Forecasting Copper Prices with Dynamic Averaging and Selection Models

Abstract

We use data from the London Metal Exchange (LME) to forecast monthly copper returns using the recently proposed dynamic model averaging and selection (DMA/DMS) methodology which incorporates time varying parameters as well as time varying model averaging and selection into a unifying framework. Using a total of 16 predictor variables that include traditional fundamental indicators such as excess demand and inventories as well as indicators related to global risk appetite, momentum, the term spread, and various other financial series such as exchange rates and stock prices, we show that there exists a considerable predictive component in copper returns. Covering an out-of-sample period from May 2002 to June 2014 and employing standard statistical evaluation criteria we show that the improvement in mean squared forecast error (MSFE) over a simple random walk (RW) benchmark can be up to 17 percent with the DMA framework, and up to 5 percent when using a simple time-varying parameter model. A visual assessment of the cumulative MSFEs shows further that a substantial part of the improvement in the forecast (relative to the RW model) is realised during the peak of the financial crisis period at the end of 2008.

Keywords: Copper forecasting, time varying parameter model, state-space modelling, dynamic averaging and selection models

JEL Classification: C11, C52, C53, G17