A Ugandan bank needed a model to calculate their twelve-month Earnings at Risk (EaR) figures and allow them to apply non-parallel shocks to the interest rates influencing their Income Statement.

Solution:
Riskworx developed a model using as inputs:
the monthly positions of the Balance Sheet’s interest-bearing products &
user-defined projected interest rates for the next 12 months – the “House View”

The model generates Monte Carlo interest rate scenarios based on historical correlations between the rates that influence the Income Statement. To ensure that the Monte Carlo rates produced for each path remain correlated as per historical correlations, the model uses Cholesky decomposition of the correlation matrix of the historical data as part of the generation of the normally distributed rates.

To speed up the model run time, and to make the Monte Carlo simulation converge with fewer paths, antithetic variance methods are used to produce the rate paths. Interest income & interest expenses are forecast based on the House View as well as the Monte Carlo simulated rate projections.

“The model generates Monte Carlo interest rate scenarios based on historical correlations between the rates that influence the Income Statement. ”

Value
Projected monthly Net Interest Income (NII) results are produced for a 12-month horizon & the EaR figures are calculated for a number of confidence intervals. The model also enables the user to shock different rates by a different number of basis points, and at different time periods, to represent the lead or lag between interest rates, thereby enabling a non-parallel rate shock.