| Summary: | 碩士 === 南華大學 === 財務金融學系財務管理碩士班 === 97 === The accuracy of Value-at-Risk (VaR) estimation is important for banks to predict their market risk under the framework of Basel Committee on Banking Supervision since 1996. However, using different structural risk model might generate different VaR forecasts. Therefore, it’s crucial to verify the most appropriate VaR model actually fitted for financial institutions in practice, particular in terms of different bank type and scale. Unfortunately, the previous studies on bank VaR estimation most focus on selective bank sample or single country while international investigation on this issue is very sparse and yet addressed on cross-country comparison. For this purpose, using market trading data at public listed banks around the world, this study performs a global comparison analysis on risk model selection for parametric and non-parametric approach for evaluating the best VaR estimation in 962 banks with respect to different categories of bank size, bank type and region. More specifically, we further empirically identify key risk determinants of 99%, 95% and 90%VaR forecasts proxy as market risk at listed banks with regard to bank characteristics, risk factors on capital market, cross country differences in macroeconomic conditions, country risk, economic freedom and government supervision. The scoring system on VaR method is initially developed and used to assess the most appropriate method for our banks in sample. The results in context of international comparison indicate that the parametric VaR method using GARCH model at 99%, 95% and 90% confidence interval performs better VaR forecasts than the non-parametric one using Historical Simulation and Monte Carlo Simulation. This implicates that parametric VaR method could better capture the time-varying risk and show better risk forecasts for bank risk measurement. Furthermore, banks with higher liabilities ratio, better profitability and higher liquidity are significantly and negatively related to the VaR forecasts while they operate in financial market with lower volatility of market return and interest rate, as well as in a country with higher fiscal freedom, better financial freedom, much monetary freedom and more explicit deposit insurance.
|
|---|
