New York - Analytics provider Numerix said it had published a paper that establishes a new algorithmic method for calculating counterparty exposure for exotic portfolios, with important implications in the context of Basel III.

The method has applications for computing Monte Carlo value at risk, expected shortfall, potential future exposure and credit value adjustment (CVA), the company said. Numeric said CVA, which concerns the price of any financial instrument due to the possible default of a counterparty, is recognised as a more advanced measure of counterparty risk and required by Basel III.

"This paper introduces a theoretical framework that can be used to reconcile various approaches to computing risk with the same level of speed and accuracy as pricing, helping to create a common language for the front and middle offices," said Serguei Issakov, senior vice president, quantitative research and development.

"While there is a consensus on how to compute price, there are various approaches to achieve risk computations - which don't always agree with each other. We found that by calculating exposure in parallel with pricing a unified, more efficient approach can be taken to computing complex risk measures."

Numerix said the concept of exposure-centric analytics, which generalises the existing price-centric analytics, was fundamental to the approach. "As such, this method also naturally lends itself to computing economic scenario generators by applying economic variables to the scenario generation framework."

Steven O'Hanlon, president and chief operating officer at Numerix, described the research as "groundbreaking".

"Under Basel III CVA changes the most fundamental assumptions in OTC derivatives pricing requiring a portfolio and counterparty level view of trades," O'Hanlon said. This, he added, can be costly and challenging.

Numerix provided the following abstract of the research paper:

"We develop the algorithmic approach for Counterparty exposure calculation and automate its application to arbitrary complicated instruments. Assuming that the portfolio is priced by the backward (American) Monte-Carlo method, our approach allows calculating the credit exposure as a pricing by-product, essentially without modifications in the usual pricing procedure. In particular, for the exposure calculation of callable instruments we manage to avoid a cumbersome aggregation of exercise indicators, applying them sequentially in parallel with the main pricing. We explain how the obtained exposure can be integrated into the Credit Valuation Adjustment (CVA), based on the extension of the pricing model with a Counterparty credit component. The presented approach to the exposure computation is formulated in an arbitrary probability measure. To perform the measure change we use the cross-currency model semantics and calibrate the model to the real-world measure using indexes projections."