THE FORMATION OF EXPECTATIONS AND THE DEMAND FOR CAPITAL (UNCERTAINTY, RISK, INVESTMENT)
This study uses the pattern of business investment expenditures in fixed plant and equipment as the focal point of a test of two possible ways in which firms form expectations about the future. Fixed investment was used because the anticipated profitability of an investment, and therefore the decision to invest, is crucially dependent upon forecasting method used by the firm. There are basically three aspects to the study. First, is a derivation that acknowledges the existence of uncertainty, an investment function is developed which reflects how investment expenditures are dependent upon: (a) expected, or forecasted, values of relative prices, tax parameters, interest rates, and the general level of demand, (b) the level of uncertainty regarding the ability of the firm to forecast well, and (c) the degree to which forecasts are borne out by events. Then two sets of forecasted values are collected for the 1971 to 1984 period. One set is based on a sophisticated "trending out" method, in which the forecasted values of each economic variable is based solely on the history of the variable itself. The second set of forecasted values is the history of the forecasts of Wharton Econometric Forecasting Associates. The Wharton forecast method utilizes an econometric model which is based largely on Keynesian theory, and therefore it represents the prevailing Keynesian view of the behavior of the economy. The third step was the use of the forecasted values as explanatory values in a regression analysis of investment to determine which of the two sets of forecast values, given the specification of the investment function determined in the first step, best explain the pattern of investment expenditures in the 1970s and early 80s. The forecast data set with the best explanatory value is the set that best corresponds to the forecast method that is used by firms. The results indicated both that the specification of the investment function was appropriate, and that firms use a forecasting method that includes much more information than a simple trending of historical data. This implies that the coefficients of investment equations based on historical data alone are erroneous, and that the use of such equations to evaluate proposed investment policies is misleading.