Do Imports of Intermediate Inputs Generate Higher Productivity?: Evidence from Ecuadorian Manufacturing

Accesible PDF image
Author
Camino Mogro, Segundo;
Pinzón, Kathia;
Date
Jun 2020
International trade is one of the most important ways to improve firms' productivity because of technology transfer from foreign companies. Particularly, imports can generate an improvement in the firms' productive performance by incorporating better inputs in their production processes. This research analyzes, by using an augmented Cobb Douglas production function, the effects of imported intermediates on the production level of manufacturing formal firms in Ecuador for the period 2007-2018, and the causal relationship between the import decision and firm productivity. We estimate the total factor productivity (TFP) at firm level using a flexible Levinsohn-Petrin semiparametric method, in such a way that we solve the problems of endogeneity and simultaneity in the selection of inputs (inherent to this type of estimations). The results show that (1) the elasticity of substitution between foreign and domestic intermediates is 2.44 using the flexible Levinsohn-Petrin estimator in the manufacturing sector, which suggests that they are substitute inputs, that (2) a 100% decrease in the share of domestic intermediates in total intermediates could increase productivity by 16% in all the industry; this result has similar path in Pavitt and technological intensity industries classification, and that (3) when we use an alternative continuous variable of imports such as the intensity of imported inputs on total intermediates, the effect is that a 100% increase in the share of imported inputs increases firm productivity by 2%. In addition, we found robust evidence that the decision to import foreign intermediates improves productivity (the productivity of continuing and entering importers is significantly higher than the productivity of nonimporters), and that there exists self-selection of more productive firms into the import market. Finally, robust evidence was found in favour of the "learning-by-importing" hypothesis.