Oil prices: Breaks and trends

Breaks, trends, and unit roots in spot prices for crude oil and petroleum products

“…Rejecting unit roots in high frequency data by employing tests that are of superior power can be considered genuine evidence of stationarity (Ghoshray and Johnson, 2010). Such studies are few, with Ghoshray and Johnson (2010) and Noguera (2013) perhaps being the closest to this paper. Ghoshray and Johnson (2010) look at the monthly observations of Brent, WTI, natural gas and coal prices over 1975-2007.…”

“…Ghoshray and Johnson (2010) find evidence of stationarity in the monthly prices of crude oil from 1975 to 2007, using unit root tests of Zivot and Andrews (1992) and Lee and Strazicich (2003). Noguera (2013) finds multiple breaks in monthly WTI oil prices over 1861-2011, and shows evidence against unit roots using the ADF and PP tests.…”

“…Furthermore, we extend our "one-break" analysis to the "multiplebreaks" case and test the unit root hypothesis with allowing for multiple breaks. Motivated by Noguera (2013), we employ the Kejriwal-Perron test to determine the number and time of breaks; then, we perform the unit root test proposed by Carrion-i-Silvestre et al (2009) on the price series with multiple breaks, allowing for breaks in both the null and alternative hypotheses. Because we now include both major and minor breaks in this unit root test, we can use the results to check the robustness of the Kim-Perron test.…”

“…More recently, by extending the testing procedures to the case of multiple breaks, Kejriwal and Perron (2010) propose a sequential procedure to endogenously determine multiple structural breaks, without knowing whether the noise term is stationary. Noguera (2013) uses the Kejriwal-Perron test on monthly real oil price data, and determines twelve structural changes, both in level and in trend, covering 1861 to 2011. Because our data cover a shorter time span of approximately 28 years, we primarily base our analysis on the Perron-Yabu trend test to detect the most important break (if any exist).…”

“…Because our data cover a shorter time span of approximately 28 years, we primarily base our analysis on the Perron-Yabu trend test to detect the most important break (if any exist). We also employ the Kejriwal-Perron test to include the minor breaks and determine the number of breaks used for unit root test with multiple breaks, as in Noguera (2013).…”