Benefit sanctions,social citizenship and the economy

Michael Adler: Cruel, Inhuman or Degrading Treatment? Benefit Sanctions in the UK, Cham, Switzerland, Palgrave Socio-Legal Studies, 2018.

Beth Watts and Suzanne Fitzpatrick: Welfare Conditionality, Abingdon, Routledge, 2018

Welfare Conditionality Project 2013–2018: Final Findings Report, York, University of York Department of Social Policy and Social Work, 2018, at www.welfareconditionality.ac.uk/wp-content/uploads/2018/06/40475_Welfare-Conditionality_Report_complete-v3.pdf (accessed 15 May 2019).

It is often said that the UK National Insurance system has always had conditionality and associated benefit sanctions. But although strictly true, this is misleading. For most of the time since its start in 1913 the system has had comparatively little emphasis on benefit conditions or on sanctions (previously called disqualifications) for breaching them. Such conditions as there were arose from insurance principles. There was a notorious campaign in the later 1920s to enforce a requirement on the unemployed to be ‘genuinely seeking work’, but after an outcry, an official inquiry and a Labour Party backbench revolt, the requirement was abolished in 1930. Thereafter there were not many disqualifications – typically a few tens of thousands per year – except for giving up a job voluntarily or being sacked without a good reason. The latter meant a delay to the start of benefit, typically of four weeks, on top of the traditional three ‘waiting days’.

The lack of emphasis on conditionality reflected the public attitudes of the time; in 1937 Clement Attlee could write

I can well remember the time when it was assumed that everyone unemployed, was so through his own fault…. to-day unemployment is realised to be in the majority of cases a misfortune due to the maladjustment of the economic machine instead of a failure of character…

The dominance of Keynesianism after 1945, and the actuality of full employment until the 1970s, ensured that there was rarely any political focus on conditionality and that the exceptions, such as the Wilson government's ‘four week rule’ of 1968 affected only relatively small numbers of people (Meacher, 1974). But over the past three decades, in the UK and across the developed world, there has been a dramatic shift to increased conditionality, accompanied by increased harshness in the penalties. This has been linked to a decline in commitment to social insurance as a collective protection against loss of income caused by unemployment or sickness, and a rise in the belief that social security (now increasingly called ‘welfare’) should be used as a means of coercion for unemployed or workless people who are thought to need to be ‘activated’. Along with these changes has gone a hardening of public attitudes towards the unemployed.

In the case of the UK, the process began under the Thatcher government in 1986 with the introduction of Restart interviews and an increase in the maximum length of an unemployment benefit disqualification from 6 to 13 weeks. It continued under the Conservatives with further lengthening of the maximum penalty to 26 weeks; reintroduction of detailed job search requirements (‘actively seeking employment’); replacement for disqualified claimants of the previous entitlement to reduced Supplementary Benefits on the normal criteria by discretionary ‘hardship payments’, harshly assessed; introduction of a two-week wait before even these reduced payments could be claimed; use (from 1994) of administrative targets for referrals for disqualification or sanction; and introduction of Jobseekers’ Agreements. The Labour Party fought these innovations in opposition. But in office from 1997 it not only kept all of them, but after scrapping independent adjudication on sanction/disqualification referrals through the Social Security Act 1998 (which had already been drafted by the Conservatives), it went on to increase job search requirements, extend work-focused interviews to all claimants, expand conditionality and sanctions to cover sick and disabled people and lone parents of school age children, and increase the effective length of the sanctions for a missed interview from a few days to two weeks. Blairite rhetoric on ‘rights and responsibilities’ also seriously undermined the insurance principle, by promoting the idea that benefits are a gift from taxpayers to claimants, rather than an entitlement which people pay for when they are in work (Hills, 2015).

The system reached its peak severity under the Cameron governments, with yet further increased job search requirements; the maximum length of sanction increased to three years; repeat sanctions made to run consecutively rather than concurrently; ‘hardship payments’ made repayable; sick or disabled claimants losing the whole of their personal allowance rather than the lesser ‘work-related activity’ component of their benefit; job search and preparation requirements extended to lone parents with below school age children; and introduction of ‘workfare’. Claimants’ access to a Tribunal to challenge sanctions was also drastically reduced through the system of ‘mandatory reconsideration’, and the official language was altered to increase the stigmatisation of claimants. This was done through use of words such as ‘failure’, ‘transgression’ and ‘offence’, and also through abolition of the distinction which previously existed between ‘entitlement’ decisions, i.e. whether a claimant fulfilled insurance conditions, and ‘sanctions’, i.e. penalties for not meeting behavioural requirements subsequent to making a claim, such as attending an interview or training course or applying for a set number of jobs. Henceforth there would only be ‘sanctions’.

Even this recital of changes to the rules (which is not exhaustive) does not tell the whole story, because the actual scale of sanctioning is obviously critical. The shift of the official statistics from paper to digital in 2000 has made it difficult to track the growing scale of sanctions back to 1986 and earlier, but at the cost of some effort it can be done (Webster, 2018). Up to the 1990s only a few tens of thousands of unemployment benefit disqualifications per year were for breaches of behavioural requirements (‘active’ conditions). Most were for ‘voluntary leaving’ or ‘losing a job through misconduct’. But from the late 1980s onwards, sanctions/disqualifications in relation to ‘active’ conditions escalated dramatically. From under 3% of claimants per quarter during the later 1980s, they rose to a typical 5% under New Labour, reaching almost 10% following a sanctions drive under John Hutton from 2006 onwards, and then rising further under the Coalition to reach a peak of 21% in 2013. One quarter (24%) of all those who claimed Jobseekers Allowance (JSA) over the six years 2010–2015 were sanctioned (National Audit Office (NAO), 2016). On reasonable assumptions, this was the highest rate of benefit withdrawal for breach of conditions in the history of UK unemployment insurance. Sanctions on sick or disabled claimants in the Employment and Support Allowance (ESA) ‘Work Related Activity Group’ only began in October 2008, but they too saw a surge under the Coalition, related to non-performance of ‘work related activity’. Sanctions on claimants of JSA and ESA have since fallen, but the very inadequate statistics so far published by the Department for Work and Pensions (DWP) suggest that sanction rates on unemployed claimants of the new Universal Credit (UC) remain at historically high levels (Webster, 2019).

Because benefit conditionality was seen as a relatively minor issue for so long, the academic literature relating to it has been correspondingly thin. The ‘genuinely seeking work’ episode of the 1920s inspired Alan Deacon’s (1976) classic study, which was able to present it as an aberration – an exercise in irrational policy making which at that time appeared unlikely to be repeated. The Conservatives’ introduction of the closely similar ‘actively seeking employment’ requirement in 1989 aroused widespread fears and prompted the Nuffield Foundation to fund Bryson and Jacobs’ (1992) detailed study of the working of the system as it stood in 1990–91. There was a minor upsurge of debate in 1996–97 around the Labour Party’s announced intention to use sanctions in its ‘New Deals’, and the subsequent Labour government itself went on to commission studies by Paul Gregg and David Freud and by Gordon Waddell and Mansel Aylward. But the intensity of the 2010–15 sanctions drive and the severity of its impacts have started to spawn a substantial new literature, with seven official inquiries since 2013 and a large volume of reports from the voluntary sector as well as a number of academic journal articles. Michael Adler’s new book however, which this reviewer had the opportunity to read and comment on in the first as well as the published draft, is the most comprehensive study to appear to date.

While the punitive sanctions regime was being built up, from 1986 to 2012, discussion was dominated by supporters of sanctions, largely those coming at the issue from an economist’s standpoint. Sanctions advocates such as Richard Layard (a major influence on the New Labour approach), Paul Gregg, David Freud, Matthew Oakley and the OECD had little interest in the implications of conditionality for human or citizenship rights, or the rule of law. None of them seems to have given any thought to the problems involved in running what is in effect a huge, secret parallel penal system which, as Adler points out, at its peak in 2013 was levying more fines than the mainstream justice system (and, as a matter of fact, on a more severe scale). They do not seem to have wondered how far the state is entitled to go in substituting its judgment for that of the citizen, nor do they seem to have imagined that there could possibly be abuse of the invasive powers that they urged the state to take on. With their commitment to ‘active labour market policy’ or ‘welfare to work’, their focus was narrowly on the question whether sanctions would get people into employment. Any administrative issues arising could safely be left to the civil service to deal with, and unwanted side effects, if there were any, would be for others to sort out.

The catalogue of administrative negligence and incompetence revealed by the Oakley (2014) report and the many reports of abuse from the voluntary sector gave the lie to this complacent outlook, and Adler’s book now provides a systematic critique of the assumptions behind it. A veteran researcher on administrative justice with an extensive knowledge of the relevant literature, Adler takes up a clear position from the outset on the disastrous implications of benefit sanctions for human rights. His underlying premise is that ‘human rights include the right to a social minimum, i.e. to possess the basic necessities of life that enable people to live autonomous lives and maintain their self-respect’. The current UK sanctions regime breaches this right.

The book is rigorously empirical in its approach. It takes the reader through the main changes in policy since the mid-1990s, deploying the useful device of comparing snapshots of the conditionality regime as it was before 1998 and as it is post-2012, thus avoiding getting bogged down in the intricacies of policy development. It also charts the changes in the scale of sanctions from 2001 to 2016, and in the main reasons for their imposition. Latterly these have been ‘not actively seeking employment’, non-participation in a training or employment scheme, and missing an interview. (In practice these are all misnomers: ‘not actively seeking work’ usually only means not doing precisely what is demanded by the Jobcentre, non-participation in a scheme means missing a single session and missing an interview often means just being late.) For evidence on the adverse impacts of sanctions, Adler is able to draw on case histories compiled by the Dundee-based Scottish Unemployed Workers Network, and on the work of the NAO and of an Oxford University Sociology Department team (Loopstra et al., 2015). The NAO found that while JSA sanctions do make claimants more likely to enter employment, they also drive many people off benefit but not into work, and are likely also to push them into lower paid jobs than they would otherwise have got. The move off benefits but not into work had previously been found by Manning (2009), and job worsening, in terms of both pay and quality, by Petrongolo (2009), Arni et al. (2012) and Van Den Berg and Vikström (2014). The NAO also found that sanctions make ESA claimants less likely to get into work. Loopstra and her team found that for each 100 sanction decisions only 7.4 claimants moved off benefit and into work in the next three months. Meanwhile a Rowntree-sponsored study (Fitzpatrick et al., 2016) found that 30% of destitute people had been sanctioned. A raft of other adverse consequences have been reported, including worsened health, hunger and resort to food banks, relationship breakdown, debt, homelessness and survival crime; the House of Commons Work and Pensions Committee has recently embarked on an inquiry into ‘survival sex’ in which it identifies sanctions as one of the potential drivers of prostitution. ‘Hardship payments’ do not prevent these effects and under UC, because they become repayable, will help even less.

The distinctive contribution of Adler’s book is the assessment of the sanctions regime in the light of the literature on administrative justice, the rule of law and human rights. He points out that the administration of sanctions has shifted away from a juridical model, which emphasises rights, legality and independence of decision-making, towards a bureaucratic and market-oriented model. Unlike court fines, benefit sanctions are imposed before any hearing; they take immediate effect, i.e. no time is allowed to pay; they do not take account of the penalised person’s circumstances; they are disproportionate to the seriousness of the ‘offence’; and they often apply for extended periods. These points and an evaluation against the principles set out by the former Lord Chief Justice Tom Bingham lead Adler to conclude that the sanctions regime is not compatible with the rule of law.

However Adler also shows how depressingly weak the law has in practice been in defending claimants’ rights. The applicability of Article 3 of the European Convention on Human Rights (ECHR) relating to ‘inhuman or degrading treatment or punishment’ has never been tested in the courts. The International Covenant on Economic, Social and Cultural Rights (Article 9 on the right to social security and Article 11 on the right to an adequate standard of living) and the European Social Charter (ESC) (Article 13 (1) on the right to adequate assistance) also provide potential defences against the excesses of the sanctions regime. But although the European Committee on Social Rights (the supervisory body for the ESC) has ruled that sanctions should ‘not deprive the person concerned of his/her means of subsistence’, in practice these provisions cannot be enforced because the UK Parliament, in contrast to the position on the ECHR, has not chosen to make them justiciable. The same applies to the UN Guiding Principles on Extreme Poverty and Human Rights, in spite of the Special Rapporteur’s strenuous criticisms issued at the end of his visit to the UK last November (UN Special Rapporteur on Extreme Poverty and Human Rights, 2018).

The Welfare Conditionality Project was an ESRC-funded collaboration between six UK universities, which ran between 2014 and 2019. The book on Welfare Conditionality by Beth Watts and Suzanne Fitzpatrick started life as the literature review undertaken for the project, and it gives what many people will find is an indispensable account of the concepts, methods and impacts of conditionality, and its theoretical justifications, within a very systematic framework. Its scope is wider than Adler’s, with some consideration of the use of conditionality in health, education and housing services as well as social security, and it is more international. On the other hand, it does not have any descriptive material about the actual UK legal framework, incidence of sanctions or reasons for them, or individual case histories, and it does not discuss the administrative procedures or appeal processes. While also concerned with the ethics of sanctions, it takes a very much more neutral approach than Adler. In this account, Adler’s ‘rights’-based perspective is just one of six possible ethical standpoints, the others being utilitarianism (‘greatest happiness of the greatest number’), contractualism (‘rights bring responsibilities’), communitarianism (a collectivist version of contractualism, influential in the Blairite ‘third way’), paternalism (‘we know what is good for you’) and social justice (‘who deserves what?’). However, it is striking that Watts and Fitzpatrick nevertheless end up in very much the same place as Adler. Having set out the possible ethical criteria, they construct an admirably logical framework for assessing whether existing forms of conditionality can be justified, and reach a very clear conclusion: ‘the recent intensification of sanctions-backed conditionality in UK out-of-work benefits, far less the much longer-standing and extraordinarily harsh regime in some parts of the US, are a very long distance from meeting the systematic tests’.

There has not been a comprehensive empirical study of the impact of benefit sanctions since Peters and Joyce (2006 – fieldwork 2005), which related to the less harsh pre-2012 regime and itself almost certainly gave too rosy a view because the one-third of the sample selected from address records that the researchers failed to contact will have included a disproportionate number of the most disadvantaged claimants, for instance homeless people. DWP is still resisting the idea of any comprehensive study of the post-2012 regime, in spite of repeated calls by the House of Commons Work and Pensions and Public Accounts Committees. Definitive empirical studies of the UK sanctions system require access to the data on individual claimants held by the DWP. But DWP has so far agreed to grant external access only to the University of Glasgow, for a yet-to-be-conducted study of the health impacts of sanctions. The NAO alone has been able to avoid this restriction, because of its special powers. All other researchers have had to use less satisfactory data. The Oxford University team, for instance, have conducted several studies using aggregate data for local authority areas, and this has produced striking results, though not fully conclusive because of the potential problem of ‘ecological fallacy’. Others have had to use self-selected samples obtained via the internet, or purposive samples obtained by standing outside Jobcentres or visiting service providers such as food banks or homeless shelters and recruiting people with the desired characteristics.

The empirical research within the Welfare Conditionality project, reported in its Final Findings Report, falls into this latter category, although it does have the particular advantage, unlike others on this topic, of being a longitudinal study, with a sample of 481 people who were each interviewed three times where possible, at intervals of a year (339 had a second interview and 262 all three). The longitudinal design meant that the effects of conditionality on behaviour, particularly as perceived by the claimants, could be tracked over time. Using a purposive sample also had the advantage that it could be chosen to include representation from each of the main claimant groups of interest – jobseekers, lone parents, disabled people, homeless people, social tenants, families subject to anti-social behaviour orders or family intervention projects, offenders, migrants and UC claimants. The timing was fortunate in that the fieldwork took place between 2014 and 2017, thus catching the later part of the surge in sanctions promoted by the Coalition government. The survey was supplemented by interviews with policy stakeholders and focus groups with frontline welfare practitioners, although DWP vetoed the participation of Jobcentre and Work Programme staff.

The Final Findings Report provides the concrete evidence which is missing from the Watts and Fitzpatrick book. Its conclusions are similarly emphatic. While it does note some counterexamples, it finds that

Benefit sanctions were ineffective in moving people nearer or into paid employment. They were routinely experienced as punitive and more likely to undermine the likelihood of engagement or advancement in paid work. In certain cases the experience of a benefit sanction led to individuals disengaging from the social security system…. It was extremely rare … for a respondent to report that … a sanction helped trigger a movement into paid work …. The most commonly held view was that the balance between sanctions and support was out of kilter.

Adler and the Welfare Conditionality project authors together effectively destroy the case for the current UK conditionality regime, but they do not tackle the economic assumptions behind the drive to benefit sanctions. While there have been dissident voices within the profession, it is mainly economists who have promoted the drive to conditionality, on the basis of what are really quite simple-minded assumptions. The typical economic journal article on benefit sanctions takes a sample of unemployed people at a point in time and analyses whether sanctions increase the proportion in employment, or reduce the proportion on benefit, at some subsequent date. If the desired effect is found, then it is concluded, with or without some caveats, that sanctions are a good thing. But this begs a whole raft of questions.

First of all is the question whether there is really a significant problem at all to which sanctions might be a solution. Demand for labour in the economy goes up and down. When labour demand falls as in the major recessions starting in 1979, 1990 and 2008, there is bound to be a rise in unemployment. But the additional benefit claimants are people who are used to being in work. Why should it be thought that they need ‘activation’? In the 1980s and 1990s a group of British economists, supported by the OECD, constructed a theory of ‘state-dependence’ or ‘hysteresis’ which held that being unemployed itself makes people less ‘employable’, by sapping their motivation and degrading their skills. This was supported by the observation that in the recovery from a recession, the proportion of unemployment which is long term is higher than in the approach to it. This reasoning was undoubtedly a major factor driving increased conditionality at that time. However, it turned out that the whole theoretical edifice had been erected on the basis of a failure to perceive that long-term unemployment always lags short-term unemployment – by around nine months if long-term is defined as over a year – at all stages of the economic cycle (Webster, 2005). More unemployment means more long-term unemployment. But this is not a separate problem requiring ‘activation’ programmes.

There has also been a reluctance among many economists, and in government, to accept that deindustrialisation has produced areas where there is a persistent lack of demand for labour. It has been assumed that migration or commuting ought to resolve any localised problem. Therefore, concentrations of unemployment must be due to the attitudes or behaviour of the unemployed – producing, once again, a rationale for ‘activation’ as well as for the virtual abandonment of regional policy. The acute awareness of the ‘left behind’ produced by the shock of the Brexit referendum has undermined this belief, but it has undoubtedly played a big role in the promotion of conditionality over the last three decades.

In the pro-sanctions literature there is a general assumption that more job search is better than less job search, and that unemployed people need to be pushed into doing more. But job search has costs to the jobseeker, in terms of time, stress, phone charges, transport, paper, postage and loss of opportunity to save money by doing things that would otherwise need to be paid for. It also imposes costs on employers: they have to consider more applicants, who are more likely to be unsuited or unmotivated for their vacancies. The most recent official study of job search (McKay et al., 1999) found that for almost two-fifths of JSA claimants, the amount of search was limited by costs (such as fares and phone charges) and for one quarter, by other factors, such as sickness or unavailability of transport. Despite this, it also found that the average JSA claimant spent 7 hours per week looking for work. Moreover, many studies have shown that the most common way that lower skilled workers find jobs is through word of mouth; state jobseeking requirements do not recognise this route as it cannot easily be documented. Patacchini and Zenou (2006) found that left to themselves, people search more when labour demand and incomes are higher. This makes sense, as the returns to search will then be greater. Chetty (2008) found that people with greater financial resources take longer to find a job, indicating that choosiness pays off. These findings confirm the rationality of jobseekers’ own decisions. British employers have been rather quiet on the subject, but the chief executive of the Council of Small Business of Australia recently said that the requirement for unemployed jobseekers to send out 20 job applications a month was particularly hated by business (Guardian, 8 January 2019). All of this evidence strongly suggests that jobseekers typically choose something like the optimal amount of job search for themselves; the state has no serious evidence that they do too little.

Perhaps the most important question is the quality of the match between the job and the jobseeker. The central function of the state employment service in its original Beveridgean conception was to improve the flow of information about jobs to jobseekers and about jobseekers to employers. This can only improve the matches that are made. But introduction of compulsion changes the position completely. It must result in people accepting jobs which they would otherwise not have done, thus losing the opportunity of a better match later on. Forcing square pegs into round holes is bad for the jobseeker, as Petrongolo, Arni et al. and Van Den Berg and Vikström showed, in lowering their earnings and destroying their human capital. It is bad for the employer, who gets a less motivated worker. And it is bad for the economy.

It has become increasingly clear that the UK is suffering from a serious crisis of low productivity growth. Output per hour grew by 2.2% per year between 2000 and 2007, but since 2010 only by 0.5% per year (Giles, 2019). Benefit conditionality will not be by any means the only cause, but it is also obvious that pushing people into low wage jobs which they are not good at and/or do not want to do is going to reduce productivity, as is undermining people’s health and wellbeing. Acemoglu and Shimer (2000) found that by encouraging workers to seek higher productivity jobs, and encouraging firms to create those jobs, unemployment insurance increases output and productivity. Benefit sanctions undermine this benign role. Because of the high turnover of unemployed claimants – 8.2 m people claimed JSA at some time between 2010 and 2015, even though total unemployed claimants never exceeded 1.6 m – the effect on the economy is potentially large.

Making training compulsory is also highly problematic. Of course it is undoubtedly true that many unemployed, sick or disabled people need support in overcoming barriers to finding a job. But there is no evidence that making these employment services compulsory actually improves their effectiveness. In any other area of life it would be taken as axiomatic that a consumer compelled to take a product will get a worse one, while few people would dispute that the commitment that comes from voluntary choice is vital to the process of learning. Reports from claimants – including those in the Welfare Conditionality sample – regularly comment on the inadequacy or inappropriateness of compulsory training. In any case under the current regime the UK provides little in the form of substantive training. Most of it is simply focused on writing C.V.s and on the process of job application. The Social Mobility Commission (2019) recently reported that as a proportion of GDP, public expenditure on training in Great Britain was among the lowest of the G7 countries between 2004 and 2011 (the latest figures available), with only Japan at comparably low levels; spending in France was some 30 times higher. They also found that disadvantaged adults with the lowest qualifications are the least likely to access adult training despite being the group who would benefit most.

In relation to workless people with disabilities or health issues, UK policy in the past two decades has been driven by the assumption that the main issue has been about claimants’ attitudes and behaviour. The Blair government promoted the ‘biopsychosocial model’ of Waddell and Aylward (2005) in furtherance of this belief. But it is striking how the economic literature has largely ignored the obvious fact that under the neoliberal order employers have become much less tolerant of sickness absences or limited capabilities among their staff. There has been little research on this. But the government-funded Skills and Employment Survey 2017 (Green et al., 2018) reported that Britons are under more pressure at work now than at any time over the past 25 years. Almost a third of respondents said they had to work at very high speeds ‘all’ or ‘almost all’ of the time. More than 45% said the same for the requirement to work very hard and meet tight deadlines. These metrics were substantially higher than in 1992. About 55% of women and 47% of men said they ‘always’ or ‘often’ went home from work exhausted. The proportion of people who have a lot of discretion over how they do their jobs has declined massively, from 62% in 1992 to 38% in 2017. Baumberg (2014) did not look directly at the question of employers’ tolerance of staff disabilities, but he was able to show that people in low-control jobs are more likely to claim incapacity benefits in the following year, and concluded that ‘deteriorating job control seems to be a part of the explanation for rising incapacity, and … high levels of incapacity should be seen as partly a result of the changing nature of work’. The implication is not only that sanctions address the wrong problem where sick or disabled people are concerned, but also that they will have contributed to worsening health through encouraging employers to offer bad conditions.

All of these considerations confirm how weak is the rationale for the current benefit sanctions regime even in its own terms. But in contrast to the labour market economists and ‘welfare to work’ advocates, Adler and Watts and Fitzgerald rightly insist that welfare conditionality raises more than technical questions about the costs and benefits of labour market policies. At the heart of the issues lie questions about values and philosophy. Conditionality and sanctions are essentially based on an assumption that unemployed people are not full citizens suffering a temporary misfortune, but are a different and lesser kind of person who does not deserve to be treated the way the rest of us expect to be treated. Their situation is their own fault; they are takers, not givers; shirkers, not workers; they do not know what is good for them, but need to be told by the state; they do not respond to suggestions made reasonably, so they must be compelled to do so. These beliefs can be shown to be false, but they are not just about facts; they stem from a lack of commitment to equality and social inclusiveness. It is no coincidence that the rise of sanctions has gone step by step with the rise of the neoliberal order. Effective reform of the system therefore depends on a reassertion of the concepts of social citizenship that underlay the development of National Insurance in the 20th century, and on securing their renewed acceptance by a large enough section of the population for reform to be durable.