Risk optimisation: the noise is the signal

Benedict Burnett, Simon O’Callaghan and Tom Hulme introduce a new method of optimising the accuracy and time taken to calculate risk for an XVA trading book. They show how to make a dynamic choice of the number of paths and time discretisation focusing computational effort on calculations that give the most information in explaining the P&L




The make-up of bank trading books can be highly variable. At one end of the spectrum are flow books, consisting of a fairly homogeneous set of quick-to-value trades. At the other end, we have valuation adjustment books, referencing a heterogeneous set of counterparties, each with different underlying trade populations. These trades, and hence the XVAs for different counterparties, will vary enormously both in the time needed for calculation and the magnitude of the r