14� Wilmott�magazine

Wilmott: In terms of the develop-ment of F3, was portfolio-level analy-sis defined as a key aspect from the outset or did it become more impor-tant after the events of 2008?

Webb: Yes, from the beginning F3 was designed so that handling a portfolio was just as easy as han-dling a single trade. In fact, any function that requires a trade object as input will work just as well with a portfolio object. Once a portfolio object has been created, there are various analyses that are avail-able on that object. The first, most basic one is the value of it, nothing too complicated about that, just the sum of the value of the trades within the portfolio. Secondly, you can calculate aggregated cash flows – the set of all of the future cash flows from all of the trades in the

portfolio, sorted chronologically. Another type of basic analysis for portfolios is what we call a Risk Report. This is the first-order sensi-tivity of the price of the portfolio to all relevant market data; and it also outputs the hedge factors which tell you how much of each market data instrument you need to hedge the portfolio. Those three portfolio-level analyses are the basic ones which are accessed through a single func-tion called ValueProduct.

The technology behind the first-order sensitivity is very efficient; we are using a proprietary approach which was built into the software architecture right from the begin-ning. We do not use bumping or finite difference, which is a very common approach, although not very efficient. Rather, our approach

involves the automatic application of the chain rule of calculus at a low level within each pricing computa-tion. It’s also very cool that these first-order sensitivities are carried – or propagated – through a calibra-tion process, so in F3 we guarantee to calculate the sensitivity of the portfolio values to all the market data even if there is a calibration step involved. With other systems you can only obtain sensitivities to

model parameters, such as a mean reversion parameter, or vol of vol, but then if the model parameters themselves depend on a calibration step what you really want to know when you’re hedging a portfolio is what’s the sensitivity to the mar-ket data. The technology we have underlying F3 propagates these risk sensitivities even through the cali-bration step, without bumping and recalibrating.

Wilmott: By what order of mag-nitude would you say the efficiency has improved in comparison to tradi-tional techniques such as bumping?

Webb: When you’re doing bump-ing you will probably increase a market quote and revalue, then decrease the quote by a small bump and revalue, so you get a centered difference. You have to do three valuations: a middle valuation to give you the price, and the other two to give you the slope. With the technology in F3, practically for the same cost that we can calculate the value we can at the same time get the sensitivity, so there’s a factor of three. It’s also more robust; when you are using a bumping technique, you have to be careful about bump size to make sure you are still within permissible ranges and so forth. Tony Webb

fincad

Portfolio-level�analysis�is�a�key�risk��management�function�in�the�current�climate�of�regulatory�transition.�Tony�Webb,�Director�of�Analytics�at�FINCAD,�talks�to�Dan�Tudball�about�how�F3�delivers�solutions�in�this�vital�area�…

On�the�

�Wilmott��magazine 15

fincad

So there’s also a robustness aspect involved here.

Before, we were talking about portfolio-level analysis, and I was talking about valuation, cash flows, sensitivity, and I talked a little bit about the underlying technology. Once the portfolio has been defined as an object, it’s quite easy to start thinking about other analyses, like Credit Value Adjustment. If there is an ISDA Credit Support Annex in

place saying that the exposure can be netted across all of the trades within a netting set – well, if you define that netting set as a portfolio then that type of analysis fits well within the F3 framework. Generic CVA is now offered for an arbitrary portfolio in the most recent release of F3. Other higher-level measures are Value-at-Risk, and scenarios – what’s the impact of various scenar-ios on portfolio value. All of these analyses are available in the most recent release. We’re finding that now the architecture is in place, a lot of dominos are starting to fall.

Wilmott: With regard to CVA and the greater focus on the handling of counterparty risk as a result of regu-latory changes, what was the inter-nal view at FINCAD with regard to the effectiveness of previous Mark-

to-Market approaches which did not consider counterparty risk?

Webb: Before 2008 we were not seeing a lot of demand for CVA. For many of our buy-side clients the counterparty was a bank or other financial institution, and the counterparty risk associated with the bank was not often considered. But obviously now the Credit Value Adjustment can be significant if the credit spread of the counterparty

is not negligible. And it increases as the volatility of the exposure increases, or if the lifetime of the exposure increases. With F3 we have a very efficient generic approach that involves internally creating pseudo-trades, such that the value of each pseudo-trade is equal to the contribution to the portfolio CVA for a particular time interval. The time intervals are partitions of the total lifetime of the exposure. Once the pseudo-trades have been cre-

ated, we can reuse the F3 generic valuation engine to value the set of pseudo-trades just like any other portfolio. The value of each pseudo-trade gives you the contribution to the CVA for its corresponding time period. The value of the portfolio of pseudo-trades gives you the total CVA. We have taken this generic approach in F3, which has turned out to be very important after the events of 2008, and will be useful for

both sell-side and buy-side clients. Previously, FINCAD products provid-ed calculations for credit exposures and CVA, but they were asset class-specific and also model-specific. For example, you could analyze the credit risk of a portfolio of inter-est rate swaps in FINCAD Analytics Suite, but we did not offer hybrid portfolios or portfolios of mixed asset classes.

A neat thing happens when you combine our CVA methodology with

the sensitivity technology that I was describing before. We can calculate the CVA for the portfolio, but we can also, at virtually no extra cost, say what’s the sensitivity of that; in other words, how much more CVA there will be for an extra dollar of notional in any of the trades which are in the portfolio. That is useful for traders who want to know what’s the marginal CVA cost of a new trade, in near real time.

Wilmott: How straightforward is the transition away from the tradi-tional approach with F3, is it a pain-less process?

Webb: It’s quite easy in F3 actu-ally. Once a portfolio object has been created, the portfolio object con-tains the constituent trade objects, and it can be passed through the generic ValueProduct function that will calculate the traditional value that ignores counterparty risk. It’s also just as easy to pass the portfolio

The value of each pseudo-trade gives you the contribution to the cVa for its corresponding time period. The value of the portfolio of pseudo-trades gives you the total cVa

Level

object to the generic CVA calculator, which will calculate the adjustment needed to the traditional valuation. The flexibility rests in the object-ori-ented nature of F3 and the ability to reuse objects.

Wilmott: One of the challenges of risk management is that you have to link into a possibly large number of different systems in order to gather the information required to make these sorts of reports. How does F3 face up to the challenge of dealing with legacy systems?

Webb: What I would say is that once the data is imported or trans-

mitted from the other systems, it’s well organized within F3. The data is automatically used where and when needed in F3. So, the data organiza-tion is taken care of very well for you. One of the strengths of F3 lies in the fact that you can reuse the data and there are very clear concepts within F3 as to what the classes represent. This really helps when interfacing with other systems. Furthermore, multiple operating systems are sup-ported, and there are interfaces for multiple programming languages. Also, there are various modes of deployment depending on what strength of coupling with other sys-

tems you are looking for.Wilmott: Where do you see the

additional demands on technologi-cal assets coming from as a result of the latest trends in regulation?

Webb: There are a lot of aspects to the changes in regulation and a lot of the detailed regulations still need to be written. One aspect that we’re following is the recent SEC/CFTC study on algorithmic descriptions of derivatives, which investigated the feasibility of requiring computer-readable descriptions of derivatives. We are following this with interest because one of the powerful features

of F3 is its generic internal trade representation, which not only describes the cash flows but also specifies decision points related to embedded options or trigger points for any instrument. The representa-tion automatically induces a pricing algorithm for an arbitrary trade at run time. There is no code genera-tion involved, and it’s not necessary to compile or rebuild the executable for a new instrument. Any domain-specific languages that might emerge from the SEC/CFTC initiative can be parsed or converted to our internal representation, so we have some technology in F3 that is going

to be very useful for dealing with computer description of trades. The ultimate process would be to scan any deal sheet, use OCR and natural language processing to represent the deal in a Domain Specific Language, and have the pricing engine auto-matically perform the valuation with little or no human interven-tion! A vision that is a bit futuristic, but we do have the technology for the final step.

As for other aspects of the regulations; I think there is a gen-eral theme in terms of transpar-ency. Transparency means different

things to different people in the industry. For us as an analytics ven-dor, it means having well-document-ed models, and being able to check intermediate calculations such as calibrations or implied cash flows. It means not having a black box, rather having all calculation results available for reporting and checking. Another related aspect is that there’s a lot more recognition of model risks these days. With F3 it’s very easy to substitute one model for another. Once you have created a trade object, you can just as easily price that trade with model A as with model B. It is just a simple substitution of one

model object with another model object. This allows you to get a han-dle on model risk quite conveniently.

Another area that I think is rele-vant is the issue of ongoing financial innovation. We were seeing a move away from exotics after the crisis, and a move toward vanillas to allow central counterparty clearing, and ultimately, derivatives trading. But people are returning to structured products in the quest for yield. And financial innovation isn’t going away. For example, take contingent convertible bonds, or CoCo bonds, which are being issued by the banks to raise capital and boost their bal-ance sheets as required by the regula-tors. New instrument types like this will continue to emerge and evolve. Analytics providers need to be flex-ible enough to handle them.

Wilmott: The move back to vanilla products is also requiring a reassess-ment of the way in which the vanillas are actually priced and valued in the market. There’s the new challenge as well. What about variations in mar-ket practices? Things like fragmenta-tion of liquidity within the market. Are those amongst the things pre-senting development challenges to FINCAD at the moment?

Webb: Definitely, in terms of liquidity modeling, these are things that we are following closely, and the F3 framework is flexible enough that we can relatively rapidly implement new approaches. Speed is always high priority; to make sure that the calculation is as fast as possible we always seek the most efficient implementation. It’s essential to handle multiple curves to handle dif-ferent funding rates, and the spread between different reference rates and tenors. It’s true that pricing even vanilla products is not as simple as it used to be!

fincad

“The ultimate process would be to scan any deal sheet, use OcR and natural language processing to represent the deal in a domain Specific Language, and have the pricing engine automatically perform the valuation with little or no human intervention!”

W

16� Wilmott�magazine

Foreign Exchange Option PricingA Practitioners Guide Iain Clark Foreign Exchange Option Pricing is a practitioner’s guide to the latest, cutting edge option pricing formulas for foreign exchange. It contains everything a quant or trader working in a bank or hedge fund would need to know about the mathematics of foreign exchange – both the theoretical mathematics and comprehensive coverage of implementation, pricing and calibration.

With content developed with input from traders, and examples using real-world data, the book introduces many of the more commonly requested products from FX options trading desks, together with the models that capture the risk characteristics necessary to price these products accurately. Crucially, the book describes the numerical methods required for calibration of these models – an area often neglected in current literature, which is nevertheless of paramount importance in practice.

978-0-470-68368-2 • Hardback • November 2010 • £60.00 / €72.00 £36.00 / €43.20

Book ClubShare our passion for great writing – with Wiley‘s list of titles for independent thinkers ...

SAVE40%

When you subscribe to Wilmott magazine you will automatically become a member of the Wilmott Book Club and you’ll be eligible for 40 per cent discount on specially selected books in each issue. The titles will range from finance to narrative non-fiction. For further information,

call our Customer Services Department on +44 (0)1243 843294, or visit wileyeurope.com/go/wilmott or wiley.com (for North America)

NewcitsInvesting in UCITS Compliant Hedge FundsFilippo Stefanini, Michele Meoli, Silvio Vismara, & Tommaso Derossi Due to their strong regulatory frameworks, UCITS (Undertakings for Collective Investment in Transferable Securities Directives) compliant hedge funds have seen a boom in recent years and are considered by many as the only way out for the hedge fund industry after the crisis.

Newcits: Investing in UCITS Compliant Hedge Funds is a one-stop resource for investors who want to get the best out of their UCITS investments. There is a large and increasing range of UCITS compliant funds out there, but despite their tighter regulation and frameworks, investors still need to understand the risks they are undertaking, the structures of the funds and their differences and similarities to mutual funds and hedge funds.

978-0-470-97627-2 • Paperback • December 2010 £29.99 / €36.00 £17.99 / €21.60

Expected ReturnsAn Investor’s Guide to Harvesting Market RewardsAntti Ilmanen Expected Returns is a one-stop reference that gives investors a comprehensive toolkit for harvesting market rewards from a wide range of investments. Written by an experienced portfolio manager, scholar, strategist, investment advisor and hedge fund trader, this book challenges investors to broaden their minds from a too-narrow asset class perspective and excessive focus on historical performance. Coverage includes major asset classes (stocks, bonds, alternatives), investment strategies (value, carry, momentum, volatility) and the effects of underlying risk factors (growth, inflation, illiquidity, tail risks). Judging expected returns requires balancing historical returns with both theoretical considerations and current market conditions. Expected Returns summarizes the state of knowledge on all of these topics, providing extensive empirical evidence, surveys of risk-based and behavioural theories, and practical insights.

978-1-1199-9072-7 • Hardback • February 2011 £45.00 / €54.00 £27.00 / €32.40

Exotic Options and HybridsA Guide to Structuring, Pricing and TradingMohamed Bouzoubaa & Adel Osseiran Exotic Options and Hybrids is a practical guide to structuring, pricing and hedging complex exotic options and hybrid derivatives that will serve readers through the recent crisis, the road to recovery, the next bull market and beyond. Written by experienced practitioners, it focuses on the three main parts of a derivative’s life: the structuring of a product, its pricing and its hedging.

Divided into four parts, the book covers a multitude of structures, encompassing many of the most up-to-date and promising products from exotic equity derivatives and structured notes to hybrid derivatives and dynamic strategies. Based on a realistic setting from the heart of the business, inside a derivatives operation, the practical and intuitive discussions of these aspects make these exotic concepts truly accessible.

978-0-470-68803-8 • Hardback • April 2010 £60.00 / €72.00 £36.00 / €43.20

Investing in Emerging MarketsThe BRIC Economies and BeyondJulian Marr & Cherry Reynard For more than a decade, emerging markets have proved one of the most exciting areas of investment, but the sector has not been without its dangers. Private and professional investors alike have continuously been attracted by the promise of riches on offer from countries such as China, India, Brazil and Russia but, as often as not, have been left with their fingers burnt.

Investing in Emerging Markets: The BRIC Economies and Beyond tempers the undoubted causes for emerging market optimism with a healthy dose of reality, illustrating how for every argument in favour of investing in the sector there are one or more reasons to tread very carefully indeed.

978-0-470-74825-1 • Hardback • August 2010 £34.99 / €42.00 £20.99 / €25.20