Modelling time-varying volatility in non-ferrous metals markets
Conference Paper
Recently the modelling and forecasting of volatility has received much attention in the literature. As volatility is generally unobservable, it must be estimated. The GARCH(1,1) specification remains the most widely used time-varying financial volatility model in practice. This paper evaluates the adequacy and effectiveness of AR(1)-GARCH(1,1) in modelling and forecasting volatility in daily price returns on futures contracts for the two most important metals traded on the London Metal Exchange, namely aluminium and copper. The empirical analysis examines the properties of parameter estimates, robust t-ratios, moment conditions,and forecasts derived from rolling regressions.
Keywords: GARCH, Futures, Volatility, Rolling regression, Metals