Geography Reference
In-Depth Information
discussion to the data foundations to
reinvigorated urban modelling.
1
A digital copy of the 1989 domestic rates
register. This provided historical information
on the 'rateable values' of every property—
that is, the annual rent that a property would
historically have commanded in a rental
market. Until the late 1980s, such values
provided the basis for local taxation in the
UK. Such valuations can provide only a rather
crude estimate of capital values in a housing
market, which today is dominated by owner-
occupied housing, supplemented by local
authority rented housing (which historically
have been built by the state and let at
subsidised rents, and today have often been
bought under 'right to buy' legislation: see
Balchin et al . 1998). The domestic rates
register also contains limited information on
property type.
Modelling property values using GIS
Our first example concerns the variation in
property values across an urban area in a single
time period. Although much has been written on
so-called 'bid rent' models (Alonso 1964; Muth
1969), we actually seem to know remarkably little
about the detailed ways in which value is
constituted in the built environment (Orford
1997). The classic bid rent model postulates that
households make trade-offs between the distance
that they live from the centre of the city and the
size (in terms of, say, floorspace) of the properties
that they occupy. Yet in reality, today's cities are
not monocentric: tenure subdivides housing
markets into owner-occupied, private rented and
public rented sectors; residents perceive distance
in a wide variety of metrics (e.g. travel time, public
transport cost); and there is much more to urban
living than the journey to work (e.g. availability
of good schools, proximity to parks). The bid rent
formulation of property values seems intuitively
plausible yet insufferably simplistic in such
circumstances. Moreover, in many countries
(including the UK) we know rather little about
the distribution of property values across towns
and cities, in the absence of publicly available land-
use information systems.
In this context, Longley et al . (1994) developed
an urban model of the capital values of all 45, 889
properties in the inner area of Cardiff, UK—the
extensive and heterogeneous core area of the city.
The UK has no national comprehensive and
publicly accessible land-use/valuation information
system, and UK readers should note that at the
time of their study the council tax valuation list
was only in the process of being compiled for the
first time. (This valuation list classifies properties
into eight broad categories only for local taxation
purposes and is not usually made available to
researchers for geographically extensive areas).The
following disaggregate data sources were available
to them:
2
A digitised network of all streets in the study
area.
3
Digitised boundaries of eighty-three 'house
condition survey areas', devised for a local
authority-funded house condition survey and
designed to identify small areas within which
there was general homogeneity of house
structure and type (but which in practice were
often far from homogeneous).
The study area, together with the digital GIS
layers, is shown in Figure 44.2. Together, these
represented a rich range of data sources and
provided the data foundations for development of
a simple spatial model of house price variation.
The basis for the model is essentially that of
Tobler's (1970) famous 'First Law of Geography':
'call things are related but nearby things are more
related than distant things'. Thus, for example,
values of properties around an attractive park are
likely to be uniformly high, and values will tail off
with distance from this 'positive externality'; and,
by contrast, property values around a waste tip are
likely to be low, and values will recover only with
distance from this 'negative externality'. But this
assumes that we hold other variables such as
construction type (house, flat, etc.), size and so
forth constant, and in reality we will have to
model the effects of more than just space alone.
Search WWH ::




Custom Search