Where Sustainable Transport and Social Exclusion Meet: Households Without Cars and Car Dependence in Great Britain

Abstract

A secondary analysis of the British National Travel Survey for the years 2002–2010 shows that the composition of the group of carless households is a good indicator of the level of car dependence in a local area: indeed, while non-car ownership in peripheral and rural areas very often correspond to a marginal socio-demographic situation, this is less and less true as one moves towards larger urban areas. Similarly, while in sparse areas most households without cars are either virtually immobile or reliant on car lifts, in large urban areas the ‘mobility gap’ between car-owning and carless households is considerably smaller, as the latter are able to use modal alternatives to the car. These findings are interpreted with reference to an integrated theoretical framework, showing how changes in land use and the environmental and social impacts of increasing motorization are intimately linked. Notably, the consequences of the self-reinforcing cycle of car dependence on two forms of car-related transport disadvantage (car deprivation and forced car ownership) are highlighted. Overall, the article highlights how the socio-demographic composition and the travel behaviour of carless households vary systematically across different types of area: this has interesting implications for sustainable transport policy and research.

Key Words:

• Car ownership
• households without cars
• transport disadvantage
• sustainable transport
• car dependence

Acknowledgements

This article is based on the PhD thesis ‘Where sustainable transport and social exclusion meet: Households without cars and car dependence in Germany and Great Britain' defended in July 2013 at the University of Milan-Bicocca, in the context of the European doctoral programme in ‘Urban and Local European Studies'.

Funding

The scholarship was funded from 2009 to 2012 by the Italian Ministry of Education, University and Research.

Supplemental Data

Supplemental data for this article can be accessed at doi:10.1080/1523908X.2013.858592.

Notes

1. To be sure, the fact that this process is self-reinforcing does not exclude external determinants of motorization such as rising income and decreasing motoring costs (see de Jong et al., 2004).

2. The notion of ‘transport disadvantage’ (Currie et al., 2007; Dodson et al., 2004; Hine & Mitchell, 2003) can be defined as the lack of access to services and opportunities arising from the interaction of three sets of factors: land-use patterns, the transport system and individual characteristics (Currie & Delbosc, 2011a, p. 15). It is assumed that this has a potentially negative impact on social exclusion and/or well-being at least.

3. The figures include individuals under 16 who by definition cannot be drivers.

4. Focusing the analysis on individuals who do not have access to a car as drivers would be equally arbitrary, as it would mean assuming that non-drivers in households with cars have less car access than drivers in households with cars—a questionable assumption.

5. The NTS 2002–2010 was conducted by the National Centre for Social Research on behalf of the Department for Transport, which owns the data. The data-set is kindly provided by the Economic and Social Data Service through the UK Data Archive at the University of Essex, Colchester.

6. In the NTS, the HRP is defined as ‘the householder with the highest income, or their spouse or partner' who answered the household questionnaire (Rofique et al., 2011, p. 16).

7. In this and in the following section, I illustrate differences between areas with reference to the variable ‘type of area’. However, every trend has been double-checked using two other geographical variables: population density in the Local Authority and in the Primary Sampling Unit. The results broadly confirm the findings illustrated here.

8. The detailed results for the models are not reported here for the sake of brevity. The models include the following independent variables: number of household members (simple and squared terms); number of members under 16; female HRP (dummy variable); age group of HRP (categories: 16–29; 30–39; 40–49; 50–59; 60–69; 70+); number of employed members; at least one member with mobility difficulties (dummy); income quintile. In addition, a ‘survey year’ predictor was included to control for differences between waves.

9. The clustering was conducted using a k-means algorithm, Euclidean distance as dissimilarity measure and standardized input variables. A four-cluster solution was retained, representing the most distinct clustering, as attested by the maximum value of the Calinski–Harabasz pseudo-F-statistic.

10. The results of a similar study on German data (Mattioli, 2013a) confirm this conclusion.

11. It must be remembered, however, that the analysis focuses exclusively on individuals in non-car-owning households. In fact, some members of ‘car-deficient households’ might also be described as ‘car reliant’, insofar as they depend on lifts from others.

12. However, they might make a difference for members of ‘car-deficient households’.