Two Methods of Backtesting for Evaluating Value-at-Risk Models

Please enable javascript in your browser.

This page was sent successfuly

Bookmark

836 viewed

ارائه دو روش پس آزمایی برای بررسی عملکرد روش های محاسبه ارزش در معرض ریسک

836 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 45863 (44)
University: Sharif University of Technology
Department: Management and Economics
Advisor(s): Zamani, Shiva
Abstract:
This thesis proposes two methods for backtesting VaR models. The first is the combination of saddlepoint technique with Berkowitz backtesting and the second is based on maximum loss which uses Fischer-Tippet theorem to backtest VaR models. Monte Carlo simulation studies show that the power of these new backtests, especially the latter which is easy to use, is not less than complex Backtests that are well-known for their accuracy
Keywords:
Backtesting ; Value at Risk ; Saddle Point Approximation ; Maximum Loss ; Berkowitz Backtesting

No TOC