IMPLICATIONS OF DEFLATING COMMODITY PRICES FOR TIME-SERIES ANALYSIS

The choice of deflators of commodity prices can change the time-series properties of the original series. This is a specific application of the general phenomenon that various kinds of data transformations can create spurious cycles that did not exist in the original data. Different empirical models of expectations result from nominal and various deflated series that have distinct time-series properties, and these models, in turn, produce varying estimates of supply response and measures of price risk. The foregoing is illustrated by annual grain prices, monthly milk prices, and a milk supply analysis. Annual prices of corn and soybeans, for example, appear to vary around a constant mean, but when deflated by general price indexes such as the CPI, the deflated prices are autocorrelated around a declining deterministic trend and/or have a stochastic trend. The quasi-rational expectations hypotheses assumes that farmers base expectations on forecasts from time-series models, but forecasts of real prices, that ultimately become negative, are not rational.