Scenario Generation for Asset and Liability Management Models Applied to a Saudi Arabian Pension Fund

In Asset and Liability Management (ALM) models, there are parameters whose values not known with certainty at decision time, such as future asset returns, liability contribution values. Simulation models generate possible “scenarios” for these parameters, which used inputs in the optimisation help thus making decisions. These decisions can be evaluated sample, on same scenarios that were decision, out-of-sample, a different, usually much larger, scenario set. With return simulation, major difficulty lies multivariate nature of data. We propose to capture this via historical copula, no distributional assumptions. suggest use univariate sample generation allows different returns modelled by distributions. The liabilities contributions have main source uncertainty population numbers; we model adapting biology (BIDE). resulting generator four ALM using dataset from largest Saudi Arabian pension fund market index.