What explains changes in full-time and part-time employment in Western Germany? A new method on an old question

From 1992 to 2005, part-time employment in Western Germany has grown by 82 percent, whereas full-time employment has shrunk by 14 percent. Behind these general figures there is substantial variation of employment schemes across industries. Beside this, the share of the service industries in gross value added has grown, whereas the importance of manufacturing and construction has decreased considerably. We analyse the extent to which the changes in part-time and full-time employment can be explained by changes in the sectoral composition of the economy or by other factors. Using West German yearly data from 1992-2005, we estimate a regression analogue shiftshare model. It allows us to divide the overall development of employment into the business cycle effect, the sector effect and the employment status effect. Moreover, we control for sectoral gross value added, unit labour costs and working time. As a methodological contribution we extend the shift-share approach into a dynamic panel model. We use a bias-corrected least squares dummy variable (LSDVc) estimator which is appropriate for our data structure. As a second step, we decompose the fixed effects of the LSDVc estimation into parameters for part-time, full-time, and self-employment as well as six sectors. Our results confirm previous deterministic shift-share analyses: Characteristics inherent in full-time or part-time employment dominantly explain changes in employment patterns in Western Germany. The sectoral composition of the economy plays a significant but minor role. The model extensions reveal that much of the status and sector effects in the simple shiftshare analysis can be captured by determinants of labour demand.