AN APPLICATION OF PRINCIPAL COMPONENT ANALYSIS TO INTERNATIONAL COMPARISON OF ECONOMIC ACTIVITIES
There is an acute need for a suitable composite measure of national economic performance insensitive to subjective preferences of different attributes of a multidimensional social pattern. The question is how to aggregate partial measures using appropriate weighting coefficients that will not rely on subjective judgments concerning their relative significance. In line with a number of studies we propose a rather simple but comprehensive approach based on principal component analysis widely used in multidimensional statistics. Though the first principal component or the set of principal components are known to be used as an aggregate economic indicator, this approach usually implies significant loss in variance of initial factors. However, we argue that the problem has a rather straightforward solution. The whole set of principal components weighted by the corresponding proportions of explained variance can serve as a universal "natural" aggregate measure of various types of economic activities. The proposed methodology is applied to evaluate the national competitiveness and to build the corresponding ranking of the countries in Eurasian region.