Are the Garchmodels Still Best in Out-of-Sample Predictive Performance When Estimated for Gold Bullion Return?

By : Prasant Sarangi

Page No: 194-203

Abstract
India as a fast growing prospective economy requires strong future markets like any other developed economics. Bullion trading, more particularly, gold is of high interest for analysts, researchers, more particularly to investors as it will pave them a way with an opportunity to benefit financially by investing their limited resources. Gold is considered by many, to be the best investment one can make to protect self and family during stock market declines and inflation. But the greatest confusion is when to invest? One alternative may lie with a search of an appropriate forecasting model. In this light, this paper is an attempt to make a selection between the most successful and widely used GARCH types of models with that of Artificial Neural Networks (ANN) and regression models. For this twenty various specifications of GARCH types of models, same numbers of ANN models and one multiple regression model are constructed to predict gold return series by using 2962 days daily data ranging from 11 January 2003 to 20 August 2013. Forecasting errors when calculated by using six forecasting error measures, GARCH (3,3) model emerges as the best suitable model for forecasting gold series. MSE is found to be best suitable for estimating forecasting errors.

Author :
Prasant Sarangi was awarded B.A (Eco.-Hons.) and also M.A.(Economics) first position in order of merit. Presently working as Faculty at Apeejay Institute of Technology, Greater Noida. Published so far more than 40 research papers in several journals of national and international repute. He has the credit of authoring books entitled ‘Research Methodology, Taxmann Publications Pvt. Ltd’; ‘Consumption, Poverty and Inequality; and Editor, Innovations in Management Sciences. Dr. Sarangi has also been associated with a number of reputed professional bodies in India. His areas of interest include Research Methodology, Business Economics, Derivative Markets and Marketing India.