Remaking urban segregation: processes of income sorting and neighbourhood change

Why examine how segregation changes?

Spatial segregation – the division of cities into richer and poorer neighbourhoods, for example – is a key feature of urban areas. Many studies look at how much segregation there is and at how this changes over time but few examine the processes which underpin these changes. 

In general, people choose to live in neighbourhoods with others like themselves. Theory suggests that, where they see a change in their incomes (social mobility), this will tend to lead them to change neighbourhood (residential mobility) to find a new place where they fit in so we expect these two processes to balance out. However, the few studies which have looked closely at these processes found them to be more complex. 

This analysis aimed to extend previous work by using better data (incomes for all residents in two Dutch cities over a seven year period), and analysing change not just at the city scale but also down to the level of individual neighbourhoods. We chose two cities with contrasting situations: Amsterdam, the centre for finance and business as well as creative industries, with a relatively young population but growing only slowly; and The Hague, a centre for law and government, with an older population but growing faster. 

Key findings

There is enormous churn in city populations. In our seven-year period, one third of those present at the start had died or moved away by the end, to be replaced by a similar number of in-migrants. Of those who remained, one third had moved neighbourhoods. So only around 1-in-2 people were living in the same neighbourhood by the end. In the face of this level of churn, it is perhaps most surprising how little segregation alters. Nevertheless, income segregation did rise in both cities so we looked at how this happened.

Our first finding is that the processes which produced these changes were rather different. In Amsterdam, the main driver was residential mobility or movements within the city but in The Hague, it was in-migration from outwith the city which brought about the changes. 

When we drill down to the neighbourhood, the picture gets more complex. In some, changes were as expected: richer areas where incomes rose, or poorer areas where they fell, both of which produce more segregation. These we term polarising neighbourhoods. In others, changes ran in the opposite direction, tending to reduce segregation: richer areas where incomes fell and poorer areas where they rose, termed reordering neighbourhoods.

Furthermore, the different types of neighbourhood have a distinct geography. Looking at Amsterdam for example (Figure 1), the poorer reordering neighbourhoods (hatched) are in inner locations surrounding the already-affluent ‘Canal Belt’ – classic gentrification locations.

Figure 1: Polarising and re-ordering neighbourhoods in Amsterdam

Remaking urban seg fig 1












Note: The map shows neighbourhoods as one of four types: polarising (lines/hashes) and re-ordering (solid shading), each sub-divided by whether incomes above or below average in 1999.


This study adds additional detail to our understanding of how spatial segregation changes. We show that processes can vary between and within cities. 

Want to know more?

Bailey, N., van Gent, W., and Musterd, S. (2016) 'Re-making urban segregation: processes of income sorting and neighbourhood change', Population, Space and Place. 

Monday, October 10, 2016
Nick Bailey

Research Strand: