Neglected Implications of Neoclassical Capital-Labour Substitution for Investment Theory: Another Criticism of Say's Law

Recent mainstream macroeconomic models take Say's Law for granted. This paper argues that the justification for this assumption is not found in general equilibrium theory but in the 'neoclassical-synthesis' (and then monetarist) criticism of Keynes, which relies in a fundamental way on a treatment of investment that turns out to depend not only on neoclassical capital-labour substitution (called into questioned by the Cambridge controversies) but also on an assumption of full labour employment that on the contrary should be a result of the analysis. This paper first criticizes the attempt to justify a negative interest-elastic investment function through adjustment costs, that is, without relying on traditional neoclassical capital-labour substitution, with special attention on the treatment in Romer's advanced textbook. Then it proceeds to its new contribution, the demonstration that an endogenous determination of labour employment raises questions about the capacity of decreases in the real wage to raise employment even accepting neoclassical capital-labour substitution, because when the latter is correctly understood the rate of interest does not suffice to determine investment; hence, there is an inevitable role for the accelerator (and Say's Law is thereby undermined). This was perceived by Dornbusch and Fischer but they did not realize that then reductions of real wages will reduce investment instead of raising it. Thus, the 'neoclassical synthesis' reply to Keynes is undermined even apart from the inconsistencies of neoclassical capital theory. So the paper exposes a deficiency of the neoclassical arguments in support of a tendency toward full employment, additional to the inconsistencies revealed by the capital critique.