Spatial pricing in interdependent markets: a case study of petrol retailing in Sheffield

This paper reports the development of empirical models to explain spatial price variation in an urban area. Models are constructed for petrol price data collected in 1995 and 1997 in Sheffield, England. The 1995 data are modelled by using only supply-side predictors following the collection of supply-side information from field surveys of the retail sites, a site questionnaire survey, and interviews with site managers. The 1997 data are modelled by using supply-side predictors and demand-side predictors that relate to the economic characteristics of the population of consumers. This modelling is based on field surveys of the sites, a new site questionnaire survey, and a household survey. The purpose of this work is to assess supply-side and demand-side factors in explaining spatial price variation. Supply-side predictors are classified into site characteristics, location characteristics, and measures of spatial competition. We examine the relative importance of these different groups of supply-side variables in explaining price variation, with a particular interest in location and competition effects as these relate directly to the spatial and geographical aspects of the problem. Another contribution of the paper is to observe the stability of findings by contrasting the best-fitting models obtained for the 1995 price data to the best-fitting models obtained for the 1997 price data. We find that no demand-side factors are statistically significant. For 1995 a spatial competition variable and a location variable (whether a site is attached to a supermarket) are the consistently important supply-side variables. For 1997 all three categories of supply-side variables are important.