Common dynamic factors in driving commodity prices: Implications of a generalized dynamic factor model

Abstract

Identification of the price drivers of commodity prices is difficult because economic indicators reflect commodity prices with lead or lag, and some commodities have spillover effects to other commodities. A generalized dynamic factor model is capable of accounting for these characteristics and can be applied to panel data of monthly returns of a vast variety of commodities. The empirical results indicate that four common dynamic factors exist that account for much of the variation in the commodity returns. The identification of the common dynamic factors is conducted by interchangeably creeping an economic indicator into the commodity return panel data and examining the ratio of variance explained by the common factors. The four common factors correspond to the U.S. inflation rate, the world industrial production, the world stock index, and the price of crude oil.