Bayesian Estimation of a Dynamic Partial-Equilibrium Model for Investment

This paper revisits the question if the user cost of capital plays an important role for investment decisions using Bayesian estimation techniques. These methods offer advantages over classical econometric tools in this area: The most important are that prior distributions offer a convincing way to confine the support of model parameters and that confidence intervals are more reliable when model parameters approach the bounds of their support. I use aggregate investment data from six industrial sectors in the UK to estimate a parsimonious partial-equilibrium model. The Kalman Filter is used to evaluate the likelihood and MCMC methods are employed to draw from the posterior distribution. The main finding is that the real interest rate accounts for less than 10 percent of the variance in investment under the 99- percent confidence level; this result is robust across sectors.