How important are common factors in driving non-fuel commodity prices? A dynamic factor analysis

This paper analyses the importance of common factors in shaping non-fuel commodity price movements for the period 1957-2008. For this purpose, a dynamic factor model is estimated using Kalman Filtering techniques. Based on this set-up we are able to separate common and idiosyncratic developments of commodity prices. Our estimation results show that there exists one common significant factor for most non-fuel commodity prices and that this common factor has recently become increasingly important in driving non-fuel commodity prices. However during the seventies and early eighties, the co-movement was much higher. In a next step, we then rely on an instrumental variable approach to uncover which variables could be linked to the common factor. We find that the main statistically significant variables are the oil price, the US dollar effective exchange rate, the real interest rate but more recently also global demand (as measured by a proxy for global industrial production).