MODEL APPROACH OF AGGREGATE RETURN VOLATILITY: GARCH(1,1)-COPULA VS GARCH(1,1)-BIVARIATE NORMAL

Abstract

Aggregate risk is an aggregation of single risks that are both independent and interdependent. In this study, aggregate risk is constructed from two interdependent random risk variables. The dependence between two random variables can be determined through the size of dependence and joint distribution properties. However, not all distributions have joint distribution properties; the joint distributions may be unknown, so motivating the use of the Copulas in this study is needed. Sometimes, the Copula model is introduced to construct joint distribution properties. The Copula model in this research is used in financial policies such as investment. In the investment sector, the aggregate risk comes from the sum of the single risks and returns. The model used in aggregate return is the Generalized Autoregressive Conditionally Heteroscedastic (GARCH) model. The data used in this study is the closing price data for Apple and Microsoft stocks from January 01, 2010, to January 01, 2024. The best model selection is the model with the GARCH-Bivariate Normal approach with the smallest MSE value. Model GARCH(1,1)-Bivariate Normal is the best model for the volatility model of aggregate return.