Estimasi Nilai Value at Risk Portofolio Menggunakan t-Copula

Value at Risk is a tool used to measure the risk of loss on a specific portfolio investment. Value at Risk is the maximum loss of investment of financial products with a given confidence level over a given period of time. Several methods have been developed to estimate Value at Risk, such as historical data simulations, Monte Carlo simulations, GARCH(1,1) method, EWMA etc. However, those methods are considered to be unable to explain the structure of dependence among random variables constructing the portfolios. By using copula functions, the dependence among assets in multivariate distributions can be modeled for which the behavior of the marginal can be observed in more detail.In this paper, t-copula function is used to model the dependence structure of the multivariate distribution of asset return portfolio. t-copula is able to model and estimate the t-student multivariate distribution without need to assume the normality of the random variables. t-copula function needs i.i.d. (independent and identically distributed) characterization of data. The empirical data used in this research is Jakarta Stock Exchange index dan Kualalumpur Stock Exchange index recorded during the period of 30 May 2008 to 30 May 2013 (1270 observations). The Value at Risk estimated uses the time horizon 22 days with confidence levels of 90%, 95%, and 99%. Comparing with normal (Gaussian) copula, t-copula performs better than Gaussian copula. This result agrees with the literature, even the difference in our result is not quit significant