In a 2012 paper for the United Nations/DESA – Super-cycles of commodity prices since the mid-nineteenth century – Bilge Erten and José Antonio Ocampo review the literature on the theory of Commodity Super-Cycles and go on to suggest that the current cycle began in 1999. Here is an extract from their concluding remarks:-
The decomposition of real commodity prices based on the BP filtering technique provides evidence of four past super-cycles ranging between 30 to 40 years. For the total real non-fuel commodities, these cycles have occurred (1) from 1894 to 1932, peaking in 1917, (2) from 1932 to 1971, peaking in 1951, (3) from 1971 to 1999, peaking in 1973, and (4) the post-2000 episode that is still ongoing. These long cycles, which possess large amplitudes varying between 20 to 40 percent higher or lower than the long-run trend, are also a characteristic of sub-indices. Among the agricultural indices, the tropical agriculture exhibits super-cycles with much larger amplitude relative to non-tropical agriculture. The amplitudes of super-cycle components of real metal and crude oil prices are comparable to those of agricultural products in earlier parts of the twentieth century, but they become much more pronounced and strong in the latter parts of the century. The presence of co-movement among non-fuel commodity indices is supported by the correlation analysis across the entire sample, and a marked co-movement between oil and non-oil indices is present for the second half of the twentieth century.
Another important finding of the paper is that, for non-oil commodities, the mean of each supercycle has a tendency to be lower than that of the previous cycle, suggesting a step-wise deterioration over the entire period in support of the Prebisch-Singer hypothesis*. This finding applies especially to tropical and non-tropical agricultural prices, as well as metals in previous cycles. An exception to this rule is that of metals during the current super-cycle, when the mean last cycle is higher than the preceding one; however, the contraction phase of this cycle has not even began yet, which can lower the mean of the whole cycle in the upcoming years. Another way of capturing these trends is through long-term trends, with tropical agricultural prices experiencing a long severe long-term downward trend through most of the twentieth century, followed by non-tropical agriculture and metals. The duration of the long-term downward trends across all non-fuel commodity groups is on average 100 years. The magnitude of cumulative decline during the downward trend is 47 percent for the non-fuel commodity prices, with recent increases of around 8 percent far from compensating for this long-term cumulative deterioration. In contrast to these trends in non-oil commodity prices, real oil prices have experienced a long-term upward trend, which was only interrupted temporarily during some four decades of the twentieth century.
The recent commodity price hike of the early twenty-first century has commonly been attributed to the strong global growth performance by the BRIC economies, and particularly China, which is particularly metal- and energy-intensive. Based on the VECM results, it is found that super-cycles in the world output level are a good predictor of the super-cycles in real non-fuel commodity prices, both for the total index and sub-indices. This finding confirms that the global output accelerations play a major role in driving the commodity price hikes over the medium run. Therefore, the ongoing commodity price boom could last only if China and other major developing countries are capable of delinking from the long period of slow growth expected in the developed countries.