Employer-to-Employer Transitions and Time Aggregation Bias

The rate at which workers switch employers without experiencing a spell of unemployment is one the most important labor market indicators. However, Employer-to-Employer (EE) transitions are hard to measure in widely used matched employer-employee datasets such as those available US. We investigate how lack exact start and end dates for job spells affect level cyclicality EE using Danish data containing daily information on employment relationships. Defining based quarterly overestimates transition by approximately 30% compared data. bias procyclical reduced more than 10% recessions. propose an algorithm that uses earnings not just jobs redefine transitions. Our definition performs better definitions literature.