Occupational affiliation data and measurement errors in the German Socio-Economic Panel

This paper shows that there are severe measurement errors regarding the occupational affiliations in the German Socio-Economic Panel. These errors are traced back to the survey structure: in years where occupational information is gathered from the entire employed population instead of only from those declaring job or labor market status changes, average occupational mobility is around five times higher. In order to construct reliable occupational affiliation data, a correction method based on related job or labor market status changes is proposed. The corrected occupational mobility patterns are then analyzed for different samples.