Ex Ante and Ex Post Data in Inventory Investment

Abstract

Abstract Expectational data on manufacturers' inventory investment and sales are analysed. These data are provided by Fortune Magazine's survey of corporations. It is found that firms tend to underanticipate investment changes and that entrepreneurs tend to be conservative in anticipations when economic conditions are depressed and less conservative when conditions are favorable. A theory of investment behavior is then derived and two realization functions are tested, one assuming perfect competition and cost minimization and the second assuming profit maximization and an oligopolistic market. The second performs best. Certain ex ante and ex post estimating forms are tested; the ex post form yielding excellent statistical results with real inventory stock as dependent variable involves the explanatory variables of real unfilled orders, real sales and a distributed lag variable. A major conclusion is that the Fortune ex ante data appear to be only marginally useful in explaining inventory investment; and until observation errors are substantially reduced (inter alia), the marginal work units involved in using this data may not be warranted. A positive conclusion is that the real unfilled order variable appears to be highly useful.