Prediction of housing prices: an application of the Arima model

Abstract

This study uses an ARIMA model to provide out-of-sample forecasts for United States housing prices as represented by the Case Schiller Index during the financial crises timeline ( between 2005 and 2009). The major findings are that the model fails to predict the peak/turning point of the financial crisis but successfully predicted declining prices since 2006:6; therefore the magnitude of the loss realized during that period could have been reduced had the models prediction been considered. The model predicted extremely negative one year ahead prices in 2008:2, 2008:3 and 2008:9 which explains the timeline of the collapse of the Bear Stearns and Lehman Brothers as well as the subsequent global financial meltdown.