Forty-Second Report - The Restart Scheme for long-term unemployed people

The Department has ended up paying more per Restart participant than for previous similar schemes, because it had to pay providers to rapidly build up capacity from a standing start and then did not need all that capacity. From 2010–11, when it was responding to the last economic crash, to 2020–21, the Department reduced expenditure on employment support from £2.9 billion a year to £300 million a year. So, when it announced details about Restart in November 2021, designed as a £2.9 billion scheme for 1.4 million people, it inevitably had to work with the market to build capacity and could not achieve as much competition on price as it otherwise might. However, the Department had significantly overestimated demand for Restart, partly because it expected there to be more eligible participants, and partly because it expected more of those eligible participants to be found suitable to start on the scheme. Providers had quickly built capacity for the high number of participants and now had costs they could not recover for things they did not need, such as leases on larger buildings than they would require given the reduced volumes. The Department did not plan for this scenario of lower demand, and it gave limited consideration to contractual mechanisms in the original Restart contracts to reduce costs or expand eligibility without renegotiations. As a result Restart is now expected to cost £2,429 per participant, making it significantly more expensive than the Work Programme, which cost the Department around £1,760 per participant. Recommendation 2: In its Treasury Minute response to this report, the Department should set out what lessons it has learnt from Restart about how it can better expand and reduce capacity for employment support as it is needed, at 6 The Restart Scheme for long-term unemployed people better value to the taxpayer. This should include an assessment of what standing capacity and capabilities it needs and whether it can better maintain the market be

The Department and providers are not working together and sharing information as effectively as they might to support participants into work. Both the Department and providers aim to provide coaching support to help participants into work. While it would have taken too long to link up the providers systems with the Universal Credit computer system so that all this information flowed automatically, the current information sharing arrangements make it harder for the Department’s work coaches and the providers’ Restart advisors to work together to move participants into work and may leave participants questioning the value of Restart. As a result, Restart participants face a duplicative and inefficient experience, where they have to tell Department work coaches in their fortnightly meetings what they are doing on Restart, and their Restart advisor what they are doing with the Department. Participants also have to give their Restart advisor lots of information that they have already given to the Department and have similar conversations with both their work coach and Restart advisor to identify barriers to work, and produce separate lists of actions they will take to overcome these barriers and move into employment. This list of actions is called an ‘action plan’ on Restart and a ‘claimant commitment’ with the Department. Recommendation 3: • The Department should ensure work coaches and Restart provider advisors always have access to key information about participants and their barriers to work, as well as the activities that participants have agreed with either their Restart provider or the Department to help move them into work. • The Department should undertake a review into how frequently participants are required to attend jobcentres while they are on Restart, to ensure attendance requirements on participants achieve the maximum value for money.

The Department did not know enough about its claimants to understand what support they needed and how many eligible claimants would go on to participate in Restart. The Department’s work coaches know a lot about their claimants, but much of this information is recorded as freeform text on a person’s benefit claim. This makes it difficult for the Department to collate and consider this information when it designs employment support. As an example, the Department said it had recently learned from providers that around 45% of people going onto Restart in some areas have English as a second language. Individual work coaches will know this about their individual claimants, but this is not information that the Department can aggregate to inform providers or use to ensure it is purchasing the support that claimants most need. Similarly, the Department does not have a vulnerability flag in the UC system, or a flag to denote mental health problems. The Department had expected its work coaches to find 82% of system-suggested claimants to be suitable for Restart and in reality, work coaches have only found 43% to be suitable. The Department did not run trials to test these assumptions about how many people The Restart Scheme for long-term unemployed people 7 would be found suitable as it felt it did not have time, that work coaches did not know their claimants as well because of COVID-19 lockdowns, and because claimant circumstances would change between it running a trial and launching the scheme. Recommendation 4: • The Department should set out how it will improve its record keeping so that barriers to work faced by a claimant, such as language difficulties or health conditions, are recorded and can be aggregated in the Universal Credit system to understand the type and scale of support the Department needs to provide, and so that providers have the best possible understanding of how they can help participants. • The Department should seek to establish the level of capacity

The Department does not understand how well each of the individual 77 providers are delivering Restart compared to their peers. The Restart scheme is delivered by a complex network of prime contractors and their subcontractors across 12 different contract areas in England and Wales, with the aim to help participants into work against the backdrop of the labour market in their area. Half of the prime contractors, while being responsible for their own contract area, are also subcontractors in at least one other area. The largest providers have their own systems which they use to assess participants barriers to work, as a directory of available support, and as a system to manage meetings and interactions with participants. Despite this, the regular management information the Department collects about Restart is all based on the performance in the 12 contract areas and does not cover how well individual providers are contributing to this complex network of support. The Department can collect information about how well individual providers are performing but it does not do this routinely or collate such information. As a result, it does not know how well individual providers are adhering to the customer service standards it introduced to stop providers focusing efforts on participants who are easiest to help, or how well individual providers are performing in getting people into work. Recommendation 5: While Restart is running, the Department should do more to collate and assess how individual providers are performing to increase transparency and competition between providers, and to identify pockets of best practice that might otherwise be lost when performance is compiled into a package area level. The Department should then seek to use this information as part of its evaluation.

Many claimants have complex barriers that prevent them from finding work, and some of these barriers may be better addressed through other means than an employment support scheme such as Restart. The Department’s work coaches must decide whether people from a system generated list of claimants are eligible and suitable for Restart. Work coaches decided not to refer 57% of the claimants on such lists. For the majority of these people, work coaches did not refer them to Restart because they felt they were unsuitable for the programme, most often 8 The Restart Scheme for long-term unemployed people because they had ‘complex barriers’ which can include homelessness, childcare needs and physical and mental health issues. The Department has identified mental health issues as a particular barrier that many claimants must overcome to move into work. We are concerned about whether this represents ‘cost shunting’ to the Department, where failures in public services such as long waiting times for mental health appointments are effectively passed onto the Department to deal with as part of its efforts to get people into work, with neither the appropriate funding nor expertise to manage these challenges. The Department expects to publish a new White Paper on health and work early in 2023, and that part of this will include how the Department can work more closely with the Department of Health & Social Care (DHSC) and others to help people stay in work, and the Department also told us that it is working with DHSC on mental health provision. Recommendation 6: The Department should set out, in its Treasury Minute response: • Its understanding of how complex barriers such as mental health problems and homelessness, which might not traditionally sit with the Department, impact on people’s ability to find work and the associated cost of this to society and the exchequer. • How it will develop and use its knowledge of claimants to help government as a whole to take a joined up and effe

On the basis of a report by the Comptroller and Auditor General, we took evidence from the Department of Work & Pensions (the Department) about the Restart Scheme.1

In the remaining 43% of instances a claimant was not referred, this is because the claimant was no longer eligible by the time they were considered by the work coach. This could be, for example, because they had found work or because they no longer needed to search for work as part of their benefit claim.11 Discussing this, the Department told us that the mix of people coming into Universal Credit was unusual, that the labour market was very dynamic, and that people’s circumstances were “changing all the time”.12

Work coaches use their knowledge of claimants to decide what support a claimant might benefit from and whether to refer people onto particular schemes. However, the Department is not able to use this information when planning the scale of support it needs to provide because it is not always recorded on the system, and when it is recorded this is as free-text notes on a person’s benefit claim rather than something that can be easily aggregated and used to predict demand.13 Work coaches would know, for example, if claimants had English as a second language and so might benefit from language support as part of Restart. However, while the Department told us that it had learnt from providers that around 45% of recent Restart participants in some areas had English as a second language, it could not confirm whether such information was available from its own systems.14 This presumably makes it harder for providers to plan that provision. The Department also did not run any trials to assess the number of claimants that work coaches would find suitable for Restart but said it was a “genuinely important challenge” about “whether we [the Department] could have done more to test the assumptions”.15

We asked the Department whether it could have had any better information or better data to predict demand, and it told us that it was keen to learn from its experience of Restart about how to predict demand, and later spoke about how some of its data mining work may be able to help. However on this topic it also acknowledged that Restart has “clearly taught us [the Department] a lesson”.16 Impact of public service failure on the employment support the Department must provide

Many of the complex barriers that claimants face, which can mean that somebody is less likely to be referred to Restart and that they might find it harder to move into work, are 9 Q 6; C&AG’s Report, figure 6 10 Q 7; C&AG’s Report, figure 6 11 C&AG’s Report, figure 6 12 Qq 2–3 13 Qq 34–36, C&AG’s Report para 2.12 14 Qq 26, 36 15 Q 4; C&AG’s Report para 2.3 16 Qq 7; 35–37 The Restart Scheme for long-term unemployed people 11 related to issues that the Department would not historically have the role of addressing. Homelessness and physical and mental health problems, for example, would all primarily be the responsibility of other government Departments (the Department for Levelling up, Housing & Communities and the Department of Health & Social Care, respectively). The Department told us that it is finding the mental health of claimants a particular barrier for people trying to get into work.17

Discussing mental health, the Department did tell us about a number of initiatives arising from a joint Department for Work & Pensions and Department of Health & Social Care unit which looks to target some of the mental health challenges claimants face. The Department told us that it had “paid for employment advisers to be actually embedded” in the NHS mental health service for England which “frees up the therapist’s time to spend on mental health, and then the individual also gets the job help”.18 Additionally, the Department told us that an upcoming White Paper on health and work, due “early in 2023”, will discuss how it can “work more closely with” the NHS, the Department of Health & Social Care and employers to support people before they fall out of work, or to give “lower-level support” before people get into “a more acute situation further down the track”19

The Department also told us about some other cross government work it is involved in, including working with the Department for Levelling up, Housing & Communities to avoid benefit recipients paying higher rents for poorer-quality accommodation in the private rented sector.20 17 Q 71 18 Q 75 19 Q 73 20 Q 72 12 The Restart Scheme for long-term unemployed people 2 Participant experience on Restart and scheme performance Participant experience on Restart

Restart participants continue to attend a jobcentre to meet their work coach every two weeks while they are on the scheme. Participants also have fortnightly meetings with their Restart provider employment advisor, half of which are in person.21 Participants have to repeat themselves in these meetings and tell their work coaches at the Department what they are doing on Restart, and their Restart advisor what they are doing with the Department. On the Work Programme, a similar previous employment support scheme, the Department ran a trial where participants were not required to meet with their work coach while they were on the scheme. The trial showed a reduction in the number of participants moving into sustained work, though this trial was conducted on a different employment support scheme and a different benefit regime, as it was before Universal Credit.22

Shortly after starting on the Restart scheme, participants undertake a ‘diagnostic assessment’ with providers, to understand the extent to which things like the physical and emotional demands of work, travel, literacy, numeracy, debt, housing, criminal convictions, and family life challenges are barriers to that individual finding work. Following this, providers will work with the participant to produce an action plan of activities they will undertake to move into employment.23 Work coaches work with claimants in a similar way, understanding their barriers and producing a claimant commitment that sets out actions that they will take to move into employment. Despite the similarities between the claimant commitment and the action plan these documents are normally developed independently, and the Department told us that providers would like more access to the claimant commitment, and more information about the barriers that work coaches know their participants face.24

The Department acknowledged that while an IT system that allows information to be shared between provider employment advisors and Department work coaches would be “perfect” and is something it would think about for future programmes, it also told us that it had looked to do this previously and decided not to because of concerns about funding, about whether such a system might act as a barrier to innovation, and because of difficulties in ensuring such a system was kept up to date.25 The Department told us that it needs to reflect on how it can ensure the experience is not “lumpy” for participants, and that Restart does not make people “grumpy”.26 The Department also acknowledged that it can do better in setting out the information that needs to be shared between jobcentre and Restart provider and that the warm handover, which was introduced to improve engagement between the jobcentre and the provider, is “not quite where we [the Department] would like it to be”.27 21 C&AG’s Report, paras 1.9, 1.10 22 Q 13; C&AG’s Report, paras 1.10, 1.12 23 C&AG’s Report, figure 4 24 Qq 18–19, 73; C&AG’s Report, para 1.12 25 Qq 19–20 26 Q 13 27 Q 20 The Restart Scheme for long-term unemployed people 13 Monitoring the performance of Restart and evaluating its impact

We have repeatedly recommended that the Department be more transparent and publish more information about the performance of its employment support schemes.28 However, until recently, the only information made available about Restart has been in response to Parliamentary Questions. While the Department did publish an ‘ad hoc’ statistical release in December 2022 about how Restart is performing, and this was reasonably detailed, it did not give any indication that this was to be the first in a series of regular and scheduled performance data releases about the scheme.29 The Permanent Secretary told us that the ‘ad hoc’ statistical release was “the start” and noted the value in transparency, while pointing to remarks made by the Secretary of State to the Work and Pensions Committee on 30 November 2022, where he said that he was “more minded to be transparent than not” but also spoke about the importance of “safe spaces” for Ministers and officials to develop policy.30

The Department used its experience of delivering and evaluating the Work Programme, a similar previous scheme which ran between 2011 and 2017, to help it to design and set up Restart.31 The Department’s evaluation of the Work Programme showed that, on average, scheme participants spent 46 additional days in work and off benefits in the two years after the programme finished, compared to people who did not participate. The Department used this evaluation when it put together its business case for Restart, which it estimates will bring sustained jobs for an additional 6 in 100 participants above what would have happened without the scheme.32

So far more Restart participants have moved into sustained jobs than the Department had expected, but this is in the context of historically low unemployment and high vacancies. The Department will not know the impact of Restart in increasing the number of people moving into sustained work until it has completed its evaluation, due in 2025, but told us that it “is reasonably confident” it would have the expected impact because the Department had been “cautious” in its assumptions.33 The Centre of Public Data and the Blavatnik Institute provided written evidence stressing the importance of transparency and the need for stakeholders and academics to be able to peer review the evaluation and stay up to date on the progress of the scheme.34 28 Committee of Public Accounts, DWP Employment support, Fifteenth Report of Session 2021–22, HC 177, 8 September 2021; Committee of Public Accounts, DWP Employment Support: Kickstart Scheme, Thirty-Ninth Report of Session 2021–22, HC 655, 25 February 2022. 29 Department for Work & Pensions, Official Statistics Summary: Restart Scheme to September 2022, 15 December 2022, available at: https://www.gov.uk/government/statistics/restart-scheme-statistics-to-september-2022/ summary-restart-scheme-to-september-2022 30 Q 29; Work and Pensions Committee Oral evidence: The Work of the Secretary of State for Work and Pensions, HC 549, 30 November 2022, Q2 31 Qq 28, 38, 47, 66; C&AG’s Report, para 1.8 32 Q 28; C&AG’s Report, paras 9, 1.16 33 Q 28; C&AG’s Report, para 12, 4.6 34 Written evidence submitted by Dr Eleanor Carter, Government Outcomes Lab, Blavatnik School of Government, University of Oxford, dated December 2022; Written evidence submitted by the Centre for Public Data and John Penrose MP, dated December 2022. 14 The Restart Scheme for long-term unemployed people 3 Managing the employment support market Building capacity of providers to deliver Restart

From 2010–11, when it was responding to the last economic crash, to 2020–21, the Department reduced its expenditure on employment support from £2.9 bn to £300 million per year.35 The Department told us that when it was designing Restart, it believed that “well over 2 million” people would be eligible and suitable for the scheme but that the maximum capacity the employment support market could build in the time available was 1.4 million people, and so it signed contracts with providers for that level of capacity.36 The Department’s focus on building capacity meant that it focused less on competitive pressures on price and performance and it set a minimum price and maximum performance expectation. All successful bidders offered the best performance they were allowed to offer, and the lowest price they were allowed to offer.37 The Department also looked to reduce the risk providers faced in delivering the scheme, by changing the payment mechanism so that providers received more income for delivering Restart at the front end of the programme.38

However, shortly after launching Restart, the Department realised that had it had significantly overestimated the demand, partly because it expected there to be more eligible participants, and partly because it expected more of those eligible participants to be found suitable to start on the scheme.39 To increase the numbers of people going onto Restart, the Department relaxed the eligibility criteria so that more claimants were eligible. It also revised downwards its expectation of how many claimants would go onto Restart, from 1.4 million to 692,000, and renegotiated contracts with the providers.40

The Department used the forecasts from the Office for Budget Responsibility to build its expectations about how much demand there would be for the Restart scheme, but acknowledges that it did not consider volumes falling as far as they did.41 The Department told us that it designs programmes to “cope with volume susceptibility”, but that the reduction for Restart was “well outside of that boundary”.42

The Department explained to us that the reduction in volumes meant providers had prepared for a “different scale of service”, and had committed to costs such as staff and building leases that reflected the scale of the programme they had originally been asked to deliver, but now did not need.43 This, coupled with a desire to maintain an uninterrupted service and a lack of flexibility in the contracts to manage such a low volume scenario, meant that the Department had little leverage in the contract renegotiations.44 In the end, 35 C&AG’s Report, para 3.2 36 Q 1 37 C&AG’s Report, para 3.4 38 Q 43 39 Qq 1–3, C&AG’s Report, para 2.8 40 Q 10; C&AG’s Report, para 13, 22 41 Q 51 42 Q 9 43 Q 61 44 Qq 55–56; C&AG’s Report, para 3.20 The Restart Scheme for long-term unemployed people 15 the Department could make only limited savings beyond those already achieved through the reduced number of ‘job outcome’ fees it would have to pay to providers, arising from the lower volumes.45

The Department set out in the Restart business case that it expected to pay around £1,800 per participant but, following the renegotiations, it now expects to pay around £2,429 per participant. The Work Programme, which ran from 2011 to 2017, cost around £1,760 per participant in 2021–22 prices.46 The Department spoke to us about the challenges it faces with contracted employment programmes such as Restart, and the requirement to suddenly build and reduce capacity dependent on how the economy is performing. The Department acknowledged that there is a question about whether to invest in long-term infrastructure for contracted employment schemes, which would make it easier to build capacity when it is needed.47 How the Department monitors performance on Restart

Although the Department signed contracts with eight prime providers to deliver Restart across 12 contract areas in England and Wales the prime providers can, and do, subcontract much of the scheme delivery to other providers.48 In total, there are 77 providers involved in the delivery of Restart. Providers have their own systems which they use to assess participants barriers to work, as a directory of available support, and as a system to manage meetings and interactions with participants.49

However, the Department’s routinely collected management information only shows the performance in each of the 12 contract package areas, rather than detailing the performance of individual providers. Its management information shows how many people are moving into work in contract areas and how regularly the customer service standards, which aim to stop providers focusing on participants who are easiest to help, are met.50 While a greater proportion of Restart participants are moving into work than the Department had expected at this stage of the contract, customer service standards have not generally been met by providers.51 When we asked the Department whether it could see which subcontractors were doing the best job, it told us that it can “see some of that data”, but set out its view that it is “it is the primes’ job to drive that management” of performance in their own contract areas.52

Several of the Restart subcontractors work in multiple contract areas, and half of the prime contractors also work as subcontractors in areas where they are not the prime. This has helped to facilitate the sharing of best practice, but it reduces competition between contract areas.53 45 C&AG’s Report, para 22 46 C&AG’s Report, para 23 47 Q 17 48 Q 26; C&AG’s Report, figure 8 49 C&AG’s Report, paras 4, 1.11 50 Qq 30–33; C&AG’s Report, para 3.5 51 Qq 65–66, 30–33; C&AG’s Report, para 12, 20 52 Qq 32–33 53 C&AG’s Report, para 3.5 16 The Restart Scheme for long-term unemployed people