Academic journal article South Asian Journal of Management

Dynamic Forecasting: Efficacy of Rolling Symmetric and Asymmetric GARCH Models

Academic journal article South Asian Journal of Management

Dynamic Forecasting: Efficacy of Rolling Symmetric and Asymmetric GARCH Models

Article excerpt

(ProQuest: ... denotes formulae omitted.)


With the rise in the importance of financial markets and growing cross-border equity investments, volatility measurement and prediction through modeling has attracted increased attention of researchers over the years. Researchers in pursuit of reliable and accurate risk or volatility forecasting techniques have proposed several types of models. In fact, volatility measurement and estimation has become the focal point of financial time-series econometrics in the recent past. Some of the frequently used models for volatility modeling are Autoregressive Moving Average (ARMA) model, Autoregressive Conditional Heteroskedasticity (ARCH) model, symmetric and asymmetric generalized autoregressive conditional heteroskedasticity (GARCH) models and regime switching models. Most volatility models use past data to predict future volatility. It can be countered that future is unforeseeable and therefore future events are unpredictable, but there exists sufficient evidence to support the contrary. Research findings have proved beyond doubt that financial volatility exhibits considerable autocorrelation with the past values and has a distinct tendency to cluster (Tsay, 2005). A phase of high volatility seems to sustain for a considerable period before the market settles into a more stable movement. This observed behavior of volatility creates a scope for use of autoregressive approach to build reasonably reliable volatility models used in the current study.

This paper explores various autoregressive volatility models to develop a volatility estimation equation for Jakarta Stock Exchange Composite Index (JCI). The primary objective of this paper is to extend the research on volatility modeling further by evaluating the effectiveness of various GARCH-family models for dynamic forecasting of equity markets' future values. The current research also aims to bridge the gap between extreme academic interest in GARCH models on one hand and their very limited use by practitioners in practical applications on the other hand by assessing the efficacy of various GARCH-family models in explaining the future volatility of the JCI. Before applying the GARCH-family models to develop dynamic forecasts for JCI, the author has studied the causality relationship of JCI with Mexican Bolsa IPC Index (MEXBO) from Mexico, Nigerian Stock Exchange All Share Index (NGSE) from Nigeria, and the Borsa Istanbul 100 Index (XU100) from Turkey using Vector Autoregression (VAR). These four indices have been chosen as they represent a new grouping-Mexico, Indonesia, Nigeria and Turkey-abbreviated as MINT, ideated by Terence James O'Neill in 2014. O'Neill had started a new trend in cross-country groupings when he coined the acronym BRIC (the acronym BRICS was originally coined as BRIC for Brazil, Russia, India and China, and became BRICS after South Africa was included as well) in 2001. MINT is definitely going to excite the interest of researchers and investors who have seen BRICS evolve into a powerhouse combination during the past decade (Wright, 2014).

MINT countries represent emerging market economies and make an appealing subject for research study. In the current study, the author has studied MINT as a group in the context of their equity markets represented by their leading indices. Episodic nonlinearities observed in the stock markets indices on account of political and financial instability (Bonilla, Romero-Meza and Hinich, 2006; and Romero-Meza, Bonilla and Hinich, 2007) make emerging economies most suitable for applying various market models for ex-post analysis and forecast of future values. By using the MINT markets' data to evaluate the efficacy of dynamic equity forecasts using GARCHfamily models, the author intends to augment the otherwise sparse research on the MINT group in this domain.

To achieve the objectives of the study, the author has forecast the stock market volatility for Indonesia by its returns using GARCH-family models and compared their forecasting performance. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.