GARCH based VaR estimation: An empirical evidence from BRICS stock markets

• Main Authors: Siva Kiran GUPTHA. K, Prabhakar RAO. R
• Format: Article
• Language: English
• Published: General Association of Economists from Romania 2019-12-01
• Series: Theoretical and Applied Economics
• Subjects: emerging markets; garch models; volatility and leverage; var estimation; backtesting
• Online Access: http://store.ectap.ro/articole/1430.pdf

This paper examines the adequacy of GARCH based VaR models in risk estimation for BRICS emerging stock markets. This study uses the daily data of stock indices in these markets for the period 25th September 1997 to 30th March 2018. Here we employ SGARCH, EGARCH and GJR-GARCH models to test volatility persistence and leverage effect of these markets. It is observed that the volatility persistence and leverage effect is present in all these markets. In GARCH estimation the error distribution - students t is found to be suitable for Brazil, Russia, India, and South Africa whereas GED for China. From the backtesting results of Kupiec and Christoffersen test, it is found that these models are appropriate for Brazil, Russia, India, and South Africa in risk estimation at 99% one day VaR.