Measuring Labor Market Transitions with Time Series of Cross Sections

We consider the problem of measuring transition probabilities across employment, unemployment and inactivity when longitudinal data is not available and/or the available retrospective data is measured with error. We establish nonparametric point-identification conditions based on time series of cross sections from the European Union Labor Force Survey, both with and without retrospective information. We propose an estimation approach that imposes restrictions satisfied by the data, and that is accurate and fast