The Impact of Residential Mobility on Measurements of Neighbourhood Effects

Abstract

Neighbourhoods and cities are dynamic; their characteristics and relative positions change over time due to constant moves in and out. However, neighbourhood effect theory and most attempts to quantitatively estimate neighbourhood effects seem to treat neighbourhoods as if they were static. This paper argues that such a view is not only strange but may also result in biased estimates. Four methodological challenges are highlighted that are directly related to mobility: (1) measures of exposure time; (2) neighbourhood change; (3) selection bias; and (4) endogeneity. These are all topics worthy of scholarly interest in themselves, but also challenges that all neighbourhood effect studies must address to convincingly argue that their results are indicative of causal relationships—results of neighbourhood transmission mechanisms—and not just statistical correlations. The paper discusses how and to what extent these challenges have been met by the quantitative neighbourhood effect literature and gives directions to future research.

Keywords:

• Residential mobility
• neighbourhoods
• neighbourhood effects
• holistic framework
• methodological challenges

Acknowledgements

The author wishes to thank Jürgen Friedrichs and George Galster for their critical reading and many helpful suggestions to an earlier draft of this paper, presented at the PhD course Poverty Neighbourhoods and Neighbourhood Effects, 3–7 August 2009, University of Oslo. Thanks are also due to three anonymous referees for valuable comments that helped improve the manuscript.

Notes

¹ Reference is made to both neighbourhoods with high shares of ethnic minorities and to neighbourhoods inhabited by an ethnically homogenous majority population.

² The author is grateful to George Galster for making this point.

³ However, the direction of the potential bias is unclear; Jencks & Mayer (1990) and Tienda (1991) argue that effects are over-estimated while Brooks-Gunn et al. (1997) suggest that the opposite could also occur.

⁴ A ‘Heckman correction’ variable; see various papers on selection bias by James Heckman from 1979 and onwards.