Empirical Tests of the HOS Hypothesis: The Cross-Section Studies

This chapter presents the cross-section econometric tests of the HOS hypothesis. Cross-section studies include both the cross-industry and the cross-country investigations. We begin with the cross-commodity studies which regress a variable measuring trade performance, usually net trade, on variables which measure determinants of comparative advantage. To test the HOS hypothesis by this method, one would normally regress net exports on capital per worker at the sectoral level on the assumption that there are two generic factors of production (capital and labour). If the results are poor, then a common practice is to bring in additional regressors (such as human capital) in order to obtain some statistically significant results. Within limits, this procedure of including extra relevant explanatory variables can be justified on the basis of a multi-factor approach. However, with the practice of adding explanatory variables ad hoc, a test could degenerate into an empirical search for significant explanatory variables till one finds econometrically acceptable results.