Gini’s transvariation analysis: an application on financial crises in developing countries

The damage and the recurrence of financial crises have increased the concern of investors and policymakers on one hand and the interest of macroeconomists on the other. This paper presents an original non parametric methodology, whose aim is to give a very intuitive and rigorous method for variable selection in order to analyse financial crises. Transvariation analysis compares the distributions of two different groups of countries (sound and distressed) with respect to a single macroeconomic variable and selects the indicators on the basis of a low transvariation probability index. The current account deficit to GDP ratio, differently from other studies on financial crises, seems to be a suitable variable in discriminating distressed countries from sound ones, and the case of Argentina and Turkey confirms this finding. Copyright Springer Science+Business Media, LLC. 2013