Numerix: Hedging CVA and DVA
Posted: 1 October 2013
The practical technology, quantitative and regulatory implications of CVA have taken center stage over the past several months, however the debate is now shifting towards CVA hedging due to the impact it can have on the bottom line. Particularly relevant in today’s financial reporting requirements and regulatory environments, banks are now considering the effects of hedging to reduce earnings volatility from changes in CVA, while also considering the utilization of hedging to reduce regulatory CVA capital charges. This article explores the complex hedging decisions banks are faced with today and the diverse range of instruments and costs needed to determine the potential impact of front office earnings volatility hedging versus regulatory capital reduction.
It is worth first distinguishing the calculation of CVA that is used for financial statement reporting (fair value CVA). Under the financial reporting standards governed by IAS 39 (IASB1) and FAS 157 (FASB2), banks must recognize expected losses associated with counterparty defaults on OTC derivatives trades, but neither board prescribes a specific calculation. Furthermore, both the default of the counterparty (unilateral CVA) and default of the bank (DVA) are recognized.
It is likely that banks actively hedging CVA or looking to create CVA hedging programs are using an advanced methodology for CVA calculation3. This methodology uses Monte Carlo simulation to estimate every market risk factor that impacts the derivative future values as well as taking into account market observable information for counterparty credit quality (CDS spreads, CDX proxy spreads, bond spreads). In this methodology, the bank is accounting for all contributors to the future exposures of each trade as well as all counterparty credit information in estimating potential losses from counterparty credit deterioration and default.
On the other hand, the Basel III framework4 prescribes two approaches for calculating the CVA capital charge (CVA VaR), the Standard Approach and the Advanced Approach. The Standard Approach utilizes a formula for the calculation of CVA with three different methods for estimating Expected Exposure (Current Exposure Method (CEM), Standard Method (SM) and Internal Model Method (IMM)). Of the three approaches for estimating Expected Exposure, only the IMM method uses a Monte Carlo simulation approach. The Advanced Approach utilizes a simulation-based approach for estimating Expected Exposure but in calculating CVA VaR, only counterparty credit spread volatility is taken into account. Critics of the Advanced Approach have raised concerns that it only takes into account credit spread volatility and ignores the market risk volatility from other factors, like Rates, FX, Equity, Commodity and Volatility that can affect the Expected Exposures.