For whom are cities good places to live?

ABSTRACT

In this paper we use survey data to examine heterogeneity in the urban gradient of life satisfaction. Are some sociodemographic groups more satisfied in cities than others? We find that young persons with tertiary education generally report higher levels of life satisfaction in Norway’s largest city, Oslo, whereas the elderly and the less educated are more satisfied elsewhere. These results may shed light on the ‘urban paradox’: the tendency of large cities in developed countries to have low levels of average subjective well-being and also why the received literature has produced mixed results, as the sociodemographic composition of cities varies.

KEYWORDS:

• subjective well-being
• life satisfaction
• cities
• sociodemographics

JEL:

• I31
• J10
• R23

ACKNOWLEDGEMENT

We thank Giorgio Fazio, Anders Barstad, Raul Brey Sanchez, the editor and the referees for useful comments. We are also grateful for comments from the seminar participants at the Norwegian University of Science and Technology, Statistics Norway, ValueChange workshop at Lillehammer 2020, and the 59th North American meeting of the Regional Science Association International (RSAI 2012). We are grateful to Simen Hustad at the Norwegian Government Agency for Administration and Financial Management for sharing the research version of the NGAAFM data set.

DATA AVAILABILITY

The survey data are available under licence from TNS Gallup, Statistics Norway and the Norwegian Government Agency for Administration and Financial Management. 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. Burger et al. (2020) carried out a Blinder–Oaxaca decomposition to examine how the urban–rural gap depends on place and respondent characteristics. However, the regression results underlying the decomposition are not reported.

2. Four surveys ask about life satisfaction, but we omit the 2020 survey due to the pandemic.

3. The EU-SILC is a panel survey where individuals are present in a maximum of four consecutive years. For this reason, some individuals are present in both the 2017 and 2018 surveys. To handle this issue, we cluster standard errors on individuals.

4. Income variables are price adjusted and expressed in 2019 NOK.

5. The interaction term with age squared is not statistically significant and therefore omitted.

6. Psychological traits vary geographically and are correlated with reported subjective well-being (Diener et al., 1999; Rentflow et al., 2008).

7. The TNS Gallup survey includes the question we use to calculate the response scale proxy, whereas the EU-SILC and NGAAFM surveys do not. Therefore, we extend the analysis of place satisfaction to control for individual variations in response scale usage.

8. People with a positive evaluation of the local climate, for given meteorological conditions, also express a positive evaluation of other local attributes, including scoring high on place satisfaction.

Additional information

Funding

This work was supported by the Research Council of Norway [grant numbers 295989, 255509 and 280393].