Luc Anselin and Nancy Lozano-Gracia 1215Spatial patterns in the housing market are expected to arise from a combina-
tion of spatial heterogeneity and spatial dependence (Anselin, 1998). For example,
spatial heterogeneity may originate from spatially differentiated characteristics
of demand, supply, institutional barriers, or racial discrimination. This system-
atic variation in the behavior of economic agents across space warrants special
attention, since any model that imposes homogeneity will be misspecified.
Spatial autocorrelation may appear when either the prices or characteristics of
houses that are closer together are more similar to each other than those from
houses that are farther apart. Alternatively, it may also stem from measurement
problems in explanatory variables, omitted variables, and other forms of model
misspecification (Baumont, 2004). A major class of such misspecifications pertains
to so-called neighborhood effects, which are typically unobserved and modeled
as part of the error term. Importantly, spatial heterogeneity and spatial autocor-
relation may be observationally equivalent (Anselin, 2001a), which may lead to
difficulties in isolating the two effects in practice. Spatial autocorrelation may also
result from spatial heterogeneity not being modeled correctly (Anselin and Griffith,
1988; Baumont, 2004).
The consequences of ignoring spatial autocorrelation and spatial heterogeneity
when they are, in fact, present in the data-generating process have been widely
discussed in the literature and has led to the separate field of spatial economet-
rics (Anselin, 1988; Anselin and Bera, 1998). A recent comprehensive review of
the field can be found in Anselin (2006). Also, after some initial work by Dubin
(1988) and Can (1990, 1992), among others, the explicit consideration of spatial
effects through the application of spatial econometrics has become more common-
place in empirical studies of housing and real estate markets. Reviews of the basic
specifications and estimation methods applied to these spatial hedonic models
are provided in Anselin (1998), Basu and Thibodeau (1998), Paceet al. (1998),
Dubinet al. (1999), Gillenet al. (2001), and Pace and LeSage (2004), among
others.
The literature on hedonic models is vast, both theoretical as well as empirical.
We do not attempt to review this in the current chapter, but instead focus on the
methodological aspects related to the implementation of spatial hedonic house
price models in empirical studies. We illustrate how different spatial econometric
approaches have been applied and their implications for model specification, esti-
mation and interpretation. We do not attempt to provide a comprehensive review
of the empirical literature, but consider a wide range of articles, illustrative of the
different perspectives taken in applied work.
We begin the remainder of the chapter by setting the stage with a brief review of
the principles underlying the hedonic house price model, followed by an extensive
discussion of spatial econometric aspects due to spatial models and spatial data,
specific to house price applications (for a more comprehensive technical review,
see Anselin, 2006). We then review in turn the treatment of spatial dependence
(including space-time dynamics) and spatial heterogeneity, with selective illustra-
tions from the empirical literature. We close with a discussion of policy implications
and some conclusions.