Unemployment in Britain

Male unemployment in Britain has risen from around 2 per cent in the 1950s to around 17 per cent in 1985 (see Figure 1). (The figures are for male unemployment because there is no consistent series for women.1) Even more remarkably, unemployment has fallen in only three years out of the last twenty (1973, 1978 and 1979). To account for this, we need a model that explains both changes in the natural (or non-accelerating inflation) rate of unemployment (NAIRU) and deviations from it. We use a three-equation supply-side model, centred on the labour market. This has two main features. The first concerns the determination of employment in the short run. The labour demand function that we use cuts through the fruitless debate now raging (especially in Europe) as to whether current unemployment is 'classical' or 'Keynesian'. According to the 'classical' view, employment is too high because real wages are too high. According to the 'Keynesian' view, real wages are not binding, and unemployment is high because the product market does not clear-with prices too high relative to nominal demand. The whole debate is set in the framework of perfect competition. Yet in perfect competition prices are set by impersonal forces, and it is not clear what could possibly stop prices clearing the market. It is much more reasonable to think of prices as being set by imperfectly competitive firms, existing prices being the best they can think of, given the demand they face. In this context, firms' demand for labour will depend on both the real product wage and the level of real aggregate demand. This is the demand function we estimate, and it conforms both to common sense and to the data. However, this does not imply that employment can be made to grow without limit by pumping up real demand. For in the medium term, when price surprises are eliminated, our model determines three variables (employment, real wages and real demand) on the basis of three equations (an employment equation, a price equation and a wage equation). Thus in the medium term there is a 'natural' level of employment and, corresponding to this, a 'natural' level of real aggregate demand. The second key feature of our model concerns the medium-term determination of unemployment. In the medium term the planned mark-up of wages over prices in wage settlements must be consistent with the mark-up of prices over wage costs in employers' pricing behaviour. For if wage-setters try to set real product wages higher than is consistent with employers' pricing behaviour, this generates ever-increasing inflation. Thus the key to understanding unemployment in the medium term is the behaviour of wage-setters. If events occur that push them towards too-high real wages, then unemployment has to rise to offset these influences. We shall call these influences 'wage pressure variables' or 'push factors,' and they are clearly crucial in understanding unemployment. The variables here include the social security system, employment protection