Factor Productivity and Global Trade

The Heckscher-Ohlin-Vanek (HOV) model performs poorly in explaining the factor content of global trade. Previous studies that introduce Hicks-neutral productivity differences in the HOV model produce mixed results on the model’s improvement in fit. We adopt an approach that uses factor earnings to measure effective factor quantities, which intends to capture both neutral and non-neutral factor productivity differences between countries. Applying this approach to a data set of 78 countries or country groups, we find that the model’s fit to data improves significantly. Despite the improved fit, the model still shows large deviations in its predictions. We detect some systematic patterns in the deviations and explain them with a model of multiple diversification cones. Results from splitting the sample into income groups support our explanation.