Do all imports matter for productivity? Intermediate inputs vs capital goods

This paper looks at the question of how import activities and productivity are related. Using detailed production and cost data from a panel of Mexican manufacturing plants between 1994 and 2003, the paper is able to differentiate between two types of imports, materials (or intermediate inputs) and machinery and equipment (or capital goods), and to separately identify self-selection and learning effects of each type of imports at the plant level. The key findings are that there is evidence of both self-selection into importing and learning-by-importing, but not all imports seem to matter for productivity. Not only more productive plants tend to become importers of machinery and equipment rather than materials, but plants that start importing machinery and equipment also experience an increase in productivity, while the same does not occur when plants start importing materials. There is also evidence of productivity gains following entry into export markets and complementarities between exporting, importing materials and importing machinery and equipment. These findings seem to suggest that capital goods from abroad are more likely to embody technological improvements than intermediate inputs.