THE USE OF ENDOGENOUS VARIABLES IN DYNAMIC MODELS OF INVESTMENT.

This paper focuses on the use of endogenous variables in dynamic models of investment. Section I explains how the theory of the firm's investment in physical capital is frequently conceptualized and discussed. Section II discusses some of the approaches which have been used to define desired capital stock in terms of variables which are at least partially endogenous to the firm. This section also contains an elementary example which illustrates that definitions of desired capital stock involving such endogenous variables are not consistent with profit maximization and as a result can lead to prediction errors in dynamic models. In Section III the results of Section II are extended to a more sophisticated version of the investment model which Jorgenson and others have developed and applied in recent years. Section IV contains a brief review of the rational distributed lag schemes which have generally been used as adjustment mechanisms in investment studies and indicates the conditions under which these may represent optimal adjustment paths. Section V applies the lag schemes developed in Section IV to obtain some comparative simulations of models in which desired capital stock has been correctly and incorrectly defined. Section VI contains some comments on estimation and prediction, and Section VII concludes with some suggestions as to how the problems raised by this paper may be avoided.