Inter-ethnic segregation across British labour markets
Genna Kik, Megan Blake and Paul White, University of Sheffield
[Project number 30034]
My use of the LS would form part of a larger PhD that examines whether the occupational segregation of certain ethnic groups across British labour markets is changing. It is hoped that the PhD will allow a better understanding of the temporal and spatial dynamics of the labour market, and a more contextualised understanding of minority ethnic experiences of the labour market.
Much of the British literature on employment suggests that occupations are gender-typed. Work in the US has demonstrated that similar trends can also be identified linking certain occupations to minority ethnic groups (Wright and Ellis 1996, 1997). Therefore this research studies the changing levels of occupational concentrations for different ethnic groups within Britain between 1991 and 2001. Further, by examining three different age cohorts, the LS will allow this research to identify any generational differences in the labour market transitions made by different ethnic minority groups in the 10-year period.
In addition to identifying cohort change, the research also intends to establish how patterns of occupational concentration of ethnic groups have changed in different labour market areas within Britain. In this respect the research hopes to overcome existing gaps in the labour market literature, which tends to be blind to group and place-by-place variation (Peck 1996).
After examining the temporal and spatial dynamics of occupational concentration
by ethnic group, the study hopes to move towards a contextual understanding
of the processes shaping the labour market transitions identified. In
line with the spatial approach being taken to study these labour market
dynamics, the research will consider the role of space and place in shaping
any labour market changes. In particular, the role played by place-based
and national economic change (examined using shift-share equations on
the LS data), and social and cultural norms (qualitative work) will be
The inclusion of the ethnic origin question in the 1991 and 2001 census offers an excellent opportunity to study the changing occupational segregation of minority ethnic groups. The LS in particular has a number of advantages pertinent to this study over other datasets, including:
" Large sub-samples of individuals from minority ethnic groups
" Different age-cohorts can be studied over time
" The possibility for the comparison of sub-national areas
" Longitudinal analysis
The first stage of the research identifies national patterns of over- and under-representation within the labour market. The project will use the LS to examine how labour market segregation has changed over the past ten years for different ethnic groups, comparing the trajectories of 3 age cohorts and males and females. Any changes observed will then be related to wider occupation and national employment trends in order to establish how the patterns diverge from more general trends. The main methods used at this stage will be:
" Derivation of a transition matrix
" Location quotients for 1991 and 2001
In the second stage it is anticipated that analysis of the LS will allow an assessment of whether the labour market segregation of different groups was spatially variable between 1991 and 2001 (Large urban areas with sufficient ethnic minority populations are needed for the project, therefore Metropolitan Counties will be used). The same methods will be used as in stage one, however it will be decided whether linked data or cross-sections will be used for this stage once the sample sizes of the national-level analysis have been examined.
- Small numbers may be a problem with the cross-tabulations requested. However, a number of strategies will be employed to overcome this.
- At the national level the cross tabulations split by sex and cohort will be undertaken independently of each other to increase sample sizes.
- At the Metropolitan level it is likely some ethnic groups may be eliminated from the study in some areas if the numbers are too small. (This step is not seen as a problem to the research as (a) the research is interested in pockets of concentration in the labour market and, (b) because there is an uneven spread of certain ethnic groups across Britain)
- Aggregations can be made to increase sample sizes.
Finally, as 13.8 million people at the 2001 Census had one or more items imputed on their records (ONS 2003), a further advantage of the LS is that members with imputations are flagged. This project will compare two versions of each requested output, the first with imputed records and the second without. This will allow an assessment of the extent to which imputed records make a difference to the final results and sample sizes.