Damning proof that the government has no evidence benefits sanctions work


The National Audit Office says the government has failed to measure whether sanctioning benefit claimants represents value for money.

Does anyone remember evidence-based policymaking? For the DWP, it appears from today’s National Audit Office (NAO) report on sanctions, it is at best a dim and distant memory.

When the Department made substantial changes to sanction rules in 2012 – marking a step-change in their scope and severity – it could not quantify the financial impact of the changes, and it said it could not predict whether the changes would create savings. Since then, it has made no attempt to track the actual costs and benefits of the changes.

As one reads through the NAO’s report, it becomes increasingly clear how their task – to assess the value for money of sanctions policy – is thwarted at every turn by lack of evidence. The words "the Department does not know", and mentions of data that the Department does not analyse or collect, recur throughout. The government is evidently operating blind, on an issue that could scarcely be more important: the decision actively to remove from already-poor individuals and households the basic means of their subsistence.

Disturbingly, there are several indications that this ignorance is wilful. The DWP has administrative data on individual benefit histories, sanctions and employment, and data on local sanction rates and performance, but it chooses not to use this body of evidence to evaluate the impacts of sanctions. The government, via the Economic and Social Research Council, has funded a £2m research project from 2013 to 2018 to understand the role and impact of conditionality in social security. In 2015, the DWP advised its Work Programme providers not to take part in focus groups for the project. And, in March 2015, the Work and Pensions Committee called on the DWP to commission "a broad independent review of benefit conditionality and sanctions, to investigate whether sanctions are being applied appropriately, fairly and proportionately". After taking seven months to respond, the Department refused.

 What we do know beyond doubt is that sanctions cause immense hardship to those who are subjected to them. This is in a sense a question of simple logic – take away a person’s primary, meagre source of sustenance, and they will suffer. Indeed, that is the Department’s stated intention: its own guidance to decision makers acknowledges that ‘it would be usual for a normal healthy adult to suffer some deterioration in their health’ if left without income for two weeks (JSA sanctions start at twice this duration). Decision makers assessing potential hardship payments should be looking only at those who would "suffer a greater decline in health than a normal healthy adult" [original emphasis]. But it is also reflected in a range of direct evidence, not least of which is the link between sanctions and food bank use: research by Child Poverty Action Group and others found that between 19 and 29 per cent of visits to the food banks we studied were caused by sanctions.

What evidence there is, the NAO finds, does not suggest an overwhelming case in favour of sanctions. The crutch upon which the DWP, when pushed on its policies over the last few years, has leant so heavily – international evidence on the impact of sanctions – is found by the NAO to be "mixed". This research found that claimants who are subject to sanctions are equally likely to move into employment and to move out of the system altogether – to an unknown destination – while both earnings and hours worked fell compared to those not subject to sanctions. The one UK study that was analysed by the DWP showed no evidence of sanctions increasing claimants’ probability of leaving benefits for work. The NAO’s preliminary analysis of the DWP’s own Work Programme data was consistent with the mixed findings of international studies.

Meanwhile, for Work Programme providers, on average, higher use of sanctions is associated with lower performance in terms of employment outcomes. Though this does not prove causality – it could be, for example, that weak Work Programme providers may use sanctions more because they are ineffective with their other approaches – it may suggest that differences in deterrence effects of sanctions are weaker than other factors explaining performance. Again, no evidence in favour of sanctions here.

The report finds substantial variation in the imposition of sanctions, both across time and between different geographical areas. So what is driving the operation of sanctions policy? The short version is that – again – the DWP doesn’t know, for sanctions administered by Jobcentres at least. Here, it could be that variation is due to differences in the types of claimants in each area, but the DWP has not assessed the causes of the variation, so it has no idea whether the variation is within acceptable limits. For Work Programme sanctions, because claimants are randomly assigned to a provider in their area, differences in referral rates are likely to reflect differences between providers rather than claimants’ behaviour. Taking geographical variation in the rate of sanctions alongside big changes over time – sanctions rose rapidly from 2012 to 2013, and then fell away so that they are now back roughly where they started – the NAO conclude that the imposition of sanctions seem to be driven by management culture at any given time and in any given place, rather than by claimants’ behaviour. In other words, sanctions are, to a greater or lesser extent, arbitrary.

Overall, the picture painted by today’s report is disturbing. Not only do the DWP not know how effective – let alone cost-effective – sanctions are at achieving their aims, they give every indication of not wanting to know. The NAO’s conclusion is damning: "Until the Department can show greater consistency in its use of sanctions and demonstrate that their effectiveness is proportionate to their costs, we cannot conclude that the Department is achieving value for money."

For a policy that is causing unprecedented hardship to hundreds of thousands of people each year, that’s not really good enough.

This blog first appeared in the New Statesman on 30 Nov 2016.