In this paper we exploit the longitudinal element of the 1990 and 1998 Workplace Employee Relations Surveys for Britain to investigate the effect of unionism on establishment closings. Contrary to both recent U.S. research and British work using information from the earlier workplace surveys, we find a robust positive association between two measures of unionism - union recognition for collective bargaining purposes and union coverage - and plant closings. This association survives the incorporation of very detailed industry controls but is driven by plants that are parts of multi-establishment enterprises. There appears to be little or no statistically significant association for single plant enterprises. In explaining our findings, we address their consistency with the widely perceived reduction in the "disadvantages of [British] unionism" in recent years.