Correlation in volatility among related commodity markets

 

Related commodity markets have two characteristics (i) they may follow similar volatility processes; and (ii) such markets are frequently represented by a market aggregate or index. Indices are used to represent the performance and time series properties of a group of markets. An important issue regarding the time series properties of an index is how it reflects the time series properties of its components, particularly with regard to volatility. In this paper, correlation matrices are derived from rolling AR(1)-GARCH(1,1) model estimates to examine the second and fourth moment properties of ARMA processes with GARCH errors, and are also compared with the properties of the individual returns series. The correlations between the volatility of returns on several 3-month non-ferrous metals futures contracts traded on the London Metal Exchange are examined for aluminium, copper, nickel, lead, tin and zinc. Relationships between the volatility of individual metals returns and returns on the London Metal Exchange Base Metal Index are also examined.

Keywords: Volatility forecasting, GARCH, Rolling models, Futures contracts, Industrial metals, Commodities, Metals, Cross-sectional aggregation