Identification problems in Granger causality tests based on the net oil price increase

The net oil price increase is one of the most popular models used to study the relationship between changes in the price of oil and macroeconomic activity. The model postulates that an increase in the price of oil has negative consequences for the economy if the new price exceeds the maximum price observed over a reference period of arbitrary length. The relationship between the net oil price increase and other economic variables is often evaluated with Granger causality tests, the results of which are sensitive to the choice of the reference period. If the reference price is chosen to best fit the data, it becomes an unidentified nuisance parameter under the null hypothesis, causing standard tests to over-reject the null. This article proposes a simple method to obtain correct critical values. Using US data for the period 1954 to 2012, it is found that these corrected critical values reduce, but do not eliminate support for the proposition that the net oil price increase Granger causes real USGDP growth.