In this paper, the interdependence between aggregate commodity prices and world gross domestic product (GDP) is studied by performing two empirical exercises with long-run data that starts in the nineteenth century. Long−term and medium-term cycles are computed and their degree of synchronization measured for different leads and lags. Causality tests are performed on the frequency domain. Both exercises deepen understanding of these macroeconomic relationships by disentangling them on the time and frequency dimensions, respectively. The results show evidence of cycle synchronization only in the case of super cycles. There is causality evidence from GDP to aggregate commodity prices mostly for long-run frequencies. Therefore, commodity-price trends and super-cycles are demand driven. There is causality evidence between oil-prices and GDP in both causation directions. However, oil-price fluctuations are predictive of GDP for business-cycle frequencies only. Overall, a frequency-domain approach is useful for identifying significant variation in the interdependence between commodity prices and GDP across fluctuation horizons.
