Unemployment in the Netherlands: a criticism of the den Hartog-Tjan vintage model

Abstract

Unemployment in the Netherlands has increased rapidly since 1971, reaching about 222,000 man-years, or 5-3% of the working population, in 1976. In 1974 an article by H. den Hartog and H. S. Tjan of the Dutch Central Planning Bureau (Hartog and Tjan, 1974) appeared, in which it was argued that part of this unemployment is struc tural in a particular sense: in their opinion the demand for labour at full capacity utilisation had not kept pace with the supply of labour. A high rate of increase in labour productivity at full capacity, caused by an upward shift in the trend rate of growth of real wage costs in 1964, made capacity demand for labour fall behind labour supply. This proposition is based on the application of a fixed coefficients vintage model to the private sector of the Dutch economy. The proposed relationship between wages and employment is introduced by a scrapping condition, which acts as principal behavioural mechanism in the model. In more general terms, this notion of the determinants of employment comes down to the view (as summarised by Van de Klundert, 1976) that 'unemployment had become a structural phenomenon, because the price of labour was above its equilibrium or full employment level.' In other words, it corroborates the neoclassical conception that real wage flexibility is a sufficient condition for labour market equilibrium. The corollary for economic policy is not surprising. A policy directed towards reducing wage costs—defined as wages plus employers' social security contributions—is taken to be the major means of reducing unemployment. Thus, the wage freeze decided upon by the Dutch government for 1976 was supported by reference to the beneficial effects of wage restrictions on employment. The intended reductions in government spending and social benefits for the years to come are justified by the same arguments. In this paper, we shall challenge the tenability of this application of the fixed coeffi cients vintage model. We shall argue that the particular assumptions on which this application is based are invalid and lead to highly implausible patterns for the endogen