Correlation structure and principal components in global crude oil market

TL;DR

This paper analyzes the correlation structure of the global crude oil market from 1992 to 2012, revealing geographical clustering, economic insights from eigenvalues, and proposing an effective market index based on principal component analysis.

Contribution

It identifies geographical clusters in the global oil market and links eigenvalues to economic factors, proposing a new market index based on principal components.

Findings

Six geographical clusters identified in the correlation matrix.

Eigenvalues reveal economic information and market-wide effects.

Proposed index outperforms the $1/N$ benchmark in tracking the market.

Abstract

This article investigates the correlation structure of the global crude oil market using the daily returns of 71 oil price time series across the world from 1992 to 2012. We identify from the correlation matrix six clusters of time series exhibiting evident geographical traits, which supports Weiner's (1991) regionalization hypothesis of the global oil market. We find that intra-cluster pairs of time series are highly correlated while inter-cluster pairs have relatively low correlations. Principal component analysis shows that most eigenvalues of the correlation matrix locate outside the prediction of the random matrix theory and these deviating eigenvalues and their corresponding eigenvectors contain rich economic information. Specifically, the largest eigenvalue reflects a collective effect of the global market, other four largest eigenvalues possess a partitioning function to distinguish the six clusters, and the smallest eigenvalues highlight the pairs of time series with the largest correlation coefficients. We construct an index of the global oil market based on the eigenfortfolio of the largest eigenvalue, which evolves similarly as the average price time series and has better performance than the benchmark $1/N$ portfolio under the buy-and-hold strategy.