As discussed in previous chapters, a key point of divergence between Jorgenson’s neoclassical model and accelerator models of investment lies in the assumptions about the nature of the production process and the capital stock. Investment models can be separated into the three categories of putty-putty, putty-clay and clay-clay models according to the assumptions made about the responsiveness of investment to changes in relative factor prices. The difference between these three types of model is captured by the parameter describing factor substitutability, i.e. the elasticity of factor substitution (σ), which captures the responsiveness of capital inputs to relative factor costs. In putty-putty models, such as Jorgenson’s model, the elasticity of factor substitution is assumed to be equal to one, meaning that a given rise in the relative cost of capital will lead to a proportionate fall in the demand for capital. In clay-clay models, such as accelerator models, the elasticity of factor substitution is equal to zero; demand for capital does not respond at all to changes in the cost of capital. In putty-clay models the elasticity of factor substitution ex ante and in the short-run is equal to zero. Ex post and in the long-run the elasticity of factor substitution is equal to one. This is capturing the fact that production processes are more likely to be flexible in planning stages and over longer time periods because time is needed to allow the capital stock to adjust to changes in factor costs. During the 1960s and 1970s a lot of econometric work was conducted to assess the relative predictive power of accelerator theory versus Jorgenson’s model and its variants. The empirical evidence discussed in this chapter confirms the finding from accelerator theory — that investment keeps pace with output growth. However, the evidence on the magnitude of the elasticity of factor substitution is very mixed with different researchers estimating a wide range of possibilities from close to zero to close to one. This suggests that the econometric modelling of Jorgensonian models has not provided any substantial empirical evidence about the validity of simple neo-classical models.
Unable to display preview. Download preview PDF.