ABSTRACT
The rise of oil prices is a main issue in contemporary economics. This study examines the monthly, weekly and daily structure in several oil prices series using a modeling approach based on fractional integration and long-range dependence. The results indicate that oil prices series are highly persistent, with orders of integration equal to or higher than 1. Breaks in the series do not alter the main conclusions of this study. That means that shocks have a permanent nature and strong policy measures must be implemented to return the series to their original long-term projections.
Notes
1 See, for example, Lean and Smyth (Citation2009) for the relevance of testing for unit roots.
2 An I(0) process is defined as a covariance stationary process with a spectral density function that is positive and finite at the zero frequency. This includes the standard cases of white noise, stationary AR, MA, stationary ARMA, etc.
3 Note, however, that fractional integration may also occur at some other frequencies away from 0, as in the case of seasonal/cyclical (fractional) models (see, Arteche and Robinson, Citation1999; Arteche, Citation2002).
4 For the purposes of simplicity we have only considered here the case of a single break, though multiple breaks are also feasible with this procedure.
5 Here we employ a grid of 0.001 values for d. That is, we test Ho: d = do in (6) with do = 0, 0.001, …, 2.
6 In fact, when the estimates of d are close to 0 (as in the case of the daily data), the AR parameters are then found to be extremely close to 1.
7 Convergence was not achieved in the case of the monthly data.