The full Bayesian treatment of error component models typically relies on data augmentation to produce the required inference. Never stricly necessary a direct approach is always possible though not necessarily practical. The mechanics of direct sampling are outlined and a template for including model uncertainty is described. The needed tools, relying on various Markov chain Monte Carlo techniques, are developed and direct sampling, with and without effect selection, is illustrated.
More papers in Working Paper Series in Economics and Finance from Stockholm School of Economics Address: The Economic Research Institute, Stockholm School of Economics, P.O. Box 6501, 113 83 Stockholm, Sweden Contact information at EDIRC. Series data maintained by Helena Lundin ().