Statistical Methods in Finance 2018

Dec 17 - 20, 2018


Investigating effect of global crisis index on historical comovement of stock returns in the G7 countries: An FDA approach

by Sonali Das

In this paper we consider the comovements (or correlations) of the stock markets of the US with the other six G7 countries. We use historical data, which in some cases spans from the 1800's. Our objective is to ascertain the effect of global crises on these comovements in a Functional Regression framework, by regressing comovement (monthly frequency) with the Global Crises Index (annual frequency). A major advantage of the Functional Data Analysis (FDA) framework is that it allows us to perform this mixed frequency regression, as aggregating comovements data would make us loose information. The regression coefficient functions are analysed, and considerable knowledge on the effects of crises on comovement of data is gained. Some preliminary analysis indicates that stock comovement between the US and UK has been significant since the early 1900's; while for the other countries (France, Japan, Germany and Italy), the comovements became significant after the mid 1950's. Most of the regression coefficient functions were significant, indicating the validity of most of these results. Another noteworthy point, is that the comovement between the US and Canada has always been significant, indicating perhaps that these two neighbouring economies have always been highly integrated.