Why do empirical tests tend to accept the NEG? - An alternative approach to the 'wage equation' in European regions

This paper posits a new approach to the ?wage equation? of the New Economic Geography (NEG). On one side, it is shown that the NEG provides a spatial explanation of marginal costs, instead of wages. On the other side, and focusing on the statistical properties of the data, it is explained why the wage-type equations tend to be accepted in empirical tests for European data. Proxying the dependent variable by gross value added per capita and Market Potential by a Harris?s (1954) index, the paper proposes two tests that are not allowed by the NEG, but useful to identify the key elements conditioning the results. It is shown that the way in which Market Potential is built generates similar estimation results for variables that do not measure market size. Market Potential captures locational information but the empirical wage-type equation is also shown to be similar when the focus of attention changes from global to local spatial patterns. The results reveal a high degree of uncertainty when interpreting an empirical wage-type equation.