Loading...

A New Approach in Value-at-Risk (VaR) Estimation by Forecast Combination Methods

Seraj, Mostafa | 2012

503 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 42844 (44)
  4. University: Sharif University of Technology
  5. Department: Management and Economics
  6. Advisor(s): Barakchian, Mahdi
  7. Abstract:
  8. Value-at-Risk (VaR) is the most commontool for risk management. This tool is used to measure market risk and also used as a basis in determining financial standards for international financial institutions. VaR is the maximum loss of the asset portfolio at the specified confidence level and certain time horizon. Many parametric, nonparametric and semi parametric methods have been invented for VaR estimation. Each one of these methods has its advantages and disadvantages and different methods may perform better in differnet situations.When estimating VaR, we can choose one of these methods or we can combine the VaRs estimated by different methods. There are few researches conducted on VaR combination. In this research, we evaluate the performance of combination methods such as simple linear combination and clustering in the VaR forecasting field. In order to propose a model which is immune to the disadvantages of the current models,which has more prediction power in measuring the market risk and also follows the market changes, the clustering methods are employed based on four criteria: accuracy, conservativeness, aggressiveness, and efficiency. Our results show that the VaR combination methods perform more efficient and more accurate than individual parametric and nonparametric methods. Besides, this approach enables risk managers to choose appropriate methods with regard to their desired level of conservativeness, accuracy and efficiency
  9. Keywords:
  10. Value at Risk ; Forecast Combination ; Clustering ; Backtesting ; Risk Measure Evaluation

 Digital Object List

 Bookmark

No TOC