The ins and outs of UK unemployment

Publication type

Conference Paper


BHPS-2009 Conference: the 2009 British Household Panel Survey Research Conference, 9-11 July 2009, Colchester, UK


Publication date

June 1, 2009


What drives fluctuations in employment over the cycle? There is continuing debate as to
the driving factors particularly, so far, among US researchers. The answers should
inform both policy and search-matching models of labour market dynamics. Shimer
(2005, 2007) and Hall (2005) have suggested that cyclical changes in unemployment
are driven by changes in the job-finding rate, and that the separation rate has little role. On
the other hand, Davis, Faberman, Haltiwanger and Rucker (2008) and Fujita and Ramey
(2007) suggest that the separation rate is countercyclical, with particular impact on
changes in unemployment at the start of a downturn.
Three of the first papers to apply these ideas to non-US data are Petrolongo and
Pissarides (2008), who look at the UK, France and Spain; Elsby, Hobijn and Sahin
(2008), who investigate 14 OECD countries; and Gomes (2008) who focuses only on
the UK. Elsby, Hobijn and Sahin (2008) find using LFS data that the split in the
explanatory power of fluctuations in inflows versus outflows in accounting for
unemployment changes is about 20:80 for the UK. Petrolongo and Pissarides (2008) and
Gomes (2008), both using the LFS (and claimant count in the former), also conclude
that the job-finding rate is dominant, except for early in (some) recessions when the
separations rate plays a big role. Recent US research (Davis, Faberman, Haltiwanger,
and Rucker, 2008) has also placed more emphasis on separations.
Yashiv (2007) concludes that the disagreements about key features of labour market
dynamics can only be resolved by further work using micro data. This paper aims to
contribute to our understanding using monthly data on labour market transitions and
related wages from the BHPS. Monthly inflows and outflows into unemployment,
employment and inactivity are calculated using BHPS lifetime history data and recall
data obtained at each annual interview. Unlike some previous research this paper
investigates a three-state model in which flows to and from inactivity can play a critical
role in employment and unemployment dynamics (Yashiv, 2007; Jones and Riddell,
2006). Unusually, the BHPS includes information concerning reason for job loss, which
allows separations to be divided into quits and layoffs, so that their individual
contribution to unemployment changes can be investigated. Employment-toemployment
transitions are also measured.
Following Shimer (2007), changes in unemployment are decomposed into changes
resulting from changing separation and job-finding rates. The technique used varies
depending whether the steady-state unemployment rate approximates actual
unemployment or not (Elsby, Michaels and Solon, 2008). Account is taken of various
aspects of measurement error, including spurious transitions and seam effects as a result
of annual interviews requiring recall of past labour market status (Paull, 2002), and time
aggregation the likelihood that monthly data will understate true transitions occurring
in continuous time. The method of data construction is based on Paull (2003).
Both inflows into and outflows from unemployment show significant cyclical variation.
BHPS data confirm the importance of fluctuations in the hiring rate. But the previous
finding (e.g. Shimer, 2007) that the overall separation rate is relatively constant over the
cycle, and hence plays little part in unemployment fluctuations, is found to hide
important differences between layoffs and quits. Layoffs move strongly
countercyclically (corresponding with results using US CPS data in Elsby, Michaels and
Solon, 2009); quits less so. These findings remain whether or not we adjust data in
various ways, including matching representative employment changes (as in Davis,
Faberman, Haltiwanger and Rucker, 2008), or allowing lags of unemployment change
to affect current unemployment change out of steady-state (as in Elsby, Hobijn and
Sahin, 2008).
The research summarised above forms the first part of a multi-stage project. A further
stage will investigate whether the data confirm rigidity in relevant pay rates.
Countercyclical layoffs imply rigid real wages, while procyclical hires imply flexible
real wages for new employees. In addition, work on BHPS will be repeated with
GSOEP. It is anticipated that comparison of these data relating to labour markets with
different characteristics in terms of pay determination and institutions will elucidate the
roles of the different flows and the influence of wage





not held in Res Lib - bibliographic reference only



Latest findings, new research

Publications search

Search all research by subject and author


Researchers discuss their findings and what they mean for society


Background and context, methods and data, aims and outputs


Conferences, seminars and workshops

Survey methodology

Specialist research, practice and study

Taking the long view

ISER's annual report


Key research themes and areas of interest