May 7, 2013

FVA in Practice – Taking Classical Theory Further

Watch the video: FVA in Practice – Taking Classical Theory Further

Dr. Serguei Issakov, SVP & Global Head of Quantitative Research and Dr. Alexander Antonov, SVP Quantitative Research join Jim Jockle, Numerix CMO to discuss recent contributions to FVA (Funding Valuation Adjustment) theory and provide their insights from the recent Global Derivatives conference held in Amsterdam.

Dr. Antonov also discusses his presentation at the conference and new research on the funding valuation adjustment for general instruments, where he reviews the classical theory based on the work of Dr. Vladimir Piterbarg and proposes an efficient approximation to the FVA instead of a direct computation which achieves a must more practical implementation.

Numerix Global Derivatives Presentation: FVA for General Instruments, Alexander Antonov

Weigh in and continue the conversation on Twitter @nxanalyticsLinkedIn, or in the comments section below.


Video Transcript: FVA in Practice – Taking Classical Theory Further

Jim Jockle (Host): Hi welcome to Numerix video blog I’m your host Jim Jockle. With me today Dr. Serguei Issakov, and Dr. Alexander Antonov coming live via Paris, how are you gentlemen?

Alexander Antonov & Serguei Issakov (Guests): Good. Weathers fine!

Jockle: Excellent! So the topic of this week is the Global Derivatives Conference that just concluded in Amsterdam, and really one of the biggest most highly debated topics is around FVA and this debate has been ongoing for quite a while, but not seeing a lot of resolution especially as the role of funding in pricing. So a lot of panels were discussed, maybe perhaps you could just take a moment to share some the arguments that were being made at the conference.

Issakov: Okay Jim let me start with that. Wanted to mention that one of the core streams at the conference was about FVA and around FVA. That was one of the essential topics of the conference involving derivatives. And people realized now that one of the main updates, from the first year back, plenty of those banks decided to include FVA calculations in their reporting. There are still debates on how to calculate it exactly but there was a decision that it needs to be calculated and needs to be accounted for. I think that one of the most important pieces of news from that conference. Regarding the presentations themselves, there were several discussions, and many of those were theoretical discussions sort of clarifying the fundamental structure. I think the presentation from Numerix was very practical and different in that respect, because what we presented was our approach which was based on a very well understood approximations and we showed that those approximations do work.

Jockle: I want to turn to your paper. In your presentation, you highlighted elements of a general approach versus a more advanced approach in terms of the calculations and one of the slides you really go into difficulties with the general approach, perhaps you could speak to that a little bit, in terms of the calculation.

Antonov: We’ve done two major things in this presentation. The first one was to generalize the FVA itself. If we take a value of portfolio and agree that several parts of this portfolio will be funded with different rates, then we will have general equation which can be used for almost all the situations. The second new thing is that we have applied this for general instruments (not only swaps or something): we’ve considered it for callable instruments (which are one of Numerix traditional strengths). So we have prepared an example for a Bermudian option and calculated its real (exact) value using nonlinear discounting. At the same time, we’ve calculated its FVA by our approximate formulas, and compared the answers: the fit was excellent. As explained in the presentation it is important to understand a difference between Bermudan instrument prices (continuation values and future values) which we plug into the final FVA formula.

Jockle: Now speaking through the full approach, what is the impact for fully collateralized deals versus partially collateralized deals? Is that taken into consideration?

Antonov: Yes. In our experiment, the deal is partially collateralized. The collateral is supposed to be posted by the counterparty but not us, so one “half” is collateralized and the other half is not. The adjustment can be quite significant, probably on the level of 1 percent or something (of course, it depends on the portfolio at hand and on a spread between a collateralized and funding curves).

Jockle: And in your presentation, you walk through an example of a Bermudian swap I believe, why that particular instrument type, and what was unique about that that exemplified the differences between the general approach and the more advanced approach that you’ve been implementing.

Antonov: Okay so we have done this because nobody has considered Bermudian options before, so the instruments considered by the scientists were simple like, zero bonds, swaps or Europeans swaptions. So nobody has done it for Bermudian options, so we’re first to pay attention to this and at the end of the day it is quite simple, we don’t need more information which we get from the portfolio for the CVA: all we have already in CrossAsset will be reused for the FVA (we don’t need any extra information from the model). This advantage makes our valuation adjustment framework complete, we can do CVA, DVA and FVA, based on the same procedure: future price (or exposure) calculation with a proper aggregation.

Jockle: So I’m going to wrap up in just a second here, but last question gentlemen, any other observations from the event that you wish to share on today’s blog?

Issakov: I could say a few words. It was stressed several times that now the regulatory environment is changing and more and more attention should be paid to the validation of models and confirmation that the models work appropriately. So more and more attention will be paid to the quality of the models in the industry.

Jockle: Thank you Serguei and Alexandre any final observations and thoughts from yourself?

Antonov: Well, I spoke with Piterbarg, Brigo, Mercurio and other leaders in the subject. At the end of the day again we have agreed that our approaches are close to each other. So in theory, there is a consensus. However, small details can still be in the game. Having said that, the impact of this details to the FVA and the CVA is not that big. Imagine we calculated an FVA which is one percent of the deal. Other adjustments could be one percent of this one percent. So with respect to the total price, we can neglect such second order adjustments. We hope that our approach is general enough to cover all the possible observations and simple enough not to get into all the small details in order to compute all these things in a measureable and fast manner.

Issakov: Jim let me also add a few remarks to what Alexander just said, maybe to express our approach in comparison to the existing ones. What exists previously in the literature, it was step by step from simple instruments to more and more complex. It started from simple swaps, but zero coupon bonds, etc. Our approach is the different because it addresses complex instruments from the beginning. It allows us to do the systematic calculation for general instruments, and it allows us to control observations we prove on that path and those approximations are some serious and powerful spreads between different rates, so you know exactly what errors exist in this approach. I think this is a very important point for practical implementation.

Jockle: Well gentlemen I want to thank you so much for your time today and I hope the weather is nicer in Paris than it is here in New York. I’m sure it always is Alexander. If you’d like a copy of the presentation it’s available on numerix.com, you can go up and download it right there, and any additional questions we’ve love to always hear your feedback. Happy to make connections with you, follow us on LinkedIn, we will be doing more video blogs on the topic as the debate absolutely continues and gentlemen you have a great afternoon.

Antonov: Thank you Jim and thank you for inviting us.

Issakov: Thank you.

Jockle: Thank you.

Antonov: Goodbye.  

 

Blog Post - Jun 23, 2011

Improving Risk Management and Transparency for Structured Products

Need Assistance?

Want More From Numerix? Subscribe to our mailing list to stay current on what we're doing and thinking

Want More from Numerix?

Subscribe to our mailing list to stay current on what we're doing and thinking at Numerix

Subscribe Today!