We study the interdependence between real commodity prices and world real GDP using long-term annual data since 1870, by performing two empirical exercises. First, we compute long-term and medium-term cycles and measure their degree of synchronization for different leads and lags. Second, we perform several causality tests in order to better understand the nature of their interdependence. Our results show that GDP and commodity-price cycles are correlated, and there is evidence of short-term causality between them. However, there is no evidence of Granger causality from GDP to medium and long term cycles of commodity prices. This finding is consistent with the technology-based theories of commodity-price cycles. Searching for a supply-side determinant, we study the interdependence between oil-price and the remaining commodity-price cycles. Our results imply that oil prices are key drivers of metal price cycles for all fluctuation frequencies.

We study the interdependence between real commodity prices and world real GDP using long-term annual data since 1870, by performing two empirical exercises. First, we compute long-term and medium-term cycles and measure their degree of synchronization for different leads and lags. Second, we perform several causality tests in order to better understand the nature of their interdependence. Our results show that GDP and commodity-price cycles are correlated, and there is evidence of short-term causality between them. However, there is no evidence of Granger causality from GDP to medium and long term cycles of commodity prices. This finding is consistent with the technology-based theories of commodity-price cycles. Searching for a supply-side determinant, we study the interdependence between oil-price and the remaining commodity-price cycles. Our results imply that oil prices are key drivers of metal price cycles for all fluctuation frequencies.JEL Classification: C22, E32, Q02Recent literature has documented the relation between the recent high-growth periods in developing economies and the dynamics of CP. This relation is discussed, among others, by Garnaut (2012), Byrne, Fazio andFiess (2013) andFarooki (2009). Furthermore, Erten and Ocampo (2013), using a vector error-correction approach (VEC), find evidence of causality running from global output to CP by performing significance tests on the magnitude of the speed of convergence of these variables to their long-run equilibrium.This document extends the analysis of Erten and Ocampo (2013) by studying the interdependence between CP and global economic activity using alternative econometric methods. Namely we apply tests for instantaneous causality, standard Granger causality and Granger causality on the frequency domain.According to Cuddington and Jerrett (2008), the importance of the long-term fluctuations