XClose

Institute of Epidemiology & Health Care

Home
Menu

A social analysis of ethnic segregation

Eric Kaufmann, Birkbeck, University of London

(Project no. 0301317, previously 30131)

The general aim of the study is to ask which factors predict 2 aspects of ethnic segregation: 'white flight' (including white avoidance of minority areas) and 'minority flight' (basically upward social mobility of minorities out of lower income areas). Aim is econometric analysis of the characteristics of movers and stayers from both groups.

We already know from Simpson (2007) that more whites than nonwhites leave largely nonwhite areas, but that the difference is much less substantial than some believe. Simpson suggests both may be leaving poorer areas for better ones. There may be more to the picture than this. What we do not know are the characteristics of whites and nonwhites who leave such areas. The project will test an argument that successful, better-educated nonwhites tend to leave their areas of immigrant settlement for 'whiter' areas; while whites who leave these areas are more working-class, less educated and perhaps have a different age or family profile. That the two groups of leavers move to different types of area (something not captured in the static area census analysis) and also that among those moving in to nonwhite areas, whites tend to be better educated and often without families (i.e. students or 'gentrification') as compared to the nonwhites who move in to these areas - often from abroad.

1) ONS LS members present at 2001 will be sorted into residential quintiles of districts, wards and output areas based on % nonwhite population where quintiles are based on dividing geographies such that each quintile contains 20% of the nonwhite population of the country, as per Simpson (2007). Note that these are not total population quintiles but quintiles based on dividing the minority population into 5 even bands. In other words, three analyses - one at each level of geography. In each case LS members will be sorted into quintiles based on the % white population in their district/ward/OA of residence as above.

2) For each quintile the migration characteristics and the background variables for age, sex, religion, class, marital status, children, education, student will be compared for movers and stayers from minority and majority groups.

3) It should be possible to do the analysis at three levels of geography: output area, ward and county district.

4) This procedure will be repeated for 1991 LS members when a direct comparison of their place of usual residence at the two time points may give a larger sample of movers and stayers. We will then see if there is a pattern of origin and destination quintile that varies between whites and minorities, and also, whether ethnic change or merely ethnic levels have an effect on the decision to migrate, by ethnic group, with controls for age, sex, religion, class, marital status, children, education, student. It is hoped to expand the project into a comparative study with the USA, Canada, and possibly Austria, Sweden and Switzerland.