Forecasting Real Estate Prices From a PCA Repeat Sales Index

"Case & Shiller repeat sales indices is a means of constructing real estate price indices only based on repeated observations of property transactions. No forecasts may be considered because of the method itself that lies on information entirely contained in the property market. Baroni M., BarthÈlÈmy F, Mokrane M. (2007) proposed an alternative repeat sales index using real estate capital returns and comparing them to other returns calculated from economic or financial variables. To elaborate this index PCA factors are computed and can be used for prediction if forecasts are available for the explanatory variables. Assuming the stability of the multi-factorial model elaborated to construct the index, we try in this paper to forecast the 2009 index for the housing market in Paris and its suburbs, from data or estimations on the explanatory economic or financial variables. In a first part, the methodology of the index and the dataset are described. In a second part the assumptions and estimations on the variables are discussed and various scenarios are elaborated and commented. In a last part, the sensitivity of the model to variations in the most influent variables is tested. Our main finding is that the decrease of the index during 2009 is explained by the factors that can be considered as ìfundamentalsî of the market.""