Estimating Country Heterogeneity in Capital-Labor Substitution Using Panel Data

I develop a flexible framework to jointly estimate the aggregate capital-labor elasticity of substitution and the labor- and capital- augmenting technologies from a panel of production functions and use it to shed light on the... [ view full abstract ]

I develop a flexible framework to jointly estimate the aggregate capital-labor elasticity of substitution and the labor- and capital- augmenting technologies from a panel of production functions and use it to shed light on the global decline of the labor share and development accounting. In contrast to previous studies, my framework considers unobserved country heterogeneity in the parameters. I use nonlinear factor models to control for time-varying unobserved labor- and capital- augmenting technologies that might be correlated with the production function’s inputs. Estimation is based on posterior distributions in a Bayesian fixed effects framework. I find evidence of substantial country heterogeneity in the elasticity of substitution, with a mean of 0.90. The bias in technical change is the dominant mechanism in explaining the labor share decline in most countries. However, the increase in capital accumulation is also an important mechanism for some countries.Finally, I find evidence to support directed technical change models.