Joan Robinson’s Phillips curve

In an overlooked essay extending John Maynard Keynes’ The General Theory, Joan Robinson set out a theory of the Phillips curve based, in line with Keynes, on a realistic account of human behaviour that acknowledges the difference with which people value gains and losses. People are happy with a gain, such as an increase in money wages, and are very unhappy, indeed indignant, with a loss, such as a cut in money wages. This realistic behaviour anticipated the concept of loss aversion, developed 40 years later by Kahneman and Tversky. Robinson inferred from her ‘realistic’ account that the inflation–unemployment relation had a more or less vertical slope at full employment and a negative-to-horizontal slope at high rates of unemployment. Phillips’ empirical investigation, 20 years later, confirmed this pattern, although he did not refer to the link with Robinson’s description of realistic behaviour. However, 30 years later, Friedman, Phelps and Lucas derived a natural or equilibrium rate of unemployment from the unrealistic account of human behaviour in which there is no marked contrast between gains and losses so that the inflation–unemployment relation is significantly negatively sloped at both low and high rates of unemployment. By making this unrealistic assumption, they and their followers went backwards.