Research on the interaction patterns among the global crude oil import dependency countries: A complex network approach

Abstract

The major goal of this paper is to focus on interaction patterns among global crude-oil-import-dependent countries (OIDCs) and to study global crude oil supply risk from the perspective of crude oil import dependency. Import values of crude oil of 38 OIDCs were selected as sample data. To analyze and visualize the global evolution and transmission characteristics of interaction patterns, we propose a method by using complex network theory. In this method, the holistic time is divided into different small-scale fragments by sliding windows, and then the series of interaction networks of OIDCs (OIDCIN) is obtained. In order to characterize the global evolution of such series of interaction networks, indexes of dynamic topological structure and the long-term global recurrence plots are introduced. Underlying dynamics of time-evolving interaction networks are revealed. A formula to measure the risk in crude oil supply is designed from the perspective of crude oil import dependency. To analyze the transmission characteristics of the interaction patterns, the interaction modes are defined. The interaction modes transfer into each other over time, forming a process of transmission. Then the interaction modes transmission network of OIDCs (OIDCIMTN) is obtained. Numerical simulations show that OIDCIN has co-movement, group-occurring and small-word property. The interaction patterns have both a short-range correlation and a long-range correlation. The evolution process of the interaction patterns can be divided into four different periods. Dates of 2008/05, 2010/09, and 2012/10 are broken dates. Dates of the inflection point of the risk in crude oil supply transformation are June 2007, June 2010, and February 2013. Risk in crude oil supply presents an increasing trend from February 2013. The strength of OIDCIMTN basically satisfies a power-law distribution. In OIDCIMTN, modes with large strength almost have large clustering coefficient. The interaction modes are transferred into each other conveniently, although the crude oil import dependency system is highly complicated. The transformation of the interaction patterns mainly revolves around three core interaction modes and each transformation needs to undergo 5 non-core modes.