CMStatistics 2016: Start Registration
View Submission - CFE
Title: Likelihood shape and regularization priors for econometric models with reduced rank Authors:  Herman van Dijk - Erasmus University Rotterdam (Netherlands) [presenting]
Nalan Basturk - Maastricht University (Netherlands)
Lennart Hoogerheide - VU University Amsterdam (Netherlands)
Abstract: `Reporting the shape of the likelihood and its properties is an important task for a Bayesian econometrician', was argued by Hildreth in 1963. This viewpoint is based on the idea that scientific evidence should be reported in such a way that the information specified in the econometric model dominates with respect to other sources of information. Shape and features of the posterior probability density function of the parameters of interest of reduced rank econometric models under regularization priors are analyzed. Given the indicated relative prior ignorance, the likelihood information strongly dominates. We obtain, to the best of our knowledge, novel results.