Advanced Space-Time Analysis: Constructing a Real Estate Price Index

Real estate data exhibit autocorrelation and heterogeneity across both space and time. The literature is beginning to advance methods that account for the four components. This article accordingly introduces the Bayesian Maximum Entropy (BME) method to real estate analysis. In addition to controlling for spatiotemporal autocorrelation and heterogeneity, BME allows for probabilistic and missing data and does not require a known distribution as assumed by linear and log-likelihood techniques. We apply BME to a dataset of house prices, which illustrates the use of the technique for developing a price index on a small geographical area. JEL Classification: C33, R12, R32