Feb 6, 2013

Modern Approach to Counterparty Credit Risk: Algorithmic Exposure for Counterparty Risk & CVA Across All Asset Classes

Watch The Video: Modern Approach to Counterparty Credit Risk

Serguei Issakov, Global Head of Quantitative Research at Numerix, joins host and Numerix CMO Jim Jockle, to discuss his recent research on the modern approach to calculating counterparty risk, presented at the November 2012 RiskUSA conference and featured in two recent papers available on SSRN (vanillas and exotics).

Issakov discusses the modern approach to counterparty risk, known as the modeling approach, and contrasts how this new approach differs from the scenario approach traditionally applied in the market.  He will also introduce the concept of algorithmic exposure, along with the operational risk benefits to payoff language.

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

Video Transcript: Modern Approach to Counterparty Credit Risk 

Jim Jockle (Host): Hi and welcome to Numerix Video Blog, I’m your host Jim Jockle. With me today, Dr. Serguei Issakov, Global Head of Quantitative Research at Numerix. Welcome Serguei.

Serguei Issakov (Guest): Thank you Jim.

Jockle: Serguei joining us to talk about a presentation that was given in November last year at Risk USA around the topic of counterparty credit risk, and one of the things that you discussed at this presentation was the introduction of really the modern approach for counterparty credit risk. Can you share with us what is that approach especially because so many people are thinking about it, and then what are other approaches in the market and why is Numerix following this approach?  

Issakov: Okay, basically there are two main approaches to counterparty risk currently in the market. One can be called the traditional approach which appeared, well long ago, and we call it the scenario approach. It consists of two steps. On the first step, you first generate scenarios, and then on the second step you evaluate the portfolio for counterparty risk to compute all exposures. But recently an approach appeared which is the modern approach. It appeared around 2010, and it’s based on, it can be called the modeling approach. It generates scenarios by using arbitrage free models, and the main components of that approach are hybrid modelsbecause there is a requirement that all their correlations between all asset classes have to be incorporated when computing counterparty risk and you have to have hybrid models to do that. The second component of that approach is the universal Monte Carlo method that could be applied across all these hybrid models, and that method is American Monte Carlo methods, so these are the two components.

Jockle: So one would assume in terms of the approaches that there are always tradeoffs in terms of time, performance, and especially indifferences of instruments where structures products are more exotic, wider risk factors that require a hybrid model versus something that would be just more plain vanilla, straightforward in terms of analytic pricing. Within this approach how are you handling both the vanillas and the more exotic types of instruments?

Issakov: Yeah, there are very different optimizations that can be applied to these types of instruments, even though the modern approach is universal across all the instruments. That’s a very important point. Of course for the vanilla part of your portfolio, it’s better to have some methods that perform much faster than the general hybrid models. And in that context, we implemented the super swap optimization, which means aggregation of all cash flows across a large portfolio of vanilla swaps and representing this portfolio as a single instrument. That leads to better performance and much faster computation without sacrificing any accuracy.

Jockle: And on the structured products side, you’ve introduced something called an algorithmic exposure, can you walk us through that a bit?

Issakov: Right, within the modern approach it is what we’ve introduced in our working paper available online, we call it algorithmic exposure. That means the following, for structured products you normally have to express the payoff of the instrument in some kind of payoff language or script language. There are different names for it but it it’s present in every trading system in some form. And the first version called standard version of modern approach used modification of the script to compute exposures. That adds a lot to operational risk, and what we came up with is what we call algorithmic exposure. In our approach, you don’t have to change all the existing scripts for structured deals, exotic deals, you can reuse those scripts and you can compute those exposures of as a byproducts of price computation. It involves non-trivial analytics to be able to do that at the time of price computation, but it can be proved rigorously that the result will be exactly the same, but you don’t have to change anything and it’s a very fast way to market to compute exposures. 

Jockle: In terms of the Monte Carlo simulation itself, there’s been advancements in terms of the American Monte Carlo in order to do the simulations and computations within that framework, is that correct?

Issakov: Yes that’s correct

Jockle: Okay and that was something that you’ve implemented a while ago correct?

Issakov: Yes, well at Numerix, more than ten years ago.

Jockle: Well Serguei thank you so much for joining us today. The paper that you’ve referenced as well as his presentation from Risk USA, are available on numerix.com. I would encourage everyone to take a read and download and of course please feel free to look at our blogs and follow along @nxanalytics on twitter and we’ll see you next time. Thank you!

Read More

For more detailed information regarding the above, read the recently published technical papers listed below:

[1] Alexandre Antonov, Serguei Issakov and Serguei Mechkov (2011),
     "Algorithmic Exposure and CVA for Exotic Derivatives"

Available at SSRN.

 [2] Alexandre Antonov and Dominic Brecher (2012),
      "Exposure & CVA for Large Portfolios of Vanilla Swaps: the Thin-out Optimization"

 Available at SSRN.

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!