Estimation of level and change for unemployment

Using multilevel and structural time-series models for unemployment estimation.
In this paper time-series models are used to estimate monthly provincial unemployment based on the Dutch Labour Force Survey (LFS). The models account for rotation group bias and serial correlation due to the rotating panel design of the LFS. We consider two approaches to time-series modelling: structural time-series modelling and multilevel modelling. The structural
time-series models are fit using a Kalman filter and smoother whereas the multilevel models are fit using a Gibbs sampler approach. Monthly unemployment estimates of level and change and accompanying standard errors based on the two model approaches are compared for the twelve provinces of the Netherlands.