New Kids on the Block
By Peter Marcus
The derivatives software market shows no signs of slowing down, judging by the ambitious plans of three new entrants into this already crowded field. The newcomers have different assessments of what the market needs. One company has designed an environment that allows quants to actually write the code necessary to price new exotics. Another claims to offer something most of its older competitors have failed at—real-time VAR reporting. A third claims to offer a front-to-back-office system that supports deal-structuring and model building—complete with math libraries and interface connections.
Tools for Exotics
One big problem facing the derivatives industry is the lag time between the conception of a new derivatives instrument and its adoption into the business flow. Months can pass before an IT department or outside vendor can build the functionality so that financial houses can trade the new product. And once initial margins decline, the instrument may not be as lucrative and therefore not as attractive.
Enter Gradient. Founded in 1993 by Roland Conybeare and Jesse Jones, the New York-based company is set to provide “high-quality support to the full scope of the derivatives business,” says Jones. Medici, the firm’s new flagship product, will be targeted to the fixed-income and exotics derivatives business.
Jones is a mathematician who has worked in finance and has helped technology firms develop foreign exchange systems; he also worked in the financial strategies group at Prudential-Bache, where he built systems to hedge portfolios of collateralized mortgage obligations. Conybeare, who first used his degree in computer languages to produce music synthesizers, moved from his native Australia to Stanford’s Ph.D. program, where he was recruited by Jones to come to Wall Street.
|“Medici’s development environment offers deal-structuring, model-building, a math library and interface connections. In short, Medici is a system that develops financial systems.”
Gradient’s clients work in Medici’s development environment, which, Jones explains, offers deal-structuring, model-building, a math library and interface connections. In short, Medici is a system that develops financial systems.
Composed of four modules, Medici contains DealDesign, which can structure a derivatives deal; ModelDesign, which designs curves and other models; HedgeDesign; and BookDesign, a trade blotter for manipulating portfolios. According to Gradient, a deal constructed in DealDesign can be valued by any existing model in Medici, while HedgeDesign will find the optimal hedge for a portfolio from a prespecified universe of deals.
Written in C++ and available on SUN, UNIX and Windows NT/95, this front-to-back-office application allows users to create deals quickly. That means, says Jones, that “someone sitting at a trading desk can create deals himself or herself in Medici without the help of the IT department.” Medici offers the quantitative analyst—the person most likely to be building deals in Medici—the language as well as the financial tools to construct deals that used to take hundreds of person-hours.
The development environment, adds Jones, allows users to put together a system in a short period of time or to build a system that works with a particular kind of deal. For example, says Jones, one client used Medici to build a system that could structure deals to trade municipal guaranteed investment contracts. Pricing an exotic derivative can be done in as little as two minutes.
Medici offers users a spreadsheet interface application that allows information in Medici to be imported into other spreadsheet applications such as Excel. Medici also offers an information distribution capability via a browser, so deals can be shared with other desks and offices.
With no competition currently in the marketplace, Gradient sees Medici’s flexibility as its strongest component. Jones emphasizes that Gradient’s target is not the high-volume, low-margin “vanilla” business, but the “introductory end of the S curve—derivatives that are innovative, difficult to model, difficult to price and difficult to hedge. Essentially, high-margin deals in new markets and new asset classes.”
Gradient sells Medici with a perpetual software license. Although the price depends on Medici’s configuration, a single site with fewer than 10 users would cost around $500,000. Maintenance and other consulting options are sold individually at the time of purchase.
Portfolio Modeling Made Easier
||609 Fifth Avenue, 9th Floor
New York, NY 10017
It’s a sign of the times. Derivatives strategists are taking risk reporting so seriously that separate value-at-risk reporting software may soon become one of the most important tools on a quant’s desk.
BPJ Technologies is answering the call with WatchDog, a suite of products that manage portfolio risk. Launched earlier this year, New York-based BPJ is a joint venture between Blaylock & Partners, a broker-dealer and investment bank, and Hedge Systems Inc., a software firm run by Dunmu Ji, who now also heads BPJ.
Scott Sussman, executive vice president of BPJ, says that Blaylock had been looking for an innovative product to sell and Ji’s software firm needed a marketing arm—“creating a great fit on both sides, because Blaylock could help sell what it considered a powerful global risk management system.” Sussman adds that the BPJ system is currently in place at a major foreign bank, a U.S. primary dealer and a U.S.-based money management firm.
The software’s main selling point, according to Ji, is that it offers clients VAR reporting in real time. The system handles international equities, fixed-income products, swaps, options, futures and other derivatives—in all currencies.
WatchDog covers risk analysis, performance valuation, scenario analysis, value-at-risk and optimization. The VAR module, for example, measures maximum potential change in the value of a portfolio, with a given probability, over a preset time horizon. Variables that a user can insert are interest rate changes, interest rate volatility, exchange rate volatility and commodity volatility. The module also allows users to break down reports by sector, portfolio group and overall portfolio. The BPJ system can be purchased in its entirety or separately by product.
Comparing BPJ’s customization process to the way Dell Computer builds machines to order, Sussman adds, “We don’t have a lot of parts lying around; we build and customize for each new client.”
Ji adds that BPJ consultants can customize client reporting along the parameters a client specifies—whether this involves value-at-risk, global scenario analysis or the optimizer product. He emphasizes BPJ’s ability to cater to all client parameters and specifications. Sussman adds that instead of a prepackaged system, what BPJ Technologies offers is the ability to purchase software that “the client can use to 100 percent of its capability.”
BPJ currently starts its software package at $50,000 and raises the price based on license and module requirements. Prices vary by client, depending on size and how WatchDog will be used.
Three or four years ago, says Ji, this product did not exist, but as the investment community has become more sophisticated, risk has been taken more seriously. “Overall, there is a greater emphasis on risk,” he says, “and statistical portfolio managers are using computer models that help them measure that risk.” Sussman notes that in addition to the demand for internal VAR reporting, large corporations are including VAR in their annual reports to shareholders. Ji and Sussman envision BPJ succeeding because of changes in risk perception across the derivatives marketplace, as well as the financial marketplace as a whole.
Indeed, Sussman predicts that over the next five years value-at-risk reporting will become standard in the industry, and VAR software will be basic to every investment company’s operation. Sussman sees BPJ’s client base expanding to include plan sponsors, large corporations, mutual funds and insurance companies.
Code-Crunching for Option Pricing
||5806 Mesa Drive, Suite 250
Austin, TX 78731
When you pair a Ph.D. in applied physics with an engineer whose Ph.D. is in software synthesis, you get powerful software that thinks for itself and that, with directional signals, builds more software. Curt Randall, a physicist and the vice president of Austin-based SciComp Inc., and Elaine Kant, SciComp’s president and a software engineer, recently announced the release of their first product—SciFinance, an option-pricing package for quants.
SciFinance focuses on pricing models for exotic options, and uses a brief description of the derivative product to automatically generate the complicated software that prices those models. A quant, for instance, can write about 20 lines of code specifying the option structure to be priced, put it into the system and “then go have lunch or work on another project,” notes Randall. The person will return to find about 5,000 lines of C code written by the software. Randall says that SciFinance uses a partial differential equation (PDE) approach to pricing, as opposed to a Monte Carlo or lattice method of pricing options. “Anything you can write in terms of a PDE, the software can price,” he says.
|“Anything you can write in terms of a partial differential equation, the software can price.”
In terms of derivatives, the company believes there are three current ways to do the pricing—either by in-house programmers, by purchasing a library of executables, or by software synthesis. SciComp says it has the only synthesis product in the financial market that it knows of, although some more limited forms of synthesis are used in universities for engineering applications. One major advantage to synthesis is that while a code library is limited to the codes within that library, software synthesis is virtually unlimited. Randall notes that SciFinance is able to be highly customized by a client, creating “one-off” codes that would not typically be in an off-the-shelf library.
|“Pure programmers will have a lot less work to do when it comes to writing the code.”
Providing a tutorial and a group of examples that show the user how to price a set of options, the software illustrates the many different ways of pricing options using PDEs. Kant and Randall recognize that PDEs can be difficult to program, but with SciFinance the programming is replaced by a short description of the option the client wants to price—“and you end up with an extremely accurate and efficient model,” says Kant.
“There is a bottleneck in pricing,” Randall says, explaining that banks and other financial institutions could be doing more business at a faster turnaround. Rather than eliminating in-house programmers, SciComp envisions making them a lot more productive. With SciFinance, the task of writing pricing code goes away. Kant adds that financial institutions will always need analysts who understand how to model complex deals, but, she explains, the people who are pure programmers will have a lot less work to do when it comes to writing the code. Although software synthesis takes over the routine aspects of programming, Randall adds that it does not eliminate the need for people to interpret and refine models and assess risk.
SciComp charges about $100,000 per year per SciFinance license, with a two-year minimum contract. The company says that SciFinance is an evolving product and that rather than selling and supporting multiple versions, it will lease the software, giving each client regular upgrades. “As we work with our clients, we see the different things they need to do, and we constantly improve SciFinance,” says Kant. “Because of that, we want to establish ongoing relationships.”
|Was this information valuable?|
Subscribe to Derivatives Strategy by clicking here!