Partial Differential Equation Representations of Derivatives with Bilateral Counterparty Risk and Funding Costs
Barclays Investment Bank
November 23, 2010
We derive a partial differential equation (PDE) representation for the value of financial derivatives with bilateral counterparty risk and funding costs. The model is very general in that the funding rate may be different for lending and borrowing and the mark-to-market value at default can be specified exogenously. The buying back of a party's own bonds is a key part of the delta hedging strategy; we discuss how the cash account of the replication strategy provides sufficient funds for this.Download This Paper
First, we assume that the mark-to-market value at default is given by the total value of the derivative, which includes counterparty risk. We find that the resulting pricing PDE becomes non-linear, except in special cases, when the non-linear terms vanish and a Feynman-Kac representation of the total value can be obtained. In these cases, the total value of the derivative can be decomposed into the default-free value plus a bilateral credit valuation and funding adjustment.
Second, we assume that the mark-to-market value at default is given by the counterparty-riskless value of the derivative. This time, the resulting PDE is linear and the corresponding Feynman-Kac representation is used to decompose the total value of the derivative into the default-free value plus bilateral credit valuation and funding cost adjustments.
A numerical example shows that the effect on the valuation adjustments of a non-zero funding spread can be significant.
The Addendum for this paper is available at the following URL: http://ssrn.com/abstract=2109723
Number of Pages in PDF File: 19
Keywords: Counterparty risk, Credit Valuation Adjustment, Funding costs, PDE, Feynman-Kac Theorem
JEL Classification: G13, C63working papers series
Date posted: May 13, 2010 ; Last revised: July 19, 2012
© 2013 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo5 in 1.375 seconds