Persistence and cycles in historical oil price data

Abstract

This paper deals with the analysis of two observed features in historical oil price data; in particular, persistence and cyclicity. Using monthly data from September 1859 to October 2013, we observe that the series presents two peaks in the spectrum, one occurring at the long run or zero frequency and the other at a cyclical frequency. These features can be well described in terms of a long memory model that incorporates both peaks in the spectrum. It is found that the order of integration at the zero frequency is about 0.6, and the one at the cyclical frequency is substantially smaller (of about 0.3) with the length of the cycles being approximately of about 74 periods (months), which is consistent with the length suggested by the business cycle theory.