Few of us make a bigger financial decision than buying a house. Where we live and choose to bring up our children is something that implies emotional commitment and investment. So house moves can be a powerful way to reveal what people really think and value – including about their neighbours.
Previous research on the UK has tended not to find evidence of “white flight” – a term borrowed from mid-20th century North America, where the large migration of people of colour from the rural south to metropolitan areas resulted in white households leaving inner cities for the suburbs.
While there has so far been limited evidence of white flight in modern Britain, there have been some shortcomings with the previous studies that have looked into it. One of the problems is that the areas studied have been too large to detect what impact people’s immediate neighbours have on their decisions to move house.
In recent research that I published with my colleague Sue Easton, we wanted to take a closer look at what was happening when people with ethnic minority names moved next door or very close to homeowners with white British names. While you might be oblivious to who moves into a house a street or two away, you’re likely to be very conscious, and much more concerned about, who moves in next door. It is at this level that our true conviviality or prejudices are revealed.
What’s in a name?
In our study, we used property registration data for the 40% most deprived census areas of Glasgow in Scotland. We created a longitudinal dataset – one that follows individual homeowners over time – constructed from the population of homebuyers recorded in all property transaction records from 2003 to 2014.
Crucially for our purposes, the property registration data we used recorded the names of both buyers and sellers in each house transaction. How people respond to particular sorts of names can reveal certain prejudices.
One 2009 study based on in-depth interviews with “white” and “non-white” participants in Glasgow, for example, found that white residents responded with racialised or prejudiced attitudes when faced with the hypothetical scenario that a person with a stereotypically Muslim name was thinking of moving into the house next door.
Respondents associated such names with Islamic terrorism and with stereotypes of Asian neighbours being loud, inconsiderate, and living in overcrowded conditions.
Analysis of particular name groups has become a lot more feasible as a result of the growing literature on combining forenames and surnames to impute ethnicity. And now sophisticated software is being designed to do this too, such as the Onomap software which can assign ethnicity to names in a large dataset based on linguistic and cultural analysis.
We used Onomap in our study to categorise the names of the house buyers in the millions of transactions in our dataset. Doing this allowed us to monitor whether the likelihood of somebody moving house is associated with buyers with particular ethnic name groups purchasing houses in the immediate neighbourhood.
Our results suggest that, on average, the length of stay of homeowners with white British names is significantly reduced if house buyers with Pakistani or other Muslim names move within 50 metres.
For example, for every household moving in with a “non-white other” name, primarily Muslim in origin, the average length of stay of existing homeowners with white British names halved. We also found evidence, however, that the effect diminished somewhat as the number of those moving in with Pakistani or Muslim names increased.
In our statistical analysis, we were careful to control for the housing market conditions in the neighbourhoods. We did this by including the monthly change in mean property price in the surrounding area. So the effect we found is unlikely to be explained by fluctuations in local demand.
Nevertheless, we cannot say for sure that those homeowners with white British names who chose to move did so because of anti-Pakistani or anti-Muslim sentiment. Our study was observational, and we did not interview people about the choices they had made.
It is also worth bearing in mind that the effect is very likely to be partly driven by “neighbourhood churn”. When people move into your neighbourhood, it usually means others are moving out.
If you have built up strong friendships with your neighbours, when any of them move, you might feel less tied to the neighbourhood. So even when newcomers have white-British names, we found that the chances of existing homeowners with white-British names moving out increased. Still, this did not happen as much as when the people moving in had Pakistani or other “non-white” (primarily Muslim) names.
Another possible explanation is that some of the buyers with Pakistani and Muslim names are landlords, rather than owner occupiers. We know, for example, that the length of stay for homeowners is more than seven times that of renters, meaning they have a higher turnover of residents.
If homeowners are averse to living near rental properties because of this impact on neighbourhood churn, they may be more likely to move out when a neighbouring house is purchased by a landlord. Although we could not guarantee that we’d remove all landlord purchases from our analysis, we did go to considerable lengths to exclude them.
One alternative explanation of “white flight” is that white owners are moving out in order to move up the housing market. But this does not explain why we found white British homeowners were more likely to move when the people buying nearby have Pakistani or Muslim names rather than white-British names.
Another explanation is that those moving out are doing so not because of the name or ethnicity of the new house buyers, but because those moving in are from a lower social class. To address this, we controlled for the deprivation level in the neighbourhood and the property price at the time of purchase. The more expensive the house, the more affluent the buyer needs to be in order to purchase the property. So it is unlikely that house buyers with Pakistani names were significantly poorer or wealthier than existing residents.
This left us with the uncomfortable possibility that some homeowners with white British names were more likely to relocate if homeowners with Pakistani and Muslim names moved in within 50 metres.
In comparison to the white flight of 1950s America, the effect is probably pretty small and much more localised. But it nevertheless suggests house moves in Glasgow may be partly driven by latent aversion to having neighbours with Pakistani or Muslim names.
Gwilym Pryce, Professor of Urban Economics and Social Statistics and Co-Director of the CDT in Data Analytics & Society, University of Sheffield.
This article first appeared on The Conversation.