Mobility and urban quality of life: a comparison of the hedonic pricing and subjective well-being methods

ABSTRACT

There is no consensus about the relation between urban scale and quality of life. Quality of life in Oslo and the rest of Norway is compared using two commonly employed methods to measure geographical variation in quality of life: hedonic pricing based on the Rosen–Roback model and analysis of subjective well-being. Since the hedonic pricing approach assumes that households are perfectly mobile, the population is divided into mobility quartiles. The methods arrive at the same conclusion for the most mobile population group – quality of life is higher in Oslo – but different conclusions are reached for other mobility groups.

KEYWORDS:

• quality of life
• cities
• mobility
• hedonic pricing
• place satisfaction

JEL:

• I31
• J17
• R23

ACKNOWLEDGEMENTS

The authors thank the editor, three referees, David Albouy, Giorgio Fazio and colleagues at Department of Economics for useful comments. The authors are also grateful for comments from seminar participants at the Norwegian University of Science and Technology, the 59th North American meeting of the Regional Science Association International (RSAI 2012) and the 52nd Annual Meeting of the Western Regional Science Association (WRSA 2013). The authors appreciate the research collaboration with Rolf Aaberge and Audun Langørgen at Statistics Norway. Register data are available under licence from Statistics Norway and the survey data under licence from TNS Gallup. Contact the authors for access to Stata codes on model specifications.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Notes

1. This excludes the investigation of very disaggregated place attractiveness. Analyses of, for instance, neighbourhoods, gentrification and ‘right to the city’ are therefore beyond the scope of this paper.

2. Age and other sociodemographic characteristics are registered in 2012.

3. In a given year, employees are included if they have at least one full-time contract, no more than two contracts and work three months or more. We exclude workers in the primary and public sectors, as the wages in these sectors are to a substantial degree determined by national regulations rather than by market forces. We have no information on overtime work. Hourly wages will thus be overestimated for workers with many overtime hours.

4. With three exceptions, definitions of sociodemographic variables are identical in the relocation and survey data sets. First, in the survey data set, marital status does not distinguish between marriage and cohabitation, whereas the relocation data set, which is based on administrative registers, does not have information about cohabitation. Second, the relocation data define children as 18 years and younger, while the survey data set operates with a threshold of 17 years of age. Third, the relocation data set defines tertiary education as at least one year of completed college/university, while in the survey data set the respondents are asked to state if they have higher education, but the questionnaire does not provide a definition of higher education.

5. Goods sold close to where they are produced should in principle be cheaper because of transportation costs. The prices of services should reflect local factor prices.

6. Municipal taxes are irrelevant to the present exercise since they are used to pay for public services to the residents of the municipality.

7. Since answers to survey questions are discrete, the reported regressions were also estimated using ordered probit models, and all conclusions carry over.

8. Studies have found associations between sociodemographic variables and happiness/life satisfaction (Diener et al., 1999).

Additional information

Funding

This work was supported by the Research council of Norway [grant number 255509].