The Geography of Employment Polarisation in Britain

By Ioannis Kaplanis

There has been a growth in the number of high-paid and low-paid jobs in Great Britain relative to middle-ranking occupations. This paper examines the geographical pattern of employment polarisation across the British regions - and is the first piece of empirical research to look at this specific issue.

How can job polarisation be explained? Autor et al (2003) argue that technology can substitute for human labour in routine tasks but not in non-routine tasks. Non-routine tasks are found predominantly in high-skill and low-skill, rather than medium-skill, occupations. Building on this, Goos and Manning (2003) argue that the non-routine tasks that cannot be substituted by technology are found increasingly in high-paid 'cognitive' jobs in areas like management, financial services and the creative industries but also in low-paid manual jobs like cleaning and bartending. These theories have been very influential but do not have a specific spatial aspect to them and thus do not lend themselves easily to research on the geographical aspects of employment polarisation.

It is necessary to ask whether London is a special case in terms of having stronger employment polarisation than anywhere else in Britain, or if my analysis just picking up an urban effect. I investigate this issue by breaking down the NES data differently, into metropolitan and non-metropolitan areas, and running the regression separately for each group of areas. The results suggest that London is a special case - I am unable to find any significant evidence of stronger employment polarisation for urban areas as a whole, compared with non-urban areas.

Further analysis of the differences between London and the rest of Britain according to different worker characteristics produces two additional findings. First, for women, there was strong employment polarisation between 1991 and 2001 in London, but no statistically significant increase in polarisation outside London. Second, the growth in low-paid part-time jobs that occurred in the national economy between 1991 and 2001 does not explain the increase in employment polarisation over this time. Employment polarisation is overwhelmingly driven by changes in the composition of full-time job occupations. Overall, employment opportunities in the lowest paid jobs, mainly associated with consumption and leisure-related services offered at the local level to affluent workers, are growing faster in London than in the rest of Britain.

Looking at the earlier period of NES data between 1975 and 1990, there is some evidence that employment polarisation increased over this period for Britain considered as a whole, but not for London. The London labour market seems to have operated very differently during the 1970s and 1980s than during the 1990s. This is an important area for future research and plausible explanations such as increased international migration to London in the 1990s should be investigated further.

(IPPR, July 2007)