A Variance Ratio Test of Random Walk in Energy Spot Markets

This study tests the random walk hypothesis in the spot prices of the petroleum products markets. Under the variance ratio test, a less restrictive random walk process namely the martingale process is examined over the period 1998-2008. The variance ratio methodology is capable of providing information regarding the linearity of multi-period variances, serial correlation as well as possible conditional heteroscedastic effect in the selected spot markets. Due to the long spanning daily data, CUSUM and Andrews tests of structural change are conducted to avoid any possible misleading statistical inferences caused by the unstable parameter in the spot markets. Our empirical findings can be summarized as follows: (1) All the energy markets reject the independent and identically distributed random walk; (2) The WTI crude oil spot prices evidence the presence of autocorrelation and conditional heteroscedastic increments; (3) The Brent crude oil and New York Harbour conventional gasoline spot prices provide strong evidence of conditional heteroscedastic increments martingale process. As a conclusion, although the energy resources returns are martingales, the heteroscedastic increments can still be used to measure the market risk and to earn a risk-adjusted abnormal return in the energy spot markets.