MVA and Capital Efficiency: Accurate Dynamic SIMM Simulation via AAD by Justin Chan

MVA and Capital Efficiency: Accurate Dynamic SIMM Simulation via AAD by Justin Chan


Click to view no log in required!

Presented by Justin Chan, Quantitative Strategy, FIS

Image

Starting in September 2016 and phased through 2020, banks globally are required to start posting initial margin for bilateral trades. The aggregate value of initial margin needed worldwide is estimated to exceed $300B which will have a profound impact on the OTC derivatives market: transforming counterparty credit risk into funding and liquidity risk.

Initial margin is creating a genuine new cost of doing business, over and above the traditional funding costs.

It is crucial to accurately capture a portfolio’s lifetime funding cost of initial margin (i.e. MVA). The industry standard for calculating bilateral initial margin is the ISDA SIMM calculation, but forward simulating ISDA SIMM until the portfolio’s final maturity date is not a trivial problem. A single SIMM calculation can involve thousands of portfolio sensitivities as input, and a direct simulation at every Monte Carlo simulation node is typically not computationally feasible. However, a novel application of adjoint algorithmic differentiation (AAD) to calculate forward portfolio sensitivities makes direct forward simulation of SIMM attainable.  The technique is transparent and maintains consistency with other XVA’s.

In this webinar, we will discuss:

  • The importance of initial margin simulation in pricing, capital, and risk management,
  • The impact of volatilities calibration and simulation measure,
  • Adjoint algorithmic differentiation (AAD) and forward portfolio sensitivities,
  • Forward simulating SIMM initial margin,
  • MVA: ISDA SIMM and CCP initial margin.