Researchers used different models to predict housing prices in 20 US states by analyzing a large amount of data. They compared models that included various economic factors and found that models using common factors generally made more accurate predictions than models that used Bayesian shrinkage.