American Monte Carlo: A Practitioner Approach

Giovanni Cesari

Since the 2007–9 global financial crisis there has been growing interest from financial institutions and regulators in computing, monitoring and possibly hedging or optimising counterparty credit exposure, credit valuation adjustment (CVA), risk weighted assets (RWA) and funding valuation adjustment (FVA). All these quantities can be obtained from transformations of price distributions computed at different time steps, from trade inception to maturity. As CVA, RWA and FVA need to be evaluated at portfolio level rather than at trade level, analytics are in general based on Monte Carlo simulations.

Conceptually, there are two steps in computing price distributions by Monte Carlo: simulation followed by pricing. In Figure 5.1 we show a schematic view of this approach. First, we need to simulate scenarios from the distribution of the underlying processes that drive the price of the products concerned. Second, the prices of these products need to be evaluated at each time step in the simulation schedule for each of the simulated scenarios.

The choice of models for generating scenarios requires a balance between sophistication for exotics products and simplicity to facilitate use