We develop new evidence on job recruitments and vacancy durations using a rich source of data on individual hires. Our core data set contains 55,000 recruitments into vacant job positions for stratified random samples of German employers from 2000 to 2010. We have information about the employer, the job position and the newly hired worker for all recruitments – including firm size, occupation, qualification requirements, previous labour market status of the new hire, and whether the job is a new position. We measure recruitment duration and the lag from recruitment to first day of work (start lag), which sum to the full vacancy duration. In addition, we link our micro data on recruitments and new hires to additional data on contemporaneous labour market conditions at the regional, occupation and industry levels. Our analysis finds a mean recruitment duration of 49 calendar days or 34 working days and a mean start lag of 27 calendar days and 19 working days, for a total vacancy duration of 76 calendar days and 53 working days, strongly varying between occupations. Hazard functions fit to micro data reveal longer recruitment durations in Eastern Germany and in larger firms and shorter recruitment durations under slack labour market conditions. Highly relevant for the length of the start lag is whether the hiring process goes as planned: If the recruitment duration is longer than the intended total vacancy duration, the start lag is significantly shorter, reflecting the specific efforts of employers in this case to fill the position as soon as possible. The use of Public Employment Services and the hiring of a person previously unemployed show significant effects on the start lag.
