Document Search

Author | |
---|---|

Title | |

Keyword | |

Abstract | |

Date: | |

To | |

Recent Comments

Accuracy versus Computational Time - A Comparison of Monte-Carlo-based Value-at-Risk Methods

**Company:**Universitat St. Gallen

**Year Of Publication:**2003

**Month Of Publication:**September

**Pages:**246

**Download Count:**4059

**View Count:**12440

**Comment Num:**0

**Language:**EN

**Source:**

**Who Can Read:**Free

**Date:**11-16-2003

**Publisher:**Administrator

**Summary**

Time is an important factor in the finance industry. Therefore it is essential to implement efficient risk calculation procedures. This thesis analyses the trade-off between accuracy and the referring computational time of Monte Carlo based Value at Risk (VaR) methods. The shown methods can also be used to calculate alternative risk measures, e.g. Lower Partial Moments or Shortfall Value at Risk. The analysed Monte Carlo methods differ in the way they approximate the profit&loss functions of portfolios. In addition to the delta and the delta gamma approximation, this work also presents other, modern grid based models, which apply techniques like principal components analysis or partial least squares methods. The accuracy of each method is tested for archetypical profit&loss functions, as the portfolios consist of variably complicated interest rate instruments (bonds, caps, floors and digital options). Furthermore, the computational times of the risk calculations are measured. The results show the relations of accuracy to computational time of the methods for the different portfolios. They can be used to find the appropriate Monte Carlo method.

**Author(s)**

This document's citation network:

Similar Documents: