Serving as a risk measure, Value-at-Risk (VaR) has achieved wide acceptance in the financial market risk management and made its way into the issue of capital adequacy in Basel rules. Using the basic Monte Carlo simulation to calculate VaR inherently suffers from sampling variability which can be mitigated only at the cost of expanding computational efforts, thus being confronted with a tradeoff between accuracy and tractability. Hence, one is motivated to research more efficient methods which are capable of achieving the required accuracy while using less simulation scenarios. This thesis deals mainly with the evaluation of various variance reduction methods in daily VaR run of a financial institute. To begin with, potential techniques such as control variates, antithetic variates, stratified sampling, Latin hypercube sampling, and importance sampling are investigated. Next, some of them are applies to VaR estimation and tested with the help of a test portfolio. Finally, the techniques are evaluated and recommended.
«
Serving as a risk measure, Value-at-Risk (VaR) has achieved wide acceptance in the financial market risk management and made its way into the issue of capital adequacy in Basel rules. Using the basic Monte Carlo simulation to calculate VaR inherently suffers from sampling variability which can be mitigated only at the cost of expanding computational efforts, thus being confronted with a tradeoff between accuracy and tractability. Hence, one is motivated to research more efficient methods which a...
»