House Price Expectations : Unbiasedness and Efficiency of Forecasters

The understanding of house price expectations formation is quite limited in the housing literature. This is the first paper to rigorously test the rationality of expectations of house price change using survey data. Using a panel dataset of the Wall Street Journal economic forecasting survey from 2007 through 2012, I test for unbiasedness and efficiency by implementing the econometric methodology proposed in Davies and Lahiri (1995) in the setting of a "three-dimensional" panel dataset. I find that, after controlling for aggregate shocks, 9 out of the 47 forecasters have statistically significant biases, and their biases are all negative, indicating that they persistently predict too high of a change in house prices. The hypothesis of efficiency cannot be rejected, suggesting lack of evidence for inefficient use of information. When the year 2012 is excluded, the unbiasedness test shows that 25 out of the 47 forecasters systematically over predicted house price changes. Again, the hypothesis of efficiency cannot be rejected