Exogeneity Testing of Variables Used in Foreign Trade Modelling – the Example of the Visegrad Group Countries

Abstract

Some concepts of modern econometrics do not use an arbitrary split of the variables into endogenous and exogenous ones. One of the most popular is the Granger causality test, in which it is assumed a priori that there is no distinction between the exogenous and endogenous variables. On the other hand, from the point of view of an econometric model, estimation or prediction may be important to test weak or strong exogeneity of variables. In modelling an economy, including international trade, there are often variables between which feedback may occur. The causality of the variables in the classical sense is not always so obvious here, and it should be tested. This will facilitate the proper model specification of foreign trade. The aim of this paper is to study the exogeneity of selected macroeconomic variables used in the models of foreign trade using the example of the Visegrad Group countries. The results may be helpful in determining the structure of the actual relationships between variables, in choosing the estimation method of appropriate models and in forecasting.