Counterparty Credit Risk at a Globally Significant Investment Bank: Regression pricing and the rapid simulation of complex exotic derivatives

Challenge and Regulatory Context

  • The calculation of future exposures uses a Monte Carlo simulation (MCS) which requires thousands of calls to derivatives pricing functions
  • For complex exotic products the pricing function itself can be based on another MCS, leading to immense computational requirements that are often infeasible
  • Tactical overrides and simplistic models used to mitigate this problem are often inaccurate and infrequently updated
  • In this case the regulator required a reduction in the number of such overrides for the capital calculation of these exotics

Success

  • Exotic derivatives valued in a fraction of the time of conventional approaches
  • Tactical overrides minimised, so reducing significantly operational complexity and risk
  • New product types require very little customisation, so shortening development and model validation times
  • Technique allows efficient exposure calculation of a host of risk parameters

Approach

  • Design and development of regression pricing model
  • Integration of pricing engine with a calculation farm
  • Testing to ensure consistency between front-office models and fast regression technique
  • Resolution of differences by e.g creating a single market data source for front office and risk systems
  • Onboarding of each instrument type to the pricing framework
  • Providing evidence for, and liaising with, the model risk management teams
  • Monitoring of the framework and results

More on Fintegral’s Counterparty Credit Risk Solutions