A Bootstrap Method to Calculate Value-at-Risk in Emerging Markets under Stochastic Volatility Models

• Main Authors: Yu-Lin Yang, 楊育綾
• Other Authors: Cheng-Der Fuh
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/42244367543040247740

碩士 === 國立臺灣大學 === 財務金融學研究所 === 95 === Nowadays, risk management is an important issue. A standard benchmark used to measure and to manage market risks is the Value-at-Risk (VaR). Emerging markets have drawn considerable interest in recent years. Since it is very popular for financial institutions to have long positions in emerging stock indices, this article applies bootstrap method to calculate the VaR estimate of nine emerging market stock indices. And we also conduct the US S&P 500 composite index and MSCI EM (Emerging Markets) Index for comparison. Previous studies showed that the returns in emerging markets are leptokurtic and the volatility is higher and time-varying. Since stochastic volatility models have properties of fat tails, high and time-varying volatility, we use this model with different distributions of epsilon to fit these indices. A back-test is then employed to see which model is more proper for each index. Simulation results show that the VaR estimate is not far from the true VaR. A back-test tells that stochastic volatility models with epsilon ~ N(0,1) or ~ t(6) or ~ t(4) can fit different indices undertaken in this article. The VaR estimates are relatively high in Turkey, India, Mexico, Russia and Indonesia; while Thailand, Korea, Taiwan and Malaysia have relatively low VaR estimates. As we expect, S&P 500 index has relatively low VaR estimate. But the result that MSCI EM Index has relatively high VaR estimate indicates that the diversification effects are not significant between emerging markets.