On the Role of Expectations in the Pure Theory of Investment Review of Economic Studies

A cursory glance at the business section of a newspaper, or at the recent speeches of the Chancellor of the Exchequer, would give the impression that expectations of future demand are a crucial factor influencing the investment decisions of businessmen. The purpose of this paper is to give some theoretical justification to the above contention, and to study with some precision the role of expectations in determining the optimal investment programme of a firm. As Arrow [1] has made clear, the traditional neo-classical investment rule, where a perfectly competitive firm accumulates or decumulates capital to the point where its marginal product is equal to the real rate of interest, leaves no role for expectations and is essentially myopic. An important aspect of the world that is ignored in the formulation of the myopic rule is the fact that, once a firm has acquired capital stock and " bolted down the machinery ", possibilities of subsequent decumulation are severely limited. One way of capturing this aspect in a model is to make investment an irreversible process and this was the approach that Arrow used in [1] in the context of a perfectly competitive firm. Once this assumption is made, then it is clear that the investment rule is no longer myopic and expectations matter. In this paper, the irreversibility assumption is used in the context of a firm which is not perfectly competitive, but instead faces a known downward sloping demand curve for its product. This demand curve is expected to vary over time, but the firm's expectations are held with certainty. The way in which the firm behaves over time is then as follows. At the present moment the firm formulates a production and investment plan based on its expectations, it follows this plan until such time as its expectations prove incorrect. It then reformulates its expectations, computes a new plan and proceeds as before. In order to bring out clearly the possibility of excess capacity, and further because the problems considered are fairly short run, the firm is assumed to have a fixed coefficients technology. Under these circumstances, it is possible without great loss of generality to assume there is only one other factor apart from capital, which will be termed " labour ".3 In Section 2 of the paper the general problem is formulated and the solution presented. In Section 3, two specific problems are considered, one concerning the optimal investment plans of the firm over a demand cycle, the other concerning expectations of a change in the price of capital. In Section 4, the problem of how the desired investment rate of the firm over a demand cycle is affected by unexpected changes in demand conditions, cost conditions and in expectations themselves, is considered. This gives some idea of how the firm will react to changes in government policy, specifically monetary policy. One drawback of the above type of analysis is the fact that it may take time for a firm to adjust to its desired capital stock if this moves upward quickly and unexpectedly. Strictly speaking, therefore, adjustment costs should be incorporated into the optimal