Determinants of trade union membership in Great Britain during 1991-2003

We analyse the determinants of union membership in the UK using data from the BHPS (1991-2003). Employing three alternative methodologies to control for the problem of initial conditions we find that union membership remains persistent even after controlling for the unobserved effect. There is evidence of a considerable correlation between the unobserved individual heterogeneity and the initial membership status. Ignoring this overstates the degree of state dependence of union membership greatly. The extent of state dependence in union membership status is notably higher in the (1991-1996) period estimates and appears to be more pronounced in the case of male employees for the entire period under analysis. The second period estimates reveal that unobserved heterogeneity has a more prominent impact in determining future unionisation probability versus past union membership. Finally, the estimates suggest that an individual´s propensity to unionise is determined by a mixture of industrial and personal characteristics. This is at odds with earlier studies, such as Booth (1986) and Wright (1995), failing to control for unobservable effects and concluding that personal attributes do not have a significant impact on unionisation propensity.Union membership, Initial conditions, Unobserved individual heterogeneity, State dependence, Dynamic random effects probit models