5th International Research Conference
on Social Security
Warsaw, 5-7 March 2007
"Social security and the labour market: A mismatch?"
Exploring the effects of integrated benefit systems
and active labour market policies: Evidence from
Jobcentre Plus in the UK
Eleni KARAGIANNAKI
London School of Economics
United Kingdom
International Social Security Association Research Programme
Case postale 1, CH-1211 Geneva 22 Fax: +41 22 799 8509 e-mail: issarc@ilo.org Web: www.issa.int
Exploring the effects of integrated benefit systems and active labour
market policies: Evidence from Jobcentre Plus in the UK*
Eleni Karagiannaki
Centre for Analysis of Social Exclusion, London School of Economics
(a revised version of this paper will appear in the Journal of Social Policy)
Abstract: In April 2002, Jobcentre Plus started to operate in the UK bringing together
the service of the Benefits Agency and the Employment Service. Offering a fully
integrated benefit claiming and work placement/job-seeking service for people of
working age, this new organisation aims to strengthen the link between welfare and
work. In this paper we examine how changes in the level of integration (measured as
the percentage of offices within districts offering the integrated Jobcentre Plus
service) within districts over time affected performance with respect to job entry,
benefit service and customer service delivery. Our analysis suggests that Jobcentre
Plus has a clear positive effect on job entry outcomes for all client groups, a negative
effect on business delivery while it has neither a positive nor a negative effect on
customer service outcomes.
JEL classification: I38
Keywords: Jobcentre Plus, welfare-to-work, non-jobseekers, policy evaluation.
Address of correspondence: Centre for Analysis of Social Exclusion, London
School of Economics, Houghton Street, London WC2A 2AE
* Acknowledgements
Eleni Karagiannaki is a Research Officer at the ESRC Research Centre for
Analysis of Social Exclusion. The author would like to thank Tania Burchardt
for providing advice and guidance for this research and Howard Glennerster for
editing this CASE paper. Any remaining errors and ambiguities are the author's
responsibility.
Introduction
Welfare systems throughout Europe and other OECD countries face significant
pressures (Finn, 2000). The main contributing factor for this situation has been the
increase in the number of people of working age who are dependent on benefits (such
as lone parents and the sick and disabled people). Faced with increases in the number
of benefit recipients and growing costs of their social assistance and unemployment
compensation systems, governments in many countries introduced radical reforms to
increase sustainable levels of employment. A major reform strategy in many countries
has been the introduction of "welfare-to-work" policies (or activation policies as are
alternatively known). In general these policies transform the social security systems
from simply providing income replacement cash benefits towards combining this
form of provision with training, job search and/or work (OECD, 2003). The main aim
of these policies is to create active benefit systems which improve employability,
reinforce work incentives and reduce costs and welfare dependency (Finn, 2000).
Traditionally "activation" policies have been applied to working age unemployed
people (i.e. to those available for and seeking work). However, faced with the
financial pressure caused by the increasing number of inactive benefit claimants,
some countries have introduced policies which expanded the number of groups
targeted by activation policies (Gilbert, 1992; Gilbert, 2002; Lorentzen, and Dahl,
2005; Scarpetta, 1996; Thornton et al., 1997; OECD, 1996, 1998, 2003). Among the
targeted groups are the non-working lone parents, disabled people and other benefit
claimants for whom benefit entitlement is not conditional on availability for work.
Recent developments in Australia, New Zealand and the UK which required claimants
of inactive benefits such as lone parents to attend work-focused interviews can be
viewed as part of such "activation" policies. During these work-focused interviews
possibilities of work are discussed, job search assistance is provided and referrals to
training programmes are made but with no obligation claimants to take-up job offers.
In addition, some governments introduced radical changes to the structure and the
management of public employment services (Clasen et al., 2001; Finn, 2000; Martin,
2000). Such changes involve attempts to integrate the operation of agencies
responsible for job placement and social assistance services into what are described as
"one-stop" or "single gateway" agencies (Finn, 2000). The aim of such attempts is to
improve the effectiveness of labour market programmes by creating an integrated
gateway to benefit and employment services (Finn, 2000; Clasen et al., 2001).
Despite the potential benefits of combining these functions under a unified agency
and extending the activation policies to a wider group of working age claimants, such
attempts carry some potential risks. First, inappropriate pressure may be brought to
bear on lone parents and sick and disabled clients. Second, there is a risk that attempts
at applying activation strategies to inactive benefits claimants will divert the energies
and resources of the Public Employment Services- which are limited, both in terms of
staff and in terms of job vacancies- away from unemployed people as traditionally
defined. Third, organisational changes which involve the integration of employment
assistance and benefit administration services carry the risk that those working in one
or other (or both) of the previously existing agencies will perceive their services as
being "taken over", thereby reducing staff morale and commitment to the service
(Finn et al., 2005).
1
It is the latter two issues which are the focus of this paper, and which are going to be
addressed based on evidence from Jobcentre Plus, the agency which started to operate
in the UK in April 2002 integrating the service of the pre-existing Benefits Agency
and the Employment Service. Being part of the welfare-to-work policy agenda the
main aim was to strengthen the link between welfare and work for a wide range of
working age benefit claimants including the unemployed, lone parents, disabled
people and carers. Reflecting the work-focus of the policy ,all new and repeat
claimants claiming a benefit under the administration of Jobcentre Plus are required to
attend a work-focused interview for their benefit claims to be processed. Unlike the
group of unemployed clients in receipt of Jobseekers Allowance who already have
been part of an intensive work-focused service, the changes introduced by Jobcentre
Plus significantly altered the experience of non-JSA clients such as lone parents and
the sick and disabled.
Prior the creation of Jobcentre Plus, the integration of the benefit claiming and work
placements/job-seeking activities was piloted by an agency named ONE. ONE
operated in 12 areas of Great Britain during the period from June 1999 to October
2001. Building upon experience gained by ONE, in October 2001, 56 Jobcentre Plus
Pathfinder offices started to operate in 17 areas of Britain. The national
implementation of Jobcentre Plus started in April 2002. However, because of the
magnitude of the associated organisational changes the national roll-out of the new
organisation is being implemented gradually over a transitional period which is
planned to end in 2006.
During this transitional period some local offices are fully integrated while for others
functions remain split between pre-existing Social Security Offices and Jobcentres.
Benefit claimants are offered the new service provided that they live in the areas
where integrated Jobcentre Plus offices are operating. For clients not living in such
areas the benefit claiming and job placement services remain split and are provided by
their local Social Security Offices and Jobcentres respectively. Jobcentres in such
areas provide integrated benefit claiming and job-seeking/job placement services for
JSA clients only, while for non-JSA client groups such as the lone parents and the
sick and disabled, benefit services are provided by Social Security Offices. In a
separate development, since April 2001 mandatory LPWFI (Lone Parent WorkFocused
Interviews) were introduced for lone parents claiming Income Support in
both integrated and non-integrated Jobcentre Plus offices (Knight and Lissenburg,
2005).
The combination of benefit and work-placement/job-seeking activities means that the
service that the Jobcentre Plus has to deliver is multi-dimensional. On the one hand it
has to deliver an efficient, accurate and speedy benefit service to all benefit claimants
as previously the Benefits Agency had to deliver, whereas at the same time it has to
deliver an efficient work-focused service, increasing job entries and the labour market
attachment ­ traditionally an objective of the Employment Service. It is obvious that
in such a multi-tasking agency the service delivery depends on the availability of
resources and the importance attached to different elements of the Jobcentre Plus
service.
2
Our aim in this paper is to examine the performance of Jobcentre Plus with respect to
different elements of service delivery. In particular we seek to examine its
performance with respect to three different elements of service:
i) job entry outcomes: measuring performance with respect to job entry
ii) customer service: measuring performance with respect to the general
standards of service that is provided to customers,
iii) benefit service: measuring the speed and accuracy of the benefit claiming
process. In the next section we discuss in detail what each of these
indicators is measuring.
The hypotheses that we want to test are the following:
i) Has the integration of benefit claiming and job-seeking activities under a
unified agency improved job entry outcomes for a wider group of
claimants i.e. those traditionally classed as inactive? If yes, were there any
negative effects with respect to job entry outcomes for JSA that would
suggest that the extension of activation policies to inactive benefit
claimants diverted energies and resources away from unemployed people
as traditionally defined?
ii) Is there any evidence that some elements of service have improved at the
expense of others?
The latter hypothesis relates to concerns that have been raised by ex-BA staff that the
Jobcentre Plus performance targets put an explicit priority to "front office" in contrast
with the treatment of the "back office" function of benefit processing (Johnson, 2003;
Finn et al., 2005). The prioritisation of "front office" under the new organisation to a
large extent reflects the high importance attached to the delivery of an efficient workfocused
service. In the context of the three elements of service that we examine here,
a prioritisation of work-focused service might increase job entry outcomes whereas it
could have negative effects on benefit service delivery. The effects on customer
service depend on the extent to which staff perceive this element of service to be a
prerequisite for the efficient delivery of the work-focused service. Examining
Jobcentre Plus performance with respect to different elements of service can thus
inform us about the extent to which all elements of the Jobcentre Plus service are
effectively accomplished or whether the work-focus of Jobcentre Plus is operating to
the detriment of other elements of the service.
Before presenting the results of our testing we review some of the studies examining
Jobcentre Plus service delivery, we describe briefly the targets upon which the
performance of Jobcentre Plus is measured and we give details on the data and the
methods used in our analysis.
Evidence on Jobcentre Plus service delivery
The Department of Work and Pensions (DWP) has commissioned a series of
evaluation research projects to monitor the extent to which Jobcentre Plus is meeting
its service delivery and labour market objectives. At the time of writing, DWP has
published five research reports based on this evaluation research (Lissenburgh and
Marsh 2003; Davies et al., 2004; Coleman et al., 2004; Corkett et al., 2005; McKenna
et al., 2005). The various elements of Jobcentre Plus service delivery that have been
analysed can be categorised within the following themes: delivery of the work-focus,
customer satisfaction, labour market outcomes, office environment, staff safety and
fraud. Highly relevant to our analysis is the evidence concerning labour market
outcomes and customer service performance presented in Corkett et al. (2005).
3
Comparing the job entry performance of integrated against non-integrated districts,
Corkett et al. reach the conclusion that integrated Jobcentre Plus offices performed
better in terms of the total number of job entries for disabled people. However they do
not find any difference between integrated and non-integrated districts for lone
parents or for jobseekers. They attribute this lack of difference between integrated and
non-integrated districts to the fact that for these client groups, work-focused
interviews are in place in both integrated and non-integrated districts (for lone parents
the lone parent work-focused interview has rolled out nationally in April 2002 while
jobseekers have always been required to attend a work-focused interview as a
condition for the benefit claim). Following a similar approach Corkett et al. find that
integration had a negative effect on customer service. This negative effect however
was mainly attributable to the negative transitional effects that were present during the
rollout period, with performance recovering a few months after the initial rollout.
Similarly to Corkett et al., we use differences in integrated status in order to examine
the effects of the Jobcentre Plus service. However, the approach that we followed to
address our research questions is very different. Corkett et al. define an integrated
district as any district in which at least one office is integrated. In this paper, rather
than focusing on differences between integrated and non-integrated districts, we use
the actual level of integration (defined in terms of the percentage of offices within
districts operating as integrated Jobcentre Plus offices) that each district has achieved
at each point in time. In addition we examine all three elements of the Jobcentre Plus
service delivery (job entry, customer service and business delivery) providing an
overall picture of the service delivery. Finally, recognising that the national
implementation of Jobcentre Plus can be a disruptive process, we present results
isolating possible negative transitional effects resulting from the implementation of
Jobcentre Plus.
The Jobcentre Plus targets
Since 1997, when New Labour came to power the welfare state in the UK has
undergone significant reforms. These reforms had an explicit welfare-to-work
orientation and were designed to promote employment, foster inclusion and reduce
poverty (Taylor-Gooby et al., 2004; Walker and Wiseman, 2003 outline and discuss
the effectiveness of various welfare reforms introduced under new Labour). Jobcentre
Plus has been a key element of the reform of the welfare state in the UK. Reflecting
the ideal "work for those who can, security for those who cannot" (DSS, 1998),
which has been central to the New Labour reform of the welfare system, the
objectives of the new organisation as they are outlined in the 2002/03 Annual Report
and Accounts (Jobcentre Plus, 2003), are the following:
* to put more benefit recipients in touch with the labour market through the
intervention of their Personal Adviser.
* to increase sustainable levels of employment by getting more benefit recipients
into work.
* to ensure that people experience effective and efficient service that is tailored to
their personal needs.
* to change the culture of the benefit system and the general public towards
independence and work rather than payments and financial dependence.
4
In order to promote the successful delivery of its objectives, a series of targets have
been set out covering several elements of the Jobcentre Plus service. The five
components are: Job Entry, Customer Service, Employer Outcome, Business Delivery
and Monetary Value of Fraud and Error. Each of these targets is briefly outlined
below:
The Job Entry Target: According to 2003/04 Jobcentre Plus Business Plan each
district should achieve a total points score of 7,681,000 based on the job entry
outcomes that it achieves. The job entry target is based on a point system that signals
priority groups (see Table 1). Each client group and employment programme was
allocated a point score with highest weight attached to job entries for clients from
priority groups and programmes. Client groups with highest weights are jobless lone
parents including people on the New Deal for Lone Parents, those on New Deal for
Disabled People, people with disabilities in receipt of a specified primary benefit and
other people in receipt of a specified benefit. There are additional scores for:
* job entries achieved for unemployed customers who live in disadvantaged areas,
defined on the basis of a high proportion of ethnic minorities or the poorest labour
market status and low income (2 additional points)
* Jobseeker Allowance clients who remain off benefit 4 weeks after starting the job
(1 additional point) (Johnson et al., 2003).
Table 1: Job entry target point system 2003/04
Priority Client Group 1: Job entry point score 12
Jobless Lone Parents including people on the New Deal for Lone Parents
Those on the New Deal for Disabled People
People in receipt of a specified primary benefit (Income Support, Incapacity
Benefit, Severe Disability Allowance, Invalidity Care Allowance and Bereavement
Benefit)
Priority Client Group 2: Job entry point score 8
People on the New Deal 50 plus
People on the New Deal 25 plus
Those on the New Deal For Young People
Employment Zones
Other people with Disabilities not included in Priority Client Group 1
JSA (Jobseeker's Allowance) long term claimants (over 6 months)
Priority Client Group 3:Job entry point score 4
JSA short term claimants (under 6 months)
Priority Client Group 4:Job entry point score 2
Unemployed non claimants
Priority Client Group 5:Job entry point score 1
5
Employed people
Note: Job entry target for 2003/04 was 7,681,000 points (www.jobcentreplus.gov.uk)
The Business Delivery Target: The Business Delivery target was designed to ensure
that Jobcentre Plus undertakes its activities in an efficient and effective manner. This
target consists of five components:
* accurate processing of claims of Income Support (assessed by checking a sample
IS claims),
* accurate processing of claims for Jobseeker's Allowance (measured by checking
a sample of JSA claims),
* accurate and timely processing of Incapacity Benefit Medical Testing decisions
(timeliness is measured through IT systems while accuracy is assessed by
specialised teams for a sample of IB claims)
* booking appointments and holding interviews with clients and following up cases
where clients have failed to attend interviews (measured by reviewing regionally a
sample of cases)
* identification of people in certain client groups who have literacy, language and
numeracy skill needs through a face to face screening process (measured as the
proportion of specific client group that have been screened. In 2003/04 the latter
component of the Business Delivery target changed slightly so that it now
measures attendance of a client at independent assessment following a basic skill
referral. The screening element is implicit to this version of the target as it needs
to happen before a client can be referred for an assessment).
For 2003/04 the target was that the Jobcentre Plus business processes are delivered
efficiently, accurately and to the specified standards in 88 per cent of the cases
checked.
The Monetary Value of Fraud and Error Target: This target was set to reduce
customer fraud, customer mistake and official error. Specialist teams visit each district
3 times a year to examine a number of randomly selected IS and JSA cases (Burgess
et al., 2003). The target for the 2003/04 operational year was to reduce losses from
fraud and error in working age Income Support and Jobseeker Allowance to no more
than 6.9 per cent of the monetary value of these benefits paid during this year.
Employer Outcome Target: The two main indicators making up the employer
outcome measure are:
* resolution, which indicates that the vacancy was filled from any source (making
up 75% of the whole target) and
* responsiveness, which indicates whether the vacancy was filled in a timescale that
met the employers' need (representing 25% of the target).
For 2003/04 the target was to achieve a positive outcome for at least 82 per cent of
employers who place their vacancy with Jobcentre Plus.
Customer Service Target: The Customer Service target was designed to measure the
performance of Jobcentre Plus in meeting the standards and commitments outlined in
the Jobcentre Plus Customer and Employer Charters. The four elements upon which it
is measured are the following:
* speed - how quickly staff answer telephone, greet customers and deal with
customers on the telephone and face to face
* accuracy- the accuracy of information given over the telephone and face to face,
6
ˇ proactivity- how well the organisation understands customers' requests,
anticipates their needs and how successfully services are tailored to meet their
individual needs,
* environment- the quality, facilities, and accessibility of the premises.
This target is divided into two areas: service to clients (assessed in terms of all four
elements outlined above) and service to employers (assessed in terms of speed,
accuracy and proactivity). Service to clients accounts for 75 percent of the whole
target while service to employers accounts for 25 percent of the target. For the client
service component performance is assessed by independent research companies and is
measured via a mystery shopping approach (Burgess et al., 2003). For the employers'
service component performance is assessed by independent contractors through an
employer telephone questionnaire (Burgess et al., 2003). For 2003/04 the target was to
achieve an 83 per cent customer service level in the delivery of the standards set out
in the Customers' and Employers' charters.
Data
Each year Jobcentre Plus publishes data on performance indicators for each of the
above listed targets. For our evaluation we use district level performance indicators on
job entry, customer service and business service for 2002/03 and 2003/04.
Job entry performance indicators are published on a quarterly basis and are used to
evaluate the effect of the integrated Jobcentre Plus service on labour market
outcomes. Two types of data are offered to describe performance with respect to job
entry:
* the total job entry points score which is an aggregate indicator of performance
derived using different weights for job entries achieved for different client groups
(according to the prioritisation presented in Table 1) and
* the actual number of job entries for each client group.
We use the absolute number of job entries that was achieved for each client group
(unemployed, lone parents and disabled people) as a percentage of the total number of
clients that comprised each group. There are two main reasons for choosing job
entries as a percentage of the relevant customer base rather than using total job entry
point score. Firstly, using the percentage of job entries for each client group allows us
to examine whether Jobcentre Plus had a different impact for different client groups
(something that would not be feasible looking at job entry point score). Secondly, job
entry point score does not take into account differences in the customer base across
districts or within districts over time. Thus, relying on job entry point score we would
probably pick up differences in customer base among districts or within districts over
time.
Performance indicators with respect to customer service and business delivery are
used to examine our second hypothesis. Benefit service delivery is examined
exploiting information on performance for three of the components of the business
delivery target i) accurate processing of claims for Income Support, ii) accurate
processing of claims for Jobseeker's Allowance and iii) accurate and timely
processing of Incapacity Benefit Medical Testing decisions.1
For customer service we
1
Performance data for the job entry and customer service targets are available from the Jobcentre Plus
website. Data on business delivery and its components were kindly provided by Jobcentre Plus
Research Division Directorate.
7
use an aggregate indicator summarising the performance with respect to all
components of the Customer Service target. Similarly to job entry, customer service
performance data are available on a quarterly basis, while the three elements of
business delivery target are available on a four-monthly basis.
Methodology
As discussed earlier the Jobcentre Plus national rollout has been a gradual process
which was planned to be implemented in several stages. In April 2002, when
Jobcentre Plus started to operate, only a limited number of offices offered a fully
integrated work placement and benefit claiming service. Among the 90 districts only
the 17 Pathfinder districts had any integrated Jobcentre Plus office in April 2002.2
The
second stage of the national implementation of Jobcentre Plus was planned to be
implemented between October 2002 and March 2003 with the rollout of new
integrated Jobcentre Plus offices across twenty-four districts (some of which included
Pathfinder offices). The third (2003/04) stage of the roll-out began in April 2003. The
planning horizon is that by the end of 2005/06, new Jobcentre Plus offices will
operate in all districts. In the meantime, services continue to be provided by local
Social Security Offices and Jobcentres, which form part of the Jobcentre Plus network
but otherwise continue to provide the same service that they provided before the
introduction of Jobcentre Plus.
Figure 1: Percentage of districts by degree of integration
Proportionofdistricts
.25
.5
.75
1
0-25% Integrated offices 25-50% Integrated offices
50-75% Integrated offices 75-100% Integrated offices
2002/03-Q1
2002/03-Q2
2002/03-Q3
2002/03-Q4
2003/04-Q1
2003/04-Q2
2003/04-Q3
2003/04-Q4
To provide a graphical description of the integration process Figure 1 shows the
percentage of districts with different levels of integration for each quarter from April
2
The average number of offices within districts is about 15 offices with a minimum of 5 and a
maximum of 42 offices.
8
2002 to March 2004.3
As it is evident from this figure in the first quarter of 2002/03,
the majority of districts had only very few offices offering the integrated Jobcentre
Plus services. By the end of the observation period (fourth quarter of 2003/04) there
had been significant increases in the percentage of districts with a higher proportion
of integrated Jobcentre Plus offices. However, even then a high proportion of districts
had a relatively low level of integration (in about 49 percent of districts less than 25
percent of offices operate as integrated Jobcentre Plus offices).
The gradual roll-out of offices offering the new integrated service can be used as a
natural experiment to test the effect of integration on the performance outcomes.
Ideally in order to test the impact of integration upon job entries, customer service and
benefit service delivery we would need information on how changes in the integrated
status have affected the performance of individual offices. Unfortunately data
providing information on performance of individual offices are not publicly available.
Our testing is therefore based on a series of regressions on a dataset that includes
quarterly (four monthly) district level performance statistics. We include a variable
measuring the districts' integration level at the beginning of each quarter (four month
period), measured as the percentage of offices within district offering the integrated
Jobcentre Plus services. In addition we include a dummy variable indicating
Pathfinder districts. Given that Pathfinder districts first piloted the Jobcentre Plus
service we might expect that they would adjust earlier than other districts to changes.
For each of the performance indicators that we examine here we present results from
two models. The first is a pooled ordinary least squares (OLS) model which treats
each observation as an independent observation. The second model is a fixed effects
model which accounts for the fact that there are multiple observations per district. The
latter model analyses differences within a given district over time. Its main advantage
is that it differences out any fixed district characteristics across the time periods and
thus eliminates the heterogeneity bias, which may result if there are unobserved
district specific effects that may affect both the degree of integration and performance
(e.g. having an enthusiastic district manager). This may be of particular importance if
the selection of districts for early integration has not been random.
These are very simple models, focusing attention on how differences in degree of
integration can account for differences in the performance among districts or within
districts over time. Due to their simple nature these models explain a small proportion
of the overall variation. More sophisticated models, using more explanatory variables
(such as local labour market conditions, and measures of staffing levels) would
increase their explanatory power, as indicated by the R-squared values. Unfortunately,
labour market data are not available at Jobcentre Plus district level. Nevertheless,
these simple models provide useful insights for the hypotheses that we seek to test
here.
Our approach has several advantages over the one followed by Corkett et al. (2005).
Firstly, whereas Corkett et al. examine differences in the performance of integrated
and non-integrated districts (defined in term of whether there are any integrated
3
The level of integration of districts is defined as the proportion of offices within districts that have
been rolled-out as integrated offices by the end of each quarter. To construct this measure we used
information on dates that offices have been rolled out as integrated offices and the total number of
offices within districts. Data were kindly provided by the Department of Work and Pensions.
9
Jobcentre Plus offices within districts) we test the effect of integration taking into
account the level of integration within districts (% of offices operating as integrated
Jobcentre Plus offices). Secondly, our fixed effects models allow us to take account of
unobserved district level characteristics. In particular if those districts selected for a
greater degree of integration early on in the national roll-out period were not typical
of all districts, a pooled OLS regression may over or understate the true impact of
integration on service delivery.
Results
Job entries
We begin our analysis by looking at the effects of the level of integration on job entry
outcomes. The hypothesis that we want to test is whether the integration of benefit
claiming and job-seeking activities under a unified agency has improved job entries.
The OLS and fixed effects estimates for the effect of the degree of integration on the
overall performance with respect to job entry are presented in Table 2.
Table 2: The impact of Jobcentre Plus on job entries outcomes
Pooled OLS Fixed Effect
Integrated status 0.109 0.618***
Pathfinder 0.116
Constant 3.595*** 3.567***
Obs. 720 720
Number of districts 90 90
R-squared 0.005 0.003
Note: *** significant at 1%; ** significant at 5%; * significant at 10%
Examining first the effect of integration using OLS estimation we find that the
integrated status variable has no significant effect on job entries. This result suggests
that districts which are more fully integrated do not achieve higher job entries than the
less fully integrated districts. In contrast to the pooled OLS model, the estimated
coefficient on integrated status from the fixed effect model suggests that the
integrated status has a statistically significant positive effect on the job entry
outcomes: a 1 percentage point increase in the number of integrated Jobcentre Plus
offices within a district increases job entries by about 0.62 percent. The difference in
the effects of integrated status variable captured by the pooled OLS and fixed effect
model is consistent with the idea that there are unobservable differences between
integrated and non-integrated districts. As we discussed earlier, in addition to the
integrated status variable in the regressions we include a variable indicating districts
with Pathfinder offices. Although we might expect that Pathfinder districts would
adjust earlier than other districts to changes given that they first piloted the Jobcentre
Plus service, the pooled OLS estimators does not suggest that Pathfinder districts
perform any better than non-Pathfinder districts with respect to job entry outcomes.
Pathfinder status is time-invariant so it is dropped from the FE estimation.
10
Table 3: The effect of Jobcentre Plus on the job entry outcomes for different clients
groups
Unemployed Lone Parents Sick/disabled
Integrated status 0.22 1.112*** 0.22 1.01*** 0.41*** 0.48***
Pathfinder 0.53* -0.02 -0.04
Constant 9.25*** 9.28*** 3.15*** 3.07*** 0.25*** 0.23***
Obs. 720 720 720 720 720 720
Number of districts 90 90 90 90 90 90
R-squared 0.008 0.005 0.001 0.151 0.146
Note:*** significant at 1%; ** significant at 5%; * significant at 10%
As stressed by Corkett et al. (2005) the effect of integration should be larger for the
group of sick and disabled clients and smaller for the jobseekers and lone parents given
that for the latter two client groups there is no major difference in the
intervention regime between integrated and non integrated offices.4
Our results concerning the effect of integration on the job entry outcomes for different
client groups are presented in Table 3. Similarly to the results presented in Corkett et
al. (2005), the OLS estimates suggest that the effect of integration is significant for
the group of the sick and disabled clients while it is insignificant for jobseekers and
lone parents. However, the fixed effect estimates reveal a somewhat different picture:
the degree of integration has a significant positive effect on the job entry outcomes for
all the three client groups with unemployed clients experiencing the highest increase
in terms of job entry outcomes. For this group the fixed effect estimate on the
integrated status variable suggests that a 1 percentage point increase in the number of
offices within districts offering the integrated Jobcentre Plus services increases job
entries by about 1.11 percent. Although Jobcentre Plus has a significant positive effect
on the job entry outcomes for lone parents and the sick and disabled people, the
magnitude of the increase for these groups is lower than the one we find for
jobseekers: the increase implied by the coefficient on the integrated status variable is
1.01 and 0.48 percent for the client groups of lone parents and the group of the sick
and disabled people respectively. Increasing job entries, albeit with different degree of
effectiveness for different client groups, is a positive outcome signalling that the
integration of benefit claiming, job-seeking and work placement activities under a
single agency is an effective policy for increasing employment for a wider group of
benefit recipients. We do not find any support for the conjecture that the extension of
activation policies to a wider group of inactive benefit claimants would have negative
effects on the job entries of JSA clients. On the contrary, the results of the fixed effect
models suggest that as the percentage of integrated Jobcentre Plus offices within
districts increases the job entries among JSA clients increase as well. Differences
between the pooled OLS and the fixed effect estimates suggest that there are
unobservable district effects that affect both the level integration and performance.
Eliminating these effects using a fixed effect model increases the effect of integration
level on job entries outcomes for all client groups.
4
As discussed earlier the lone parents work-focused interview has rolled out nationally in April 2002
while for jobseekers have always been required to attend a work-focus interview as a requirement for
the benefit claim.
11
Benefit delivery service
Having examined the impact of Jobcentre Plus on job entry outcomes we now turn to
investigate its effects on outcomes relating to the delivery of benefit services. The
hypothesis that we want to test here is whether the emphasis attached to the delivery
of a more efficient work-focused service has been at the expense of benefit service
delivery. Table 4 presents results on the effect of integrated status on the three
components of the business delivery target that relate to the accuracy of three
different benefit types: JSA, IS, IB (the main benefits for the unemployed, lone
parents and the disabled people respectively).
Table 4: The impact of Jobcentre Plus on business delivery
JSA accuracy IS accuracy IB Accuracy
OLS Fixed
effect
OLS Fixed
effect
OLS Fixed
effect
Integrated
status
-3.05** -3.37** -2.71** -3.93*** -6.71*** -8.87***
Pathfinder -0.39 1.60*** 1.77**
Constant 92.93*** 92.89*** 91.70*** 92.14*** 98.86*** 99.46***
Obs. 539 539 539 539 485 485
Number of
districts
90 90 90 90 90
R-squared 0.017 0.017 0.015 0.001 0.053 0.041
Note: Missing values for the IB accuracy performance for the first two quarters results in a
significantly lower number of observations for IB accuracy equation.
*** significant at 1%; ** significant at 5%; * significant at 10%.
For all the three benefits under examination results from both the pooled OLS and
fixed effect models suggest that higher degree of integration is associated with lower
accuracy of benefit processing. According to the fixed effects model a one percentage
point increase in the percentage of integrated offices within the district decreases
accuracy performance by 3.37, 3.93 and 8.77 percent for JSA, IS and IB benefits
respectively. This finding supports the hypothesis that the work-focus of the new
organisation is to the detriment of business delivery.
These results concerning the negative impact of integration on business delivery have
to be treated with caution since they might reflect transitional problems caused by
disruption and reorganisation following the integration process. The possible
importance of such transitional effects is reflected in the OLS estimate on the variable
indicating Pathfinder districts. This estimate suggests that Pathfinder districts achieve
a better accuracy performance for IS and IB than do non-Pathfinder districts. In the
next section we examine the importance of such transitional effects more closely by
providing results for the integrated status after eliminating the importance of
transitional effects.
Customer service
The final performance measure which we examine is customer service (Table 5).
12
Table 5: The impact of Jobcentre Plus on customer service
Pooled OLS Fixed Effect
Integrated status -2.101** -3.227***
Pathfinder 1.105**
Constant 83.656*** 83.988***
Obs. 720 720
Number of districts 90 90
R-squared 0.0087 0.0191
Note:*** significant at 1%; ** significant at 5%; * significant at 10%
Given that good customer service is a precondition for the delivery of an efficient
work-focused service we might except that the work-focus would have positive spillover
effects on customer service. Despite its importance, results from both OLS and
the fixed effect models suggest that integrated status has a significant negative effect
on customer service. When we eliminate the unobserved district effects a 1
percentage point increase in the number of offices within districts offering the
integrated Jobcentre Plus services is found to decrease the customer service
performance indicator by 3.23 percent. Unlike the results concerning the job entry
outcomes, the results on customer service suggest that Pathfinder districts achieve a
slightly better performance than non-Pathfinder districts lending support to the idea
that transitional effects may be important.
Accounting for the importance of transitional effects
In the analysis of the previous section we have used differences in the percentage of
offices within districts operating as integrated Jobcentre Plus offices in order to
investigate the impact of integration on various performance indicators. Although
differences in the percentage of integrated Jobcentre Plus offices provide important
statistical information to identify the impact of Jobcentre Plus on various performance
indicators it may produce misleading results if there are negative transitional effects
associated with the creation of new integrated Jobcentre Plus offices. Such
transitional effects are to be expected given the large organisational changes that were
necessary in order to deliver the new service.
In order to disentangle any negative transitional effects, in this section we present
results from models estimated after eliminating those observations in which the
degree of integration has changed since the previous observation. Basing our
estimation on those observations for which the integration status is constant for at
least one quarter (or in the case of the benefit service delivery indicators a four
months period) we eliminate the potential negative transitional effects associated with
the change.
Results from the fixed effects model on job entry and customer service outcomes are
presented in Table 6. When we eliminate transitional effects associated with periods
during which new integrated offices have been rolled-out, the coefficient on
integrated status for job entry outcomes increases to 0.99 (compared to 0.618). The
increase in the effect of integrated status compared to the model that did not account
for transitional effects indicates that there are indeed negative transitional effects. In
13
contrast to the results presented in Table 4, which indicated that degree of integration
had a significant negative effect on customer service, the second column of Table 5
suggests that after eliminating transitional effects the impact of integrated status
variable on customer service is insignificant.
Table 6: Fixed effect models estimating the impact of Jobcentre Plus on job entries
and customer service and benefit service delivery after eliminating transitional effects
Job entry Customer
Service
JSA
accuracy
IS
accuracy
IB Accuracy
Integrated status 0.987*** 0.774 -4.96** -6.76*** -7.39***
Constant
3.576***
83.872***
93.11*** 92.13*** 99.33***
Obs. 657 657 474 474 420
Number of
districts
90 90 90 90 90
R-squared 0.014 0.003 0.005 0.004 0.020
Note: *** significant at 1%; ** significant at 5%; * significant at 10%
On the other hand, for the benefit service delivery the results presented in Table 6
suggest that even when we isolate transitional effects, the degree of integration has a
significant negative effect on performance with respect to each of the three benefits
under consideration. Thus, for the benefit service delivery, the negative effect of
integrated status does not seem to be a transitional but rather reflect more permanent
features of the system.
Discussion
The importance of integrated management of benefit systems and active labour
market policies under an agency which combines the three core functions of job
placement, benefit payments and placing participants on active labour market
programmes has been widely stressed by international organisations. It is argued that
combining these functions can extend the activation policies to a wider group of
working age claimants who otherwise will be difficult to reach. Despite the potential
benefits of combining benefit and work placement activities, such attempts carry the
risk that the emphasis put on the delivery of a work-focused service will have
negative effects on the delivery of the benefit service.
In this paper we evaluate the impact of integrating the benefit and job seeking
activities under a unified agency on various aspects of service delivery, based on
evidence from the Jobcentre Plus which started to operate in the UK in April 2002.
Due to the magnitude of the required organisational change, the rolling-out of new
Jobcentre Plus offices is being implemented gradually. During a transitional period
which is planned to end by 2006, services are provided by new integrated Jobcentre
Plus offices as well as by pre-existing local Social Security Offices and Jobcentres,
which continue to provide the same service that they provided before the introduction
of Jobcentre Plus. Here we exploited changes in the degree of integration within
districts over time in order to evaluate the impact of the integrated Jobcentre Plus
service on three performance measures: job entries, customer service and benefit
service delivery (evaluated in terms of accuracy of three different benefits). Our
14
analysis suggests that after eliminating significant transitional effects, the integrated
Jobcentre Plus service has a clear positive effect on job entry outcomes for all client
groups while it has neither a positive nor a negative effect on customer service
outcomes. In line with the argument that the new integrated service and the workfocus
of the new organisation would have negative effects on elements of the service
relating to the benefit claiming process, we find that the accuracy of processing
claims for the JSA, IS and IB benefits (the principle benefits for unemployed, lone
parents and sick and disabled people respectively), are negatively affected by
Jobcentre Plus.
Overall our analysis suggest that despite the positive labour market outcomes, there is
an unmet need for an adequately resourced "back office" functions, in order for the
benefit service not to be affected negatively by the change and the emphasis placed on
the delivery of an efficient work-focused service. In addition, given the negative
transitional effects that we found for Jobcentre Plus for job entries and customer
service as well as benefit delivery governments and policymakers in other countries
should take special care when designing multi-tasking organisations such as Jobcentre
Plus to guard against the disruption to services associated with major organisational
change.
15
References
Burgess S. C. Propper, M. Ratto and E. Tominey (2003), "Incentives in the public
sector: Evidence from a Government agency" CMPO Working Paper 03/080
Coleman N., N. Rousseau and H. Carpenter (2004), Jobcentre Plus Service Delivery
Survey, IAD Social Research Division Research Report No. 223
Corkett, J., S. Bennett, J. Stafford, M. Frogner and Kim Shrapnell (2005) Jobcentre
Plus evaluation: Summary of evidence, IAD Social Research Division
Research Report No. 252
Davies V., J. Taylor, Y. Hartfree and K. Kellard (2004) Delivering the Jobcentre Plus
vision: Qualitative research with staff and customers (Phase 3), IAD Social
Research Division Research Report No. 222.
DSS (Department of Social Security), (1998), New ambitions for our country: A new
contract for welfare, Cm 3805, London: The Stationery Office.
Clasen J., G. Duncan, T. Eardley, M. Evans, Pascal Ughetto, W. Oorschot and S.
Wright (2001) "Towards `single gateways'? A cross-national review of
changing roles of employment offices in seven countries." Zeitschrift fuer
Internationales und Auslaendisches Sozialrecht, vol. 15(1): 43-63
Finn, D. (2000) "Welfare to work: the local dimension" Journal of European Social
Policy, vol. 10(1): 43-57
Finn, D., M. Knuth, O. Schweer and W. Somerville, (2005) Reinventing Public
Employment Service: The changing role of employment assistance in Britain
and Germany, Anglo-German Foundation for the Study of Industrial Society
Gilbert, N. (1992), `From entitlement to incentives: the changing philosophy of social
protection', International Social Security Review, 45(3): 5-17
Gilbert, N. (2002) Transformation of the Welfare State. The Silent Surrender of
Public Responsibility.Oxford: Oxford University Press.
Jobcentre Plus (2003), Jobcentre Plus Annual report and Accounts.
http://www.jobcentreplus.gov.uk
Johnson S., Jobcentre Plus Performance Targets: A Review of the Evidence 2000-
2002, Policy Research Institute Leeds Metropolitan University, 2003.
G. Knight and S. Lissenburg, 2005, Evaluation of the extension of to Lone Parent
Work-Focused Interviews eligibility: Administrative data analysis, IAD Social
Research Division Research Report No. 237.
Lissenburgh, S. and A. Marsh (2003), Experiencing Jobcentre Plus Pathfinders:
Overview of early evaluation evidence, DWP In-House report No. 111
Lorentzen, T. and E. Dahl (2005), "Active labour market programmes in Norway:are
they helpful for social assistance recipients?" Journal of European Social
Policy, vol. 15:1 pp. 27-45
Martin, P. J. (2000), "What work among active labour market policies: Evidence from
OECD countries' experiences", OECD Economic Studies Paper No. 30
K. McKenna, A. Slater, J. Steels and H. Walton (2005) Delivering the Jobcentre Plus
vision: Qualitative research with staff and customers (phase 4), IAD Social
Research Division Research Report No. 253
Scarpetta, S. (1996), `Assessing the role of labour market policies and institutional
settings on unemployment: a cross-country study', OECD Economic Studies,
26(1): 43­98.
16
Thornton, P., Sainsbury, R. and Barnes, H. (1997) Helping Disabled People to Work:
A Cross-National Study of Social Security and Employment Provisions, SSAC
Research Paper 8, London: The Stationery Office
OECD (1996), The OECD Jobs Strategy: Enhancing the effectiveness of Active
Labour Market Policies, Paris: OECD.
OECD (1998), Key Employment Policy Challenges Faced by OECD Countries,
Labour Market and Social Policy Occasional papers 3, Paris: OECD.
OECD (2003), Benefits and employment, Friend or Foe? Interactions between active
and passive social programmes, in OECD Employment Outlook: 2003
Taylor-Gooby, P. T. Larsen and J. Kananen (2004) "Market Means and Welfare
Ends: The UK Welfare State Experiment" Journal of Social Policy vol. 33(4):
573-592
Walker, R. and M. Wiseman (2003) "Making welfare work: UK activation policies
under New Labour" International Social Security Review, vol. 56: 3-29
17