The purpose of this paper is to determine whether the way competitiveness is measured matters, as well as to analyze the relationship between tax burden and economic competitiveness using a variety of model specifications.
This paper uses statistical models aimed at finding the relationship between taxes and different measures of economic competitiveness, such as gross domestic product per capita, employment and a third-party competitiveness index. Other variables are also considered, and statistical procedures such as lag specifications and “white period” errors are used to address problems of endogeneity and serial correlation.
The models find no robust relationship between taxes and competitiveness. Certain models find correlations between the tax burden of specific income groups with economic competitiveness, but these vary in direction and are difficult to interpret. This follows past research, which shows different results depending on the period analyzed, measure of competitiveness and other variables used.
This paper looks at many of the different measures of competitiveness, control variables and periods that are used in the previous literature and shows how any changes to these model specifications cause inconsistent results. This paper highlights that, because results can vary greatly depending on the model, researchers and policymakers must be careful when drawing any conclusions from relationships between taxes and economic competitiveness.
