.
.--.
Print this
:.--:
-
|select-------
-------------
-
Brady's Trinity

By Andrew Webb

Trinity is Brady PLC's solution for trading and managing derivative products. At present, coverage includes interest rate options, foreign exchange options, options on futures, metals and swaps. Apart from providing front-to-back-office trade processing in these product areas, Trinity also includes risk management functionality at both desk and firm-wide levels.

History and background

Trinity has its roots in the pricing algorithms that Brady developed for its first client, Midland Bank, in 1985. The original Brady MS DOS products were Brady Capital Markets (IROS), Brady Pricing Screens and Brady FX and Commodities (CAM-OS). The functionality of these products was integrated into the single Trinity Windows-based product in the autumn of 1994.

The original version of Trinity, first implemented by South African trading house Investec in 1995, ran on Windows 3.11, though the product is now principally Windows NT-based. Although the current version of Trinity permits the use of Windows 95 and Windows 3.11 clients, future releases will require Windows NT throughout.

Currently Trinity is deployed in some 12 sites around the world, with an additional four sites currently in implementation projects. The product is primarily aimed at small- to medium-sized trading operations that have fairly sophisticated needs for trade and risk management and pricing. Although Trinity is not targeted at the largest sites, the analytically demanding nature of its potential client base means that it is typically competing with much larger risk management vendors.

The latest version, 4.7, should be in beta by late April and will incorporate substantial enhancements to the data manipulation and display functions.

Independent Evaluators

Sanjay Dighe
principal,
Vega Risk Consulting, London

Brady's new Trinity is a brave attempt to create a truly integrated system for off-balance-sheet interest rate products. It allows deals to be priced by the trader, captured by the system, settled and then reported in various ways. While it has a fairly narrow brief, it succeeds admirably within its limited aims.

The pricing tools are fairly comprehensive. Brady has a good pedigree in derivatives pricing, and this product is no exception. The analysis screens are extremely flexible and allow for input of most off-balance-sheet products. A cash-flow-based approach means that even nonstandard products are easily accommodated. This does not apply equally to some hedging instruments, however. While most standard bonds are handled, callable Euros are not included (though they can be created synthetically).

The deal pricing and input facilities are well-designed. Users can set up templates for standard transactions to minimize input times. For nonstandard deals users can define fields for unique attributes. This includes flagging related trades so that a strategy can be tracked, even across portfolios. This is one of the most attractive features of the system. This flexibility also means that exposures and risks can be sliced and diced many ways when producing reports. This is particularly important for heads of trading who need to visualize risk and exposure in many different ways.

Yield curves are another area of strength. Brady provides a choice of standard yield curve calculations from raw data that can include deposits, futures, forward rate agreements and swaps. Users can export the calculated curves, though they can't import curves from other systems—an unfortunate omission. The assumption is that Trinity will be the source of a company's yield curves for all systems. As such, it would have been nice to have a plausibility check facility built in. To do this a user would have to filter it through an Excel spreadsheet.

Trinity is also strong on allowing users to view and manage data. Each transaction can be included in many different views. A view is any combination of trades that can either have predefined relationships or be selected using specific criteria. Performance analysis is also easy, using a combination of mark to market and (where necessary) mark to model. Revaluations can also be archived in order to track profitability.

Trinity also incorporates a lot of market risk functionality. In addition to the usual risk matrices and sensitivity analyses, users can also carry out scenario, Monte Carlo and historical value-at-risk calculations. The link with RiskMetrics allows reports to be created using standard inputs. Unfortunately the interface is not as user-friendly as it should be. In particular, the third-party report generator feels and acts very much like a bolt on. This is another area that needs more work before the next version is released. The same applies to the graphical representations of positions and risk.

The system shows every sign of having been built by academics and technologists without much input from end-users. It is a classic example of function over form. For example, it is easy to carry out a hedge calculation on a trade or group of trades. But it is then quite difficult to take the hedge and simulate its impact on the portfolio. (Evidently this function will be possible in the next release of Trinity.) It would be nice to see the next release address this kind of problem—a little effort to make this product more user-friendly would have a disproportionate impact on its marketability.

This product is extremely specialized, aimed at a niche market. It allows a small or medium-sized financial institution to run an interest rate derivatives book without a great deal of customization. I suspect, however, that users would soon find themselves irritated by the little niggles that make it feel clunky.

In its favor is the fact that it is a true integrated package, which means it uses the same algorithms and databases for all purposes. There is a lot of flexibility, which means handling new or nonstandard instruments is easy. Trinity offers a lot of functionality but it is not a complete "black box” solution. It is much more difficult than it should be to make full use of the system. The final presentation detracts from what has the potential to be quite an attractive system. I would certainly put Trinity on the short list when evaluating potential systems for any financial institution other than the very largest. If the next release addresses the drawbacks I've outlined, I can see it developing into a world-beater.

Sanjay Dighe is a London-based independent consultant specializing in Derivatives, Risk Management and Internet Applications. His previous positions include global head of equity derivatives at Paribas, head of financial engineering and risk management at Standard Bank and head of OTC equity derivatives at Citicorp. He is a member of the London Steering Committee of the Global Association of Risk Professionals and a Fellow of the Royal Society. He maintains the Vega Risk Consulting web site, which can be found at www.riskreward.co.uk.

Pat McConnell
principal,
Risk Trading Technology Ltd.

As an IT consultant to international banks, one of the questions that I am frequently asked is "What is the best risk system on the market?” The answer, of course, depends on where one is coming from and where he or she wants to go. If a large global financial institution does not already have a system that provides risk management capabilities better than those provided by Trinity, then I would consider the firm to be in serious trouble. But for a medium-sized bank with active Treasury or derivatives operations and sizable books in foreign exchange and interest rate derivatives, the Trinity system should certainly be on its short list of systems for consideration.

Trinity is designed as a "plug and play” solution for banks that need the latest derivatives technology without the overhead of a large IT department. The product is based on solid client/server technologies, such as Windows NT and Oracle, which are fast becoming standards in the financial industry. The system is easy to use and the concepts would be familiar to experienced derivatives traders and risk managers. End-user reporting is provided through the Crystal Reports package, Microsoft Access and, of course, Excel.

As with its main competitors, Summit and Infinity, Trinity can be viewed as a number of subsystems built around an integrated data object model (although the value-at-risk subsystem has been decoupled to allow data to be imported from legacy systems). Deals in the model consist of streams of cash flows that may be combined in flexible ways to represent real flows of cash, such as in an interest rate swap, or contingent flows, as in an option. More complex derivatives, such as swaptions, can be modeled by combining streams of real and contingent flows.

Deals are combined into portfolios and profit centers for tracking profitability, and flexible selection features are provided to view risk and profitability across portfolios. An API is provided by Brady that allows banks to integrate deals from existing legacy systems into the product's data object model for risk management purposes. While extremely powerful for capturing complex deals in the foreign exchange, interest rate and metals markets, however, such a flow-based approach is difficult to use in situations where cash flows are uncertain (rather than variable), such as with equities and path-dependent options.

Deal-capture screens in Trinity are conventional, even slightly boring, using standard pull-down menus and spreadsheet-like columns. Facilities are provided for locally defined fields to be added to generic deal-capture screens, but this seems to be an afterthought rather than a coherent attempt to extend the data model. (For example, there are few capabilities for adding local definitions where they might be most useful, such as on counterparty data for credit risk management purposes.)

Pricing of deals within Trinity is quite flexible. Each stream of cash flow in a deal can be tied to a specific discount curve; multiple discount curves can be defined for each currency or market traded and blended curves can be constructed from real-time price feeds or from user-supplied models. Likewise, multiple volatility curves can be created against underlyings and option strike prices. Consistent valuation of deals across products and markets is, of course, one of the absolute essentials of firm-wide risk management.

Having captured the cash flows underlying each deal, these data can be used in the back office to make the necessary payments. Trinity provides confirmation, settlement and accounting capabilities that were not covered in detail in this test drive but, from a cursory inspection, appear to provide adequate back-office support.

The value-at-risk (VAR) subsystem is designed to support the latest set of BIS proposals, and thus can be used by banks to provide their regulatory reporting. Trinity provides two levels of VAR computation. The simpler approach follows the JP Morgan Riskmetrics variance/covariance methodology and uses the data sets published for it. The more complex approach employs Monte Carlo simulation techniques. Input to the subsystem is in the format specified for Riskmetrics and hence data can be imported relatively easily from legacy systems to calculate firm-wide VAR numbers.

Given that the banks purchasing Trinity are likely to have a complex derivatives book, it is unlikely that the simpler Riskmetrics approach, which does not handle options well, will satisfy regulators, so some form of simulation will be needed.

The Monte Carlo subsystem consists of two modules: the scenario builder and the VAR calculator. The mechanism is simple yet powerful. Starting with a predefined set of prices, volatilities and correlations, the scenario builder shifts each data point for the number of simulations specified, using advanced stratified sampling and variance-reduction techniques. These scenarios are saved for later use by the VAR calculator. To calculate VAR, a selected set of deals (either firmwide or subset) is run against a set of saved scenarios, and VAR, at the required confidence level, is computed. Tabular and graphical reports of the complete profit or loss distribution are provided.

The major problem with the Trinity VAR subsystem is that, while extremely data-rich, it is information- poor. Backup reports contain too much detail and there is a need for better summary graphics and reports. For example, the system holds copious data on each scenario to allow risk managers to investigate potential risk hot-spots, but there are no reports that sift through this vast amount of data to summarize such hot spots. To support back testing, as required by the BIS, some mechanism will have to be provided to summarize data along the organizational lines typically used by a bank's profitability systems. Likewise, some form of VAR limit-checking will have to be added to the system to implement the qualitative requirements of the BIS proposals. Betraying its MS DOS origins, the system is not as automated as it should be; each stage of the process must be initiated by a user from a workstation and hence the daily grind of producing VAR numbers is likely to be error-prone and manually intensive. These are relatively minor problems, however, and the system should mature quickly when used in anger by experienced, creative risk managers.

Trinity provides a level of basic credit risk management, called group risk in the system. Multilevel credit limits can be set up and risk calculations can be defined by deal type, including worst-case risk computation for delivery risk. Exposures can be netted and over-limits reported. Brady has also announced that it will be providing an interface from Trinity to JP Morgan's Credit

Manager system, which uses the recently announced CreditMetrics methodology. In a similar vein, Brady already provides a VAR interface to the Elmbridge package, which is widely used in Europe for CAD reporting.

Trinity is not cheap. Brady estimates that an entry-level system will cost around $200,000, while an average system will run to $400,000, plus implementation. For a small to medium-sized bank that needs to employ sophisticated risk management techniques to keep regulators on-board yet are strapped for IT resources because of the year 2000 bug and EMU problems, Trinity certainly merits close consideration.

Patrick McConell is principal at Risk Trading Technology Ltd., a small consultancy that specializes in developing IT strategies for international banks, in particular for risk management systems and advanced trading room technologies. He holds degrees in business administration, operational research and mathematics, and has conducted research on the organization of risk management functions in international banks.

Client Opinion

Jon Howell
assistant treasurer (technology),
Alliance & Leicester PLC
(Formerly Alliance & Leicester Building Society)

We've been dealing with Brady since 1989 and started upgrading to Trinity in the autumn of 1996, with the system going live in the spring of 1997. In broad terms, the reason we had to move to Trinity for the flotation as a PLC was that the Bank of England wished to see off-balance-sheet items consolidated into one system. At the time, we had a lot of instruments on individual spreadsheets, so Trinity was important in that it allowed us to button-down any loose ends in the way we controlled our off-balance-sheet items. It also integrated directly with our IBIS accounting and general ledger system.

At the time of the implementation, the Society was undergoing a period of considerable administrative upheaval with the creation of a separate treasury subsidiary, so the circumstances were far from ideal. Because of the other changes that were taking place, we often only had a small window within which to complete certain parts of the implementation. We were generally pleased with the way that Brady was able to work with us to accommodate that sort of situation.

We currently use Trinity to manage some 700 open positions. A big addition to our activity level developed when the treasury operation here started to handle the hedging requirements of all the other subsidiaries within the Alliance and Leicester group. The acquisition of a finance company (primarily involved in leasing) last November involved taking on a total of roughly 140 existing trades and an ongoing activity level of some six trades a month, which are almost invariably amortizing structures.

On a day-to-day basis, the swaption and cap/floor pricing tools see a lot of use, primarily as control on other things that we're watching. We tested them against the existing Brady MS DOS models and also benchmarked them against figures that we obtained from third parties, and we found them accurate. At present, the risk tool that we use the most is the hedge calculator, particularly in calculating the effective hedges required for the option portfolio that are included in the risk report. We also use GroupRisk for calculating our credit exposures on a daily basis, and we upload the figures to our main system to update limits utilization. This has worked well as we have been able to configure the ways in which individual instruments are treated.

Trinity is user-friendly and, as it is upgraded, is becoming even more so. Features such as the ability to pull in revaluation rates directly from Reuters, without the need to re-key, are obviously a big plus. In terms of specific functionality, Trinity's ability to price and handle the amortizing swaps has been most effective. The flexibility of the payments editor and deal capture, and the ability to tailor cash flows to exact requirements have been highly beneficial. We recently had to handle a big-ticket leasing transaction, which had the most bizarre amortizing swap structure we had ever seen. Despite this, we were simply able to put in all the cash flows, click the calculation icon and immediately obtain a price for it. Previously we would have had to ask a third party to price a transaction like that for us.

Rob Hughes
head, front office development,
DKB International, London

We've been a Brady client since 1992, when we purchased the Brady Capital Markets Trading System for over-the-counter bond option trading. We subsequently decided to use Trinity as our derivatives trading and risk management system in 1995 after conducting an extensive systems review. Since then, we've gradually extended our use of Trinity. Originally we were using it strictly for bond option pricing, but after Brady performed some proprietary tweaking for us we began using it for asset swaps, trade management and back-office settlement. We now use the trade management module for the settlement of all swaps and derivative products, as well as Trinity/RiskMetrics to calculate value-at-risk on an end-of-day basis. The Trinity general ledger interface has also become effective for us in feeding account postings into the books-and-records system.

Bond option pricing and booking are two components of Trinity that are seeing continually increasing use. It's giving us practically all we need in a derivatives system, handling 70 percent to 80 percent of our requirements. In view of the size of our operation and our preference for not building components in-house, our choice of vendors is limited. Under those circumstances, companies are lucky if they can find a supplier with a product that gets to 50 percent. Trinity does what we need, and to a large extent in the way we need it. The asset swap functionality has been particularly well-suited to our requirements, as it allows us to price and handle transactions quickly. It also allows the staff on the desk to maintain a clear view of their open positions across the board.

We normally request any additional functionality that we need from Brady, though anything that can be simply integrated with Trinity we'll add in-house. This has been made easier by Brady's commitment to open system architecture. In technology terms, Trinity has fit in well. It's Windows NT-based and is currently migrating toward all executables being 32 bit in operation, which is particularly desirable from our point of view. The architecture has allowed us to build and integrate into existing systems fairly simply. For example, the Brady RiskMetrics product that we use for our daily risk management allows us to integrate our bond trading system into it and feed data into other systems as well. We also hook quite a lot of added value-at-risk functionality into both Brady and the RiskMetrics system, mostly for feeding information back to the head office in Tokyo. In terms of integration it's simple. The system has a number of APIs that it uses for getting information both into and out of the system, so we haven't had to spend a lot of time integrating the system even though we have a lot of external interfaces connecting with other systems. Most of the tools included with the system can provide common output forms such as Excel, comma-separated values or simple text reports.

Perhaps our greatest pleasure has been in developing the relationship with Brady, working with it and finding it responsive to our needs. When we first started using Trinity, there were a few unwieldy elements for which we suggested changes, and Brady made them. It's the kind of supplier relationship that I enjoy. Brady is extremely flexible and Trinity is reasonably priced for what it delivers.

Pricing & Contacts

Pricing
Pricing of the Trinity system depends on three factors—number of users, number of sites and number of modules taken. For example, a license for three users running all the modules would have a list price of £150,000. Typically, license costs for existing clients have been in the £200,000 to £300,000 range. In addition, a standard maintenance charge of 18 percent per year applies.

Contacts
United Kingdom:
Jacqueline Clark
Brady PLC
Cambridge Science Park
Milton Road
Cambridge
CB4 4WE
UK
Tel: 44-1223-423250
Fax: 44-1223-420302
e-mail: j.clark@bradyplc.com
Internet: www.bradyplc.com

Italy and Switzerland:
Dr. Mariano Kihlgren
Brady Italia
Via Fontana 11
20122 Milano
Italy
Tel: 39-2-541-01373
Fax: 39-2-541-01380
e-mail: bradyitalia@mail.apa.it

United States:
Melany Smith
General Electric Information
Services
401 North Washington Street
Rockville, MD 20850
USA
Tel: 301-340-4665
Fax: 301-340-5250
e-mail: melany.smith@geis.ge.com

RiskBusters: F for Effort

Who are you going to call for help with BIS compliance? Not RiskBusters, say our reviewers.

By Andrew Webb

RiskBusters is a spreadsheet-based capital charges calculator produced by ARMS (Advanced Risk Management Solutions) as an aid for calculating the capital requirements for market risk. The calculations are modeled on the standard method suggested by the Bank for International Settlement (BIS). The software claims to cover interest rate, equity, foreign exchange, commodities and options risk via the Delta-plus method, the scenario-grid or the simplified approach, and will break the financial instruments in a portfolio into the appropriate time zones, mapping the cash flows in accordance with BIS requirements. The product is designed to calculate total market risk capital charges and incremental capital charges as well as an adequate capital ratio.

My view

"As the deadline of the end of 1997 approaches,” reads the introduction to the RiskBusters help files, "banks will start to feel the urgent need for an effective and versatile software tool to compute their market risks and their associated capital charges on a daily basis. RiskBusters, which covers interest rate risk, equity risk, foreign exchange risk, commodities risk and options risk, is designed specifically to satisfy that need and as such spares the users the tedious chore of meticulously calculating the required capital charges.”

While the product may, depending on one's interpretation, cover the requirements for the BIS capital charges calculations, the cited text seems to me to imply some form of risk management functionality. Judging by their reaction to the product and their specific criticisms, it appears that the independent reviewers also took this view.

We raised the specific criticisms with ARMS and discussed their response to these with the reviewers concerned. In certain cases, ARMS asserted that a specific functionality that our reviewers remarked on the absence of is not required for calculating BIS capital charges. Our reviewers contend either that they are required, or that based on the earlier description (or the national implementation of the BIS capital charges) they should have been included anyway. A case in point was the absence of facilities for repos, MBSs and the more complex option types.

We also had various versions of the product flying around and received conflicting information from the ARMS support department. We only discovered in passing from one of the client references that there was any sort of version for Windows NT and Office 97 Excel available at all. ARMS insists that most of the problems we encountered were a result of the special security measures built into our review copies. (The product usually ships with a hardware security device.) I certainly hope that this is the case. Any risk management product that automatically dumps you into the Microsoft Visual Basic (VB) editor, where one can edit the macro code every time there is a run time error, has terrible implications. There's also a rather alarming lack of sanity checks—one reviewer reported that he entered negative numbers in the yield curve fields and RiskBusters happily included them in the calculations and produced a capital charge figure. ARMS's response was that interest rates may be negative in some countries in the short term and that this field can be set to check for only positive values if the client desires. Our reviewers took the not unreasonable view that since countries don't tend to lend money and pay interest, this field should be set to do this by default.

The lack of security features were also a source of concern. ARMS says, "We can easily implement security features for the download of the portfolio files in our system integration consultancy, which is available to our potential clients. If the client wishes to buy it off the shelf, we can easily customize RiskBusters to meet the clients' back office needs and the local regulations.” This may be so, but for $35,000 we felt that this should be included, not an extra.

The inactivity of the "update from RiskMetrics” command button in the rate input screen also came in for criticism. ARMS states that the function worked fine on its machine and that possibly the software was uploading from the same RiskMetrics data set, hence the apparent inactivity. Since there was no reference in the Help files to this function we were unable to ascertain whether this was the case, though there was a RiskMetrics data file present in the directory. If ARMS's assertion is correct, then clients will have to write their own download routine for the RiskMetrics files—all the "update from RiskMetrics” will do is read the data file.

The lack of delivery calendars for markets/countries was criticized, though ARMS can supply these—as an extra via its system integration program. The lack of price or yield fields in the bond files means that price sensitivities based on yield changes cannot be calculated. ARMS points out that, according to the BIS standardized methodology, market-level data to derive price sensitivity to calculate the capital charges for market risk are not required. Our reviewers point out that the BIS accord states that the duration method shall use price sensitivities. In addition, for derivatives, the position amounts should be reported as the market value of the underlying instrument.

There was also some confusion over the credit rating field, which can only be set to sovereign. ARMS disputes this, claiming that it can be set to both government and corporate, which are "Z” and "S,” respectively. The ARMS documentation provided allows for two credit types, however: (GOV, Z), which are governments, and (S, A), which are referred to as "swaps.” More important, the BIS accord states that two credit ratings are required for netting. The FRN trade file does not contain a field for the reset rate index or a reset calendar. While this is a deficiency in the BIS accord, it does clearly state that the local national regulatory authority has the right (duty!) to proscribe such mandates. The U.S. Federal Reserve and the Bank of England both require reset dates and rates for their calculations.

RiskBusters doesn't have the hallmarks of a product intended for sales on a standalone basis. It seems much more like a promotional tool for ARMS's consultancy services—or to be included at a discounted rate as part of a consultancy package. A general criticism from the reviewers was that in order to use RiskBusters to any meaningful effect, clients will have to do the bulk of the calculation and preparation work themselves outside the program, leaving the software itself with a few relatively simple calculations to perform. A case in point: netting for capital haircuts has to be done outside RiskBusters and then imported as preprocessed position data. The company didn't make available as references any clients who had simply purchased the product alone. At $35,000, I'm not entirely surprised. To be fair to ARMS, the product is clearly not an all-around risk management package (despite the implication to the contrary in the help files), but at the current asking price it should be offering at least some of that sort of functionality. There may well be a need for this sort of product among smaller banks, but as it stands the product is substantially overpriced.

Independent Evaluators

David Little
president, Derivate Services Corp.

RiskBusters is a set of Excel spreadsheets and accompanying macros for calculating the capital needs for a bank. The software is consequently designed to implement the algorithms and options for the new BIS risk-adjusted capital adequacy regulations, which are complex and require the aggregation and evaluation of all of a bank's portfolios. Typical users of the product would therefore be members of the finance team of a bank's head office or branch treasurers.

The bank's portfolio has to be read into RiskBusters from a set of correctly referenced proprietary format text files. Each of these contains descriptions of the current trades that are on the bank's books, including counterparty names and credit ratings. Each file (which can contain one type of trade) can be referenced in a portfolio definition file for inclusion in or exclusion from the calculations. There is no security on these files other than would be provided by the operating system, and I was able to create and amend files in a simple text editor. There is no data-entry facility for trades (as opposed to positions) from within the workbook, though users can enter a set of cash flows to represent a group of trades that are not sufficiently generic for the limited file definition.

The opening RiskBusters screen provides a selection of command buttons that allow users to access the various function screens within the workbook:

Methodology. This allows users to choose various options for evaluating the capital required for the trading book in accordance with the BIS regulations, which permit a number of different methods for aggregating and evaluating risk.

Capital charge calculator. This displays a summary of the position (and the capital charges relating to it) broken down into interest rate, commodities, options, equities and foreign exchange. Of these subsections, the options area, not being fully integrated into the risk areas of the option's underlying product, is particularly complex and odious. A useful addition here would be the ability to simulate the results of the various methodologies in order to determine the lowest-possible capital charge method permitted under the BIS regulations.

Cash flow summary. This is a worksheet that holds any input cash flows and calculates the raw numbers required for the capital ratios.

Rate input. As the title implies, this is the input sheet for all exchange and interest rates required for the calculations. These can also be entered automatically via Excel's DDE link. The screen also contains a command purporting to allow downloading and updating with RiskMetrics data sets. The button for this was not functioning when tested.

Capital ratio. This is where the current bank equity is input and, on the basis of the capital charge calculations, the current capital position is displayed.

The system was quite buggy, necessitating three revisions to be sent before it would actually work. Even in the last version one of the trivial screens would not come up. Instead it would throw an error and dump me into the visual basic editor with the macro code accessible for editing.

Though it does implement a complex set of regulatory algorithms, as a standalone program RiskBusters is not a viable product. The main problem is security. Using Excel to calculate something as crucial as capital and risk measures is, I feel, dangerous, as I was able to get into the code in the VB editor. I easily could have overwritten code or hardwired answers to certain reports.

In the bigger picture, leaving aside the questionable $35,000 one would spend on RiskBusters in the first place, the software has a major problem in terms of actual implementation. The cost of creating and managing a process where all trades have to be sent on a timely basis to a single place in a single set of field formats (and getting all areas of the bank to agree on one set of rates and customer names to use in the consolidated mark to market) would be colossal. Any bank capable of effectively handling an integration and implementation task of that order of magnitude is going to find writing its own in-house tool for calculating capital charges an absolute snap.

Peter Vinella
president, PVA International

RiskBusters is a BIS capital usage calculator, and a very simple calculator at that. If you are in the market for a true risk management tool, one that understands forward, spot and volatility curves; covariance; value-at-risk; scenario and sensitivity analysis; credit derivatives; nonlinearity; or even MPT, well, forget RiskBusters.

This is not to say that, theoretically, RiskBusters doesn't meet a specific need. In every bank and brokerage firm, there is someone, usually in financial control, who must calculate capital charges for regulators. Often this is done using standalone spreadsheets, which are used to collect information from the various back-office systems, perform the necessary calculations and produce the required reports in the proper format. But if you are the poor sucker who has to compute Reg. T capital, is RiskBusters for you? Probably not. You most likely already have similar spreadsheets—and ones that work.

There are two operating modes: manual and automated. In the manual mode, you enter all the cash flow information and initiate the calculations by hand. In the case of the latter, the user is asked to supply RiskBusters with a preprocessed portfolio text file. Since I didn't receive the manual describing the format of the text file—or any other documentation—I went to the help utility, which was relatively extensive and complete. Since the help utility crashed when the find mode was selected, I was only able to uncover the file formats after some extensive manual searching. I almost wish I hadn't found them at all.

The files must be created in MS Notepad or Word format. And there are a lot of them—a trade file containing daily trade details for each RiskBusters class of instruments (12 in all and, surprisingly, no currency files) and a portfolio file containing the list of trade files. More seriously, I found in nearly every case that RiskBusters does not come close to complying with either its documentation or the BIS standards it claims it offers in its product literature.

RiskBusters' automated cash flow generation capabilities are almost nonexistent. True, some of the problems stem from shortcomings in the BIS standards themselves. But RiskBusters cannot even meet these weak standards. The FRN trade file does not even contain a field for the reset rate index, much less a reset calendar. RiskBusters (and the BIS) treats convertible trades as "debt securities if they trade like debt securities and as equities if they trade like equities.” Allowable options types are only simple European or American contracts. In terms of swaps, only simple fixed vs. floating interest rate or cross-currency are supported.

Another feature of RiskBusters is a curve/foreign exchange rate input and display function. You get one curve per country and up to 15 tenor dates per curve. Based on the program's calculations, I assume that these must be spot curves since there is no way to differentiate a spot curve from, say, a par curve. The exchange rates are all relative to the dollar and spot only. There is a button to initiate an update of the data via RiskMetrics, but it didn't work in my version, and if you use your own curves and rates, it's not helpful in any event.

RiskBusters also ignores covariance and nonlinearity. In other words, RiskBusters is not a risk management tool. Even ARMS admits in its documentation that for large banks, RiskBusters should be used as a secondary source to validate their existing capabilities. For small banks, it suggests that RiskBusters will meet their needs. But are small banks included in the BIS reporting requirements?

As far as the calculations are concerned, other than the Black-Scholes option calculator—again, no path-dependency or weighted binomial model here—once you have input your portfolio's asset concentration and cash flows, it's simple arithmetic. The capital haircut is simply the concentration in the various BIS asset/maturity bands multiplied by the BIS capital requirement. The netting conventions outlined in the BIS amendment cannot be implemented, since the RiskBusters data fields do not sufficiently identify cash flows at the security level. Generating the cash flows and consolidating the data is the hard part, not the calculation. There is no VAR and no hope of generating one given the current state of the program. Scenario and stress analysis are barely touched on in the option volatility matrix. Multicurrency is just a simple spot translation of the forward cash flows. There is simply not enough data in the trade files to compute the accrued interest from the value date. In fact, there is no date field in any of input files other than the cash flow file. In other words, there is no way to differentiate data generated on different dates. Forget as-of or time-series analyses; the only date displayed on any of the screens is today's date. There are no consistency checks.

RiskBusters is yet another example of a product trying to capitalize on the risk management craze. If it was marketed purely as a BIS capital usage tool, it might be a viable offering with some major enhancements; in its current form, it is basically useless. Its literature implies that RiskBusters is a risk tool as well, but this is clearly not the case. If ARMS wants to be a serious player in the risk management marketplace, it will need to develop a much better product than RiskBusters.

Client Opinions

Thia Ting Tiew
vice president of risk management, Keppel Bank, Singapore

We've only been using RiskBusters for a couple of months, but it's been working pretty well so far. The processing time is fairly nominal and has dramatically cut down the overall amount of work involved. It used to take us several days to perform the capital charge calculations manually, but now it takes around 15 minutes. Our trading volumes are not particularly high but the time savings we have achieved with the product have been worthwhile. Once we have extended its use throughout the bank for all the other financial instruments as well, the time savings should be even more significant.

The only setback we have discovered so far is that certain currencies are not included. These include some that we trade occasionally, though they are fairly marginal and the amounts involved are small. As a result, we have to include them manually at the moment. There is also a certain amount of preparation needed before you can run the system. For example, you have to convert your data files into the specified text form first, though that is a process that could probably be automated relatively easily.

The product had to be modified in order to meet the MAS requirements. The stability has been pretty good and we haven't had any problems. We certainly planned to extend our use of the product in the future. There are quite a number of features within it (such as the options module) that we have not as yet made use of.

Obviously, we tested the results obtained from the product against our previous figures as well as performing a number of other validation tests. The numbers produced were satisfactory and accurate. We took the package as part of a larger consultancy deal, so I think the price is not unreasonable. As we have our own access-level security policy here, we are not particularly bothered that it is written using Visual Basic and run within Microsoft Excel.

Chan Poh Wah
acting head of liability management, Perwira Affin Bank, Malaysia.

We haven't actually installed RiskBusters as yet but we have agreed to take it. We have been reasonably impressed. The concept behind the product is not particularly difficult, but the beauty of it is the time that it will save. We haven't bought the product on a standalone basis, but as part of a consultancy package. Though there is at present no requirement for us to provide capital charges for market risk, we foresee that Bank Negara (the Malaysian central bank) will adopt the BIS requirements in time. The product coverage is certainly appropriate for us, as shortly we shall be trading a range of option products and interest rate swaps.

Pricing & Contacts:

Pricing
The price of RiskBusters is $35,000 per copy with square-root pricing for multiple copies. For a 20 percent annual maintenance fee, clients receive upgrades that maintain the product in accordance with the local regulator's requirements. Any data file conversion work required before implementation costs $300 per hour. ARMS estimates that this will take no more than three to four hours in most cases.

Contacts
Juliana Yong
business development manager
Advanced Risk Management
Solutions Pte Ltd (ARMS)
400 Orchard Road
#17-06 Orchard Towers
Singapore 238875
Tel: 65-734-9803
Fax: 65-734-0392
Mailing Address:
Tanglin PO Box 391
Singapore 912414
e-mail: sales@arms.com.sg
Internet: www.arms.com.sg

New Software Purchases

BUYER SOFTWARE BRAND/PRODUCT DESCRIPTION
Donaldson, Lufkin & Jenrette Securities Corp. (DLJ) Infinity Financial Technology: Infinity 7 Trading and risk management applications
Toronto-Dominion Bank Algorithmics Enterprise-wide risk management software
Koch Industries Triple Point Technology: Tempest 2000 Integrated energy trading, risk management and operations software

New Software Releases

COMPANY SOFTWARE BRAND/PRODUCT DESCRIPTION
Kamakura Corp. KRM-mv 1.6 Option-adjusted risk/return measurement and management tool for trading and risk management
Contact: (310) 914-0180 UNIVOPT Version 8—Universal Options Add-in Option pricing and risk management system
Mamdouh Barakat Risk Management (MBRM) UNIVGARCH—Universal Garch Add-in Implements GARCH models for pricing standard and exotic options
Contact: Mamdouh Barakat 44-171-628-2007 Analytical extension to UNIVGARCHGARCH Add-in Enables analytical pricing of an exotic option in addition to GARCH pricing
Monis Optimum Interest Rates Structuring, pricing and hedging software for fixed-income and interest rate derivatives
Contact: Abha Dawesar/Michael Green 212-573-6733 Jeffries' Convertible Bond Data Monis signed an agreement with Jeffries to distribute its convertible bond data to Monis' clients, including banks, hedge funds and asset managers
Savvysoft TOPS (Tanenbaum Option Pricing Software) for Infinity FinEx Derivatives pricing spreadsheet analytics connected to straight-through trading system
Contact: In Chen 212-742-8677
IRIS Investment Support Systems ORION 2.9.5 Integrated front-office decision-supporting analysis system
Contact: Brigitte Tuma 31-20-624-09-96    

--