Forecasting Dynamic Investment Timing under the Cyclic Behavior in Real Estate

Abstract

This paper applies the Hodrck-Prescott (HP) filter to forecast short-term residential real estate prices under cyclical movements. We separate the trend component from the cyclical component. We show that each regional residential market reacts not only to previous price movements, but also that these regional markets react to previous shocks under Auto Regressive Integrated Moving Average (ARIMA) modeling. Using the S&P Case-Shiller Home Price Index, we compare our forecast to index values from the Chicago Mercantile Exchange (CME) Housing Futures and Options. Our study identifies possible systematic errors from the different price adjustments reflecting current market situations.