Violent bureaucracy: A critical analysis of the British public employment service

Between 2010–2015, the Coalition’s pursuit of a radical austerity programme saw Britain’s Jobcentre Plus experience some of the most punitive reforms and budget cuts in its history. Focusing on the outcomes of these reforms, a growing body of research has found that claiming processes became a more ‘institutionally violent’ and injurious experience for out-of-work benefit claimants. The present article draws upon ideas, developed by Bauman (1989), which focus on the processes that facilitate ‘institutional violence’. We use this framework to analyse ten interviews with front-line workers and managers in public/contractor employment services. In doing so, we expose an array of policy tools and hidden managerial methods used during the Coalition administration which encouraged front-line staff to deliver services in ways that led to a range of harmful outcomes for benefit claimants.


Introduction
In 2010, the Conservative/Liberal Democrat ('Coalition') government outlined a 'radical' austerity programme (Osborne, 2010a). This programme was designed to rescue business from the throes of an enduring North Atlantic Financial Crisis by implementing a series of unprecedented public expenditure cuts and social security reforms which redistributed power/wealth away from working class populations (Cooper and Whyte, 2017). A central feature of austerity was to galvanise political disenchantment with a comprehensive social security system. The Chancellor of the Exchequer, George Osborne (2010b), declared that excessive spending on social security services and cash benefits under the previous New Labour administration had served to undermine national economic performance by 'crowding out' investment. The Prime Minister, David Cameron (2011), declared that social security provision had fuelled a surge in irresponsible behaviour and (under-)class cultural values including voluntary unemployment and an entitlement mentality. As such, the Coalition government pursued 'social security austerity' in two broad directions (Grover, 2019). On the one hand, numerous efforts were made to slash social expenditure on key services/benefits and/or channel it into new investment opportunities by increasingly exposing services to the (quasi-) private market (Finn, 2018). On the other hand, the Coalition government built on the foundations left by their predecessors to carve out a more punitive system of employment service delivery.
These reforms transformed the operational logic driving the day-to-day running of Britain's network of public (Jobcentre Plus) and quasi-private (Work Programme) employment services. For front-line workers, reforms generally entailed delivering services with less resource, more focus on moving people off benefit as quickly as possible, and minimising the costs of fraud and error (Finn, 2018: 226). Meanwhile, for those claiming out-of-work benefits, reforms generally meant undertaking more compulsory work-related duties in exchange for fiscal support far below the income necessary to meet basic needs (Fletcher and Wright, 2018). Failure to perform such duties to an agreed standard would also potentially result in a disqualification from benefit under the rubric of an 'enhanced sanctioning regime' (Adler, 2018).
To legitimise these reforms, concerted efforts were made by politicians and the media to cultivate an increasingly hostile socio-political climate; framing Britain's poor people as a central threat to national security (Tyler, 2020). A plethora of classed and racialized groups (migrant, unemployed, disabled persons) became the prime antagonists of various pejorative discourses emphasising top-down processes of stigmatisation to garner public support for welfare reform. Hostilities of this kind -emerging from above and manifesting on the ground, relationally, in everyday social interplay (cf. Burnett, 2017: 220) -promoted expenditure cuts and punitive social security reforms as the only logical policy response to simultaneously restore competitive economic conditions and manage the perceived threat(s) of capriciously defined social 'others'.
A growing body of research has found that benefit claiming processes became a more physically and psychologically injurious experience for out-ofwork populations under social security austerity. Batty et al.'s (2015) survey of 1,013 homelessness service users in 2015 found the enhanced sanctioning regime disproportionately affected those with poor mental health, drug/alcohol dependency issues and poor literacy; leading to rough sleeping, hunger, exacerbating mental health issues and increasing foodbank usage. After conducting interviews with 481 benefit claimants between 2014 and 2017, the Welfare Conditionality (2018) project concluded that post-2010 reforms increased poverty/destitution, exacerbated ill health, facilitated movements into survival crime and/or encouraged disengagement from the social security system. The Department for Work and Pensions (hereafter DWP) carried out internal peer reviews of 49 benefit claimant deaths between 2012 and 2014, finding 'that in ten cases the claimant had had their benefits sanctioned' (Pring, 2017: 54). Between 2010-2014, statistics show that over 9,000 benefit claimants died and an additional 725,000 anti-depressant prescriptions were administered after they were declared 'fit-to-work' or their benefit was made conditional on participation in work-preparation groups (DWP, 2015;Mills, 2018;Ryan, 2019). A number of these deaths have been 'publicly attributed by family and friends to the removal of benefits by the DWP' (Clifford, 2020: 158).
The present article seeks to explain how ordinary people carrying out their daily duties in employment service offices were able to implement cruel and inhumane social security reforms by drawing upon Bauman's (1989) theory of 'institutional violence'. We contend that it is crucial to situate this behaviour in the context of policy and practice changes which have encouraged the production and delivery of 'institutional violence' on the front-line. Drawing upon interviews with ten front-line workers and managers in public and quasi-private employment services, the authors expose an array of policy tools and hidden managerial methods that were used during the Coalition administration (2010)(2011)(2012)(2013)(2014)(2015); encouraging front-line workers to deliver service in ways which led to a range of harmful outcomes for benefit claimants.

From outcomes to processes of institutional violence: The social production of moral invisibility
'Institutional violence' originated out of Galtung's (1969) germinal efforts to broaden the narrow concept of violence. Rather than understanding violence solely as a phenomenon that occurs interpersonally and involves an exertion of force by an actor who intends to inflict harm on another; Galtung (1969) argues that violence is also impersonal, produced and distributed within the power structures and institutions that govern societies. From this perspective, violence is present in 'avoidable' situations where the monopolisation of insight and/or resources by a group or class excludes another group(s) or class(es) from the resources and/or recognition necessary to realise their capabilities to the same extent. Violence in this sense can be exercised even if there are no concretely identifiable perpetrators directly attacking others -it occurs indirectly and silently (Galtung, 1969). His ideas have since been applied and developed in relation to a plethora of social injustices; ranging from slavery and racism, to poverty and gender inequalities (Farmer, 2004).
Academics have drawn inspiration from Galtung's broader notion of violence in efforts to make sense of 'social security austerity' in Britain. Grover (2019) conceptualises post-2010 social security changes as a form of 'violent proletarianisation'. He concludes that reforms led to a range of avoidable harms and fatalities; primarily by forcing claimant groups off benefit and/or into precarious wage labour in ways that 'socially murdered' some of Britain's most vulnerable individuals through penury and suicide. Relatedly, Wright et al., (2020) view this as a form of 'social abuse'. They conclude that Jobcentres have become more dangerous places, as reforms frequently inflicted material and symbolic forms of suffering which occasionally had life-threatening consequences. Meanwhile, Cooper and Whyte's (2017: 23;: 5) edited collection conceptualises post-2010 reforms as 'institutional violence'. They argue that reforms have not only delivered a range of severe physical and psychological harms, but they were organised and legitimated through licit means. That is, by 'smartly dressed . . . armies of civil servants' and frontline officials sitting behind desks in public-facing bureaucracies.
Whilst the concept of institutional violence has received a number of critiques (cf. Grover, 2019;Wright et al., 2020), for present purposes, the main shortcoming of both Galtung's work and subsequent applications of institutional violence is that authors focus overwhelmingly on outcomes and not on processes (Gupta, 2012: 20). For example, while Cooper and Whyte (2017: 3;2018: 5) profess to 'identify precisely how particular public and private organisations have delivered acute physical and psychological harm'; we contend insufficient attention is paid to the 'how' and that their edited collection instead proceeds to focus predominantly on the many violent outcomes fuelled by austerity driven policies (e.g. 'physical', 'psychological', 'symbolic', 'epistemic'). Consequently, there is insufficient focus on the specific processes which actively encourage smartly dressed civil servants in public-facing bureaucracies to perpetrate institutionally mediated acts of violence.
To address this shortcoming, we draw inspiration from Bauman's (1989) exposition of the psycho-social processes which facilitate institutional violence. While Bauman (1989) focused on a very unique case study -the Nazi genocide of European Jewish people between 1941 and 1945 -he asserts that the 'social mechanisms' which brought about the Holocaust are 'also set in motion under contemporary conditions' and hence contain 'crucial information about the society of which we are members' (Bauman, 1989: 95;xiv). As such, though we in no way seek to draw moral or historical equivalences between the Holocaust and the present analysis, we contend that Bauman's analysis possesses a number of key insights that can be excavated and recalibrated towards original, critical interpretation of employment service delivery under the Coalition government's reign. Bauman's (1989: 184-185) analysis is founded on the proposition that most of those perpetrators carrying out the genocide on a daily basis were normal people and that humans possess a number of 'primeval moral drives' -an 'instinctive . . . animal pity' (Arendt, 1963:106)-pertaining to an inhibition against inflicting harm on another human being and a compulsion to help those who are suffering. As such, Bauman (1989) identifies two relational social mechanisms that were requisitely manufactured and harnessed by Nazi elites to suspend moral inhibition and transform masses of normal people into conscious collaborators in the operation of mass slaughter.
First, Nazi propaganda drew upon and enflamed centuries worth of anti-Semitic discourses to cultivate a hostile socio-political environment, constructing Jewish people as a pestilent political-economic threat to national security. The political leadership's language and rhetoric was fraught with images of Jewish people as an economically degenerative 'parasite, a sponger who, like a pernicious bacillus, spreads over wider and wider areas' and drains those who grant them hospitality (Hitler, 1925(Hitler, [1939: 253). Propagandists also devoted huge resources to mass media portrayals of Jewish people as a sub-population that rarely engaged in productive work and was 'always living off the honest toil' of the hard-working German citizenry (Taylor, 1998: 175-176). This dehumanised Jewish people as a hostile and undesirable other; worthy at least of disassociation, apathy and punishment in the eyes of German citizenry.
Second, the Nazi's mediated the action of violence against the Jewish citizenry, enabling the perpetrators to injure and kill their victims at greater physical and/or psychic distances. This was not only accomplished through the promulgation of anti-Semitic propaganda, but also through utilising many archetypal characteristics of the modern bureaucracy (see Weber, 1948). The Holocaust relied upon a top down hierarchical 'system of super-and subordination' (Weber, 1948: 197); whereby both scientific expertise and one's immediate superior in the chain of command supplanted primeval moral drives as the principle source of moral authority. This ran in tandem with a 'meticulous functional dissection and separation of tasks' (Bauman, 1989: 100). Specifically, the practical effects of violent tasks were (mostly) optically separated from the perpetrators' vision via more clinical killing technologies; while each task was scientifically managed through quantitative (time and motion) monitoring and diagrammatic representation.
Cumulatively, when violent acts were (1) morally sanctioned by authority, (2) broken down piecemeal and shielded from visceral response, then (3) numerically measured, according to Bauman (1989), the moral concerns of the perpetrator were drawn away from focus on the suffering of their victims and the inhumane consequences of their actions. The Jewish people had been reduced to objects of bureaucratic task performance. Moral concerns were instead 'forcefully shifted in another direction -the job to be done and the excellence with which it [was] performed' (Bauman, 1989: 159). It no longer mattered how the human objects fared. Emotions were marginalised or altogether made obsolete from the bureaucratic operation. The perpetrators forgot their violent actions were a means to something other than performing the immediate task at hand in satisfaction of quantitative criteria and a moral authority. Instead, they began to occupy themselves almost exclusively with the rational task of finding more efficient means to achieve targets. In doing so, they began to neglect the moral task of evaluating the ultimate objective.
Bauman's analysis has been criticised. Some have disputed his argument that humans have innate aversions to violence; suggesting that his understanding of primeval moral drives is both 'one-sided' (e.g. neglects aggression instincts) and dismissive of the sociological orthodoxy that morals are ultimately a social product of cultural and historical processes (Vetlesen, 2005: 47). And Bauman (1989: 245) possibly contradicts his own analysis by later suggesting that it is 'the civilising process' -that is, the historical development of social attitudes, standards and stigma against behaviours typically associated with barbarism (see Elias, 1994) -which has fostered human aversion to violence. Nevertheless, his ideas on propaganda and bureaucracy remain a powerful way of explaining how ordinary people, raised in (late) modern societies, may routinely carry out institutionally mediated violence.

Methodology
All interviews were conducted in 2019 for a PhD funded project examining contextualised agency in public service interactions between front-line employment service staff (n=11) and young male benefit claimants (n=15). The authors draw specifically on the responses of ten public and contractor employment service staff possessing direct experience of working in the employment service between 2010-2015. One manager and 14 claimants did not have relevant experience in this time period and have thus been excluded from the present analysis. A strength of the present sample was that these workers cumulatively had almost 200 years' experience of working in the civil service. Some had witnessed first-hand several decade's worth of social security reform: 'It was different because in the '70s and '80s. . . It was more overall about the well-being. . . of a person'. The present sample also comprises workers occupying a variety of different roles: with one JCP manager; three JCP front-line staff; one JCP and later Work Programme front-line worker; one decision-maker, and four Work Programme front-line staff.
All participants worked in different offices across the country (no two were the same). This reflected the continued difficulty of accessing employment service workers, the data protection regulations governing their experiences, and perhaps the DWP's own increasing sensitivity to external scrutiny. This was most visibly revealed in its decision to rescind the involvement of Jobcentre Plus and Work Programme staff in a major ESRC-funded investigation of the impact of UK welfare reform (see Welfare Conditionality, 2018:13). Many of those approached declined to speak citing fears around confidentiality (one in the present sample refused to be recorded). To overcome this, contact was established with retired civil servants and adjacent service providers where individuals had migrated into new roles. This reveals another possible limitation. The overwhelming majority of the sample (n=9) had left civil service at the time of interview. Nevertheless, we thought that this provided sufficient distance for many participants to provide thick descriptions of their experiences. This was possibly reflected in the ease at which a 'structure of feeling' could be discerned across their individual experiences.
Given some participants' clear concerns around issues of privacy and confidentiality, as well as the sensitivity of some of the data revealed, we have opted to ensure all details of name, place, gender, age, ethnicity, job tenure and so forth are withheld or pseudonymised. Ethical approval for this research project was received by Sheffield Hallam University.

The production and delivery of institutional violence
The remainder of this article will outline the production and delivery of institutional violence in Britain's network of public/contractor employment service offices. We begin by first showing how political and media elites combined to intensify forms of propaganda; constructing various benefit claiming groups as a drain on public resources to secure consent for austere reforms. Second, we illuminate the contours of post-2010 reforms to the operational logic underpinning Britain's employment service bureaucracy. Finally, we draw upon primary qualitative data to reveal how core features of bureaucracy and propaganda operate on the ground and, at certain points, work in alloy, to produce 'moral invisibility' and encourage the delivery of institutionally violent practice(s).

Welfare propaganda: Cultivating a hostile environment
The Coalition's austerity programme took, in part at least, the symbolic 'form of a massive propaganda exercise' (Tyler, 2020: 194). From 2008, pejorative, stigmatising portrayals of out-of-work populations became increasingly prevalent in political discourse (Okoroji et al., 2020). Politicians intentionally rehashed and enflamed centuries' worth of 'scrounger' discourses to fortify anti-welfare common sense and manufacture consent for austerity (Morrison, 2019: 8;Tyler, 2020: 19). Britain was repeatedly decried as broken. Sexually excessive 'parasites' breeding 'feckless families' courtesy of the taxpayer (Atkinson, 2013), fraudsters in possession of an entitlement mentality (Duncan-Smith, 2012), and migrants placing increasing strain on public services and jobs (Vickers, 2019), were repeatedly portrayed as a threat to, and a drain on, the hard-working British citizenry.
Provocative political discourses tend to be widely shared in newspaper outlets and particularly by those sharing similar ideological orientations (Okoroji et al., 2020: 2). Consequently, this fuelled a huge surge in stigmatising terminology diffused through many mainstream British newspapers; reaching crescendo in 2013 (Morrison, 2019: 20-21). Not only did this bring about widespread dissemination of articles portraying migrant and out-of-work populations as a drain on public resources; but the digitalised, socially mediated methods through which these articles were disseminated enabled consumers to bolster narrative power by sharing, re-posting and adding further details or truth claims to the articles they interacted with (Morrison, 2019: 201).
This resulted in outbreaks of interpersonal violence upon migrant and benefit claiming populations (Burnett, 2017: 220;Ryan, 2019: 29). However, the main effects were a general hardening of public attitudes towards claimants and growing support for punitive measures (Okoroji et al., 2020). According to one survey, the percentage of people agreeing that cutting welfare benefits would damage too many people's lives had fallen by 17% from 59% in 2000 to 42% in 2010 (Taylor and Taylor-Gooby, 2014: 6). Support for more spending on benefits for disabled people unable to work fell from 63% to 53% between 2008-2011 (Park et al., 2012: ii). Meanwhile, 70% of 2,407 respondents in a 2011 poll felt that people receiving unemployment benefits who refused job opportunities or failed to attend interviews should lose half or more of their claim (Adler, 2018: 13). Consequently, Tyler (2020: 197) concludes that propaganda spun around austerity corroded compassion and hardened feelings to the plight and suffering of Britain's poor people.

Social security reform: Manufacturing a violent bureaucracy
A series of reforms to the operational logic driving Britain's employment service bureaucracy were also implemented. Public and contractor employment service offices have operated under the archetypical logic of a modern bureaucracy since their modernised inception in the early 1900s (Price, 2000). Britain's network of employment service offices has always tended to be monocratically organised and sensitive to government control. They operate on a top-down 'hierarchical subordination' model stretching from the Secretary of State in central government filtering all the way down to front-line service (DWP, 2019: 110;Weber, 1948: 197). Employment services have also always tended to possess a 'rational character' (Weber, 1948: 244). This means that they tend to be staffed by officials who perform formal duties within a chain of command and in line with the authority of more highly ranked officials above them (Wright, 2002: 16). This also means that staff perform according to a set of calculable rules -customarily pertaining to policing benefit claims and matching local labour supply with demand (Price, 2000) -that serve to produce workers' day-to-day norms of conduct (Weber, 1948: 215;220).
From 2010 onwards, the Coalition government radically altered this operational logic. This changed workers' norms of conduct and, in turn, transformed Britain's network of employment offices into far more dangerous places for benefit claimants (Wright et al., 2020). We contend that this was accomplished via three key (not exclusive) social security reforms.
First, the Coalition engaged far more seriously with emerging scientific expertise on utilising policy tools to alter human behaviour (see https://www. bi.team/). In this vein, they built upon the 'work-first' foundations laid by previous administrations and introduced a rapid extension/intensification of existing behavioural change policy. The introduction of Universal Credit saw new groups subject to behavioural conditionality, while the introduction of a new Claimant Commitment also saw more claimant sub-groups subject to an intensification of work-related conditionality (Fletcher and Wright, 2018). Failures to comply were also punished under the rubric of an 'enhanced sanctioning regime'. Introduced in 2012, the new regime featured a built-in tiered system of sanctions with longer disqualification periods (up to three years) for more serious non-compliance; differentiating between first, second and third violations to upscale punishment for recidivists (Adler, 2018: 37). The Coalition's sanctioning regime also featured a separate referral and decision-making mechanism; whereby front-line workers would initially raise a claimant non-compliance doubt and then refer the case with relevant evidence to an Independent Decision Maker who assesses the case according to specific legislative criteria in a separate office (see DWP, 2017).
Second, the Coalition took inspiration from preceding Labour government experimentations and uprated investment in quasi-marketised employment services operating on a competitive, payment-by-results funding model -titled, 'Work Programme' (Wiggan, 2015). Unlike its predecessors, however, the Work Programme was targeted specifically at the 'hardest-to-help' claimant populations (Finn, 2018). This required service providers to absorb risk and invest up-front, with profit largely dependent on their ability to secure initial and sustained job outcomes for long-term unemployed people and those receiving illness/disability benefits subject to work-related requirements (Finn, 2018). The Coalition also explicitly mandated Work Programme providers to operate under the rubric of its newly enhanced sanctioning regime and make sanction referrals for any breach of requirements (Webster, 2016: 2). Thus, front-line workers were not only managed according to a more stringent set of performance criteria centred firmly on job outcomes, but they also had a duty to refer claimants for sanction. This brought a wider range of some of the most vulnerable and least work-ready claimants under the ambit of a more stringent work-related conditionality and sanctioning regime.
Third, from 2010 there was an 'unannounced change of policy by ministers. . . to pressurise DWP staff to make more referrals' for sanctions (Webster, 2016: 2) and, shortly after in April 2011, a major simplification of the target regime governing front-line staff in Jobcentre Plus. Previously, worker performance was measured against a range of indicators, such as the proportion of claimants entering employment and the volume of adviser interviews (Finn, 2018: 226). By contrast, the new regime was to measure performance strictly according to 'off-benefit flows' -successful outcomes were now achieved when claimants ended their benefit claim, irrespective of whether they had entered employment (Finn, 2018: 226). Fletcher and Wright (2018) suggest that largely hidden managerial pressure combined with Coalition's new off-benefit flow target regime encouraged frontline staff to prioritise cases and actions that would most quickly and efficiently secure the termination of benefit claims. Correspondingly, Webster (2016) claims that unannounced pressures and targets, alongside breaches of requirements with Work Programme providers above-mentioned, played a major role in inducing a huge increase in sanctioning rates between 2010 and 2013 -reaching over one million sanctions in 2013 and rising approximately 345% above their 2001-2008 average level (Adler, 2018: 48). This also tallies with the DWP's own decision to raise 'it's off-flow targets for jobcentres as part of its annual review of their performance' and increase targets for those offices who were already 'meeting them consistently' in 2013 (NAO, 2016: 25).

Moral invisibility and violence as technique on the front-line
In 2013, skyrocketing sanction rates were accompanied by media concern; particularly by Guardian newspaper journalists who were openly hostile to the Coalition regime and revealed documents indicating that a secret regime of sanctioning targets existed within Jobcentre Plus (Couling, 2013:3). This prompted internal investigations to quickly deny the presence of sanctioning targets and/or the inflationary effects of off-flow targets on sanctioning rates (NAO, 2013). The Secretary of State for the DWP, Iain Duncan-Smith, commissioned former Principle Private Secretary and later Universal Credit Director General, Neil Couling, to conduct an internal investigation. Couling (2013: 9) 'found no evidence of a secret national regime of targets or widespread secret imposition of local regimes to that effect' and also 'found no evidence people [were] being wrongly sanctioned as a consequence'. In contrast, the present research found clear evidence of secret, localised sanctioning regimes and clear indications that staff were inflating sanction referrals in at least three different Jobcentres situated across the UK. One worker reported: weekly team meetings. And s/he [team manager] used to produce a table which showed how many people you've sanctioned or how many people you'd referred to a decision-maker for a sanction. (JCP Executive Officer [worker three]) While the DWP explicitly denied the existence of sanctioning targets, it was conveyed how mounting expectation to administer sanctions from above, translated into the formation and execution of local target regimes on the ground: 'certain staff would come [in the canteen] and say "well I've got my [sanctions] target for the week"' (JCP Executive Officer [worker seven]). It became clear that sanctioning and off-flow target regimes had a 'dehumanising' effect. For instance, manager one had witnessed first-hand the harsh realities of poverty and destitution experienced by out-of-work claimants s/he had worked with. Yet, this didn't prevent them from pursuing sanctions and 'off-benefit flows'. On the contrary, achieving targets was described as 'exciting': According to Bauman (1989: 102), dehumanisation 'starts at the point when. . . the objects at which the bureaucratic operation is aimed can, and are, reduced to a set of quantitative measures'. This makes it easier for workers to overlook the human consequences of their actions. In the present study, top-down managerial pressure -diagrammatically represented in the shape of sanctioning tables and off-flow targets legitimised by the moral authority of the democratic state -appeared to act as a moral anaesthetic. This frequently made invisible the needs and interests of claimants in the eyes of workers; sedating some from the possible outcomes of their actions and thereby allowing workers to view caseloads with 'ethical indifference' (Bauman, 1989:103). This was not only evidenced in competitiveness and enthusiasm around achieving targets -'you always wanted your team to be the best team' -often irrespective of their harmful outcomes (see later); but it was also evident in some respondents' repeated attempts to 'demote, ex-probate and delegitimise the ethical motivations of social action' (Bauman, 1989: 28). In this vein, respondents would sometimes absolve themselves of moral responsibility by assuring that they were 'only implementing the law which made by MP's which is voted by the people' and that their actions were 'not personal', but driven by a superior moral authority: yes people [managers] did pull the wool over my eyes [with sanctioning targets]. I'll quite freely admit it. I don't care, that's their conscience, not mine (worker seven) Moral sedation was also evident in the emphasis participants placed on remaining emotionally detached in the role: I think it's like you shut down the personal stuff, you're there to do a job. . .we can't become emotionally involved, can we? (manager one) It was also evident in the way some workers relinquished their political and moral ideals to the chain of command and the rules/procedures of the bureaucratic operation: 'The way I was treated, I had to abide by these rules that went against everything I thought and believed in' (Work Programme Welfare-to-Work Advisor [worker two]). When this occurs, personal responsibility is divorced from action and the ethical motives of those carrying out the operation become superfluous. This creates a moral vacuum through which institutional violence can thrive.
Moral sedation may also have been present in the way emotional detachment turned to 'disapprobation and censure' when 'resistance, or lack of cooperation' on part of a claimant 'slowed down the smooth flow of bureaucratic routine' (Bauman, 1989: 103). Manager one revealed an instance where s/he used their relationship with a local employer to set a work-resistant claimant up for a sanction: 'one day, this particular man, I knew he didn't want a job, so I sent them to see this employer . . . [he] turned up in flip flops and Bermuda shorts . . . I presume he got his benefit disallowed because he wasn't making himself available for work'. This could point to a limitation with the application of Bauman's ideas to the present case. It was possible that organisational conditions were not exclusively producing indifference. Some participants suggested that a small minority of colleagues throve on their legitimised ability to inflict harm: 'you had some people who it seemed enjoyed the stick. . . it's like a power trip' (Work Programme Welfare-to-Work Advisor [worker six]).
Finally, moral sedation was evident in pre-occupations with the rational task of finding more efficient methods of carrying out the bureaucratic operation, focusing 'fully on the good performance of the job at hand' and deploying 'morally abject' strategies to surpass targets (Bauman, 1989:102). In worker three's JCP office, management had set up a team specifically to achieve benefit 'off-loads' (respondent terminology) within 0-13 weeks. Because frontline staff were measured ultimately on 'off-loads', finding ways to sanction claimants and/or dissuade claims, as opposed to finding ways of facilitating transitions into work, had become the more rational option. When asked how the 0-13 week team would achieve 'off-loads', worker three proceeded to explain how they would frequently treat claimants with 'disrespect' and use psychological harm as technique: 'they were pushing them until they either just cleared off because they couldn't take the pressure or they got sanctioned' (cf. NAO, 2016:28). This chimed with worker five (JCP Executive Officer), who suggested that some staff intentionally tried to antagonise claimants in efforts to dissuade claims.
Morally abject strategies to surpass targets were also evidenced in the case of one manager: who just thought s/he could get anybody sanctioned because they weren't complying. It didn't matter that they couldn't speak English. S/he misinterpreted the role to the point that s/he felt that the team would only be successful if they had an average amount of sanctions (worker seven) Worker seven proceeded to detail how claimants who did not speak English were often unable to evidence work searches in English but, as long as they had agreed to provide evidence in English, then they were liable for a sanction under the rubric of their Commitment: s/he would still insist that if they hadn't written in their Claimant Commitments themselves in English, that they should be sanctioned.. . . So they couldn't be applying for jobs. They couldn't understand that.
Worker seven was expected to draw up Claimant Commitments with non-English speaking claimants on the stipulation that they would provide evidence of work search in English. While this caused conflict on numerous occasions, worker seven's manager proceeded to knowingly sign Commitments with claimants that s/he already knew would be highly unlikely to comply: s/he said to me, 'you think I'm racist, don't you?' I said, 'Well, I didn't say that'. I never said 'racist', but I said, 'How can you stand there and tell me that if I'm speaking a different language to the person in front of me, I can penalize them because I can be sure that they've understood what I've said?' S/he said s/he would do that. That, that was right. S/he would do this Claimant Commitment and get them to sign it. So I said, 'Well, I'm not going to sign it.' S/he said, 'Well, I'll sign it.' So I said. 'You're signing to say they've understood that?' 'Well, they've signed it, so they must understand.' This case could indicate that racial bias was driving managerial behaviour; whereby non-English speaking claimant lives were seen as less valuable and thus less deserving of fiscal support -perhaps indicating internalisation of heightened anti-migrant, racialized political and media discourse. It could also reflect an effort to secure an 'easy [sanctions] win' (see Couling, 2013:15). Sanctions have to be referred to an independent decision-making process, which includes a right to appeal an adverse decision (for claimants). In relation to work-search related sanctions, decision-makers are explicitly mandated to legitimise sanction referrals in cases where claimants undertake work-search activity 'that does not offer them any chance of getting an offer of paid work' as it 'cannot help them satisfy the work-search requirement' (DWP, 2017: 64). Therefore, non-English speaking claimants were likely viewed as low-hanging fruit because their lack of cultural resources meant they could be readily penalised on this stipulation and are likely less able to legitimately challenge any adverse decision. It does not strain credibility to suggest that managers were actively probing for loopholes within the legislative framework to achieve sanctions. On the contrary, manager one was fully aware of the importance of working with the independent decision-making process in mind: I used to send my staff to the decision-makers for training. We were particularly good at getting people off the books. It was usually because they were working anyway The 'stigma power' (Tyler, 2020) of pejorative welfare tropes often appeared in complementarity with the pursuit of off-flow and sanctioning targets. As can be seen in the excerpt above, by constructing out-of-work claimants as an undesirable and potentially threatening 'other' (e.g. as a 'fraud', 'scrounger', 'drain'), staff could efface the humanity of their caseloads and remove them from moral obligation. This provided some workers with a sufficient cause to justify punitive working practices which would likely lead to harmful outcomes: 's/he took the view that claimants are lazy, dishonest, not trying to get work and the stick was important'. In this vein, worker three explained how their manager made efforts to discipline workers and achieve sanctioning rates by harnessing 'stigma power' as a way of rousing antipathy among frontline staff and establishing distance between workers and claimants: there was a point at which s/he said, 'It's your money! It's your taxes that they're living off! You know, you should be sanctioning them!' Stigma driven behaviour is not a phenomenon unique to the 2010-2015 period. Nevertheless, for worker three, this period was somewhat unique in that stigmatising language was increasingly prevalent in formal communications meetings.
This was reported to have a significant bearing on the frontline; 'infecting the culture, practices and attitudes of welfare workers' (Tyler, 2020: 196): [it was] just work coaches sitting in the canteen at lunchtime saying how awful claimants were and how they were scroungers and liars and all the rest of it.
Pejorative, stereotypical views of claimants were also present in the perceptions of some participants: Nowadays when you go a customer's house, they all have the big TV. I know it's a stereotypical thing to say, but they do. Because that's what they do all day. They sit all day and they put Jeremy Kyle on (manager one) Stigma driven behaviour was also detected in the behaviour of G4S security guards who, according to worker seven, would sometimes intentionally try to antagonise claimants and 'make them feel uncomfortable' through use of stigmatising language: 'she said [in response to a claimant] "yeah, but at least I have a job to go to. Unlike you two.". . . sometimes it's the other people that are about that will make the customer feel uncomfortable'.
Similar processes were working to manufacture violent practices in Work Programme provider offices. Welfare-to-Work advisors frequently reported facing stringent managerial pressure -mediated through job outcome targets -to "push" mentally or physically disabled claimants into work. This was reported by worker nine as resulting in near fatal outcomes: [I had] a lovely guy who I really felt for who had mental health issues and the day after I had to reluctantly mandate him to something -he attempted suicide. I also had another lady who we pushed into work and it made her that ill she had a fit in her new job and was admitted to hospital. (Work Programme Welfare-to-Work Advisor [worker nine]) Performance targets were of pivotal importance in governing behaviour (Soss et al., 2013). Worker six went on to describe how, in spite of 'hav[ing] days where [s/he] would be in tears' due to 'forcing' people into employment who 'couldn't even function you know on like a normal level', a combination of job outcome targets and managerial pressure: made you feel competitive I think. I don't know why because I'm not like that this in role. It kind of made you really want to achieve, probably because you had a meeting every Wednesday with the whole team. . . It was very transparent on, you know, it would be highlighted the really good people. . . if people are getting flagged up for being great and then you don't. Like oh, you're going to know, I'm shit. (worker six [our emphasis]) This affirms the strength of applying Bauman's ideas; revealing how moral inhibitions to violence and suffering were subordinated to competitive, target-focused behaviours that were driven by a desire to be seen as an economically productive, efficient and diligent worker in the eyes of colleagues and superiors. Worker six had recently moved into a new role working for a mental well-being, charity-based employment service where achievement was not inextricably associated with job outcomes. This had transformed their behaviour; once again underscoring the crucial role performance metrics and managerial pressures play in shaping front-line practice.
The 'moral sleeping pill' (Bauman, 1989: 26) effects of performance metrics, moral authority and stigmatisation meant that resistance to managerial pressure was uncommon. Nevertheless, our research uncovered a few instances of resistance. Evidently some workers felt it necessary to use discretion; creating minor spaces of personal control and autonomy by allowing some claimants to take 'small liberties' (Dubois, 2010: 151). Worker three was, for example, openly hostile to the UK's sanctioning regime which occasionally led to minor acts of subversion: If I don't think their job search is good enough, I might just ignore it However, workers often had to justify their actions and could be subjected to further managerial pressure. This sometimes led to feelings of intimidation or fear which could make it more difficult to resist managerial demands: 'I allowed myself to feel intimidated by this manager who criticised my lack of sanction' (worker three). Moreover, this could encourage staff to deliver the service in a more disrespectful and psychologically harmful way: I've got my manager sitting there and they don't come up with a good story about their job search. . . So I give them a bollocking. . . They [claimant] were exposing me as a softy. . . They were putting me at risk. So I was particularly angry about that.(worker three)

Conclusion
We have demonstrated how ordinary workers have been encouraged to implement social security reforms which are frequently experienced as cruel and inhumane. We have sought to re-calibrate Bauman's conceptual armature in an effort to make sense of the processes (or 'social mechanisms') which inform violent practice in UK employment services. In so doing, we have highlighted the pivotal role played by the (re)intensification of stigmatising welfare narratives and a number of social security reforms that changed the operational logic of Britain's employment service under Coalition rule. These facilitated the production and delivery of institutional violence on the frontline of service delivery.
Nevertheless, there are a number of key strengths and limitations of applying Bauman's ideas to the operation of Britain's employment service bureaucracy. In terms of the former, we have shown how 'moral authority' (in the shape of hierarchy/chain of command) and 'quantitative measures' (in the guise of sanctioning/off-flow/job outcome targets) alongside heightened stigmatisation often worked in alloy to produce a psychic distance between workers' moral drives and their actions. This has been crucial in terms of paving the way for (institutionally) violent front-line practice.
However, there are four key limitations to the application of Bauman's ideas. First, social security reforms were not instituted to facilitate a genocidal outcome and consequently we did not detect any presence of clinical killing technologies on the front-line. This is important. The aim here has not been to simply transpose Bauman's analysis of the Holocaust onto new conditions verbatim and we are not suggesting that the Coalition's austerity programme held a genocidal intent -although 'social murder' has been an indirect outcome of post-2010 reforms (Grover, 2019).
Second, Bauman's ideas are less relevant to contemporary employment service delivery. From 2015 onwards, policy makers have once again significantly altered the logic driving the employment service bureaucracy. Maximum sanctioning periods have recently been rolled back to six months from three years. Meanwhile, according to one active worker, the managerial message has shifted from policing claims and achieving off-flows to 'focusing on Universal Credit and trying to get that to work' with reduced staffing and resources: 'it's changed now. It's completely disappeared in our office. There is no manager putting any pressure on us to sanction. There is no conversation in communication meetings which says claimants are lying scroungers'. This underlines the importance of political control over the purpose and operation of service delivery.
Similarly, the new Work and Health Programme (replacing Work Programme) is still driven by job outcomes but with less emphasis on conditionality and sanctions and considerably more focus on the well-being of claimants than its predecessor. Sanctioning rates have been in significant decline and have fallen back to an average below their 2001 level. Some (not all) of the longer-term claimants spoken to in the present study even reported a change in service delivery: 'it seems like they've actually learnt a thing or two. . . it seems they've gone, we do actually need to take care of these people'. This not only points to the importance of conceiving of the employment service as a monocratically organised bureaucracy sensitive to government control; but, perhaps most crucially, that institutional violence is socially produced from above and is thus an 'avoidable' phenomenon (Grover, 2019:339).
Third, Bauman's theory as applied to the present case downplays the agency of employment service workers and the role of discretion in front-line practice (Lipsky, 2010). Although uncommon, we encountered staff that resisted managerial authority and refused to carry out their duties in a socially harmful way. However, resistance to sanctioning/off-flow targets was exceedingly difficult for workers, especially given that instruction through the chain of command was reported to firmly assert that sanctions 'should be applied and are not a matter for individual discretion' (Couling, 2013: 9).
Finally, it is difficult to see how Bauman's idea of optically separating the practical effects of violent tasks from the perpetrators vision is relevant to the present case; particularly as front-line staff are frequently confronted with the human consequences of their actions at subsequent face-to-face encounters with benefit claimants. It is salient to note that employment service staff were subjected to increased levels of verbal and physical intimidation from some claimants over (ES, 2013. However, the potential for institutional violence mediated through optical separation may become a more prominent feature of front-line service delivery. This is because the employment service bureaucracy is at a critical techno-economic juncture in its historical development. On the one hand, the roll-out of Universal Credit alongside persistent budget cuts in the last decade have fuelled a shift from inter-personal services towards more costeffective forms of digitalised service provision (Finn, 2018). By the time of full roll-out (now 2024) over 80% of claimants will be expected to manage their interactions with the DWP online (Finn, 2018: 225). Online service provision also features a 'digitalised advisory function' that allows Jobcentre staff to monitor individual job search activity and make sanction referrals for non-compliance (Finn, 2018: 225).
On the other hand, the employment service will soon have to reduce a persistently high claimant count resulting from the COVID-19 pandemic. To this end, it has recently been announced that a major recruitment drive for new staff will take place but this will also intensify the shift toward more digitalised, socially distanced forms of service provision. Historical evidence has shown how institutional violence tends to thrive in periods of crisis management so as to dissuade claims and shift large numbers off the register (cf. Fox-Piven and Cloward, 1972). This was evident in reforms made during the Great Depression in the 1930s (Fletcher, 2015) and again in response to the North Atlantic Financial Crisis (see above). Consequently, it is possible that the employment service will revert to reducing claims facilitated via more impersonal, digitalised forms of socially distanced service provision. Together, this may create a somewhat unique situation on the front-line. One in which workers are not only encouraged to dissuade claims via (institutionally) violent practice, but, for the first time ever, they are to do so online. Accordingly, workers will be more optically separated from the human con-sequences of their actions than ever before and are thus increasingly shielded from the moral inhibitions that the consequences of any violent action might ordinarily evoke. It will therefore become more important than ever that research exposes the employment service bureaucracy to scrutiny in order to demystify front-line practices and uncover the human consequences of crisis management reforms.