Asset Allocation Based On The Filtered Historical Simulation Value-at-Risk*

By Viral Acharya, Giuseppe Arbia, Giuseppe Corvasce, Matthew Richardson

The paper proposes the Filtered Historical Simulation (FHS) Value-at-Risk computed at 95% and 99%, for ranking a certain number of stocks. FHS retains the non-parametric nature of historical simulation by bootstrapping from the standardized residuals. These bootstrapped standardized residuals are then used to generate time paths of future asset returns. The empirical analysis is based on the constituents of the Dow Jones Industrial Average index and derives the results for a simple exercise of asset allocation. The simulations assess the Value-at-Risk, relying on 50,000 and 1,000,000 independent random trials over n. 10 and n. 22 trading days.

Contact Prof. Dr. Giuseppe Corvasce with questions about implementation.