Truerisk Releases Market Simulator Module Based On Infinity Toolkit

VENDORS & SERVICES

TRUERISK, the Toronto-based financial systems developer, has released a high-end market risk management application dubbed Market Simulator.

Market Simulator is designed to help risk managers perform uniform and consistent market risk measurement for large, organization-wide portfolios.

It uses transaction aggregation and efficient sampling routines in concert with parallel processing technology to provide fast and accurate simulation-based market risk analytics.

Infinity libraries

Built using Infinity Financial Technology's data model and class libraries, Market Simulator is being targeted at bank treasuries, securities broker/dealers and major money management houses.

Only larger institutions will be able to afford the new system: purchasing and configuring Market Simulator will typically cost six-figure dollar sums, according to Truerisk officials.

The new product is currently being tested at Bankgesellschaft Berlin. Bankgesellschaft officials couldn't be reached for comment by RMO's press time.

Earlier this year, Bankgesellschaft contracted with Mountain View, California-based Infinity to use the vendor's platform to build a bank-wide trade data repository and risk management analytics platform (RMO, April 22).

The Truerisk product is available both as an application and as a class library that can be used for in-house systems development.

It is available on several Unix platforms, including Sun Microsystems' Solaris and SunOS operating systems. Truerisk expects to release a Microsoft Windows NT version of the product in the fourth quarter.

Truerisk developers invested nearly two years of effort into designing and building the application, according to Dan Rissin, one of the vendor's founding partners and its lead product developer.

"Interest in the area of firm-wide risk management has been growing steadily over the past two years," he says. "We believe that as the risk management arena continues to evolve, there will be an increasing demand for software tools providing fast, accurate simulation."

Market Simulator's framework comprises three separate components for scenario generation, risk valuation and risk analysis.

A typical scenario would use Infinity's data model to store yield curve and volatility curve shifts and changes to foreign exchange rates for all relevant rates and currencies in the portfolio, according to Truerisk's marketing literature.

Market Simulator's Infinity-based valuation engine uses the transaction and curve data saved in the data model to value the portfolio under each of the scenarios, then stores the results.

The analysis module then uses these portfolio values to calculate VAR, profit/loss distributions and profit/loss sensitivities.

Four alternatives

Market Simulator supplies risk management staff with four alternative methods of analyzing the market risks of their books.

Historical simulation uses historical market data to show how a portfolio would have performed over any given past time period.

Structured Monte Carlo simulation uses volatility and correlation data, which may be obtained from JP Morgan's Riskmetrics data set, together with random sampling to produce scenarios.

The resulting prices and rates are normally distributed with the required volatilities and correlations.

Stratified sampling also uses volatility and correlation data to produce VAR figures for portfolios. But an improved multivariate stratified sampling technique is used instead of random sampling.

Truerisk officials claim this method produces stable VAR results faster than structured Monte Carlo simulation.

Such stratified sampling methods have been the subject of much recent academic research. Theorists believe they offer a faster alternative to notoriously compute-intensive Monte Carlo methods (Derivatives Engineering & Technology, October 2, 1995).

Finally, end-users can use Market Simulator's stress-testing capabilities to analyze the market risk of portfolios under unusual or extreme market conditions of their own design.

Although each of the above risk methods employs a lognormal or Gaussian distribution in producing its output, Truerisk is currently working to enable Market Simulator to similarly evaluate non-normal price and rate distributions.

Three tricks

Rissin says developers need to get three things right in designing market risk management application systems that provide timely, flexible and accurate decision support across a global network of trading centres.

Such systems need to distribute and process prices and other market-related data across multiple workstations and servers; they need to have a mechanism to aggregate cashflows for more efficient processing; and they need to use "smart math" to reduce the number of simulations required in producing accurate output.

In designing Market Simulator, Truerisk developers first considered the global trading environment of its target customers. "We started by trying to figure out how best to perform simulations on large global portfolios," says Rissin.

The vendor worked with test portfolios of 50,000 to 100,000 transactions, primarily interest rate derivatives.

"Our initial target was to process these portfolios in four to five hours at most on an overnight basis," says Rissin.

Producing all the cashflows and sensitivities for these portfolios is a difficult task, he adds, even for linear instruments such as interest rate swaps.

Truerisk therefore built a transaction aggregation mechanism into Market Simulator. Simulations are performed only on the total cashflows rather than on individual transactions, leading to faster results without sacrificing accuracy.

Generating cashflows and sensitivities within Market Simulator as opposed to obtaining them from source trading systems provides users with greater flexibility in analyzing data from varying perspectives, says Rissin.

Adopting the latter approach forces developers to selectively choose which risk factors are obtainable from their source systems. Typically this means that some risk factors are omitted, he adds.

By employing multivariate stratified sampling rather than random sampling methods in generating alternative risk scenarios, Market Simulator further reduces processing time by reducing the number of simulations required to generate accurate VAR figures.

"Users typically incorporate anywhere from two to seven years-worth of data in constructing up to 1,500 scenarios to estimate and analyze historic volatility," says Rissin.

Fast work

"Stratified sampling produces equally robust and accurate results using less than 1,000 scenarios," he claims. This can translate into an eight-fold boost in performance, he adds.

By parallelizing Market Simulator's code, Truerisk developers enable risk routines to be run simultaneously across multiple processors or networked workstations.

In addition to significantly reducing processing time, this also makes Market Simulator extremely scaleable, says Rissin.

The combination of reduced scenarios and distributed processing can provide performance increases of 100 times or more, say Truerisk officials. Accurate simulation of large portfolios of complex derivatives is not practical without such methods, they add.

In addition to selling Market Simulator to existing Infinity clients, Truerisk also expects the software to be attract any large financial organization looking to build a robust and flexible market risk management platform.

Rissin says that by combining Market Simulator with Infinity's data model and recently released Riskview Mapper data translation and transport engine (RMO, July 1), customers will have all the tools they'll need to implement such a system.

Infinity has recently been striking deals with application developers that are interested in using the vendor's platform to build end-user applications. These third-party developed systems are then jointly marketed to global financial institutions.

Observers note that Infinity has in the past received some criticism regarding the length of time and cost required to build working applications using its platform and toolkits.

The vendor views such application development partnerships as one means of addressing this issue. To this end, it has revamped its strategic business plans emphasize such vendor partnerships (DE&T, June 26, 1995).

  • LinkedIn  
  • Save this article
  • Print this page  

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact [email protected] or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact [email protected] to find out more.

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here: