Skilled and Unskilled Labour: Are They Worth Their Weight in Growth?

Abstract

Neoclassical growth models incorporating human capital usually adopt a Cobb-Douglas output function with two main features: a labour-augmenting technology term and a linear aggregation of the labour inputs. These functional forms rely on the assumptions that (i) measured in efficiency units, different types of workers are perfect substitutes; and (ii) technical changes are neutral with regards to their effect on the relative productivity of the factors of production. Both these premises contradict a vast microeconomic labour literature which has long showed that: (i) the elasticity of substitution between skilled and unskilled workers is much less than infinity; and (ii) that recent technical changes have been favouring skilled rather than unskilled labour. This study investigates how the human capital aggregation and its parameters affect the results of the economic growth literature regarding the relevance of education. The exercise consists in extending Bils and Klenow (2000)'s model to incorporate imperfect substitution among different levels of human capital and skill-specific technology terms. The results show that altering the human capital specification leads to diverging conclusions regarding the effect of schooling on human capital and technology. When technology is considered as affecting skilled and unskilled workers differently, and these workers are imperfect substitutes of each other, the rewards for education are intensified at the macroeconomic level.