Estimation, Interpretation, and Hypothesis Testing for Nonparametric Hedonic House Price Functions

Despite well-documented shortcomings, hedonic and repeat sales estimators remain the most widely used methods for constructing quality controlled house price indexes and for assessing housing attribute capitalization into dwelling prices. Nonparametric estimators overcome many of the problems associated with these approaches by controlling for misspecified spatial effects while using highly flexible forms. Despite these advantages, nonparametric procedures are still not used extensively for data analysis due to perceived difficulties associated with estimation and hypothesis testing. We demonstrate that nonparametric estimation is both feasible for large data sets with many explanatory variables and offers significant advantages in terms of the information content of the estimates. These features are demonstrated in an application of valuing capitalization of access to a rapid transit line into surrounding dwelling prices.