We present an endogenous growth model where innovations are factor saving. Technologies can be changed paying a cost and technological change takes place only if the bene…ts are larger than the costs. Since the gains derived from factor saving innovations depend on factor abundance, biased innovations respond to changes in factors supply. Therefore, as an economy becomes more capital abundant agents try to use capital more intensively. Consequently, (a) the elasticity of output with respect to reproducible factors depends on the capital abundance of the economy and (b) the income share of reproducible factors increases as the economy grows. Another insight of the model is that in some economies the production function converges to an AK in the long run, while in others long-run growth is zero.

We present an endogenous growth model where innovations are factor saving. Technologies can be changed paying a cost and technological change takes place only if the bene…ts are larger than the costs. Since the gains derived from factor saving innovations depend on factor abundance, biased innovations respond to changes in factors supply. Therefore, as an economy becomes more capital abundant agents try to use capital more intensively. Consequently, (a) the elasticity of output with respect to reproducible factors depends on the capital abundance of the economy and (b) the income share of reproducible factors increases as the economy grows. Another insight of the model is that in some economies the production function converges to an AK in the long run, while in others long-run growth is zero.

In this study, we propose an explanation for why labor and capital shares do not seem to have a trend: an increasing trend in physical capital share is compensated by a decreasing trend in land share. Similarly, an increasing trend in human capital share is compensated by a decreasing trend in raw labor share. We also find empirical support for the claim that the elasticity of output with respect to reproducible factors, human and physical capital, is positively correlated with the income level. This result has important implications for economic growth theory and for empirical exercises related to economic growth.factor income shares, biased innovations, elasticity of output with respect to factors,

No abstract

† We wish to thank Michele Boldrin and Kim Ruhl for comments on a previous version of this paper. We also thank Stefan Krause and other participants at the Emory University seminar series; John Conlon and other participants at the University of Mississippi seminar series; Daniel Mejía, Carlos Esteban Posada and other participants at the Central Bank of Colombia seminar series.