Extreme value approach: The fourth model of value-at-risk

• Main Authors: Lucero, Joe Marcoln R., Pacumio, Margaret C., Ramoy, Andrea Gabrielle L., Sy, Jerick Mark G.
• Format: text
• Language: English
• Published: Animo Repository 2017
• Subjects: Mutual funds > Philippines; Risk management > Philippines
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/8500
• Institution: De La Salle University

This study focuses on assessing the accuracy of the existing and widely accepted value-at-risk estimation models-- traditional historical simulation, parametric method, Monte Carlo simulation and the relatively new method, the extreme value approach. Moreover, this study emphasizes the volatility of different mutual fund classes and the significance of VaR to its stakeholders. The proponents implemented the top Philippine peso denominated mutual funds that have ten (10) or more years of existence in the market and applied three (3) backtesting methods-- Kupiec test, Christoffersen test and Hendricks test to verify the accuracy of the results and to examine the reliability of the four (4) methods. Results show that in terms of accuracy and variability, the rank of estimation methods are also as follows: Monte Carlo simulation, traditional historical simulation, extreme value approach and parametric methods. On the other hand, in terms of accuracy only, the rank of estimation methods are as follows: traditional historical simulation, Monte Carlo simulation, extreme value approach and parametric method.