An Investigation into the Estimation and Reliability of Urban Shopping Models

Abstract

This paper evaluates the reliability of variants of five urban shopping models : Central Place theory, the Intervening Opportunities model, Gravity model, the Entropy model, and the Building Research Establishment (BRE) model . For each model variant parameters are estimated against data for four towns in the United Kingdom : Coventry, Derby, Watford and Telford, using alternative procedures for testing significance of the calibration . The models are then used to forecast hypothetical changes in local shopping environments, including dynamic effects, such as the progressive decline of local centres . On the basis of the results, some general conclusions are drawn as to the relative performance of the models and the factors determining their reliability in different forecasting situations.Mathematical models are used extensively in urban and regional planning to estimate traffic flows, housing demand and the likely impact of new shopping amenities . There is still, however, considerable controversy surrounding the use of models and great suspicion about the reliability of their results . Very rarely, in fact, have the forecasts made using these models been checked . In this paper we take shopping models as a typical example of the problems and possibilities with regard to the use of models in general .Other urban models (such as the traffic forecasting methodology of the Transport and Road Research Laboratory) and their predictions have been more thoroughly criticised (Gershuny, (1978) ; Turner, (1979, forthcoming) and comparisons of existing spatial interaction models made by Openshaw (1976) and Hathaway (1975)) . Comparisons of urban shopping models (e .g. Hodges and Kennett, 1973 ; Pankhurst and Roe, 1978) are referred to later . Models often conceal, in their wealth of technicalities, decisions that could substantially affect the results of using them (see, for example, Massey and Cordey Hayes, 1971) . In particular, the application of a given shopping model may lead in particular instances to quite arbitrary results . This of course has significant implications for any conclusions based R .on these models . The aim of the paper is to obtain an understanding of how reliable such models and the procedures for using them may be and how models trade off the interests of different social groups, as affected by planning decisions . Following this, we make suggestions about the use of models in planning . Although the discussion is for foodstuffs shopping models, the general arguments apply to other kinds of urban and regional models . We do not intend to consider the technical problems of building a model in any detail here . Rather, in the various stages of the construction and application of the models we shall indicate where uncertainty may arise, or judgments are made about the content of the model . These occur especially in the collection of data and their measurement, the choice of model, the form of the relationship used in a model, the procedure used to fit the data to the model (i .e . calibration) and t...

A cross-sectional model of rents in urban retail centres

“…First, the quality of transport infrastructure is relatively constant, and second where the consumer has a choice, the shortest route and fastest or most economical method is preferred. Turner and Cole (1980) observed `crow-¯ight' distance as a more commonly-used method than road distance. Distributive Trades EDC (1970) note the straight-line distance cannot allow for traf® c congestions, public transport±car ownership, or physical barriers, but the empirical research demonstrates a close correlation between such measurements and actual journey times recorded.…”

“…The pedigree of gravity models' stretches back to 1931 with Reilly' s Law which set out a simple model to allocate catchments to towns (Distributive Trades Economic Development Committee, 1970). Subsequently a family of spatial interaction shopping models under the generic title of gravity models has been spawned (see Turner andCole (1980), Huff (1963), Kivell andShaw, 1980, Lakshmanan andHansen (1965)). The gravity submodel utilized here follows a standard speci® cation and is set out below.…”

How Critical Is a Good Location to a Regional Shopping Center?

“…Detractors at that time argued that the clustering of homogeneous retailers was socially wasteful and economically unstable for retailers since there is a doubling of economic effort. 3 See LaLonde (1962), Dent (1978), Ellwood (1954), Lakshmanan and Hansen (1965), Pankhurst and Roe (1978), Turner and Cole (1980), and Okoruwa, Nourse and Terza (1996). ~ The measurement of y, requires that we have data on total retail expenditures within the trade area.…”

Evaluating a Store Location

“…The second most popular technique -also subjectively or intuitively assessed -is the analogue study: the comparison of the proposed store/site with similar existing stores in the company's portfolio in analogous areas [6]. The original spatial interaction models were the first of the mathematical approaches developed based on Reilly's Law of Retail Gravitation [7,8,9,10]. These have been developed [11,12] into intervening opportunities models [13] and multiplicative interaction models [14,15].…”