Trends and Cycles in Real Commodity Prices: 1650-2010

The Prebisch-Singer hypothesis is often popularised as implying a declining long-run trend in primary commodity prices relative to manufactures, and conventional datasets to examine the hypothesis typically commence at the beginning of the 20th century. Theoretical rationales include the juxtaposition of highly competitive primary commodity markets with oligopolistic manufacturing markets, or technological and productivity differentials between core (industrial) and periphery (non-industrial) countries. However, even if such rationales are germane for the 20th century, are they equally so for prior centuries? In particular, the extant literature notes a 19th century terms of trade boom in the primary commodity producing periphery, hardly suggestive of a declining long-run trend in relative commodity prices. To disentangle trend and cyclical components for relative commodity prices we employ new and ultra-long aggregate time series, covering the period 1650-2010, and apply recent tests robust to the order of integration. Strikingly, results show that relative commodity prices present a significant and downward global trend over almost the entire capitalist age. Moreover, structural break tests suggest the full sample can be segmented into four intertemporal regimes: 1650 to the early 1820s; the early 1820s to the early 1870s; the early 1870s to the mid-1940s; and the mid-1940s to 2010. This identification of changes in the trend function provides new characterisations of historical price behaviour - for example, the 19th century terms of trade boom is captured by a local increase in prices during the second regime, superimposed on a generic long-run downward trend. Finally, Kondratieff-type cycles around this downward trend are shown to have increased in amplitude from the early 1870s onward.