Visibility graph analysis of economy policy uncertainty indices

TL;DR

This study applies visibility graph algorithms to economic policy uncertainty indices of the US and China, revealing their complex network properties and persistent, scale-free, small-world characteristics, offering a novel perspective for economic analysis.

Contribution

It introduces a complex network approach to analyze economic policy uncertainty indices, uncovering their topological features and persistence, which is a new application in this field.

Findings

EPU indices are persistent with Hurst exponents between 0.5 and 1.

EPU networks exhibit scale-free degree distributions with power-law tails.

US EPU network shows small-world properties with logarithmic shortest path growth.

Abstract

Uncertainty plays an important role in the global economy. In this paper, the economic policy uncertainty (EPU) indices of the United States and China are selected as the proxy variable corresponding to the uncertainty of national economic policy. By adopting the visibility graph algorithm, the four economic policy uncertainty indices of the United States and China are mapped into complex networks, and the topological properties of the corresponding networks are studied. The Hurst exponents of all the four indices are within $\left[0.5,1\right]$, which implies that the economic policy uncertainty is persistent. The degree distributions of the EPU networks have power-law tails and are thus scale-free. The average clustering coefficients of the four EPU networks are high and close to each other, while these networks exhibit weak assortative mixing. We also find that the EPU network in United States based on daily data shows the small-world feature since the average shortest path length increases logarithmically with the network size such that $L\left(N\right)=0.626\ln N+0.405$. Our research highlights the possibility to study the EPU from the view angle of complex networks.