The primary purpose of this thesis is to combine two branches of empirical work concerning firms demands for factors. One set of studies involves the simultaneous estimation of money, labour and capital demand functions, where the focus is on the role of money as a factor production, and adjustment costs for capital are ignored. The other set of studies involves estimation of an investment function, together with other factor demand functions. Writers in this area typically exclude money as factor in the production process, though we wish to overcome that omission in this thesis. We also study the possible supply side effects of monetary policy that arise because of the role of money as a factor of production and deduce macroeconomic policy implications. Our empirical work is divided into three stages. In each stage we assume that firms minimize the cost of producing a given level of output subject to a production function that includes real money balances as a factor input. In the first stage we estimate a full equilibrium model in which firms can adjust their capital stocks without any lags and there are no costs associated with the adjustment of these capital stocks. In the second stage of estimation we introduce the fixity of capital stocks in the short-run into the firms' optimization problem. In this temporary equilibrium model, costs of adjustment of capital are not assumed. Finally, we estimate a dynamic model of the firm based on the assumption of non-linear internal costs of adjustment for capital. The three empirical models are estimated at the aggregate level for non-financial corporations in the United States. On statistical grounds the full equilibrium model did not fit the data well. The own price elasticity of real money balances was not significantly different from zero. In the other two models, where capital is fixed in the short-run, all the own price elasticities are significantly different from zero and have negative signs. Furthermore, in the full equilibrium model, autocorrelation seemed to be present even after making a first order correction for the errors. The temporary equilibrium model was also statistically rejected, conditional on the particular functional form of the cost function employed in the dynamic model. We conclude that the dynamic specification is most appropriate in this context. The price elasticities varied substantially across the different models. The cost minimizing interest rate elasticity of labour demand was significantly different from zero and negative in sign in the dynamic model. However, the implied profit maximizing interest rate elasticity of labour demand was not statistically significant. This suggests that earlier estimates of money's role in the production process were contaminated by the restriction that there were no costs of adjustment.

Doctor of Philosophy (PhD)