Stock returns in emerging markets and the use of GARCH models

Abstract

We use the Hinich portmanteau bicorrelation test to detect for the adequacy of using GARCH (Generalized Autoregressive Conditional Heteroscedasticity) as the data-generating process to model conditional volatility of stock market index rates of return in 13 emerging economies. We find that a GARCH formulation or any of its variants fail to provide an adequate characterization for the underlying process of the 13 emerging stock market indices. We also study whether there exist evidence of ARCH effects, over windows of 200, 400 and 800 observations, using Engle's LM (Lagrange Multiplier) test, and find that there exist long periods of time with no evidence of ARCH effects. The results suggest that policymakers should use caution when using autoregressive models for policy analysis and forecast because the inadequacy of GARCH models has strong implications for the pricing of stock index options, portfolio selection and risk management. Specially, measures of spillover effects and output volatility may not be accurate when using GARCH models to evaluate economic policy.

Spillover effect in Asian financial markets: A VAR-structural GARCH analysis

“…This index is computed on the basis of variance decomposition, and can provide a precise measurement of spillovers between markets. It has been accepted in academic literature, and used in a number of recent studies, for example, Bonilla andSepulveda (2011), Gebka (2012), Diebold (2012), and Fujiwara and Takahashi (2012).…”

Investigating Abnormal Volatility Transmission Patterns Between Emerging and Developed Stock Markets: A Case Study

“…Srinivasan (2011) conducted an empirical study on modelling and forecasting the stock market volatility of S&P 500 index using GARCH family models and concluded that symmetric GARCH model it fits much better in forecasting the conditional variance of emerging stock market return series compared to the asymmetric GARCH model. Bonilla and Sepúlveda (2011) have investigated suitability of GARCH models for modelling conditional volatility of 13 selected emerging stock market indices and have suggested an inadequacy of Generalized Autoregressive Conditional Heteroscedasticity models in the case of spillover effects and output volatility. Vidanage et al (2017) have conducted an empirical research study on various emerging Asian economies, i.e.…”

Shock and Volatility Spillovers between Stock Markets of India and Select Asian Economies

“…The outcome of these studies have produced explicit evidences which substantiate the unstable periods and qualified nature of volatility (Panagiotidis et al, 2003;Xiao and Dhesi, 2010;Bonilla and Sepulveda, 2011). Accordingly, there is a growing propensity to employ nonlinear models for the evaluation of the spillover of volatility (Talwar, 2016).…”

“…Large number of researches (Karolyi, 1995;Hong, 2001;Koopman et al, 2005;Hammoudeh and Li, 2008;Khemiri, 2011;Bonilla and Sepulveda, 2011;Fiszeder and Orzeszko, 2012;Salma, 2015;Mittal and Kumar, 2016;Demirer, Gupta and Wong, 2019) adopted non-linear ARCH/GARCH family models (Engle, 1982;Bollerslev, 1986) for probing precariousness of different types of time series data. The literature further state that predominance of the research studies in Asia have implemented vector autoregression (VAR) framework, Cointegration and Granger causality test only (Joshi, 2011;Tuan et al, 2015).…”