A characterization of oil price behavior - Evidence from jump models

This paper is concerned with the statistical behavior of oil prices in two ways. It, firstly, applies a combined jump GARCH model in order to characterize the behavior of daily, weekly as well as monthly oil prices. Secondly, it relates its empirical results to implications of Hotelling-type resource extraction models. The empirical analysis shows that oil prices are characterized by GARCH as well as conditional jump behavior and that a considerable portion of the total variance is triggered by sudden extreme price movements. This finding implies that, first, oil price signals are not reliable and, as a consequence, both finding optimal extraction paths and decisions regarding the transmission to alternative technologies are likely to be compromised. Second, this behavior is in stark contrast to the notion of deterministic trends in the price of oil.

► Strong evidence of jumps in daily, weekly as well as monthly oil prices ► A considerable portion of the variance is triggered by jumps. ► Considerable change in oil price behavior after the mid-1990s ► Results are in stark contrast with the notion of deterministic trends.