A review of benefit sanctions

Key points

• Evidence was limited by the narrow range of sanction effects considered in the literature reviewed. Most studies looked only at short-term effects; few considered longer-term impacts or the effects of warnings. None looked at wider effects of sanctions in the system (such as benefit take-up levels).
• Unemployment benefit sanctions in European systems have generally had positive effects on short-term outcomes: reducing unemployment duration and raising employment rates. However, leaving unemployment benefit earlier, prompted by sanctions, can result in poorer quality employment (lower earnings and instability). Sanctions have also impacted unfavourably on crime rates.
• US evaluations have identified short-term effects on benefit take-up; full family sanctions (where the entire grant is suspended) have reduced benefit claims. They have shown mixed sanction effects on employment, and unfavourable impacts on earnings, hardship and outcomes for children.
• Studies with benefit claimants have demonstrated low levels of awareness of sanctions. Although people realised that penalties were part of the system, they rarely knew when penalties could be imposed or how they could be reversed. Disadvantaged claimants facing multiple barriers to work were at higher risk of sanctions.
• Qualitative evidence suggests that, although the threat of sanctions may encourage participation, sanctions themselves do little to change motivation to work.

The research
By Julia Griggs and Martin Evans, Department of Social Policy and Social Work, University of Oxford

Background

The previous Government's rolling programme of welfare reform sought to change the fundamental assumptions of many social security programmes and to promote employment. This approach of increased conditionality has grown incrementally over time and been broadened to include 'new' claimant groups. The move to conditionality has also extended purely work-related issues into a policy model that sees conditionality and sanctions as tools to change other behaviours – for example, Housing Benefit sanctions for anti-social behaviour.

Different types of sanction effect

When considering the impact of benefit sanctions it is important to be aware of the different types of effect that sanctions have (or might have) on claimants. The four key dimensions to these differences are as follows:

1. Type of sanction – whether the sanction results from administrative failure (such as failing to complete required paperwork or attend a meeting), or from a behavioural ‘misdemeanour’ (for example, not ‘actively looking’ for work, or refusing employment).
2. Type of effect – whether the sanction has an impact arising directly from lower levels of benefit entitlement (i.e. fewer people claiming benefits, and lower spending), or from the changed behaviour of claimants.
3. Timing of the effect – there are three points at which sanctions can have an impact:
   a. before the claim – sanction-backed conditions may deter potential claimants;
   b. during the claim, but before the sanction – as a result of the general threat of sanctions or an actual warning;
   c. during the claim, following a sanction – a behavioural change or reaction to being sanctioned.
4. Type of outcome – sanctions have the potential to impact on claimants and their families in many ways. In the short term they may encourage/force compliance or participation (intermediate outcomes), or encourage claimants to end their claim, possibly to enter employment. In the longer term they may affect earnings, material hardship and children’s well-being.

Impacts of sanctions linked to employment-related conditionality

Studies exploring the effectiveness of sanctions have focused almost exclusively on the impacts of sanctions imposed, with a small number also looking at the impact of warnings of sanctions. Effects on take-up and the presence of sanctions on the behaviour of the general claimant population have not been considered, thus limiting the messages that studies of impacts can provide.

Consolidating the findings of Unemployment Benefit (UB) and welfare evaluations indicates that sanctions for employment-related conditions (which in US welfare systems are full family sanctions, where the entire grant is suspended) strongly reduce benefit use and raise exits from benefits. However, they have generally unfavourable effects on longer-term outcomes (earnings over time, child well-being, job quality) and crime rates. Beyond this the findings are harder to reconcile; for example, while European UB programmes tend to demonstrate positive impacts on employment (job entry), this is not always the case for US welfare schemes.

Contextual factors for employment-related sanctions

A number of contextual issues need consideration alongside evidence on the impacts of sanctions. Factors such as claimants’ understanding and awareness of their responsibilities and the consequences of not meeting them limit the potential effectiveness of sanctions. If claimants do not know what is expected of them and what will happen if they fail to meet these expectations, sanctions will 'punish' a lack of awareness rather than deliberate flouting of the rules for receiving benefits.

In addition, evidence suggests that administration of sanctions is not rational and free from bias. US studies have found persistent concerns about bias from race (Schram et al., 2008), as well as considerable evidence of geographical differences in the likelihood of sanctions being applied.

A large body of research exploring the characteristics of sanctioned claimants has demonstrated that those most vulnerable to sanctions are the most disadvantaged. Strong links have been identified among barriers to employment and opportunity – lack of education and work experience, disability and practical constraints, such as lack of transport. Exploration of demographic differences has also shown that young claimants, those with large families and those belonging to black and minority ethnic groups are at increased risk of sanctions.

Qualitative research with claimants has provided little indication of deliberate non-attendance or non-engagement with services or programmes. Failure to attend or participate was more often a product of poor information and non-intentional behaviour such as forgetfulness. Studies have also suggested that, although claimants may be encouraged to attend meetings and participate in activities in order to avoid sanctions, sanctions do little to change motivation or claimants’ attitudes to work.

Other forms of conditionality

The use of conditions and sanctions has not been confined to unemployment and social assistance benefits. Importantly, attempts have been made to make other kinds of benefit payments conditional on certain behaviours – for example, the Sure Start Maternity Grant and (other) Conditional Cash Transfers (CCTs).

Although covering very different programmes (health, education, child support, substance misuse), with very different objectives, the overriding message within each strand of policy is the inconsistency of results.

For example, the US Preschool Immunization Project (PIP) and Primary Prevention Initiative (PPI) programmes impose sanctions on parents who are unable to show that their child has been immunised; but while PIP has had very favourable results on vaccination rates, PPI has had little impact. Evaluations of ‘family cap’ policies, which impose sanctions on those having babies while claiming welfare benefits, show similarly mixed results on claimants’ birth rates. The growing international importance of CCT programmes warrants a note on effectiveness.

Evaluative evidence relating to CCT programmes operating in different South American countries has demonstrated their largely positive impact on child and adult health, school enrolment and attendance, and poverty. However, longer-term benefits have been questioned, such as whether there are beneficial effects of prolonged school attendance without corresponding improvements in attainment.

Justifications for sanctions

Earlier commentary on benefit sanctions and conditionality has tended to focus on underlying conceptual issues of welfare rights and responsibilities. However, some justifications can be based on measurable aims and therefore informed by the evidence: equality, efficiency and effectiveness.

Equality-based justifications

While sanction-backed conditionality ensures that claimants cannot 'opt out' of programmes designed to benefit them, there are clear (though under-researched) effects on benefit take-up. Although all claimants within mandatory programmes are subject to the same work-related activities and have access to the same services, those who are most disadvantaged may be more deterred from entering programmes or more inclined to leave into inactivity or informal work. These same claimants are also more likely to have sanctions applied to them (indicating inequality in the imposition of sanctions). Equal access to programmes and services does not mean equal quality in those provisions, nor does it necessarily lead to equality in outcomes. Indeed, evidence suggests that sanctioned claimants are less likely to enter sustainable employment or make longer-term gains in income.

Efficiency-based justifications

Sanction-backed conditionality is argued to be efficient, in that such an approach is best able to use available resources to maximise positive outcomes by ensuring that claimants are better informed and more realistic about opportunities. This involves managing – in most cases lowering – expectations, and reservation wages (the lowest rate at which a worker is willing to accept a job). Job-search is also ‘optimised’ and ‘deliberate’ job loss is minimised. While sanctions may be efficient in terms of shortening people’s spells of unemployment, taking a longer-term approach to impacts demonstrates the problems of such efficiency arguments, in particular the negative effects of sanctions on job and earnings progression. Inefficiency in administering sanctions is also a problem, with information being poorly communicated to claimants. Furthermore, while cutting take-up of benefits is an efficient way of reducing expenditure, other factors such as spill-over effects on crime rates, along with higher spending on in-work benefits, offset savings.

Effectiveness: optimal models for changing behaviour

The primary purpose of sanctions is to change behaviour. However, they can only function as intended with claimants’ full awareness of the possibility of sanctions and knowledge of how to avoid or reverse them. Crucially, however, qualitative evidence suggests that the majority of claimants only have a limited understanding of the sanctions system.

Conclusion

This review brings into focus the gulf between the rhetoric and evidence for the effects of sanctions in welfare reform. The gulf is not just on evidence, but also in different approaches to preventing poverty and promoting opportunity. In the US, lone parents were targeted for reduced levels of more conditional provision (and the large evidence base resulting from those reforms has dominated this analysis), but with inadequate reflection on the quality and coverage of evidence and without a systematic appreciation of what sanction effects to expect or how to measure them.

The UK, on the other hand, has committed itself to reducing and ultimately eliminating child poverty, and has invested greatly in evidence-based policy-making. This suggests a potentially different approach to ‘welfare reform’ around sanctions – one that takes a more rounded approach to assessing and using evidence. However, to date there has been little indication that this is occurring; policy-makers continue to justify extending sanction-backed conditionality on moral grounds while taking an ambivalent attitude to the evidence. Such ambivalence can be identified in policy (green and white) papers; evidence is marginalised by discussion of principles and what can be expected of claimants in return for benefits.

This review leads to recommendations to:

1. replicate the 2009 Arni, Lalive and van Ours study (‘How effective are unemployment benefit sanctions? Looking beyond unemployment exit’, IZA Discussion Papers, 4509) in the UK to test the effect of sanctions on earnings and sustainability of work;
2. put in place better, more wide-ranging cost-benefit studies of conditionality and sanctions, to look at displacement and spill-over effects;
3. look more closely at the (potential) spill-over effects of conditionality and sanctions on the grey economy and informal work;
4. ensure that longitudinal datasets can capture the changes to conditionality already put in place (for example, by ensuring that the next wave of the Millennium Birth Cohort Survey assesses the effects of new benefit conditions on lone parents according to their youngest child’s age).