Statistical properties of world investment networks

TL;DR

This paper analyzes the statistical properties of global investment networks using IMF data, revealing scale-free distributions, Weibull weight models, and universal allometric scaling relations in network structures.

Contribution

It uncovers universal allometric scaling laws in world investment networks and characterizes their degree and weight distributions.

Findings

Degree and node strength distributions are scale-free.

Weight distributions follow Weibull distribution.

Universal allometric scaling relations are observed in maximum flow spanning trees.

Abstract

We have performed a detailed investigation on the world investment networks constructed from the Coordinated Portfolio Investment Survey (CPIS) data of the International Monetary Fund, ranging from 2001 to 2006. The distributions of degrees and node strengthes are scale-free. The weight distributions can be well modeled by the Weibull distribution. The maximum flow spanning trees of the world investment networks possess two universal allometric scaling relations, independent of time and the investment type. The topological scaling exponent is $1.17\pm0.02$ and the flow scaling exponent is $1.03\pm0.01$.