Time series evidence on union densities in European countries

We follow a business cycle approach to study the determinants of union density, finding evidence ofnegative effects of labour market slack (proxied by unemployment) and some impact of compositionaleffects (negative impact of increased labour market participation, as measured by the employment rates, and positive effect of educational attainment in the population, as measured by the population with secondary degrees or more). For a subset of countries we also show that the sectoral changing composition (decline of industry and increase in private services) and increased flexibility in the labour market (as measured by temporary workers) have an adverse effect on density. This pattern is common to all 14 European countries under analysis, with national variations due to inflation rates, wage shares on value added, unemployment subsidies and government political orientation. However Nordic countries distinguish themselves from the common pattern in that they exhibit a positive long-term relationship between union density and unemployment. This may possibly be explained by the fact that Nordic unions provide services to unemployed workers (the so-called Ghent system: mainly administration of unemployment benefits, but also job counselling and general training). The general model is robust enough to appear even when we consider Europe as a single labour market and aggregate national membership as if it were a single union.