Geographic variation in claimant unemployment
Karl Ashworth and Salah Merad, Office for National Statistics
[Project number 20089]
The principal focus of the research is to provide evidence for the usability of the claimant count data added to the LS. Following the claimant count beta test, scheduled to end in October 2006, analysis will continue to provide staff development opportunities in the Newport key site specialisms of claimant count data usage and multilevel modelling techniques.
In addition to informing the beta test results and metadata for the CCC database, it is anticipated that the study results will produce outputs for presentation at various conferences and academic articles for publication in peer reviewed journals.
Local labour markets are highly influential in determining labour market supply and demand. The study aims to explore geographic variation in unemployment exit and re-entry rates, controlling for personal characteristics and local labour market conditions, incorporated into the data from JUVOS. It proposes to use event history techniques within a multilevel model framework, so that individuals are nested within travel to work areas (TtWAs), proxying the local labour market, and repeat spells will be nested within individuals to enable estimation of a between person baseline effect. A range of personal characteristics will be drawn from the 1991 and 2001 Census variables to identify characteristics that either facilitate or retard the rate of movement off benefit. Local labour market data, drawn from JUVOS, and area data such as percentage of ethnic composition and so forth, aggregated from the LS census variables, will be used to identify the characteristics of areas that are associated with the rate of movement off benefit. Explorations will be made for evidence of differential impacts of personal characteristics across TtWAs random coefficient model constructions. Changes over time will be investigated by including calendar time dummy variables into the model. Seasonal effects can similarly be incorporated using dummies for time of year.
Further analysis will repeat the above distinguishing different exit routes, e.g. moves to employment rather than simply recording a move off benefit. This analysis will be restricted to post October 1996 (JSA) data because pre-JSA data do not record exit destinations consistently with post JSA data.