Comparing different estimation methodologies of regional GDPs in Latin American countries

This chapter presents a survey of the different methods used to reconstitute long-run income estimations for the Latin American regions. The main purpose is to alert on potential biases derived from them. Although the bottom-up approaches based on the direct estimation of aggregate production, income or expenditure are the preferred option, they have barely been used. The main reason is that all of them are highly data-demanding. Instead, the indirect non-parametric approaches, combined sometimes with direct estimations for agriculture and extractive industries, have been the most recurrent way to estimate output. Out of the nine countries in the sample, seven have followed this "mixed approach" (Argentina, Bolivia, Brazil, Chile, Mexico, Venezuela and Uruguay). Meanwhile, the Peruvian estimates follow a parametric approach and the Colombian estimates backwardly discount the regional growth rates of proxy variables to current official regional estimates. Apart from other considerations, this issue should be taken into account when comparing the substantial differences across countries with regard to regional income disparities.