The effect of accessibility on retail rents: testing integration value as a measure of geographic location

Abstract

This article investigates cross-sectional variation in retail rents on the micro level, i.e. differences in rent for shops on street X compared to shops on street Y. Retail rents from the inner city of Stockholm are analysed. Free standing shops dominate the database (i.e. not located in malls). Differences in rent across locations is modelled with distance to the city centre, distance to sub-centres within inner city Stockholm, city part dummies and so-called integration values. Integration values have shown a strong correlation to pedestrian and other types of traffic. Loosely speaking, the integration value for a particular location in an urban area is the average number of turns a pedestrian must make to reach other locations in the city. In this study it is interpreted as a measure of accessibility. It is hypothesised that locations with high (low) integration values typically have high (low) retail rent. Integration values are found to complement the traditional location measures distance to city centre/sub-centres and city part dummies. Integration values can be calculated for urban areas that have not yet been built. They may, thus, be useful in the planning of new urban areas and for predicting retail rents.

Keywords:

• rents
• retail
• spatial autocorrelation
• cross section
• multiple regression analysis

Acknowledgements

The author would like to thank the people who answered the questionnaire and Lars Erik Lundberg Scholarship Foundation for funding. I would also like to thank three anonymous referees for their helpful comments and suggestions.

Notes

1. The measure was neither named nor developed by the author of this article but within the research field Space Syntax.

2. This is debatable and it would be interesting, in further research, to look into the possibility that global integration values have an effect on retail rents.

3. One may argue that I have let current economic activity determine the city centre. However, there is sound economic theory behind the variable distance-to-city-centre (travel costs are the smallest in the city centre) so there is a fundamental story to that variable. I have merely let the data decide the exact location of the centre. The sub-centre, on the other hand, is more problematic in this sense and was in fact included in part to see if including it would take away the effect of integration values. Such a result would suggest that the location of sub-centres has to do with integration values. Or, differently put, that integration values can help predict where sub-centres occur. Since the sub-centre and integration values are both significant, this does not seem to be the case.