Introduction, 117. — I. The classical economists, 120. — II. Marshall and Pigou, 123. — III. Veblen and Hobson, 128. — IV. Knight and Simons, 134. — V. The Keynesians, 135. — VI. Schumpeter, 136. — VII. Conclusion, 137.

Introduction, 295. — I. Increasing eosts and the distribution of earnings, 296. — II. Increasing costs and the rate of return, 299. — III. Implications and conclusions, 301.

This article studies the implications of complementarity on the problem of the long-term forecast. The conditions for maintaining long-run equilibrium between factor demand and factor supply are derived and illustrated with the help of a three-factor model, so as to bring out the particular problem of foreign trade. Equilibrium for all of the three factors appears possible only if the parameters of the system satisfy a set of specific relations or certain policy variables are introduced. Finally the complete system as used by the Central Planning Bureau is presented, together with the numerical values chosen for the parameters. 1. SOME IMPLICATIONS OF COMPLEMENTARITY 1.1. FOR THE PURPOSE of long-range projections a choice must be made as to the nature of the production function. Two extreme assumptions are possible, viz., perfect substitutability and strict complementarity. Most long-range projections published hitherto are based explicitly or implicitly upon the hypothesis of complementarity. Some of the implications of this hypothesis will be discussed in the first and second sections of this paper. In the third a model based on complementarity as has been used by the Central Planning Bureau for projections covering the period 1950-1970 will be presented. 1.2. The type of model to be considered will cover a period of only one to three decades. It is, moreover, not concerned with the cyclical variations of the variables nor with problems of a really secular character such as those studied by Haavelmo.' Furthermore, the implications of disequilibria due to a disproportionate development as between sectors of the economy will be ignored.2 The main problem to be dealt with is therefore the question of equilibrium for the macro variables in the medium-long run. As compared with economic statics the equilibrium concept should obviously be widened so as to allow for the dynamics of long-run development. In the following equilibrium will be defined as a development that is compatible with the equality of demand and supply for each of the factors of production. Neither stable values of the endogenous variables nor constancy of the policy parameters is required. Defined in this way, long-term equilibrium does not necessarily imply optimal development. Unless only one development-equilibrium is possible, restrictions other than that

rrHE economic evaluation of public policy 1 must be based to a considerable extent upon the way it influences the distribution of income. This is especially true for such matters as labor legislation, agricultural subsidy payments, the tax structure, the tariff structure, and public works programs. Accordingly, a systematic examination of the effects of changing income distribution seems greatly to be desired. One aspect of this problem is considered in the following analysis: the effects of the equality of income distribution upon consumption expenditures. More specifically, our concern is with the effects of income distribution upon (a) the percentage of disposable personal income spent, and (b) the allocation of the expenditures which are made. Review of theories. Thus far two theories have been advanced regarding the effects of equality of income distribution upon consumption expenditures: the comparative propensity to consume theory and the emulation theory. The comparative propensity to consume theory holds that the effect of a change in equality of income distribution may be analyzed by comparing the relative marginal propensities to consume of those incurring the income decrease and of those receiving the corresponding income increase. Thus, a reasonably symmetrical equalizing of income distribution presumably would tend to raise the aggregate consumption function for the economy as a whole, since the nch are believed to have a lower marginal propensity to consume than the poor -and vice versa.' Furthermore, it should cause the allocation of what expenditures are made to change as follows: food to increase. clothing and transportation to decrease, and shelter to remain nearly constantand vice versa.2 Veblen's emulation theory holds that people in any given socio-economic class tend to have as their material standard of living the goods and services actually consumed by the next higher socio-economic class.3 For this reason, it is argued, the spread between the standard and plane4 of living for consumers in the aggregate (hence, the pressure on them to spend) is a function of the income differences between the various socio-economic classes in the economy. From this it follows that a reasonably symmetrical equalizing of income distribution should tend to reduce the income disparity between the various socio-economic classes, thus reducing the spread between their standards and planes of living (reduce the pressure on them to spend), and consequently cause the consumption function for the economy as a whole to fall.5 The opposite reaction logically should occur as incomes become less equally distributed, but only for moderate decreases in income equality. Too great an income disparity between the various socio-economic classes presumably would cause them to lose

I. Introduction, 156. — II. Procedure, 156. — III. Results and conclusions, ifin