Samuelson’s Neoclassical Synthesis Did Not Betray Keynes: Keynes Had Offered Just Such a Synthesis in the General Theory, As Well As to Hicks and Harrod in Correspondence, in 1936

In a recent reprint of material taken from R. E. A. Farmer’s recent 2017 book, Prosperity for All, certain misrepresentations of Paul Samuelson’s work and J M Keynes’s work is presented. For example, in the presence of a high unemployment rate, following a severe recession or depression, Samuelson’s prescription is the use of expansionary fiscal policy that involves large government deficits to finance major public works, public goods, and public infrastructure spending to replace the large gap resulting from the collapse in private investment spending. This will increase p, the general price level, while reducing the real wage, w/p. This will result in an increase in employment and a fall in the unemployment. The unemployment rate will start to move back toward the full employment target rate of 4 %. The problem is not that money wages are too high, but that, due to decreased private corporate investment spending, spending on investment goods is too low, so that the overall price level is too high, resulting in a higher than optimal real wage. This higher real wage is misinterpreted as being due to money wages that are too high, while, in reality, it is due to insufficient spending on investment goods that is too low.

J M Keynes, in December 1933, invented and taught the IS-LM(LP) model. An inferior version of it was developed by Keynes and appeared in the mid 1934 draft copy of the General Theory. An improved, revised version of it was presented in the General Theory in chapter 15 and chapter 21 on pp.298-299 in 1936. The aggregate price level was not determined in Keynes’s IS-LM(LP) model. It is determined by Keynes in a separate, supporting model, which Keynes called his D-Z model, which appeared in chapters 20 and 21. Keynes developed the Aggregate Supply Curve (ASC) in aggregate Income-employment (D, Y, Z; N) space. This ASC represents a locus of all multiple equilibrium, optimal D=Z outcomes, all of which satisfy the necessary and sufficient first and second order conditions for an expected profit maximum. The first order conditions are that the expected real wage equals the marginal product of labor. Keynes presented two simplified versions of his ASC on pp.295-296 of the General Theory, which has the familiar completely horizontal, completely elastic, range of the standard textbook aggregate supply curve in (O,P) space where O=a function of N, and PO equals aggregate income. PO is Keynes’s Y; pO, expected aggregate income, is Keynes’s D. N is aggregate employment.

Both Keynes and Samuelson realized that Frisch’s reconstituted version of K. Wicksell’s rocking horse (pendulum) model could only involve negative feedback. It does not, and could not, involve the positive feedback effects resulting from the interaction of the multiplier and accelerator models, first presented by Keynes in his correspondence with Harrod in August, 1938, in discussions over his economic growth model using a first order, difference equation, and simultaneously presented by Samuelson in 1939-1941 by the use of difference equations.

Microeconomic optimizing approaches, (individual, rational utility maximizing and profit maximizing problems) based on pendulum models, which have purely negative feedback effects that are self-correcting over time, have no explanatory significance or policy application at the macro level. All that it is possible to say at the macro level is that the dynamic equilibriums are either stable or unstable. It is impossible to conclude that they are optimal or non-optimal in terms of aggregate employment or aggregate “output”.