Employee flows to improve measures of job creation and destruction and of firm dynamics. The case of Belgium

This paper presents a method for improved estimation of measures of firm dynamics and job creation and destruction. In keeping with a novel approach recently developed in the Scandinavian countries and the U.S, we use employee flow information to re-establish broken links between records of the same firm, and to identify relationships between firms in case of mergers, take-overs, split-offs, and other forms of restructuring. The study is based on a linked employer-employee data set covering all private employment in Belgium. The employee flow approach results in a substantial quality improvement of the measures discussed. In the period of observation, we find that 35 to 50 per cent of total entries and exits of firms with at least 10 employees does not coincide with the real opening or closing of a firm. After correcting for these spurious events, job creation and destruction levels are substantially revised downwards. Besides affecting average levels, correcting for employee flows also strongly reduces annual variation in estimated measures, especially at the sectoral level.status: publishe