Credit Valuation Adjustment Implementation of CVA PRMIA Credit Valuation Adjustment (CVA) CONGRESS IMPLEMENTATION UND PRAXIS Wolfgang Putschögl Köln, 20 th July 2011
CVA in a nutshell Usually pricing of derivatives does not take the possibility of default into account Exposure E E.g. Value of Option V(t,T) (1-R)e -rt E t E t Price should be adjusted for default risk Time t t i t i+1 V Risky (t,t) = V(t,T) CVA (t,t) + x + x + x + = CVA CVA is correction for credit-risk free calculations! Probability of Default expected value short rate exposure at time t Time t t i t i+1 2 recovery rate of counterpart time of default probability of default of the counterpart
CVA and CCR Model Macro Process Overview Scenario Generation Market data is received and a set of scenarios is produced with reference to each relevant risk factor/driver The output is a set of evolutions of risk factor scenarios (interest rates, volatilities, spreads ) Position Revaluation Value computed of every trade comprised in the defined perimeter at any time step for the full set of scenarios Output is a cube of mark-to-futures Output Aggregation Aggregating the mark-to-futures computed along any scenario according to netting and margining agreements. Computation of relevant risk figures/metrics (Exposure measures, CVA, ) TARGET SERVICES Exposure Measures Expected Exposure EPE Effective EPE Stressed EPE Marginal EPE PFE REGULATORY Effective EPE/ Stressed EPE (*) CVA VaR (*) RWAs (*) Back-testing Stress Testing CCR Credit Limit Mgt Collateral control CCR Monitoring Concentration Risk reporting CVA CVA desk services CVA Transfer Pricing P&L allocation and control CVA Accounting (*) Introduced or enriched by Basel III 3
CVA and CCR Model Operating Model Critical activities 1 2 Scenario Generation Position Revaluation CAPITAL MARKET (CVA) Model define and development Exposure measurement CVA Hedging Scenario Generation Intraday limit check Intraday CVA Pricing 4 5 6 Critical Cross Actitivities produc. and maint. Scenario Generation Intraday checks 7 8 9 Exposure Measurement c/p Mgt, review and aggr. Risk Reporting RISK MANAGEMENT (CCR) Model define and development Exposures measurement Regulatory reporting Scenario Generation Stress tests c/p review/mgt Risk Reporting 3 Output Aggregation Greeks Monitoring P&L measurement c/p modelling, mapping Backtesting CVA Desk risk limits monitoring Wrong way risk assessment Concentration risk DATA MANAGEMENT (INPUT/OUTPUT; REPOSITORY; AUDIT CHECKS; ) 1 2 3 Risk factor scenario are used to get possible evolution of exposures. The same framework can be used for both regulatory purposes (EoD) and intraday trading needs. Based on the risk factor scenarios, a full repricing of the portfolio is needed at each future time grid point until the last expiry date. Once the value of each trade in the portfolio is known until maturity, mark to futures must be aggregated according to the available legal agreements (netting/collateral) in order to get the exposures 4 5 6 Market data acquisition and validation Unobservable market data computations (correlations, illiquid names PDs) Historical series storage Financial models definition for risk factor evolution (scenario generation) Model calibration with historical (Limit/Regulatory) and risk neutral parameters (for CVA) Pre-deal: computation of the CVA charge to clients (incremental CVA) Pre-deal: Credit Limit check based on PFE Intraday: On demand monitoring of CVA and PFE 7 8 9 Definition of Financial pricing libraries for intraday and batch portfolio revaluation Intraday to allow CVA desk risk limits control Batch to measure risk exposures and RWAs Defines netting nodes and aggregation rules Management of counterparty specific data (Counterparty Group), Regulatory reporting Credit limits reports (PFE) Stress testing results and Wrong way risk Concentration risk (per issuer, country, ) 4
Product Eligibility, Limits Trade Data, EOD and Intra Day CVA and CCR Functional Model Trading Risk and Valuation Data Controls Fixed Income (Rates, Credit) Pre-Deal Limit and CVA Check Limits Exceptions Limit/CVA Manager, Exception Manager Results Data (Exposure, CVA,RWA) Client Onboarding Equities Incremental Credit Exposure, CVA P&L Vectors, Sensitivities FX and Money Market Pre-deal limit check, trade approval, CVA Charge Exposure CVA, CVA Sensitivities, CVA Charge RWA Internal and Regulatory Reporting Aggregation Engine Legal, Netting Agreements Repos and Lending Scenario Generator (support for risk neutral (CVA) and historic (Limit) scenarios) Credit Hierarchies Counterparty Data, PDs Counterparty Data Stress Test, Back Testing Revaluation Engines Netting Agreements Collateral Balances Collateral Management Current Trade and Positions Trade and Position Data Adjustments Controls and Adjustments Legal Agreements CSAs 5
CVA Trading Desk - Transaction Lifecycle and Position Valuation Product Control Trading Desks Position Valuation & PnL Client Transaction Risk Free MTM Price / PnL Net Trading Desk Position Hedging Hedging Valuation & PnL CVA Cash Payment Reconciliation Trading Desk PnL Financial Accounting Income Statement CVA Trading Desk Position Valuation & PnL Trading Desk PnL Change in CVA Valuation & CVA Desk PnL CVA Charge Cash Balance CVA Cash Balance Reconciliation Change in CVA Valuation & CVA Desk PnL Balance Sheet CVA Valuation CVA Hedging Book CVA Valuation & PnL CVA Hedging Valuation & PnL Net CVA Desk Cash & Hedge Position OTC Derivative Position MTM Fair Value Adjustment of CVA Total Net CVA per Counterparty ISDA 6
Implementing CVA - practical considerations Active approach vs. Passive approach Active approach (profit center ) Passive approach (utility function) Pricing new trades Active hedging What hedging strategies? Mandate of CVA desk? Pricing new trades Monitoring of CVA Accounting CVA IT Architecture (Build vs. buy) Data challenges Define relevant data necessary to calculate and aggregate CVA Business processes Forthcoming regulatory requirements 7
Implementing CVA - practical considerations cont d Methodology Unilateral vs. bilateral Risk neutral scenarios for exposure calculation blend of market implied and historic market data Modelling of probabilities of default historic vs. risk neutral choice of mapping approach Client segmentation observability of clients (liquid vs. illiquid c/p) Wrong way risk design and implementation Gap risk (for CSAs) 8
Implementing CVA - practical considerations cont d Credit Mapping - Example CDS Region Client Type Index Name Sector Rating CDS Index Proxy EUR CEE Invest. Grade Speculative Grade Itraxx EUR HiVOL Itraxx EUR Non-Fin Itraxx EUR SnrFin Itraxx EUR Xover Itraxx EUR SnrFin Itraxx SovX CEE Financials, Govt, Basic Materials, Consumer Goods, Consumer Services, Health Care, Industrials, Oil & Gas, Technology, Telecoms, Utilities AAA AA A BBB BB B CCC Western Eur Sovereigns Itraxx SovX WE Asia Pacific Itraxx SovX AP US CDX IG CDX HY Single name CDS 9
CVA Trading Market risk limits To manage CVA risks the CVA Trading Desk needs to be able to hedge Credit risk: trade any products (credit products, derivatives, ), EPE (Underlying risks on the derivatives): trade instruments relevant for hedging e.g. interest rate swaps, fx forwards, volatility products. CVA VaR economic CVA VaR vs Basel III CVA VaR CVA VaR managed by CVA Trading Desk Basel III CVA VaR only for CS01 (for computational reasons), still reliable sensitivities helpful to support what-if analysis for hedging Discrepancy between hedging economic CVA VaR (including exposure sensitivities) and Basel III CVA VaR (only CS01) No VaR limit for hedging trades on CVA desk Granular Market Risk Limits Risk metrics/sensitivities have to be produced without excessive noise and proven P&L attributive ability 10