From Keynes (1936) to the neoclassical synthesis

After Keynes, portfolio theory was developed, building microeconomic foundations of the demand for money, thus continuing the work of Hicks (1935) in his propositions for a research program designed to integrate money into the theory of choice.

It is generally accepted that macroeconomic Keynesian models in the twentieth century were the first in the line of Hicks’s IS-LM model (Hicks 1937). The central bank supply of money determines, with the public’s demand for money, a long-term rate (bond rate) on the LM curve. This demand for money includes the three Keynesian motives: transaction, precaution and speculation. It is increasing with income and decreasing with the interest rate. The LM curve gives all the solutions (both income and interest rate) that correspond to the equilibrium of the market for money. Together with the IS curve (which describes equilibrium on the goods market), it determines the global macroeconomic equilibrium.

In the 1960s there was a revival of interest in the quantity theory of money. Milton Friedman (1956) reformulated this theory as a demand for money in the Cambridge style. Money is considered as a capital good. Any real or financial asset is a substitute for money within agents’ wealth holdings. According to this new quantity theory, the velocity of circulation is no longer constant but a stable variable, because it is a function of stable factors. Friedman minimised the role of instable variables such as interest rates (returns on assets) in agents’ choices that had been stressed by Keynesians and gave great impact to the stable new variable he introduced: permanent income.

However, this monetarist monetary policy is a rule that is imposed on the central bank. The latter has to increase its money supply with a fixed coefficient independently of the economic circumstances. Monetarists were to criticise the Phillips curve and introduce adaptive expectations. So, any increase in the money supply can only reduce unemployment (below its natural level) temporarily, before agents adjust their inflation expectations. As a result, it is better not to modify the monetary policy and to follow a quantitative and constant rule in order to attain an inflation target.

However, Keynesians criticised the stability of the monetarist demand for money in the 1970s as a result of empirical studies. Moreover, it became more and more dif­ficult to define the quantity of money supply that could be the basis of monetary policy, because of constant financial innovations in liquid assets that are substitutes for money and encourage modifications of behaviour on the part of asset owners. The role of expectations was to be developed by the new classical economics, which introduced “rational expectations”. Agents use all the information at their disposal in the best pos­sible way. They cannot make systematic errors. Inflation is a function of the expected rate of growth of the money supply. So there is no possibility for a Phillips curve and no possibility to reduce (even in the short run) natural unemployment. Any discretionary monetary policy is useless.