Title: Modelling long memory dependence structure using FIGARCH-copula approach - evidence from major Asian stock markets

Authors: Pankaj Kumar Gupta; Prabhat Mittal

Addresses: Center for Management Studies, Jamia Milia Islamia, New Delhi, India ' Satyawati College (Eve.), University of Delhi, New Delhi, India

Abstract: Increased volatility in the stock markets has led the market to originate a new variety of techniques to predict markets efficiently. The aim of the study is to scrutinise the potential dependence among different Asian stock markets, using the FIGARCH copula approach. In the first step, the marginal distribution for the copula has been estimated with the best-fit approach using minimum AIC on the underlying assumptions of normal, Student-t and generalised error distribution (GED). The results indicate that Student-t best fits for the return series SHANGHAI and NIKKEI, while GED for the HANG SENG, KOSPI and NIFTY. In the next step, we have used Gaussian, Student-t, and Clayton copula to estimate the parameters and the dependence measures. The performance of the three copula distributions has been compared based on AIC and BIC criteria. We find t-copula performs better than the other two copula functions.

Keywords: long memory; volatility forecasting; stock market; fractionally integrated-GARCH; copula.

DOI: 10.1504/GBER.2024.135287

Global Business and Economics Review, 2024 Vol.30 No.1, pp.56 - 71

Received: 23 Jan 2022
Accepted: 15 Aug 2022
Published online: 04 Dec 2023 *