1244 Spatial Hedonic Models
The distinction between equations (26.45) and (26.46) is important in light of
the points recently raised by Small and Steimetz (2006). They considered a different
interpretation of welfare effects between thedirectvaluation in equation (26.45)
and the multiplier effect included in (26.46). In their view, the multiplier effect
should only be considered as part of the welfare calculation in the case of a
technological externality associated with a change in amenities. In the case of
a purely pecuniary externality, the direct effect is the only correct measure of wel-
fare change. Only in a spatial lag specification is it possible to distinguish between
these two effects.
The use of an analytical approach to deriving MWTP in impact analysis is limited
in a number of ways. It is constrained by the functional specification of the hedo-
nic model. Specifically, if nonlinearities are introduced in the hedonic model, such
nonlinearities will be transfered to possible dependencies and/or nonlinearities in
the MWTP itself. Also, the analytical derivation of MWTP is limited to marginal
changes in the characteristics. For non-marginal changes, the inverse demand func-
tion needs to be estimated, which is often difficult in practice. Moreover, nonlinear
specifications require a value for the price and/or characteristics in the calculation,
which is typically simplified by using a mean or median value.
An alternative to the analytical derivation is to use a simulation approach, as
outlined in Anselin and Le Gallo (2006). The essence of the approach is that valu-
ation is based on the computation of predicted values for individual observations
given their actual household characteristics. In essence this boils down to a discrete
approximation to the notion of marginal willingness to pay. A major advantage of
the simulation approach is that it allows greater flexibility, both in the specification
of the type of policy experiment as well as in the valuation. Since the valuation is
computed for individual house observations, the results can be obtained for any
desired level of spatial aggregation, such as by county or zip code (for an extensive
application, see Anselinet al., 2008).
In a non-spatial linear model, the change in predicted values can be expressed
as follows:
pˆ 0 −pˆ−k=(z 0 −z−k)βk, (26.47)wherepˆ 0 −pˆ−kis the change in valuation and(z 0 −z−k)is the change in charac-
teristick. If a spatially lagged dependent variable is included in the hedonic price
equilibrium, the change in valuation resulting from a change in characteristick
would turn out to be(I−ˆρW)−^1 (z 0 −z−k)βk. Any nonlinearities of the hedo-
nic price function would also be reflected in this approximation to the change in
valuation.
By ignoring this multiplier effect and looking at individual or private benefits
only, underestimation of the overall social welfare from changes in house char-
acteristics may result, such as reductions in air pollution, greater access to public
facilities, and public services. In decision making under strict efficiency rules, this
may lead to an underinvestment in such characteristics.
In sum, a spatial econometric approach yields more efficient estimates of policy
effects of interest, allows for a distinction between direct and indirect welfare effects