Kabundi, AlainMiller, Stephen M.2009-07-312009-07-312009-05Gupta, R, Kabundi, A & Miller, SM 2009, 'Using large data sets to forecast housing prices: a case study of 20 US States', University of Pretoria, Department of Economics, Working paper series, no. 2009-12. [http://web.up.ac.za/default.asp?ipkCategoryID=736&sub=1&parentid=677&subid=729&ipklookid=3]http://hdl.handle.net/2263/10885We implement several Bayesian and classical models to forecast housing prices in 20 US states. In addition to standard vector-autoregressive (VAR) and Bayesian vector autoregressive (BVAR) models, we also include the information content of 308 additional quarterly series in some models. Several approaches exist for incorporating information from a large number of series. We consider two approaches – extracting common factors (principle components) in a Factor- Augmented Vector Autoregressive (FAVAR) or Factor-Augmented Bayesian Vector Autoregressive (FABVAR) models or Bayesian shrinkage in a large-scale Bayesian Vector Autoregressive (LBVAR) models. In addition, we also introduce spatial or causality priors to augment the forecasting models. Using the period of 1976:Q1 to 1994:Q4 as the in-sample period and 1995:Q1 to 2003:Q4 as the out-of-sample horizon, we compare the forecast performance of the alternative models. Based on the average root mean squared error (RMSE) for the one-, two-, three-, and four–quarters-ahead forecasts, we find that one of the factoraugmented models generally outperform the large-scale models in the 20 US states examined in this paper.enUniversity of Pretoria, Department of EconomicsHousing pricesForecastingFactor-augmented VAR (FAVAR) modelVector autoregressive (VAR) modelBayesian vector autoregressive (BVAR) modelLarge-scale BVAR modelHousing -- Prices -- United States -- ForecastingWorking Paper