Creating and analyzing spatio-temporal financial maps in international bust and boom extreme episodes

One of the more interesting questions in the international finance literature is how to model the propagation of financial extreme episodes. This paper will study daily stock market indexes from major international stock exchanges using spatio-temporal statistics to investigate the issue of contagion in financial markets. We will use a joint multidimensional scaling and spatio-temporal geostatistical approach to analyze 25 international stock indexes. We pay special attention to the concept of “distance” in the financial markets framework and we construct financial maps with multidimensional scaling that will be further modeled with the most sophisticated spatio-temporal covariance functions, which represent the spatio-temporal dependencies of daily stock market indexes. These methods have been carried out in several important bust and boom episodes that impacted the stock exchange market of some countries. More specifically, we analyze how the financial consequences of the events in one country are transmitted to other countries by means of the structure of the spatio-temporal dependence. Our findings show that (i) in general, there is not a pattern of this contagion effects based on financial maps, and (ii) the structure of spatio-temporal dependence varies with the events, i.e., depending on the nature of the episode (if it is controlled by the policy makers or not), depending on the place where it happened (if it is in the USA, in the EU or in Asia). In this sense, it is possible to found some patterns measured through the spatio-temporal covariance funtion. Our approach and results can be of interest for further analyses on the stock exchange market linkages.