Introduction: "Men who work can be incapable of work"

"Men who work can be incapable of work," ran the headline in a Times law report in 1985, reporting on a legal case concerning two men who had been claiming Invalidity Benefit (The Times, 26th June, 1985: 25, reporting on Hunt v Chief Adjudication Officer; Merriman v Chief Adjudication Officer 1985). Both men had been certified by their doctors as incapable of work but Mr Merriman had been working as a local councillor, while Mr Hunt had been working a few hours a week, checking tickets on the gate at his local cricket club. The court's decision was that both men could be eligible for benefit. The legal case turned on the meaning of the term "work," and the benefit regulations, which a legal analyst described as "highly convoluted" (Mesher, 1986: 53). This example illustrates a question that policy makers have struggled with since the beginning of state incapacity benefits1 in the early twentieth century. The problem lies with the definition of "work" and whether people claiming incapacity benefits should be expected to do nothing which could be described as work or whether they should allowed or encouraged to carry out work-like activities. UK disability benefits systems have developed the concept of "permitted work" to allow claimants such as Mr Merriman and Mr Hunt to do small amounts of paid work while retaining entitlement to benefit.

Stone argued in the 1980s that the creation of the "disability category" within welfare states was a way of differentiating between those citizens who were required to work and those who could claim exemption from that requirement (Stone, 1984: 4). She argued that, while medical certification has become the main mechanism for differentiating between those who are capable and those who are incapable of work, welfare systems have also used what she calls "revelatory signs" as tests of claimants' legitimacy. Such revelatory signs can include collecting wider information about a claimant through home visits, local investigations and tests of what the claimant can and cannot do (101). People who claim incapacity benefits and attempt to work may find these attempts to be examples of such revelatory signs. These signs can mean different things: that the claimant is not as ill as they claim and thus is not entitled to benefit; that the claimant is genuinely trying to improve their chances of returning to the labour market and thus is deserving of benefit; or that they are unlikely ever to return to work but are making morally good use of their time while claiming. This article will argue that changing ideas about permitted work have been used to define who is entitled to claim exemption from the labour market in the UK benefits system and who is expected to make every effort to return to the labour market.

Background to research and sources

This article stems from a wider research project on the development of the concept of incapacity for work in the UK benefits systems since 1911.2 The first state insurance-based incapacity benefits were established by the National Insurance Act 1911, which made provision for short term Sickness Benefit and a long term Disablement Benefit. Versions of these benefits continued, with variations in names and entitlement rules (including insurance based and means-tested benefits) up until the current system of Employment and Support Allowance, which has both an insurance-based and a means-tested version. All of these benefits have had a common requirement that claimants must have a medical certificate of incapacity for work in order to qualify. Some other welfare states (e.g. Germany, the Netherlands) have labelled some disabled people as "partially capable of work," enabling people to participate in the labour market, while also receiving some payment from the welfare state to compensate for their lack of "full capacity" (Lindsay and Houston, 2013; Roulstone, 2012). Although the UK benefits system has not used this partial capacity approach, there have been various attempts by policy makers to provide in-work means-tested benefits for disabled people, such as Disability Working Allowance in the 1990s, Tax Credits since the late 1990s and now Universal Credit (see below for details of Universal Credit). Alongside these contributory and means-tested benefits there has been a tradition of compensatory payments for some disabled people whose impairments were the result of war or industrial injuries. These have usually operated on different principles from incapacity benefits, with payments made according to the extent of impairment caused by the injury or illness rather than on incapacity for work as such. The basic principles of the first industrial injuries scheme, established in 1897, and subsequent schemes, were to provide state regulated payments for compensation for injury at work (Bartrip, 1987). This scheme paid compensation according to the extent of workers' injuries, related to their previous earning power and could include reduced payments where the worker was considered to be fit for light work (for a detailed discussion of how this worked in practice, see Turner and McIvor 2017). The 1897 scheme continued, with amendments, until a revised industrial injuries scheme was established in 1948, as part of the post-war National Insurance scheme. The 1948 scheme had several components, but mainly required an assessment of "disablement" which was paid on the basis of the assessed extent of the claimant's impairment, rather than their incapacity for work (Ogus and Barendt, 1978: 299-303). Some elements of industrial injuries schemes have used the concept of incapacity for work, using a similar definition of incapacity to that used for incapacity benefits (Jones 2000). These have had similar permitted work rules to those discussed in the present article, relying heavily on medical opinion on the value of therapy and rehabilitation, although differing in some technical detail (Bonner et al, 1991: 366). Specific benefits for incapacity for work for industrial injuries claims were abolished in the 1980s and 1990s. The remaining industrial injuries benefits are now assessed on impairment only (Jones, 1999: 466). Compensation for war injuries has an even longer history, with particular developments in the twentieth century. As with industrial injuries benefits, war pensions have largely been assessed according to the extent of the claimant's impairment rather than incapacity for work as such (Ogus and Barendt, 1978: 352).

The main focus of this article is the mainstream state incapacity benefits, beginning with 1911 and continuing to the present day.

Access to employment for disabled people

Before turning to the concept of permitted work, it is important to consider how people with health issues and impairments are excluded from the employment market. Access to employment has been one of the most pressing demands of the disabled people's movement, beginning with the demands in the 1970s, of organizations such as of the UK's Union of Physically Impaired Against Segregation (UPIAS), which argued "the struggle to achieve integration into ordinary employment is the most vital part of the struggle to change the organisation of society so that physically impaired people are no longer impoverished through exclusion from full participation" (UPIAS and Disability Alliance, 1976: 15). The UPIAS also argued that earnings replacement benefits served to exclude people further from the labour market, and that attempts to measure people's capacity for work would lead to unacceptable levels of surveillance.

Others have written at length about exclusion from employment and what policies might best enable disabled people to access the labour market (for example Barnes, 1991, Grover and Piggott, 2015b, Lindsay and Houston, 2013, Shah and Priestley, 2011). The debate about the best way to ensure rights for disabled people to full participation in society continues today, with a strong state emphasis on the idea that work is the best way out of poverty for everyone, including disabled people. This pattern can be seen across most welfare states (OECD, 2010). In the UK this has taken the form of making benefits for disabled people increasingly conditional on preparation for work in the labour market. These policies have been underpinned by a rhetoric that work is good for people and that making benefits more difficult to claim will encourage disabled people into work (For example, Department for Work and Pensions, 2007, Department for Work and Pensions and Department of Health, 2016). The effects of these welfare reforms have been well documented (for example Barr et al, 2015; House of Commons, 2015; Lindsay and Houston 2013). These increasingly harsh developments have led to some arguing for the right "not to work" (Grover and Piggott, 2015). These debates will continue but whether state policies exclude people from work or encourage people into work, there remains the question of how state systems treat people who claim out of work incapacity benefits.

The beginning of permitted work: behaviour during sickness and rehabilitation

When the first statutory Sickness Benefit was introduced in the UK, under the National Insurance Act 1911, the scheme was administered by "Approved Societies" (for example friendly societies, trade unions and industrial societies such as the Prudential Assurance Company). The Approved Societies were subject to strict statutory regulation but were entitled to make some of their own rules, including the power to regulate "behaviour during sickness" (National Insurance Act 1911, S14, 2). Societies were able to suspend benefits for up to twelve months, fine members up to 10 shillings, or expel members from membership for breach of such rules (National Health Insurance Commission [hereafter NHIC], 1912: 18-19). In the first set of guidance provided to Societies, there was no specific rule about working while claiming but some individual Societies did include a rule prohibiting working while claiming. We can see this in one of the earliest appeal cases to be heard under the scheme. This concerned a woman whose benefit had been suspended for three months in 1915 because she had been observed by the Society's sick visitor to be "carrying coals" and was alleged to have breached the rule which prevented a member from "following any occupation" while in receipt of benefit (NHIC, 1915: 9, Case 2). The woman's appeal was successful because the adjudicators found that the alleged carrying of coals was not "an occupation." The case illustrates how "working" while claiming could be used to suspend a claimant's benefit. It also provides an example of the use of sick visitors to monitor the details of claimant's daily lives to seek for revelatory signs of claimants' eligibility.

Working while claiming could also be used as evidence of capacity for work. In 1916, a fifty-eight year old former iron ore miner had been claiming benefit for six months. His Approved Society discovered that he now held a position as a caretaker which he shared with his son. The Society decided that this constituted work and that therefore he was not entitled to benefit. The claimant himself believed that he ought to be entitled to benefit, stressing that that what he was doing was not really work. In his appeal he said:

The duties of this position are extremely light, consisting of a daily inspection of the house and the sending of a quarterly report to the Council. Even these light duties I am from time to time prevented from performing, and in the main they are performed by my son. In no sense can this be regarded as effective work (NHIC, 1916: 160, case 59).

This man's appeal was unsuccessful because the adjudicators found that even these "very light" activities counteracted his claim of incapacity. The claimant's own argument that this was not "effective work" shows that he believed it was acceptable to carry out some work-like activities while still claiming benefit. His case shows that, at this stage of the system, where no general prohibition on working was in place, even very small amounts of work could be used to provide evidence of a claimant's capacity, causing them to lose entitlement to benefit. In 1920 the Ministry of Health published new Model Rules, in which we can see the first explicit prohibition on working while claiming but also the first sign of a provision of "permitted work." The rules stated that claimants of Sickness or Disablement Benefit:

must not do any kind of work unless it is any light work for which no remuneration would ordinarily by payable ….This provision does not apply however to work undertaken as a definite part of the member's treatment in a hospital or similar institution and it may be relaxed in the case of a member who is incapacitated from following his usual occupation and is undergoing training which is necessary for the purpose of fitting him anew for remunerative work (Model Rules on Behaviour during Sickness, cited in Lawrence, 1926 Appendix 1 para 68).

The Model Rules distinguished between unpaid "light work" and work which was focussed on rehabilitation. This division between "light work" for some and "rehabilitation" for others shows how the development of permitted work was part of a general policy goal of encouraging most sick or injured people to return to the workplace as soon as possible, while allowing others to participate in work-like activities so long as they were sufficiently trivial. It marks the beginning of a heavily medicalised view of rehabilitation as a key policy goal within the benefits system. This provision was expanded in the 1930s to include people who were resident in hospitals, "certain sanatoria, colonies for the treatment of tuberculosis, institutions for the treatment of neurological cases, and colonies for epileptics," since any work carried out under these conditions "may reasonably be regarded as a part of his curative treatment, the object of which is to restore general health and working capacity" (Ministry of Health, 1933: 104).

These exemptions from the general working while claiming rules were framed in the expectation that it was the duty of disabled claimants to retrain and make themselves fit again for the labour market. The rules were heavily medicalized, based on the assumption that medical experts could judge what would and would not be therapeutic. The rules on hospitals and therapeutic institutions were fairly specific but cases often appeared at appeal hearings where people's attempts at working in less structured environments were less clear. These people, who did not have the formal endorsement of medical expertise, were at risk of being considered to be trying to cheat the system. This shows how there was an inherent ambiguity in the principle of allowing some people to work to improve their health or employability, while being mistrustful of those who might be attempting to claim benefit illegitimately. An example of this can be seen in an appeal case from 1929, where a man, claiming Disablement Benefit, had been working for a couple of hours a day for his brother. Instead of regarding this as an attempt to get himself back to full time work, the adjudicator at the appeal hearing saw the claimant's attempts at working as evidence of capacity for work and disallowed the appeal (PIN 63/4/509).

Some Societies interpreted the prohibition on working rather leniently, as we can see from a couple of cases from 1938, where members of the Ideal Benefit Society were allowed to continue carrying out "light work" while claiming benefit. In one case a claimant had been on benefit for four years:

His doctor recommends that he helps a little in his mother's fruit shop, as this would help him mentally to pass away the tedious hours and strengthen his feet. Permission was therefore granted, the case to be reviewed in 2 months.

In another, the claimant had been claiming for six years and been allowed in the past to carry out "light duties":

It was agreed to accede to his request provided the privilege was not abused in any way (both cases from PIN 24/74 Ideal Benefit Society Minute Book Appeals sub-committee Men, 1935-1941).

The claimants in these two cases were allowed to continue working but only under conditions of surveillance to ensure that they were not cheating. An example of this surveillance concerned a man who had been claiming benefit for nine years after a stroke. When he was hit by a car at seven in the morning, questions were raised as to why he was out and about at this time. It was discovered that he had been delivering newspapers. This was considered to be "work" and he was fined for breaching the Society's rules (PIN 24/74 Ideal Benefit Society Minute Book Appeals sub-committee Men 1935-1941). The Ideal Benefit Society's use of discretion, conditional permission and surveillance shows how the concept of permitted work could be used to make moral distinctions between people who were deserving of support and those who were suspected of attempting to cheat.

Carrying coals and hanging out washing: women and domestic duties

The regulations concerning breach of rules were used particularly harshly against women who were alleged to be doing housework. Elsewhere I have argued that the interpretation of the meaning of the term "incapacity for work" in the sickness benefit scheme was heavily gendered (Gulland, 2013). Decisions about whether or not a claimant was incapable of work were dependent, amongst other things, on gendered ideas about what work the claimant might be expected to do. Although domestic activities were sometimes considered explicitly as evidence of capacity for "paid" domestic work, this was not always the case. Often the underlying question was whether or not the women were really available for work in the open labour market. Societies used a system of sick visitors and local knowledge to collect information about women's domestic activities. The earliest case, regarding the woman seen carrying coals, discussed above, is an example of this. There are many of these in the appeal papers in the 1920s and 1930s. A further example can be seen in 1927, concerning a married woman in her late thirties who had a visual impairment. She had been claiming benefit for three or four years until a sick visitor spotted her "hanging the family washing on the line; the washing was done well." This observation prompted a review of the woman's benefit and her benefit was stopped. At her appeal hearing the adjudicator agreed with the Society and found her fit for work (PIN 63/1/414).

Some Societies treated housework as work for the purposes of the behaviour during sickness rules. An example of this can be found in an appeal case from 1928 where a woman had been observed peeling potatoes. At the hearing the Society argued that she had breached a rule which required that claimants:

shall not do any kind of work, domestic or other, unless it be light work, for which no remuneration is or would ordinarily be payable (PIN 63/3/493)

The case is important in drawing attention to the existence of a rule which prohibited "work, domestic or other," which was intended to remove women from benefit entitlement on the grounds that they were really housewives rather than potential workers. Appeal adjudicators did not always agree and the debate continued through the 1930s. In a decision in 1936 the adjudicator argued that the relevant question should be whether women were capable of work in the open labour market:

In setting a standard of their own, namely that of capacity or otherwise for carrying out the ordinary daily work of her home, the Respondents are in my judgement mistaken. This is not the test laid down by parliament and it is not the test which I am prepared to follow. (copy of case in MH62/201/1099)

These disputes about whether housework constituted work for incapacity benefits purposes were more or less removed after the introduction of the post-war "Beveridgean" National Insurance scheme. Although draft rules included a prohibition on domestic work (Ministry of Health draft rules under the Beveridge proposals, 1943, in PIN 8/106), this was removed in the final version. A key legal case in 1951 explicitly excluded domestic work from the definition of "work" for decision making in Sickness Benefits (National Insurance Commissioners' Decision R(S)11/51). The use of the domestic activities test to refuse women benefits is an example of how the revelatory sign of working was used to test the suspicion that women might not really be workers. The issue did not entirely go away as women continued to be subjected to higher levels of surveillance than men in investigating their domestic activities. However, in theory at least, domestic duties were removed from the formal definition of work in benefits assessments.

The Post War National Insurance scheme: therapeutic work, good cause and trivial work

The debates about permitted work continued with the introduction of the 1946 National Insurance scheme. This provided a contributory Sickness Benefit which was payable to people who were considered to be "incapable of work" in a similar way to the 1911 scheme. Benefit could be paid indefinitely so long as the claimant continued to be incapable of work. It included a specific prohibition on working while claiming, prohibiting payment of benefit on any day in which the claimant "did any work," unless it was:

work which is undertaken under medical supervision as part of his treatment in or of a hospital or similar institution and in respect of which his earnings do not exceed 20s 0d a week (National Insurance Act, 1946. National Insurance (Unemployment and Sickness Benefit) Regulations, 1948, Regulation 10(d)(iii)).

This form of words followed almost exactly the form used in the Model Rules from the 1930s except that, for the first time, there was an explicit cash limit on the amount of earnings that the claimant could receive while carrying out permitted work. Commenting on the draft regulations, the National Insurance Advisory Committee noted that if this was not very carefully supervised it might amount to subsidising claimants who were partially capable of work, a view that was voiced many times across the remainder of the twentieth century:

The object of this provision… is to enable a cash incentive to be offered to persons in receipt of sickness benefit to encourage their co-operation in rehabilitative work … it is necessary to ensure that special provision of this kind is not used merely to subsidise with sickness benefit the earnings of persons who cannot earn full wages (National Insurance Advisory Committee, 1948, para 29-30).

These apparently explicit rules were not straight forward and led to a series of appeals heard by the newly formed National Insurance Commissioners. An early issue concerned the term "patient in or of a hospital or similar institution." This was discussed in an appeal from 1952. The claimant was being treated for tuberculosis and had been claiming Sickness Benefit for four years. There was no doubt about his incapacity for work but his appeal concerned whether working one day a week in a furniture factory counted within the permitted work rules. The factory was run by a charity to provide rehabilitation for tuberculosis patients and the pay was within the 20 shilling limit. At the appeal hearing, the Commissioner agreed that the factory work came within the rules because of the clear role of medical expertise and supervision in the arrangements. (R(S)3/52, case papers in PIN 62/1541).

During this post-war period, there was an expansion of segregated workshops set up under the 1944 Disabled Persons (Employment) Act. These workshops provided low levels of pay but usually they would have paid too much to enable a worker to claim benefits under the therapeutic earnings rules. For example Borsay gives examples from the 1970s "the weekly wage at Remploy was £18.12 a week" (Borsay, 2005: 138). At that time, the permitted earnings limit was £10 a week (Ogus and Barendt, 1978: 157). Although low, the earnings of these workers can be contrasted with people carrying out piece work for similar charities, working from home. The Piercy Committee distinguished two types of homeworkers:

disabled homeworkers may be divided into two categories: those capable only of diversionary occupations or handicrafts primarily intended to be of therapeutic or morale building value … and those capable of earning their living without supplementation from other sources (Piercy, 1956: 57).

Policy makers were particularly exercised by the position of these home workers and the need to differentiate between those who required only diversionary occupations and those who might be able to earn more. The concern that people might be deliberately restricting their earning power shows us the inherent tension in the permitted earnings rules. On the one hand they were intended to encourage people to get back to work. On the other hand there was a concern that they would provide an opportunity for malingerers to stay on benefits longer than necessary. The suggestion that people might keep their earnings deliberately low is another example of the use of revelatory signs and provides a foretaste of today's in-work conditionality in the Universal Credit scheme (see below).

Pay for disabled homeworkers was linked to the permitted earnings rules, but this caused some administrative problems. In the 1960s the earnings limit for permitted work was 40s (£2). In a letter to the Ministry of Social Security in 1967, the British Legion Poppy Factory explained that there were practical difficulties in paying £1.19.11d rather than £2. This letter was one of many written by such organisations, complaining about the low level of the permitted earnings limit. Replies from the Department of Health and Social Security explained that the limit could not easily be changed and also explained the thinking behind the policy:

It could also be psychologically wrong to encourage patients (sic) to think of themselves as incapable of work by allowing them to go on receiving sickness benefit together with higher earnings instead of putting emphasis on recovery and fitness for work in the full sense. (Letter from Supplementary Benefits Commission to Association of Hospital Management Committees, dated 7 March 1969, in reply to a letter of 17 October 1968 requesting that the earnings limit be increased from £2, in PIN35/392)

This correspondence illustrates the thinking that the point at which a person's imagined earning power went above the permitted work limit, they would be considered "fit for work."

These formal therapeutic and rehabilitation schemes provided an institutionalized and medicalized form of permitted earnings. Case law from appeals against refusals of benefit shows us that there were many disabled people attempting to carry out small amounts of paid work who were not involved in such formal rehabilitation schemes. Although claimants often struggled to come within the permitted work rules, in an early legal case, National Insurance Commissioners interpreted the rules rather broadly. This case concerned a man who earned around 7s a week "attending to sick and injured cats and dogs." There was no debate about the man's medical diagnosis and, generally, the decision makers appeared to be sympathetic: "The claimant appears to have been a sick man for many years." Their main concern was whether or not his work with cats and dogs constituted "work," whether he could be earning more and whether or not he was acting under his doctor's advice. The National Insurance Commissioners decided that the work was sufficiently light to not constitute work, it was below the 20s earning limit and was clearly supported by the doctor, and so allowed the appeal (National Insurance Commissioners case CWS 2/48KL, additional papers filed in PIN 62/1533). This established a principle that work could be disregarded if the claimant had "good cause" to be doing it but confirmed the continuing need for close medical supervision of any permitted work. However, medical advice was not sufficient if the other aspects of the rules were not met. In 1952, a man who was experiencing "nervous overstrain" from his usual work took up an opportunity to carry out two weeks voluntary farm work under a Government scheme to support the agricultural industry. Work for the scheme was paid but deductions were made for accommodation and meals, leaving only 4 shillings pay. The claimant took up the work on his doctor's advice. His Sickness Benefit was stopped and at an appeal hearing, the National Insurance Commissioner concluded that this "voluntary work" constituted work under the regulations, that the calculation of pay included the cost of accommodation and meals, and that the doctor's advice was not sufficient to come within the therapeutic rules (R(S)5/52, case papers filed in CT11/56).

A rather different case concerned a young man, aged 17, who had an inoperable tumour, which had led him to give up his usual job and claim Sickness Benefit. His father had found work for him in his own workplace, a brush factory, where the young man worked for up to four hours a day. This part-time work was supported by his doctor, who said:

I am only allowing him to do this light work to make him feel that he is like other boys. He is not capable of doing a real job, but the fact that he is at work has given him a feeling of self-respect and that is why I have allowed him to take this job (National Insurance Commissioner's Decisions C(S)9/48(KL), case papers in PIN 62/1273).

Although this looks as if it would clearly be suitable for the permitted work exception, the young man's earnings of 29s a week were too high. This young man was not expected to recover and so the permitted earnings exception would not have led to rehabilitation but could have been used in a therapeutic way to help him maintain some self-respect, if he, his father or his doctor had better understood the rules.

The late 1960s and early 1970s was a time of great debate about how best to support disabled people, including the campaigns by the Disability Alliance and the Disablement Income Group for a "comprehensive disability benefit," which was opposed by the UPIAS (For detailed discussion of this debate, see Hampton, 2016; Thomas, 2011; Walker, 2010). During the passage of the Social Security Bill 1973, two disabled peers (both reported in the Daily Telegraph as "speaking from a wheelchair") unsuccessfully proposed an amendment to increase the earnings limit. The proposal is discussed in some detail in a Department of Health and Social Security memo, which reinforces the thinking behind the permitted earnings rules:

The proposed amendment attempts to introduce an entirely new concept and provide for .. a rule which would in effect allow the benefit to be paid as a supplement to part-time or even full time work. It has never been considered a proper function of the National Insurance Fund to supplement earnings and it would clearly be inconsistent with the principles of which sickness and invalidity benefits are paid to make them available to people who are capable of substantial earnings (clipping from Daily Telegraph, dated 13 June 1973 and memo, dated 26 June 1973, in PIN 35/394)

The subsequent Social Security Act 1975 did not change the basic rules although it did formalize, by regulations, the rule which allowed claimants do carry out small amounts of work, within the earnings limits, so long as they had "good cause" to do so. The interpretation of "good cause" still required clear medical evidence and still led to complex legal debate (Bonner et al, 1991: 512).

The interpretation of the law required that such work was "trivial." In a case from 1978, a man who had been claiming benefit for about 18 months was spending his time "helping out" with his wife's business running a toy shop, going in on days when he felt able to help. His apparent working while claiming was reported by a local newspaper, showing a continuing mistrust of benefits claimants. The man contacted his local benefit office to explain but his benefit was stopped and he appealed. At his appeal to the National Insurance Commissioners, the Commissioner concluded that the extent of his activities

were not, in my judgement, work, whether part-time or whole time for which an employer would be willing to pay… Intermittent tinkering in the business of the kind and degree that took place in this particular case, is not 'work' in terms of the [legislation]. (National Insurance Commissioners Decision R(S)4/79, para 9).

The rules were complicated and it is not surprising that both benefit claimants and local offices had difficulty interpreting them. In the early 1980s, the legal cases discussed at the beginning of this article confirmed that it was possible for people to be doing "work" and still be eligible for benefit, so long as the work met the strict therapeutic and earnings rules (Mesher, 1986). These kinds of cases continued throughout the 1980s. Some examples include a woman whose benefit was allowed when she worked part-time in a bar in order to raise the money necessary to buy specialist contact lenses. She was considered to have had "good cause" for working. Usually, however, the requirement that the work was "therapeutic" was applied quite rigidly, based on a medical model of therapy (Bonner et al, 1991: 511-513).

Incapacity Benefit and the beginning of work preparation

A version of the permitted earnings rule was included when Incapacity Benefit was introduced in 1995, with an entirely new mechanism for assessing capacity for work (for details of the changes, see Bonner, 1995). It retained a rule prohibiting payment for benefit on "any day …which he does work," but again allowing the exemption in certain circumstances, including the "therapeutic" reasons and working in hospital workshops. New exemptions were added, including working as a volunteer, being a local councillor and sitting on certain social security appeal tribunals (Wikely and Ogus, 2002: 549).

These exceptions began to show a new pattern of exemption from the not working while claiming rule. The specific exemptions for civic duties such as being a local councillor or sitting on a social security tribunal recognized the value that disabled people bring to civic life. A more significant change was introduced in 2002 when the rules were amended to change the focus of permitted work from "therapy" to work preparation (Wikely and Ogus, 2002: 550). The therapeutic earnings rules allowed claimants to earn up to £66 a week if the work was explicitly therapeutic but claimants could also work in any work, with an earnings limit of £66 for up to 16 hours a week for up to 26 weeks. The rules for work in specific therapeutic settings marked this as having different status from other types of work which a claimant might do. The open "any work" rules for working up to 16 hours for 26 weeks introduced the possibility of work trial for claimants, which did not require the endorsement of a medical specialist, although any extension beyond 26 weeks did require authorisation that the work was "likely to improve his capacity to engage in full-time work" (Wikely and Ogus, 2002: 550). Research on the effect of these rules showed that although some claimants used the rules as a stepping stone to come off benefit, the large majority continued to claim benefits, although some continued to work part-time (Dewson et al, 2005). While a helpful snapshot of the experiences of some claimants at this stage of the development of the scheme, this research was unable to ascertain the extent to which the permitted work rules had caused this return to work. These new rules show the beginning of a move away from medicalised permitted work to a stronger focus on preparation for the labour market.

Employment and Support Allowance, Universal Credit and the end of permitted work?

The next major change in incapacity benefits occurred with the introduction of Employment and Support Allowance (ESA) in 2008. This new benefit came with a whole new package of assumptions about incapacity benefits, with a much stronger focus on getting claimants back into work and off the benefits books (for details of these changes, see Grover and Piggott, 2010; Gulland, 2011). The key difference between the old Incapacity Benefit and the new ESA was the creation of a new category of benefit recipients who, instead of being classified as "incapable of work" were reclassified as having "limited capacity for work" and who were now required to participate in a range of work-related activities in order to keep their benefits, with the threat of sanctions if they failed to do so. The story of the introduction of this kind of conditionality in incapacity benefits, the resulting sanctions and the effect on people's lives has been well rehearsed (Dwyer, 2016; Geiger, 2017; House of Commons, 2015; Patrick, 2011).

The introduction of ESA continued the explicit permitted work rules, with very similar exemptions to those under Incapacity Benefit, for being a local councillor, serving on a tribunal, voluntary work, therapeutic work and work related to preparing to re-join the labour market. The therapeutic work rule was not time limited while the any work rule was limited to one year for most claimants. This time limit was removed in 2017 in order to "improve work incentives" (DWP, 2017: para 7.1). The permitted work rules continue an older practice of distinguishing between people who are expected to bring themselves closer to the labour market and people who are unlikely to manage to do so. The therapeutic option harks back to older ideas about allowing people, who are not expected to fully join the labour market, to do some work. This can be seen in the DWP guidance on therapeutic work which suggests that this permitted work exemption:

is appropriate for claimants whose disability has stable and established effects with a significant impact on their ability to learn or sustain a traditional job which will always or for a number of years, prevent them from working a few hours a week (Department for Work and Pensions, 2013: para V3065).

The examples given in the guidance include a person receiving treatment for cancer who gives painting lessons "as a hobby" and people with learning disabilities working in supported employment. These examples, although not based on legislation, continue the idea that some disabled workers should be allowed to work while claiming, because of their status as less effective workers, or as a diversion, rather than because working will prepare them for the open labour market.

The permitted earnings rules in Employment and Support Allowance show both a continuation from the past and a broadening of the right to work while claiming so long as that work is explicitly related to returning the labour market and closely tied with the job-preparation conditionality built into the scheme. However these rules are soon likely to be replaced for many claimants by the transfer to the means-tested Universal Credit. Universal Credit differs from past means-tested incapacity benefits by removing the distinction between being "in-work" or "out of work." It starts from a basic presumption that citizens have a duty to provide for themselves through earning an adequate wage in the labour market. The state will top up any earnings below an agreed level but claimants are expected to make efforts to increase their earnings so that this is no longer necessary. Universal Credit is new in that it transfers this duty to seek more work, and subsequent harsh penalties for failure to do so, from those out of work to those in part-time or even full-time work (Dwyer and Wright, 2014).

Although Universal Credit is intended to merge all existing means-tested benefits and tax credits into a single system, it still maintains distinctions between people who are "unemployed" and those who have "limited capacity for work." This distinction is made by requiring all claimants to sign up to a claimant commitment, which will determine the level of work-related conditionality, with which they must comply. The claimant commitments include stringent expectations of work-seeking behaviour, subject to sanctions which can lead to benefit being stopped for up to three years (Universal Credit Regulations, 2013). Levels of benefit and of conditionality are dependent upon whether the claimant has no limits on their capacity for work, has "limited capability for work" or has "limited capability for work related activity." Those with limited capability for work related activity (the support group) are entitled to a higher level of benefit and, currently, no work-seeking conditions attached to their benefit. Recent proposals suggest that greater pressure will be put on even these claimants to participate in work of some kind. The Department for Work and Pensions and Ministry of Health proposed that, for those in the "support group," there is a:

plan to introduce a 'keep in touch' discussion with work coaches… which could be explored as a voluntary or mandatory requirement (Department for Work and Pensions and Department of Health, 2016: para 114).

The assessment of capacity in Universal Credit uses a similar points based procedure to that used for Employment and Support Allowance (Welfare Reform Act 2012: s13 and 19). This means that the distinction between people who have limits and people who have no limits on their assessed capacity for work will continue to be important for Universal Credit claimants. However, because Universal Credit does not distinguish between part-time work and full time work, the rules on permitted earnings have been dispensed with. Claimants will now simply have their benefits adjusted according to what they have actually earned. There will continue to be serious implications for disabled claimants who take part in paid work while claiming Universal Credit. The in-work conditionality rules will mean that people will be subject to close scrutiny regarding whether they are making sufficient effort to work longer hours. There is also still a risk for people claiming on the basis of limited capacity for work that the fact of their part- or full-time earnings will call into question their "limited capacity" status, leading, potentially, to a loss of income, loss of entitlement to benefit altogether or to a change in the conditions attached to their claim. For those who continue to claim the non-means-tested, insurance-based Employment and Support Allowance, the permitted work rules will remain similar to those from 2013, with an upper limit on permitted earnings and exemptions for people doing particular kinds of therapeutic work or civic duties. However, the extension of Universal Credit to include housing costs as well as living costs will mean that the broad net of means-tested conditionality will be spread to many of these claimants as well. This spread of conditionality will bring an associated increasing surveillance of people's everyday lives, as demonstrated by recent work by Manji (2016).

Discussion

The history of the concept of permitted work in relation to UK incapacity benefits is important because it illustrates the flexible boundary between work and not work in relation to disabled people. Incapacity benefits systems originally assumed that a clear line could be drawn between people who were capable and people who were incapable of work. These assessments have been largely based on a medical model of disability, relying on medical professionals to certify people as incapable of working. While some social aspects of people's lives have also been considered relevant, the boundary between capacity and incapacity has usually required a rigid medical approach.

Looking at people who have worked while claiming shows us that common sense assumptions about what is and is not work are heavily structured by social class, gender, participation in the labour market and expectations of "disabled" behaviour. While some disabled people have been expected to remain outside the labour market and have been permitted to carry out small amounts of "trivial" work for very little pay, others have been expected to use their work experiences to prepare to return to the labour market, or lose their right to benefit altogether. A clear trail can be seen, linking the earliest attempts to exempt some claimants from the "thou shalt not work" rules in the early 20^th century, through the permitted earnings rules in the late 1940s, changes to social security benefits in the late twentieth century and into the increasingly conditional benefits for disabled people today. Across the first half of the twentieth century we can see that the concept of permitted work grew out of the original "behaviour during sickness" rules. These rules were intended to prevent benefit claimants from doing anything that might endanger their health, or impede their recovery and was accompanied by close surveillance of their behaviour. The assumption was that it was the claimant's duty to recover as quickly as possible from whatever health problem was limiting their ability to work. However it is clear, from the policy makers' deliberations and the cases which reached the appeals, that other considerations also came into play. These included a fear that people might be attempting to claim benefits to which they were not entitled but also a belief that people ought to be making every effort to return to the labour market. Women were seen to be particularly problematic because decision makers could never be sure that they were really intending to return to paid employment. The developing rules on permitted work were a way of enabling some people to do work under strictly observed medical supervision in order to prepare themselves for the labour market. In the second half of the twentieth century the rules became more focussed on the concept of rehabilitation and preparation for work. While the focus turned to work preparation, the system retained some older ideas about providing diversions for people who would never be able to work in the open labour market.

In the twenty-first century, the UK benefits system has been converted into an almost entirely means-tested system, where these ideas about permitted work have changed. Benefits claimants are now expected to work if they can and to lose their right to benefit if, and when, their earning power increases. The very assumption behind the Universal Credit system is that people will increase their income from employment and decrease their "dependency" on the state, however unrealistic this assumption may be (Millar and Bennett, 2017). The conditionality attached to Universal Credit means that claimants will be subjected to close scrutiny of their daily lives. Universal Credit maintains hard and fast distinctions between people with different levels of "limitation" in their capacity for work. This puts claimants at continuing risk of losing their status as people with "limited capacity for work,", leading to a potential loss of income and the even more conditional state of unemployment and precarious wage earning of the non-disabled unemployed. The constant surveillance of new benefits regimes may lead claimants to be wary of doing anything that could be described as work. At the same time, most claimants will be expected to carry out more work or work-like activities in order to meet claiming conditions. A new revelatory sign is now needed to show that claimants are making sufficient effort to return to work.

In the past any attempts to work while claiming benefits and retain "disabled" status required people to subject themselves to a high level of surveillance by medical and rehabilitation experts who could confirm that the work was therapeutic or closely linked to a return to the open labour market. Thus permitted work was another area of benefits provision, along with medical assessment, where decision makers used revelatory signs to distinguish between different categories of benefit claimant. With the diminishing value of non-means-tested benefits within the UK benefits system, this technical concept of permitted work has almost, although not quite, disappeared. It has been replaced by residual welfare state which relies on constant surveillance of claimants' incomes, behaviour and working patterns.