As Keynes remarked, ‘the ideas of economists and political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood’ (Keynes, 1936, p. 383). The natural rate of unemployment hypothesis formulated by Edmund Phelps (1967) and Milton Friedman (1968) provides testimony, having wielded a powerful influence on thought about what determines, and what can be done about, unemployment from the 1970s onwards.

The distinguishing feature of the hypothesis is not so much what it says about the determinants of unemployment, as what it says about the factors that do not influence the natural or equilibrium rate of unemployment. The standard definition of the natural rate is as depending on ‘the actual structural characteristics of the labour and commodity markets, including market imperfections, stochastic variability in demands and supplies, the cost of gathering information about job vacancies and labour availabilities, the costs of mobility, and so on’ (Friedman, 1968, p. 8). This in itself does not rule out an influence for the aggregate demand for goods stressed in the preceding Keynesian wisdom regarding unemployment: aggregate demand could influence the ‘actual structural characteristics of the labour and commodity markets’, and hence the equilibrium rate of unemployment. The innovation comes instead in the preclusion of nominal magnitudes, which of course affect aggregate demand, from the list of factors affecting the natural rate: ‘[the] crucial element is that nominal magnitudes must be sharply distinguished from real magnitudes, and that nominal magnitudes in and of themselves cannot determine real magnitudes’(Friedman, letter to the author, 2 November 1990).