Topic: Title: On a systematic fanning out of income profiles Authors:  Sarah Meyer - University of Muenster (Germany) [presenting]
Abstract: The nature of labour income risk is important for various economic decisions. Two different approaches to modelling labour income profiles have been established in the literature: the heterogeneous income profile model (HIP) and the restricted income profile model (RIP). Both models mainly differ in their assumptions about whether differences in labour income profiles are driven deterministically or stochastically. Our aim is to empirically investigate which of the two models is more suitable to describe real income data. To this end, a dynamic linear model is proposed. It allows for both individual-specific and time-varying coefficients. Estimation of the latent state vector and of the unknown variance parameters is carried out using Gibbs sampling, a Markov chain Monte Carlo (MCMC) algorithm. In our application, the Forward Filtering Backward Sampling (FFBS) algorithm is used as a building block, while further draws result from the conjugate prior distributions. Applying our framework to the German SOEP (GSOEP) data, we find evidence that the earnings data disprove the RIP approach. For validation we employ the Bayes factor, which confirms our finding.