Entropy correlation distance method applied to study correlations between the Gross Domestic Product of rich countries

TL;DR

This paper introduces an entropy-based correlation distance method using the Theil index to analyze economic globalization by studying GDP correlations among rich countries over time.

Contribution

It presents a novel entropy correlation distance method applied to GDP data, revealing insights into economic globalization trends.

Findings

Mean distance between developed countries decreased from 1960 to 2000

The method distinguishes globalized economic patterns from traditional correlation measures

Hierarchical structures like networks and spanning trees are constructed from the data

Abstract

The Theil index is much used in economy and finance; it looks like the Shannon entropy, but pertains to event values rather than to their probabilities. Any time series can be remapped through the Theil index. Correlation coefficients can be evaluated between the new time series, thereby allowing to study their mutual statistical distance, - to be contrasted to the usual correlation distance measure for the primary time series. As an example this entropy-like correlation distance method (ECDM) is applied to the Gross Domestic Product of 20 rich countries in order to test some economy globalization process. Hierarchical distances allow to construct (i) a linear network, (ii) a Locally Minimal Spanning Tree. The role of time averaging in finite size windows is illustrated and discussed. It is also shown that the mean distance between the most developed countries, was decreasing since 1960 till 2000, - which we consider to be a proof of globalization of the economy for these countries.