Comparing Between Multivariate Volatility Models in Estimation of Exchange Rate and Stock Index Relationship

Document Type : Original Article

Authors
Hosein Abbasinejad 1
Shapour Mohammadi 2
Sajad Ebrahimi 3
1 Professor of Economics, University of Tehran , E-mail:habasi@ut.ac.ir
2 Associate Professor, University of Tehran , E-mail:shmohammadi@gmail.com
3 Ph.D. student of Economics, University of Tehran, E-mail : ebrahimi_s@ut.ac.ir
Abstract
Financial variables volatility as basic factor of financial assets pricing, has been observed by many studies. In addition to GARCH model as common model for volatility estimation, stochastic volatility (SV) model is another approach that rarely is applied in research. In this paper base of daily data from 1381 until 1392 exchange rate and Tehran stock index volatilities are estimated by applying bivariate stochastic volatility (SV). In order to evaluating the result, a loss function is formed based on value at risk(VaR) and then volatility models result(including stochastic volatility(SV) and GARCH) compare to each other. According to Loss function comparing result, BEKK-GARCH with t student distribution has more accurate estimation of exchange rate and stock market index volatilities. In addition, the results of the best model show that increasing exchange rate growth leads to stock index growth, but stock market changes have not significant effect on exchange rate growth. Also rising in volatility of a market causes increasing in volatility in another one.
Keywords
Stochastic volatility (SV)
Value at Risk (VaR)
multivariate garch
stock index
exchange rate