Journal de l'économie mondiale

Global Stock Exchanges Spatial Autocorrelation Using Functional Areal Spatial Principal Component Analysis

Abstract

Richael Treeby

The functional data displaying geographical dependency are the main focus of this work. Using the functional Moran's I statistic, classical principal component analysis and functional areal spatial principal component analysis, the spatial autocorrelation of stock exchange returns for exchanges in 69 countries was examined. This study focuses on the time when the global stock market sold off and established that there is spatial autocorrelation among the stock exchanges under consideration. Prior to applying the technique, the stock exchange return data were transformed into functional data. The sell-off in the world markets had a significant influence on the spatial autocorrelation of stock exchanges, according to the results of the Monte Carlo test of the functional Moran's I statistics. Positive spatial autocorrelation is visible in the stock exchanges' principal components. Regional clusters developed. Amid the worldwide market sell-off in 2015–2016. This study investigated if there was positive spatial autocorrelation in the data from the world's stock exchanges and demonstrated the value of as a technique for investigating spatial dependence.