摘要

As a special energy commodity, oil price shocks can affect not only the energy market but also the performance of the macroeconomy. This research provides complementary explanations for nineteen major oil-related countries/regions' macroeconomic effects caused by unexpected oil price changes. It focuses on the macroeconomic performance of oil price shocks from outlier perspective, investigating the inner hidden factors of co-movements between oil price shocks and macroeconomy. Three methods called Empirical Covariance (EC) method, Robust Covariance (RC) method and one-class Support Vector Machine (SVM) method are used in the outlier detection. Empirical results show that: (a) one-class SVM method has the best environmental adaptability for detecting the outlier performance of the co-movements between oil price shocks and macroeconomy, followed by EC and RC methods; (b) according to the time axis, the outlier performances of gross domestic production, consumer price index and unemployment rate are all concentrated in 2005-2014, which is highly consistent with the oil price shock process; and (c) according to the spatial axis (countries), four categories with similar outlier performances are obtained. Outlier performance of macroeconomy of oil-related country/region in lower levels seems to be worse than that of the country/region in higher levels, which can be attributed to more attentions paid by higher-leveled oil-related countries on the oil price shock response system to relieve the severe macroeconomic impacts caused by oil price shocks.