Hedonic Pricing Model

Empirical Implementation

In empirical analysis, the implicit prices of an incremental unit of each element in vector z can be estimated by regressing observed prices on objectively measured characteristics using sample data and ordinary least squares (OLS). The OLS statistical model by researchers frequently takes the following form:

where pricei is the observed price from house transaction i for sample size I; β is the vector of coefficients, or implicit prices, to be estimated for each house characteristic in the vector x (plus the intercept term β0); x is a vector of property characteristics hypothesized to be statistically related to price; and e is a random error term. Examples of variables commonly included in hedonic price models applied to single-family housing include age of dwelling, lot size (square feet or acres), architectural style, number of bedrooms, number of bathrooms, number of fireplaces, type of heating and cooling system, neighborhood descriptors, presence of a two-car garage, proximity to open space, restrictive covenants, and other defining characteristics of the house and its location.

Estimating a hedonic price model is typically accomplished using ordinary least squares (OLS) regression modeling. OLS is the best linear unbiased estimator (BLUE) when the model conforms to the classical linear model assumptions of (a) the model is linear in the parameters, (b) the sample is a random draw from the population data, (c) none of the independent variables in x are constants and there is no exact linear relationship between any of the independent variables, (d) the expected value of the error term e equals zero for any given values of the independent variables, (e) the error term e is homoscedastic, meaning that its variance is the same for any given values of the independent variables in x, and (f) the error term is normally distributed. Any violation of these underlying assumptions necessitates consideration of the impact on the OLS estimates and may dictate a different method of parameter estimation or a different functional form of the price model such as restating price as the natural log of price or including quadratics and interaction terms for certain independent variables to ensure the model is linear in the parameters (assumption a above).

...