Investigating the dynamics of commodity markets

Abstract

Many economic time-series present significant changes in their behavior in the wake of “extreme” events such as a financial crises or war. When such changes arise in the (time-series) data, a powerful statistical tool can be employed, i.e. that of Markov switching models. This dissertation employs the Markov regime switching framework to capture the time-varying dynamics in 5 commodity indexes over the period 1995:10-2015:09. These indices are the following: (a) the Commodity Agricultural Raw Materials Index; (b) the Commodity Metals Price Index; (c) the Commodity Fuel (energy) Index; (d) the Commodity Food Price Index; and (e) Gold. Also, the return of S&P 500 is used as an additional regressor. The empirical results show that the S&P 500 and the food index have a positive and statistically significant effect on all commodity returns.