sold investors options on hedge funds, which were dubbed STAR notes. But sources say senior management believed so strongly in the fund that they insisted their investment be used in the bank's Treasury instead.
According to sources, the trade was proposed by Ron Tanenbaum, who worked in a UBS unit that specialized in fixed-income structured transactions. Tanenbaum had previously worked with a number of LTCM partners at Salomon Brothers.
|BOB, SAY IT AIN'T SO!
|“The other author of the Black-Scholes model was Fischer Black, who worked with me at Goldman. I was always told how our model showed that one of our positions would be going up while another was going down. But when trouble came, all those positions went in one direction, which was down.”
Treasury Secretary Robert Rubin
New York Post, October 4, 1998
The Collapse of Russia's MICEX Exchange
By Ted Kim
All the big global banks lost money in Russia in recent months, but the prize for the biggest loser probably goes to the Credit Suisse Group, the holding company for CSFB. Although the bank recently announced that pretax profits for the first six months of 1998 were up by 38 percent compared with the same period last year, a $1.06 billion provision was made to compensate for anticipated losses on CSFB's Russian operations. Some $637 million of the $1.06 billion resulted from losses sustained on currency forwards, which the bank purchased in order to hedge its ruble exposure. And much of that loss on forwards was sustained because the Moscow Interbank Currency Exchange (MICEX), as well as individual Russian banks, defaulted on the currency derivatives they wrote with CSFB as counterparty.
In an attempt to salvage whatever it can from the defaulted contracts, CSFB has initiated legal proceedings against MICEX. A number of other banks that have suffered from the MICEX default are also expected to join with CSFB in suing MICEX.
MICEX has long established itself as the dominant currency exchange in Russia, with daily trading volumes often exceeding $1 billion. Much of the trading on MICEX was linked to GKOs (Russian ruble-denominated short-term Treasury bills). Foreign investors, including CSFB, frequently used MICEX to move dollars into the GKO market and, upon redemption, sell their rubles and repatriate the dollar proceeds. A thriving futures and options division had developed at MICEX that offered contracts on currency, equities and Treasury bills. The one-month currency contracts were particularly popular, since foreign investors in the GKO market were required to wait one month after selling GKOs before being permitted to sell their rubles and send dollars out of the country.
|The complete loss of credibility among foreign investors in MICEX could have serious implications for emerging market derivatives exchanges worldwide.|
Although MICEX itself was never rated by an international credit-rating agency, the widespread assumption was that MICEX contracts carried quasi-sovereign risk, given the ownership structure of the exchange. Major MICEX shareholders include the Central Bank of Russia, the Ministry of Finance and the Moscow City government. Furthermore, according to the rules of MICEX, should one counterparty default on a contract, any losses to the other counterparty would be covered by the exchange members, which included most all the major names in Russian banking.
Exactly why CSFB chose to hedge ruble exposure through a Russian counterparty—be it MICEX or another Russian bank—is not clear. From a risk management viewpoint, using a Russian counterparty to hedge Russian risk would be similar to buying Brazilian sovereign default insurance through a credit derivative contract written by a Brazilian bank. One reason put forward by Moscow analysts for using Russian counterparties in ruble hedging was the thin liquidity and high transaction costs involved with other hedging methods, such as the ruble contracts on the Chicago Mercantile Exchange or an over-the-counter transaction written out of London. The CME contracts were only introduced this year, and ruble-structured products in London are still considered highly exotic.
The complete loss of credibility among foreign investors in MICEX could have serious implications for emerging market derivatives exchanges worldwide. If the local exchange could collapse hand-in-hand with the local currency, what would be the point of any investor using the exchange to hedge local currency exposure? For CSFB, by contrast, the $637 million provision for losses on currency derivatives is a minor pothole in the road. CSFB's Moscow office was among the most profitable financial operations anywhere in Russia or Eastern Europe.
Both CSFB and MICEX declined to comment on the lawsuit.
Can You Patent a Financial Model?
By Nina Mehta
Philosophically, the United States Patent and Trademark Office sits at the crossroads of private property and technological progress, ensuring that the rights of inventors and the economic interests of society are protected. One of its principal tasks is to decide what is and isn't an invention.
Computer software—and the financial models written into computer software—have thrown it for a loop. The agency and the courts have struggled for years over whether software is in fact patentable. If software can be considered a “machine,” in legal parlance, it can be patented; if it can't, it falls beyond the patent system.
Until recently, people in the derivatives business have paid little attention to patents, because it has been difficult to claim that a particular financial model or piece of software was a machine. But this may be changing. On July 23, the U.S. Court of Appeals for the Federal Circuit (in State Street Bank & Trust Co. v. Signature Financial Group Inc.) ruled that all patent claims should pass the same standards tests, regardless of their subject matter. These requirements are utility, novelty and nonobviousness (that is, that the work is indeed an invention).
The case could have lasting significance for the high-tech and financial sectors. Although some issues remain to be decided in the case, a machine loaded with software that has a practical application is now valid subject matter for a patent. If a dealer develops a computer model for a particular financial instrument, for example, it may now be be possible to patent that model and charge others who use it royalties.
William Ellis, an electronics and software attorney at Foley & Lardner, a Washington law firm, says the Federal Appeals Court decision should serve as a wake-up call for the financial industry. “The computer-dependent financial industry generally doesn't pay attention to patents, and I think it probably will now,” he says. “Companies will need to patent what they are doing, as a protective strategy and also to give them assets to cross-license.”
Gil Leistner, a Chicago Board of Trade member who has served on the strategic planning and product development committees at the Chicago Board of Trade and the Chicago Mercantile Exchange, believes the case “will create a bit of a revolution in the derivatives industry, because now, for the first time, if the decision is upheld, people will be able to patent derivatives products.” Under this new legal regime, for example, Fischer Black, Myron Scholes and Robert Merton probably could have patented their option-pricing formula, which would mean that “every time somebody issued a price on an option, there could have been, in theory, a royalty payment.” Leistner believes that the green light for companies to patent new computer-implemented financial strategies and pricing models, as well as new software applications, “will eliminate a lot of copycat products and will change the competitive landscape fairly widely.”
The derivatives and risk management software industry has shown a reluctance to patent over the years, in part because of uncertainty over whether financial models were considered mathematical algorithms—which have traditionally been deemed abstract ideas that are unpatentable—or machines, and in part because of industry practice.
But another reason for the reluctance to patent is the need to fully disclose the invention—not only the technology running the machine, but also the best method of achieving what the patent claim purports to accomplish. For software, this usually means putting the source code in the public domain, a prospect many find counterintuitive, if not perilous. Instead of doing this, software developers tend to protect their products under trade secret and copyright law.
Although these two forms of protection are useful for preventing what's known as slavish copying, neither protects the broader concept of the invention itself. As a result, says David Boundy, an intellectual property attorney at Morgan & Finnegan in New York, both leave open the possibility of “reverse engineering,” the process of working backward from a product to its design. “Reverse engineering is one of the magic incantations you can pronounce to keep the curse of a trade secret at bay,” he says.
Michael Zuckerman, a copyright attorney at SunGard Trading Systems Group, demurs, pointing out that “large software” is usually so complex that to work backward could take years. There may be an inherent trade-off between patenting a product and protecting it by trade secret, he notes, in that if a firm patents an invention, the firm has to disclose it. “And if you've disclosed it and the patent is invalidated,” he says, “my gosh, you've lost your trade secret.”
Another hurdle for a company seeking a patent is that patents are reserved only for inventions—and cannot be granted once a product has been in commercial use for more than a year. That explains why Scholes and Merton can't simply apply for a patent on their model and begin charging fees to everybody in the derivatives business.
Fast Track to Wall Street
Who needs a Ph.D. when an M.S. will do?
By Nina Mehta
More than 500 years ago in Italy, a handful of “abacus schools” were launched to give merchants, craftsmen and others a quick education in the practical and commercial applications of mathematics. Those who went to these schools learned basic arithmetic, algebra and geometry.
Now, Carnegie Mellon, Columbia, NYU and a dozen other universities have launched a new generation of high-tech abacus schools, offering master's degrees in financial engineering, financial mathematics and computational finance. Although scores of quants enter Wall Street each year from Ph.D. programs in math and physics, these master's programs focus on teaching the heavy-duty financial and computational tools now needed in bulk by the derivatives industry.
The programs are usually 12 months in duration and cost about as much as a pre-owned Lexus. They all use (or plan to use) industry professionals to teach seminars and courses that give students a hands-on understanding of financial products and their applications. These same industry contacts also inevitably give students access to companies and jobs once the program is over.
Take Jeff Miller. After graduating from Harvey Mudd College in California with a B.S. in mathematics, he cast about for a field that combined his interests in pure and applied mathematics and computer science. An article in Wired about David Shaw, the renowned hedge fund trader, caught his eye. “There are ex-scientists and super-high-IQ people at his firm doing what, to me, is enormously creative and interesting work—all with mathematics and computation!” enthuses Miller. “It made me realize that this new hybrid occupation of applying mathematical theory to the financial markets fits my academic goals perfectly.” After deciding on a career in proprietary trading, the next step was to build his credentials and make himself “more relevant to employers.” So this fall he entered the financial mathematics master's program at New York University's Courant Institute.
These schools have been attracting people with different backgrounds. Some already have careers in finance and are looking for a career shift. These candidates expect the programs to give them a competitive edge by ratcheting up their computational skills and getting them up to speed on emerging technologies and applications. Others have already completed Ph.D. programs in mathematics but want to make their resumes more attractive to Wall Street.
The programs have also been a lifesaver for academic math and science departments suffering from cutbacks in basic research, notes Neil Chriss, director of NYU's program in financial mathematics and vice president of the quantitative research group at Goldman Sachs Asset Management. “Many academic departments have needed new ways to justify themselves and keep up enrollment in their classes,” he says, and these programs provide an answer.
The University of Chicago's master's program in financial mathematics, for instance, was created by the mathematics department to service the growing need for individuals with quantitative backgrounds. Those who developed the program, says Niels Nygaard, the current director, decided to build on the reputation and skills of the mathematics department “to make a program that specialized in the kind of material that the financial industry was interested in—a program that was academically as well as industry-wise a good thing to do.”
|SCHOOLS FOR QUANTS
|Name of school
||Name of program
||Year program started
||Current number of students
||Length of program
||Cost of program (full-time)
||master of science in computational finance
||30 full-time in Pittsburgh, 30 part-time in New York, 20 part-time in London
|Claremont Graduate University
||master of science in financial engineering
||master of arts in mathematics with a specialization in the mathematics of finance
||15 full-time, five part-time
||M.B.A. or master of engineering with a certificate in financial engineering
||seven–10 in the M.B.A. program, 30–40 in engineering program
||two years for an M.B.A., one year for a master of engineering
||$49,000 for an M.B.A., $11,400 for a master of engineering
|Sloan School of Management, MIT
||50 in financial engineering track
||two academic years
|Courant Institute of Mathematical Sciences, New York University
||master of science in financial mathematics
||10 full-time, five part-time
|Oregon Graduate Institute
||master of science in computational finance
||15 full-time, 10–15 part-time
|University of Alberta, Edmonton
||master of science in mathematical finance
||two academic years
||approx. C$6,000 for Canadian students, approx. C$10,400 for foreign students
|University of Chicago
||master of science in financial mathematics
|University of Michigan, Ann Arbor
||master of science in financial engineering
||$10,000 for Michigan residents, $20,300 for non-residents
|University of Toronto
||master of mathematical finance
Carnegie Mellon was the first to launch a master's degree program in the broad quantitative financial field. Started in 1993 by Sanjay Srivastava and colleagues in the business school, Carnegie's computational finance program is now the largest in the United States, with 30 full-time students in Pittsburgh and 50 part-timers in New York and London. (Courses are taught at two campuses at a time, via video hook-ups, and Carnegie professors rotate between campuses.) The program started, says Fallaw Sowell, the current director, because “there was a need for individuals who could understand the business issues on the finance side as well as the technical dimensions of stochastic calculus.” The financial industry's quantitative focus in recent years has been driven by advances in technology, but also by a changing work force. Over the last two decades, says Srivastava, “there has been a much greater skill level in the markets, and this has made it easier for more quantitative tools and models to get assimilated and used.”
Around the time Carnegie started its program, Andrew Lo, director of MIT's Laboratory for Financial Engineering, realized that what the financial industry needed didn't always match what was being taught in traditional business schools and graduate departments. So five years ago, the Sloan School of Management designed a track system to give M.B.A. students specializations geared toward their eventual profession, “as opposed to specializations in academic disciplines such as accounting, marketing, finance and economics.” MIT's first track (and certificate for M.B.A. students) was in financial engineering, one of the school's areas of strength since the 1970s, after Robert Merton, Fischer Black and Myron Scholes had settled in Kendall Square.
While the recent development of quantitative master's programs has benefited the financial industry and students, these programs aren't identical. They draw, instead, on the strengths of schools and departments and have different disciplinary orientations. According to John Moody, director of Oregon Graduate Institute's program in computational finance, “financial mathematics programs, by and large, concentrate on stochastic calculus and derivatives pricing.” OGI's program, which is offered through the department of computer science, produces graduates “with strong programming skills and an advanced quantitative understanding of many areas of finance, including global risk management and derivatives pricing.” The University of Michigan's program in financial engineering, notes director John Birge, “tries to put the mathematics in an immediate financial context and give students those tools that are really useful.” The university's program draws on mathematics, statistics, industrial and operations engineering, and finance. Carnegie's, meanwhile, is a joint venture between mathematics, statistics, computer science and the business school.
Cornell offers a certificate in financial engineering to M.B.A. and graduate engineering students. The advantage of this dual program, says director Robert Jarrow, is that employers can opt for students with more management experience or a stronger engineering background—and engineering students are less expensive than those who have just spent two years getting an M.B.A. Another benefit, he notes, is that courses are taught by finance and engineering professors, and the mathematics department is not directly involved.
At NYU and other schools, the reverse is true. Mathematics professors from NYU's Courant Institute teach the mathematics of finance, says Chriss, giving students a visceral understanding of the fundamentals that go into building models and pricing options, and “financial practice” courses are taught by professionals specializing in the application of quantitative methods to financial markets. The program's strength is in “teaching the computational methods that go into the models,” agrees Jackie Rosner, vice president of fixed-income derivatives trading at Salomon Smith Barney, who is finishing up a master's in financial mathematics at NYU this fall. (Courses in financial mathematics have been offered at Courant for a couple years, although not in a formal program.) “The business of derivatives is quite mathematical and the market has become more competitive over the last five years,” Rosner adds. “NYU is offering courses in things that five years ago people in business school would have seen only in an occasional seminar or would have picked up only in bits and pieces on their own.”
The competition between schools is undoubtedly a measure of the field's good health. Additional evidence of this takes the form of the Association of Financial Mathematics Programs, an organization established by Chriss and the directors of the Chicago and Columbia programs to monitor the field, recruit students and promote the discipline as a whole.
But there is still more competition (and opportunity) around the corner. MIT is thinking about launching a separate master's program in financial engineering to reach students who don't need a battery of business courses. Stanford University's year-long master's program starts next fall. The University of Texas at Austin has added a finance and economics specialty to its Ph.D. program in computational and applied mathematics, and is planning a two-year graduate program in computational finance. In addition, most of the current programs have plans to grow. Abroad, there are already master's degree programs at the University of Warwick and Dublin City University, and a joint program at the University of Edinburgh and Heriot-Watt University. And Carnegie Mellon may expand its current program in computational finance to Germany and the Far East. The future, it seems, is secure for the modern abacus school.
Chase Tackles FX Settlement Risk
By Robert Hunter
Cash settlement in the foreign exchange market is like summertime in the Mississippi Delta—the constant threat of a late-afternoon storm keeps people constantly looking to the skies. When banks settle their FX deals, they expose themselves to varying degrees of risk while waiting to take possession of their respective currencies. Consider Banks A and B settling a dollar-Deutsche mark deal. Bank A pays away the required Deutsche marks to Bank B before Bank B pays Bank A in dollars in New York, because Bank A has to pay the Deutsche marks to Bank B's account in Frankfurt in Frankfurt time. Bank A thus cannot be sure that Bank B will deliver the dollars to its New York account when it relinquishes the Deutsche marks. Such uncertainty is known as settlement risk, and, given the enormity of the world's FX markets, it is considerable.
Chase Manhattan, in conjunction with the British Bankers Association, hopes to change all that. It is developing a new foreign exchange contract, known as the contract for differences (CFD), to mitigate settlement risk by eliminating cash settlement altogether. “We were sitting around trying to come up with a way to solve the Bank for International Settlements' requirements to reduce foreign exchange settlement risk,” says Dennis Oakley, managing director at Chase. “Somebody asked, ‘Why are we cash-settling all of our FX transactions? Why don't we “derivatize” it and settle only the profit of the contract?'”
The CFD does just that—counterparties settle only the profit or loss of a contract, rather than swapping the notional amounts. The key to the CFD is the way the profit or loss is determined: It is based on an index produced by the British Bankers Association. The BBA has been working with Reuters for some time to develop FX indices for the G-7 countries. They are experimenting with two at this point—a survey of 12 banks and a survey of five brokers—to determine which is more accurate. After a three-month testing phase, Reuters will publish the prevailing index on its system at 11 a.m. London time each day. Parties involved in CFDs will settle deals on a daily basis based on the index.
|“Basically, the contract for differences would be like every other derivative—you settle the profit on the contract, rather than going through the whole process of swapping the underlying amounts.”
To many, the CFD makes perfect sense. “If we were reinventing the FX market today, we probably wouldn't have a gross-settlement market,” says Simon Hills, a director at the BBA. “More than 90 percent of all FX transactions are undertaken to help financial institutions take a view on the market, rather than because they need to take possession of, say, Deutsche marks to buy a load of chemicals.”
Indeed, such a net-settlement system offers a cleanliness and simplicity that cash settlement cannot possibly approach. But some aren't so sure the CFD is the answer for the FX world. “There is skepticism about the settlement procedure and the marking procedure,” says the director of foreign exchange at a major money center bank. “If you and I agree to deliver dollars against Deutsche marks on a certain day for a certain amount, there's no ambiguity, no subjectivity and no potential for dispute. But if we settle for dollars based on a fixing, and you deem that fixing to be not objective, I'm not going to be particularly happy.”
But the CFD would dramatically lower transaction costs, says Oakley. “Basically, it would be like every other derivative—you settle the profit on the contract, rather than going through the whole process of swapping the underlying amounts.” Oakley says there has been a great deal of third-party interest in the product so far.
Because the CFD offers net settlement and no margin requirements, it has all the benefits of a future contract, but with all of the flexibility of an over-the-counter product. Chase has tried to downplay the CFD's resemblance to a currency future, given the Commodity Futures Trading Commission's recent contention that it should revisit the 1993 swaps exemption that has rendered the OTC world virtually unregulated for years. The CFD, in fact, seems to be exactly the sort of thing the CFTC would like to get its hands on.
Even if the CFD manages to escape the CFTC's regulatory grip, its future isn't certain. “The CFD certainly solves some credit issues,” says the FX director. “If you're dealing with a counterparty bank and you'd rather not take settlement beyond a certain amount, it certainly has credit utility. But I think that will have to be proven over time, and people should use this contract cautiously at first.”
|Was this information valuable?|
Subscribe to Derivatives Strategy by clicking here!