Improving Loan Portfolio Optimization by Importance Sampling Techniques: Evidence on Italian Banking Books

type="main"> The aim of this work is to explore how importance sampling (IS) techniques may improve internal banking portfolio optimization models. The current economic downturn contributes to an increase in the credit risk amount of the loan portfolios reducing the quality of the banking credit. In such a difficult economic context, characterized by the credit crunch phenomenon, robust and stable methodologies of credit portfolio risk minimization, based on ‘enhanced’ Monte Carlo simulation (MCS) portfolio models, may encourage Italian banks to finance the real economy by lending also during risky and uncertain economic conditions. For these reasons, our investigation focuses on Italian banking books composed of credit loans to the 23 Italian non-financial economic sectors. In order to reduce the high volatilities both of the estimates of the credit portfolio's tail risk and of the optimal solutions in terms of credit asset allocations, we combine a two-step IS technique with a MCS model in a context of credit portfolio's Conditional Value at Risk minimization. The performance of this improved credit portfolio optimization model, in terms of accuracy of the estimates, has been investigated for the sample Italian loan portfolio in two different phases of the economic cycle, precisely before and during the current financial and economic crisis.