1216 Spatial Hedonic Models
26.2 Hedonic house price models
In this section, we first outline the main properties of the hedonic price model
in the context of specifications for house prices. Next, we discuss some important
features of the estimation and identification of such models. We postpone the
discussion of specific spatial aspects to the next section.
26.2.1 General framework
In the second half of the 1960s, a new branch of utility theory evolved from the
pioneering work of Lancaster (1966), in which utility was defined as a function of
the characteristics of a good. Initially, the focus was primarily on consumer models,
until Rosen (1974) generalized this to a model of market equilibrium that took into
account both consumers and producers. In this, the individual’s utility becomes a
function of the characteristics of a commodity, and producer costs depend on the
characteristics of the good.
The hedonic price equation defines a market equilibrium after all interactions
between supply and demand have taken place. A considerable literature has built
upon this basic model (for a review, see, e.g., Malpezzi, 2002), and active research
pertaining to both theoretical and econometric aspects continues apace (e.g.,
Ekelandet al., 2004).
In the specific context of house price models, the basic hedonic specification
assumes that the utility of a household or an individual is a function of a compos-
ite goodx, a vector of location specific environmental characteristicsq, a vector
of structural characteristicsS, a vector of social and neighborhood characteristics
N, and finally a vector of locational characteristicsL(Freeman, 1999). One of the
main assumptions of the hedonic model is that preferences are weakly separa-
ble in housing and its characteristics. This implies that the demand for housing
characteristics can be written as a function of “expenditure and prices within the
group alone” (Deaton and Muellbauer, 1980, p. 124). Specifically, this implies
that housing demand can be written as a function of the house price, with the
prices of all other goods represented by a composite goodxas the numeraire.
Also, perfect information is assumed, in the sense that a consumer perceives all
relevant house characteristics and takes them into consideration when purchasing
a house.
The decision problem involved in a house purchase then consists of maximizing
utility subject to the usual income constraint. Formally, for housei:
Max U(
x,qi,Si,Ni,Li)
(26.1)
s.t. M=Pi+x,whereMis the household income,Piis the price of housei, and the composite
goodxis the numeraire.
For each characteristic of interest, the first-order condition defines the marginal
willingness to pay (MWTP) for changes in the levels of such a characteristic. For