Elsevier

Labour Economics

Volume 16, Issue 3, June 2009, Pages 239-250
Labour Economics

You can't always get what you want: The impact of the UK Jobseeker's Allowance

https://doi.org/10.1016/j.labeco.2008.09.005Get rights and content

Abstract

In 1996 the UK made major changes to its welfare system for the support of the unemployed with the introduction of the Jobseeker's Allowance. This tightened the work search requirements needed for eligibility for benefit. It resulted in large flows out of claimant status, but, this paper concludes, not primarily into employment. The movement out of claimant status was largest for those with low levels of search activity. But, this paper finds no evidence of increased job search activity as a result of this change.

Introduction

Although most discussions of the incentive effects of UI systems focus on the level and duration of benefits, many systems of welfare support for the unemployed make receipt of benefit conditional on the individual making efforts to seek work, to be available for work and not to impose unreasonable restrictions on the type of work they accept — what are collectively known as eligibility conditions. Grubb (2000) provides an overview of different OECD countries' regulations of this type and argues that “the enforcement of eligibility criteria may have a larger impact on behaviour than variations in replacement rates…because the income implications for the individual are larger” (Grubb, 2000, p149).

This paper is about the impact of a major change to the UK system of welfare support for the unemployed in October 1996 — the introduction of the Jobseeker's Allowance (JSA). This change had a large number of elements (described in more detail below) but the most important aspect was a stricter enforcement of eligibility conditions. Indeed the name change from the previous Unemployment Benefit suggests the new emphasis on this welfare benefit being an allowance for those who are looking for work instead of an income for those who are unemployed (that they have a right to because of previous social security contributions they have paid).

In the UK, JSA is widely believed to have been a ‘big deal’1. The reason for this can be understood from a few pictures. Fig. 1 presents the time series on the UK claimant count for the period 1984–2004. JSA was introduced at a time when the claimant count was falling but the months following introduction show a markedly higher rate of decline. The fall in the seasonally adjusted claimant count in November 1996 was the highest ever recorded (Sweeney and McMahon, 1998). The reduction in the claimant count seems primarily to have been due to an increased off-flow rather than a reduced on-flow as can be seen from Fig. 2.

It does seem that JSA had an impact on the claimant count but did it also have the other intended effects notably to encourage greater search activity among the unemployed that, presumably, leads to higher flows into employment? There is some a priori evidence that some of those removed from the claimant count did not go into employment. Fig. 1 presents a comparison of the claimant count with the numbers who are unemployed on the ILO definition (looked for work in the past 4 weeks and available to start work within 2 weeks). Prior to JSA these two series tracked each other very closely (though there were always some claimants who were not ILO unemployed and vice versa) but after JSA there is a remarkable divergence that has never disappeared. This suggests that JSA removed some claimants who were and remained unemployed on the ILO definition, perhaps suggesting that the eligibility conditions under JSA were stricter than those implied by the ILO definition of unemployment. This paper investigates these issues in more detail.

The plan of this paper is as follows. The next section discusses the changes to the welfare system associated with the introduction of JSA and discusses the existing evaluation of JSA, arguing that some (though not all) of the conclusions drawn are not really justified on the basis of the analysis done. The third section discusses what economic theory has to say about the impact of stricter eligibility conditions making the simple (but perhaps surprising point) that those who are moved off welfare as a result of these rules may actually choose to reduce their search activity. The fourth section discusses the data used in this paper and reproduces the results obtained in the existing evaluation when using its methodology. The fifth section then proposes a different methodology to get a better estimate of the impact of JSA that avoids the problems identified in the existing evaluation. It shows that JSA did have a sizeable impact on the flows out of claimant status but that this flow was largely, if not entirely, into non-claimant non-employment and that this impact was larger for those with low initial levels of job search activity. The sixth section shows, however that the best estimate of the ‘average’ treatment effect on search activity is very close to zero. The seventh section then investigates whether there is any impact on the distribution of search activity but none is found. The bottom line is that there is no evidence here that JSA had any impact on the behaviour of any of the non-employed even though it did have a sizeable effect on the numbers claiming benefits.

Section snippets

The Jobseeker's Allowance and the existing evaluation

JSA is the current system of welfare for those who are unemployed in the UK. It was introduced on 7 October 1996 replacing the previous system of Unemployment Benefit and Income Support (UB/IS). JSA comes in two forms. Contributory JSA (known as contJSA) is an unemployment insurance (UI) scheme with entitlement based on previous national insurance contributions (as the UK social security payments are known) of limited duration (6 months being the maximum), and not means-tested. This replaced

A simple model of search activity

In this section we present a very simple model of search intensity to develop some predictions about the likely impact of tighter eligibility conditions. The model is a familiar one in the literature (e.g. see Barron and Mellow, 1979, Mortensen, 1986).

There is a distribution of wage offers f(w) that arrive at a certain rate, λu(s) for the non-employed and λe(s) for the employed that can be influenced by the level of search activity, s, that has costs cu(s) for the non-employed and ce(s) for the

Data

The main data set used in this paper is the UK Labour Force Survey (LFS), an address-based quarterly survey of approximately 60,000 households that is the basis for most UK labour market statistics (broadly it is equivalent to the US Current Population Survey though it is much more detailed). Each sampled address is followed for 5 quarters so there is a limited panel element. As this is a representative national sample no-one is going to go missing and the impact of JSA is large enough to be

Methodology

The methodology for estimating the impact of JSA is the following. For a ‘treatment’ group I use claimants in the period July to September 1996 inclusive. When first observed (what will be referred to as wave 1) these individuals will be subject to pre-JSA rules and their eligibility for benefits will be defined by the pre-JSA rules. But, when they are next observed 3 months later (what will be called wave 2) they will be subject to JSA rules. Of course, any change in labour market outcomes or

The impact of JSA on job search activity: average treatment effects

The previous section demonstrated that the impact of JSA was less for individuals who had been searching intensively and using a large number of different search methods. Given this one might expect that the incentives to search were increased by JSA (as was its intention) and one should be able to see this in the reported levels of search activity.

Table 5 investigates this using the two different measures of search activity. The first row reports the results where the measure of search

The impact of JSA on job search activity: the distribution of treatment effects

This section investigates whether there is any impact of JSA on the distribution of job search activity. The theory presented in the third section suggested that there might be both an increase in the numbers reporting high levels of search activity and a low level of search activity. This obviously has the ability to explain why the average treatment effect is essentially zero at the same time as JSA still had an impact on search activity.

We take two approaches to detecting any distributional

Conclusion

The introduction of JSA in the UK was a big deal — it seems to have reduced the claimant count by about 8 percentage points. The impact was larger for those with low levels of search activity. This obviously resulted in savings in the payment of welfare benefits but it was also intended that the change increase the search activity of jobseekers and, hence, raise inflows into employment. This paper has found no evidence that moves into employment or measures of search activity were increased by

References (32)

  • AshenfelterOrley et al.

    Do unemployment insurance recipients actively seek work? Evidence from randomized trials in four U.S. states

    Journal of Econometrics

    (2005)
  • BlundellRichard et al.

    Unemployment, discouraged workers and female labour supply

    Research in Economics

    (1998)
  • MortensenDale T.

    Job search and labor market analysis

  • Abbring, Jaap H., van den Berg, Gerard J. and van Ours, Jan (1997) “The Effect of Unemployment Insurance Sanctions on...
  • BarronJohn M. et al.

    Search effort in the labor market

    Journal of Human Resources

    (1979)
  • BlackDan A. et al.

    Is the threat of reemployment services more effective than the services themselves? Evidence from random assignment in the UI system

    American Economic Review

    (2003)
  • BlundellRichard et al.

    Unemployment and female labour supply

    Economic Journal

    (1987)
  • Boone, Jan and van Ours, Jan (2000) “Modeling Financial Incentives to Get Unemployed Back to Work”, IZA Discussion...
  • Boone, Jan, Fredriksson, Peter, Holmlund, Bertil and van Ours, Jan (2001) “Optimal Unemployment Insurance with...
  • BurdettKenneth et al.

    Earnings, unemployment, and the allocation of time over time

    Review of Economic Studies

    (1984)
  • BurdettKenneth et al.

    Labor supply under uncertainty

  • DoltonPeter et al.

    The impact of restart on reservation wages and long-term unemployment

    Oxford Bulletin of Economics and Statistics

    (1995)
  • DoltonPeter et al.

    unemployment duration and the restart effect: some experimental evidence

    Economic Journal

    (1996)
  • DoltonPeter et al.

    The long-run effects of unemployment monitoring and work-search programs: experimental evidence from the United Kingdom

    Journal of Labor Economics

    (2002)
  • DonaldStephen G. et al.

    Inference with difference-in-differences and other panel data

    Review of Economics and Statistics

    (2007)
  • FlinnChristopher J. et al.

    Are unemployment and out of the labor force behaviorally distinct labor force states?

    Journal of Labor Economics

    (1983)
  • Cited by (47)

    • Risk aversion and job mobility

      2019, Journal of Economic Behavior and Organization
      Citation Excerpt :

      Although the LISS data does not contain information about acceptance or rejection of job offers, there is information available about job search activities of the respondents. In the previous literature, job search effort has been measured by the time spent on search activities (Krueger and Mueller, 2010), the number of applications in the past month(s) (van der Klaauw and van Vuuren, 2010), the number of job search channels (Manning, 2009) or a combination of some of these indicators (Bloemen, 2005; van Huizen and Alessie, 2015). Here we use three indicators of on-the-job search intensity that are available in the LISS: whether the worker applied for another job in the past two months; the number of applications in the past two months; and the number of job search channels the worker has used in the past two months.

    • Job search requirements, effort provision and labor market outcomes

      2019, Journal of Public Economics
      Citation Excerpt :

      It is interesting to compare this result to studies evaluating the introduction of a job search monitoring regime (i.e., the extensive margin). These studies tend to find negative effects on the quality of accepted jobs (e.g. Petrongolo, 2009; Manning, 2009). In turn, our results show that small intensive margin increases in the requirement level do not induce substantial additional harm to employment quality.

    • Can't work or won't work: Quasi-experimental evidence on work search requirements for single parents

      2018, Labour Economics
      Citation Excerpt :

      These findings are consistent with a simple search model that predicts that some claimants might find the search requirements too burdensome and give up search entirely. As Manning (2009) shows, this is likely to be the case for individuals with initial low level of search, for whom the marginal cost of the extra search effort might exceed the expected benefit. And there are plausible reasons to think that single parents might be disproportionately found in this group.

    • Demanding Activation

      2017, Journal of Social Policy
    View all citing articles on Scopus

    I would like to thank Olivier Marie for his help in pointing me to data sources for this project and to seminar participants at CEP, MIT, Berkeley, McMaster, the Bank of England, Policy Studies Institute and referees and Katherine Abraham for their comments.

    View full text