An econometric approach to macroeconomic risk

Abstract

A contribution to the study of volatility and country risk is made in order to achieve a successful cross country comparison. We present a methodology for the evaluation of country risk that include endogenous detection of multiple structural breaks (also identifying its different kinds), determination of persistence of shocks through their structural-break free fractional integration order and determination of the adjusted volatility which best characterizes the economy. This methodology is applied to developed and emerging countries' GDPs (taking 9 countries from each group). Although the former have fewer structural breaks than the latter, these breaks are extremely relevant in 14 of the 18 countries. This affects the calculation of the series persistence and volatility. Comparing a traditional risk indicator to our suggested one we find that the cluster of reference of 60of the countries changes. Most countries present fractional integration (long memory) being the distribution between both groups heterogeneous. Country volatility varies strongly if we isolate structural breaks that present a probabilistic distribution different from intrinsic GDP volatility. Clusters arrangement is different with some risk country evaluation methodologies.