Welcome to the Public Accounts Committee on Thursday 12 March 2026. The Access to Work scheme is a demand-led grant to help people obtain or stay in work if they have a physical or medical health condition or a disability. It is administered by the Department for Work and Pensions, and support for individuals is usually approved for three years, with a maximum total annual funding of £69,260 per individual. Both the demand and the spend on this scheme have doubled since 2018-19. The total number of applications reached more than 157,000 in 2024-25, with a spend of £321 million. Concerns have been expressed about the increase in the number of outstanding applications and payments, and the time taken to process applications, so today we will be looking to challenge the DWP on the issue and examine the impact of the delays and backlogs, which have trebled in the past three years, causing hardship to some applicants and to some small employers and charities. We will also scrutinise whether DWP’s administration of the scheme is aligned with the policy intent, exploring whether the scheme provides value for money, as well as looking at the DWP’s plan for reforming the Access to Work scheme. I warmly thank all those who have taken time to submit their evidence to the Committee: it has been a large number. We have been unable to publish all the submissions, but they have been very useful to the Committee and have helped to inform our questions today. We are very grateful. I would like the witnesses to introduce themselves. Sir Peter, you are welcome to the Committee. You are the permanent secretary, but will you introduce yourself briefly?

Thank you, Chair. I am Peter Schofield, the permanent secretary at the DWP. I am joined by colleagues alongside me—

We will come to them in a minute, but I just have a word to say: I want to thank Sir Peter very sincerely for his dedication and hard work over his eight years in this role. We think this will be the last time that you appear before this Committee, Sir Peter, so it is appropriate for me to pay tribute to you. You have always treated the Committee with the utmost courtesy and respect, and we wish you the very best in your future endeavours.

That is very kind. Thank you so much. I have to say that “enjoy” is not necessarily the word I always use, but I have always felt that this Committee has treated the issues that we deal with in DWP with great responsibility, courtesy and care. The topic that we are about to talk about is absolutely in that category, and your challenge has been fair, often getting right to the heart of the point, but holding us to account in a way that makes a real difference for the people we serve. Thank you for your kind words.

That is exactly how this Committee and Departments and their permanent secretaries should work, so thank you very much for those remarks. Neil, you are a regular flyer at this Committee, so will you introduce yourself, please?

Plenty of previous offences here. I am Neil Couling, the director general for DWP services and fraud. Access to Work is within my group in the Department.

Dr Helga Swidenbank, you are also a regular attender at this Committee. Will you introduce yourself, please?

Good morning, Chair. My name is Helga Swidenbank. I am director for accessibility, disability and disputes. Access to Work falls within my operational delivery area.

Thank you. Finally, but by no means least, Bill Thorpe, please.

Hi, everyone. I am Bill Thorpe, the policy director at DWP for working age and child disability benefits, which includes the policy lead for Access to Work.

I think this is your first appearance before the Committee.

Yes, this is my first time.

As Sir Peter has said, you will not necessarily find the experience enjoyable, but I hope that you find it beneficial. It is great to welcome you. We will now move on to our questioning, starting with a question from Chris.

Good morning, everyone. Sir Peter, let me quote from the NAO Report, paragraph 14: “The average time taken by DWP to process applications more than doubled, from 28 working days in 2021-22 to 66 working days in 2024-25, and has continued to increase. In November 2025, the average time taken was 109 days compared with DWP’s target of 25 days.” When do you expect to get the backlog of applications down to an acceptable level?

I will start where the Chair started: Access to Work is a demand-led grant scheme that is intended to provide support to customers who are looking to work, or who are in work, to top up the reasonable adjustments that their employer might have made. In figure 1 on the Report, you will see the process that we run through to assess customers for those Access to Work claims. I am sure that we will get into the detail of what it means to process the backlogs and go through those applications. I want to put it into a wider context as well. Work really matters, and Access to Work plays a massively important role not just in helping people to get into work from a fiscal and economic point of view, but in the difference that work makes. We know this: I see it when I travel around the country, meet customers and colleagues and see the difference that getting into work can make not just to their economic position, but to their wellbeing, their sense of self-worth, their sense of confidence and their sense of purpose. We want Access to Work to be a success within the wider context of the things that we are doing to help people into work and help people to stay in work, particularly people with health conditions and disabilities. The challenge that Access to Work has faced—it comes out really well in the NAO Report, as you started to draw out, Mr Kane—is the fact that the world of work has changed so significantly since the pandemic. Access to Work has been with us for 30 years or so. For much of that time, it has had a pretty stable client base and pretty stable volumes. What we have seen since the pandemic is a doubling in claims, as you point out, and we have doubled the number of people working on casework commensurately. But it is not just more claims; it is also a greater complexity of claims. Figure 3 shows that applications have doubled. Figure 4 has what we are seeing in terms of the complexity of cases: there are more people with mental health situations, and more people with neurodiversity. We are seeing a real change not only in the volumes, with the growth in the number of applications, but in the complexity. That goes back to where the Chair started: this is a grant scheme tailored to the individual. As figure 1 says, we have to go through that with a great deal of care to make sure that we are tailoring the support to the needs of the individual. As the complexity of those situations changes, as the nature of work changes and as the nature of disability changes, that whole process becomes much more difficult and more complex. The reason we have seen a backlog is that simply doubling the number of people working on Access to Work has not been enough. We have focused on throughput, so we have seen a loss of consistency in decision making. This is our process for addressing the backlog; I am getting to the heart of your question, Mr Kane, on what we are doing. We have doubled the number of caseworkers, but that is not enough. The key thing for us is to introduce greater consistency in decision making, so we have trained our colleagues to be able to assess whether employers are doing their bit to do the reasonable adjustments that they should be expected to do, to make sure that we are consistently applying the principles of Access to Work in the way we assess applications for support workers. Alongside that, once we have done that and got that consistency back, we are going to have a further increase in the number of caseworkers on Access to Work—we will recruit another few hundred into the team—and we are going to drive productivity as well. All of that, together, will start to slow the growth in the applications backlog. We have already removed the first backlog, which was the payments backlog: that has now gone. As we make the changes to consistency and decision making and as we make the changes in terms of increasing the staff complement yet again, and as we increase productivity, so we will see the applications backlog start to turn and start to improve. I am sure we can talk a bit more about what we are doing to drive productivity in the system and greater consistency as we go through.

We will come on to that.

Paragraph 18 of the Report says: “DWP told us it is currently seeking to manage the backlog of applications within the budget for the scheme, but this means it does not expect the backlog to fall significantly in the short term. It noted that it needs to balance the demand for Access to Work with its other priorities and commitments.” Could you talk a little bit about why you have not prioritised further reducing the backlog within the bigger scheme?

It is a good question about prioritisation. We prioritise, first and foremost, those customers who have a job offer, are waiting to work and should be ready to start work in the next four weeks: those go to the front of the queue. That is the key prioritisation, and I think it represents about 10% of applications. Many of the applications are from people who are already in work, but we prioritise the applications from people who are ready to start work and have a job offer. If you turn to figure 10 in the NAO Report, which is in part three—

On page 32.

Yes. If you look at that, Mr Kane, you will see the effect of the prioritisation. At the top, you can see that 40% of applications are cleared in fewer than 40 working days, and of those, 55,706 are cleared within the first 19 days. That gives you a sense that although the cases are complex to progress, the ones that we prioritise are the ones that are really time-critical, because the application is crucial for someone’s job application and their ability to start work. That is the key part about prioritisation. The second thing about prioritisation is prioritising consistent decision making. That really matters from the point of view of making sure, within the budget that we have, that we are funding within the context of the principles of the scheme. The principles of the scheme are set out in paragraph 1.4, and we have to make sure that we are using the budget that we have in a way that is consistent with that. In 2023, with a rather less sophisticated approach to trying to tackle the backlog, we drove throughput at the expense of consistency of decision making. What you therefore saw was that some of the decisions were probably not the right decisions consistent with the principles. One example of that, which you can see in figure 7, is the way that spending on job aides went up dramatically—tenfold—in just a few years. There was a real sense that we were not getting the right decisions on that, so the second thing we have prioritised is consistency in decision making. The third thing to prioritise, of course, is increasing the number of people in the team. That has not been enough on its own, but we are trying to make sure that as we grow the team, we also increase productivity. One thing we know we need to bring through—we have plans to bring it through later in the year—is a reform to the IT platform that underpins the whole process. That will help to provide the opportunity to improve productivity and avoid things like double keying, re-keying and some of the frustrations that the NAO saw—I have seen some of them myself—when they came to visit our teams not that long ago.

Bringing that future-gazing to the target, when can we expect the average time taken to process applications to fall to your target of 25 days?

I am not going to promise, for two reasons. First, I do not know what will happen to volumes. Volumes of applications have doubled, and I do not know whether that will continue. Secondly, other than for the priority group of people whose application is crucial to their starting work—I want to get that point across, and I think it comes across well in figure 10 that we are prioritising those people—it is more important to me to prioritise the right decision, as opposed to making the wrong decision more quickly. I need time to work that out, so it is a work in progress. My plan is to start to arrest the growth in the backlog over the next few weeks and months, as more people come through into the team, and then seek to see it falling over the next 18 months or so, I imagine. I do not want to be held to account on that, although maybe that is easy if this is my last time in front of the Committee; I just want to get a sense of the complexity and unpredictability of demand. The importance of getting the right decision means that ultimately I cannot be fully sure, but my plan for the next 18 months to two years is to get the backlog back down to where it should be.

You are talking about a demand-led programme, with numbers that can ebb and flow, but a target that is fixed at 25 days. Can you talk about how you look at that target? I am not asking you to tell me what you may have internally, but if the numbers treble and the target stays the same, and it is known at the start of the programme that it can change, how do you approach the value of that target?

It is an interesting one, because Access to Work is not like a benefit, where people are entitled to it and it is funded out of the AME budget. This is a grant that is funded out of the DEL budget, so each year I have to carefully budget for this with an assessment of demand. Forecasting demand has been difficult in the past, as I say, but I do have to make a forecast for demand; I then have to make a forecast for how likely I am to be able to improve productivity, and therefore how many colleagues I need to bring in to do this work, which is complex: it takes about six months to train someone to do it, so there is a lead-in time, and then I need to make sure that the decisions are consistent and of the right quality. I prioritise the timing when it is crucial to customer and whether they will get into work and be ready for a job. Subject to that, I also prioritise getting the right decision from the perspective of budgetary responsibility and accounting officer responsibility, making sure that the money goes to the people who are most likely to benefit from it, consistently with the principles set out in paragraph 1.4.

May I add one more thing, Mr Kane? I think that 25 days is achievable. We have 65,000 or 66,000 cases on the stocks at the moment. A normal head of work is about 10,000 cases, so the backlog is actually about 55,000 cases. If that were cleared, it would be possible to clear Access to Work applications for the 10,000, which would roll on as we cleared cases and new applications were made. We get about 2,000 or 1,500 applications a week, so it is possible to hit that target; it is the backlog that is stopping that at the moment. There are also some quirks in the data that are probably worth being aware of. The course of an application is taking a bit longer at the moment, as the Report sets out. If you have a change of circumstance during that, that adds to the clearance time for the overall claim. Even though we may have assessed the initial application and put that into payment, if a customer reports a change of circumstance to us, that suggests, in the way our data records things, that it takes even longer.

I get the fact that you have a justification for missing the target. I accept that that justification exists, given the numbers. What I am saying is that as the numbers in the programme can ebb and flow, a target of 25 days is an internal target to help you manage internal processes and expectation, but it is also an external factor that allows people who are applying to get a sense of when they should expect to see something. As you are sitting at an average of 109 days, compared with 25, the target seems to be a measure that is bordering on pointless. If you were sitting at 28 days and had a target of 25, I could see it, but the disparity between 109 days, which is your current average, and 25 means that the target itself seems to be causing problems—for you, for us and for people who are applying.

You are right that our customers need to know what to expect when they apply, and the backlog is a problem in that respect. We see, in other things we do across the operational delivery of the Department, that when you have a big backlog, getting through each case takes longer: by the time you get to it, the situation may have changed, or you may find that people are understandably phoning up to chase progress, so the person managing the case is answering the phone instead of processing cases. So you need to get ahead of the backlog, and our plan will help us to do that. Once you have got ahead of the backlog, it is much easier to manage the cases coming in in real time, because the process is streamlined and you are not affected by the things that I have just described. We have seen that across the board in other aspects of service delivery.

When you say that you are confident you can get to the 25-day target but you cannot tell us when and you are not prepared to put a figure on it, I understand why but it does not give us the confidence we need to know that you are moving towards the target being met. If this 25 days should be 40, to give us a realistic sense of what the target is, that would be helpful. I am left dissatisfied with the fact that, in one sense, you are saying it is achievable, but you are also saying that you are not sure when you can do it.

I understand why you are saying that, Mr Kane. I want to give you confidence in the rate of progress. The trajectory is not going down at the moment. That is why this is quite a difficult moment to be trying to reassure you, but the trajectory is turning. When the trajectory starts to go down, we will then have a trajectory that shows us when we will get down to 25 days. The confidence I want to give you is that I think we are doing the right things now. We are not only growing the number of people working on Access to Work; we are improving the consistency of decision making and we have plans to improve the productivity of what we are doing. By way of arithmetic, unless demand continues to grow again, that will inexorably reduce the backlog and get us down to where we need to get to.

As Peter said, there were actually two backlogs—it is in the Report. There was a payments backlog—those were people already granted Access to Work who were submitting receipts into us for payments of their grants. That was about 28 days in the Report; that is now back down to its standard of making payments within 10 days. We prioritised that because these were people who already had awards and had employers expecting costs to be reimbursed and so on. That is why we went there first. By the end of this month, staffing will be up to 648—it was 588 when the National Audit Office came in and completed its Report. You can see work is in hand, but I do get the point you are making that it would be easier if we could say at some point, “This case will take you this long”. The variation in the cases, and because we prioritise those 10% of cases that are actually starting work, means that it is very hard to give that kind of general assessment at any one point in time on how long cases are going to take.

I want to reinforce Chris’s excellent questions. We also have to be very considerate of those applying for these grants. As I said in my opening remarks, we are talking about 109 days—three to four months. I accept that you prioritise those—you have said this very clearly—in urgent need of them, because they want to get a job, but nevertheless, this is causing hardship. It is causing people not to get the jobs they have applied for, it is causing businesses to go out of business because they are not getting the money that they expected, and it is causing hardship to some charities. When do you think you might be able to get the backlog down, given that Neil has said that the backlog is really the problem, and not the increase in applications? I accept that you have your Department expenditure limit—the DEL—but would it not pay you to surge the number of people actually dealing with these cases to get rid of the backlog in a relatively quick time, so you can get back into normality in processing the new applications?

The things you just described, Chair—we are absolutely trying to avoid them. You talked about people missing out on jobs. The reason why we prioritise those people who have a job that they are waiting to move into is to avoid exactly that. You say that businesses are missing out on cash flow. The reason why we have prioritised eliminating the payment backlog is to remove that. I would push gently back on the assertions in your question that these are some of the effects. That is not to say that there are not cases where things have gone wrong. I am very happy to look into those individual cases but, by and large, our whole approach is to avoid the things that you just described and to make that a reality.

The thrust of my question—the crux, if you like—is this: wouldn’t it be better for everybody if you put a surge of resources into it to get rid of the backlog quicker, so that you could get back to normality?

That is what we have tried to do in the past, and it has not worked. You have seen that we have doubled the resource already and are seeking to grow it further. However, these are really difficult decisions. As figure 1 says, this is not just assessing someone for a benefit claim and checking entitlement on three criteria. You must look at the claim in great detail. You must understand the nature of the disability and the work environment they are operating within. That often requires great skill, sometimes proper external assessment, and often a lot of conversation back and forth. In that interim journey, the customer will often have regular contact with the case manager, who is asking for more information to get more detail so they can make the claim. Crucially, they are pushing back on the employer to ensure the employer is doing what it should. Many employers are not and have not done sufficiently. The danger of surging it is that you put people in who are less experienced and less enabled, or you try to make decisions too quickly without looking at all the evidence. That is what we saw when claims and expenditure went up on things such as job aides which, when we looked at them again with a consistent decision‑making hat on, were the wrong decisions. You then end up with a situation where someone comes to the end of their three‑year grant entitlement, it is up for renewal, and we say they had been awarded the wrong grant if you apply the rules consistently. When we started applying them consistently, people said, “What has happened? My circumstances haven’t changed, but I am getting a lower grant award.” That is the danger of surging. I have been doing this role for quite some time, and I have been looking at Access to Work over the years. For all the reasons I said at the very beginning—because Access to Work matters so much—we are desperately trying to see what we can do to drive down the backlog. The risk with shortcuts is that the problem gets worse. That is why we are doing this in that consistent way. That is why this is a really challenging issue. This is at the heart of the problem we are solving.

I get all that, but let’s try to get something out of this exchange. To manage expectations, would it not be possible at the beginning of a person’s claim to give them at least an estimate of how long it is likely to take? That would manage their and their employer’s expectations. If it is going to take three months, fine—but they need to know that.

That is very fair. Helga, do you want to say a little more about what we do and how we do that? It is a fair challenge.

It is a fair challenge. When a customer calls our Access to Work helpline, we have a voice recording that will tell them how long it is likely to take for a decision to be made. At the moment, we are saying 37 weeks, so that information is out there and accessible to customers—

Can I stop you there? I hear Blake complaining. He is quite right to do so—37 weeks is more than six months.

That is the longest it will take. That is what we are advising is the longest it may take for some customers. It is worth noting that some of our decisions are relatively easy to make, and others are much more complicated.

The crucial thing is that at the point you apply, we also say, “Please tell us if there is a time‑critical element to your application.” As I said earlier, if you have a job that you are ready to start within four weeks, we prioritise it. If there is some other reason why it needs to be prioritised, we prioritise it. As Neil said earlier, the majority of people on Access to Work are already in work; it is time‑critical for everyone, but we have to prioritise. It is a way of managing expectations but also trying to draw out from the applications at the start who are the ones for whom it is most time‑critical and we need to focus on.

I was going to offer a glimmer of hope. In the Report, the average time taken was 109 days. It is starting to turn. I am not saying this is acceptable or that I am pleased with this, but I am encouraged that it is now down to 106. I have now worked for over 40 years in a lot of backlog situations and cleared a number of them out; it is one of the things I have done since Peter appointed me into this role across a number of service lines. Once you start getting some momentum there, the numbers do start to fall. The situation is starting to turn, through the actions we have taken. But we prioritise those in work, or those about to start work, and then the payment backlogs first, because to us that seemed like a logical way to deal with the problem that we had.

I want to bring in Blake; Chris, you have had a good long time. Blake—I heard your frustration.

It goes back to the complexity of the caseload coming in now. The 2.8 was a standard. In fact, I will go back to what I said at the beginning. Access to Work started in 1994; for many years, there was a pretty stable caseload. The conditions coming in, in terms of the applications, were relatively stable from one year to the next. What we have seen is not only a doubling of claims but a really dramatic change in the nature of the claims coming in and of the health conditions that we are talking about—more mental health conditions and more conditions around neurodiversity. There is not a standard way of looking at what would be required in terms of reasonable adjustments. You have to think about each of those conditions individually. They could be fluctuating conditions; they could be conditions that impact in particular circumstances in a different way. Obviously, the nature of the workplace has changed dramatically as well. To get this right—to get the right decisions—is a really complex piece of work, and we have got to get it right. And by rushing it, I would say that we have made the wrong decisions in the past. The crucial thing, first of all, is to get the decisions right. What have you seen with that line on figure 12? It shows productivity going down before it starts to go up.

Yes, we can come on to that.

Remember, what we were also doing at the same time was significantly increasing the number of colleagues working on Access to Work.

No, no—sorry, you’ve got to the heart of the question. When you bring in a new member of staff, they are not as productive. And when you bring in a new member of staff and they need to be trained up, the colleagues training them up are often the most experienced staff, who you have to take off the work. What we always see when we recruit—when we build the capacity in a service line; Access to Work is no different—is that there is a downward shift in productivity before it goes up, and that is what you see. That is the trade-off we often have when we recruit. It is the right thing to do long-term to build capacity, but there could be a short-term hit on productivity. That is what you saw then. As we recruit again—Neil can say a bit more about our plans to recruit again in the future—we are likely to see a further dip in productivity. How do we get around that? You will also see in the Report a lot of the things that we have been trying to do to drive increased productivity. We have been trying to increase standardisation, with a standard operating process and some standardisation of some of the decisions that we make on things such as travel-to-work costs and support worker costs. Further on from that, what we can talk about later as well, if you like, is our plans to improve the IT support. There are a whole host of things. At the moment, we are not at 2.8. We want to do a work study to work out whether, given the change in caseload nature and what it now takes to make an appropriate decision, we can get to 2.8. We will do a work study to assess that. We will wait and see. But we clearly need to drive productivity up, because we need to make the most of the colleagues coming in to build the capacity.

Can I just add to that? The Report also mentions a lot of manual process, which, in this day and age, does seem to be quite archaic.

It really is. I have sat alongside colleagues in Harrow as they have had to do that, and it has driven me to say, “How do I accelerate our plans, to get this right?” I will get Helga to say a bit about the standard operating process in a minute, but I will cover the IT first. What the NAO found is that there are basically three IT systems. We have a more modern system for the application, to enable customers to do their applications online. Then, at the end point, we have the process that we use to assess the receipts that come in, to make the payments of the grants that have already been awarded, which is also a more modern system. In the middle, there is this legacy platform—a case management tool—that has not been developed in recent years at all, going back to when Access to Work started. That does not talk effectively to the other two systems. Over the summer—we have got to that point in our modernisation programme more broadly, and I am really keen to drive this now—we will be introducing a modern platform for that middle stage, for that case management system. We are saying now that that will come in in June this year. That will then create a platform that can connect effectively to the other two platforms. As we build the systems off the back of that, we can then introduce a modern case management system that avoids the need for re-keying, and we can build on from that much better data collection that enables us to target our work more effectively. It is a starting point; the foundations go in in June, which will enable us to avoid the problems that we have seen in the past. This is the nature of what we have inherited across DWP over the years: many, many legacy systems. It has been taking us time to address each of those. We talked about this in the hearing on customer service a few months ago, with Helen Wylie there as well. There is a programme of work that gets to Access to Work this summer. That is the platform—the foundation—and we will build from there. In time, we will no longer see some of the outdated processes that the NAO saw when it was visiting our teams.

They will be integrated, so they will work effectively. DWP has many modular systems. We are basically developing modular systems that talk to each other. Each one can then be modernised without switching off the others, which enables us to do modernisation in parallel and make progress more quickly. We talked a bit about this in regard to the customer service system. But that is only one element, and that is to come. Perhaps Helga can talk about the standard operating process now.

The standard operating process is probably one of the key planks for helping us address the productivity challenge; Peter has already talked about the volumes, the complexity and the staffing levels, and we have just talked about our digital initiatives. We introduced a new standard operating process late last year with the intention of ensuring consistency of decision making across all of our decision makers. We are rolling that out office by office, and we hope to have that rolled out by this summer. That will allow us to create a consistent platform across all our decision makers and case managers. That will allow us to do the work study that we talked about a moment ago. That work study will then allow us to understand how long it will actually take, and what our aspirations should be, in terms of clearing our cases and our backlog. That is the intention: by the summer, we will have a consistent level of delivery across all of my teams, which will allow our work study to take place, and we will then have much more clarity about what is needed to ensure that we are able to deliver for our customers.

We will see more consistency, which will then allow us to do the work study, and the work study will take place in the summer. When that is completed, we will have a sense of what our productivity targets should be and how we want to resource those.

We will still face the issue—it is important that the Committee clock this—that if I recruit new people into the service for Helga, they will initially have productivity of 0% or 0.1%. When you average things out, that will drive the average down for a while. It takes about six months to train and then six months to become fully proficient, so there will be a lag in the productivity numbers. If you were to ask a parliamentary question of us and we came back to you with an answer, it would probably say something like that. But it is the right thing to do to standardise the processes, make sure the decisions are correct and then release activity aimed at dealing with the backlog problem that we have.

No, please do.

We will do the work study with experienced Access to Work assessors so that we can judge. The new caseload mix, as Peter was saying, has changed. In 2018-19, 58% of recipients of Access to Work had a physical limitation; that is down to 28% today. This caseload has markedly changed over the last few years, so we need to reassess all that and work out what is possible. That will then inform what the correct level of staffing is for the scheme once we have cleared the backlog, so that we do not get into a new backlog situation.

You can come back to it at a time of your choosing.

I would suggest that a year’s time would be a good moment. Let’s take the different stages. We need to recruit more people. That is a six-month process, so in six months’ time, you will be able to assess whether we have built the capability, the capacity—where are we in terms of the numbers? The process of improving the impact of the planned operating process and getting productivity up you will see, I think, in six months’ time after that. So in a year’s time, you should hopefully see capacity working with greater productivity. And six months later, hopefully, you should see the backlog—this is easy for me to say, given my position—all but gone, but let’s see. As I say, it is very easy for me to say that compared with my colleagues.

Thank you!

But I think that if you come back to it on a six-monthly basis, you should see those stepped improvements, bit by bit, if we are delivering the plan that we have described this morning.

Before I move on to the next question, Sir Peter, you have heard the Committee’s frustration about the 37 weeks. I hear that you are bringing in the new system by the end of the summer. Knowing the power of AI to sift information and come up with the answer in seconds, I would have thought that if this new programme is designed properly, it should produce a step change in productivity and you should be looking at a much greater target than 2.8 cases per person per day as a result of the new technology.

Neil should come in on this, but the point I would make is that with AI—you know this—it’s all about data and data quality and whether the data tells you—

Sifting information.

I am a big believer in AI. There is a slight scepticism. We are talking about individual cases: people can have the same health condition, but it can play out in a completely different way in different work settings, so there are a lot of variables and data that you need to draw together to make a decision. I think AI augmentation will be able to help decision makers make their decisions, as it could help them sift through the evidence that is there, but I am not sure about the idea that you could press a button and just find everything out. Sorry, Neil, you should answer that.

I just think that it is not—

Very briefly, Neil, having said what I have said on AI, I do accept that it always needs the human judgment to analyse the data that it is producing. We have had a discussion on this issue, and we have discussed productivity and backlogs. We have a lot to cover, so we might just move on. I am going to ask Amanda Hack to ask the next question; in doing so, I extend a very warm welcome to her, as a guest from the Work and Pensions Committee, and also to Steve Darling—you are both very welcome at this Committee.

Thank you, Sir Geoffrey; that is very kind. Thank you for allowing us to join the Committee for today, and I also thank the panel for coming along. It is very hard not to be distracted by the 37 weeks figure—I had my first baby quicker than 37 weeks. We do not really appreciate the fact that, behind the figure, there is a person waiting for their claim, or an employer waiting to employ somebody. That is a really difficult premise—the fact there is an individual sitting behind each claim. There is clear evidence that the backlog is having an impact, whether that is on job security for those with a current open claim, or on employers taking on a new person. In my previous life as an employer, I went through the Access to Work process with somebody we employed. It is an unbelievably difficult thing for a brand-new employer, when starting a brand-new relationship with a brand-new employee, to then have a really difficult start in those opening few weeks of working with them. What evidence do you have on the impact that the delays and backlogs are having on those individuals, as well as the employers themselves?

I shall start, and my colleagues may want to come in. I completely get where you are coming from, Ms Hack, and the heart of your question goes to the impact of this. There is a numerical answer to your question, which is covered in the figure 10 graphic. Turning to figure 10, in the specific situation that you have described—where someone is waiting to start employment—they should be covered by the bar on the far left-hand side. Of the total number of applications that we are talking about, 55,000 were dealt with in the first 19 days. The tail of claims on the right hand-side often tends to cover self-employed customers, for whom there is a lot of work to be done in assessing the viability of the business plan. It is about getting to the heart of not just the aspects of the disability, and how it impacts on the person’s work, but the viability of the business and whether it is a business that we should be supporting through Access To Work. That is a really skilled job that needs to be done. It is about getting a sense of the different types of claim. When we talk about the average number—we often talk about this issue in the Work and Pensions Committee—it hides a tail to both the left and right. I would just urge customers, if there is a time-critical aspect to their claim—such as when we describe someone who is waiting to start work but cannot do so until the claim comes through—we need to identify that at the beginning of the claim process. Also, please get in touch with the teams in Access to Work to see whether we can prioritise you, as we should prioritise those claims where someone is waiting to start work. That is the human side. It is difficult—I go back to my answers to Mr Kane in terms of how you manage the doubling of the caseload and the additional complexity of cases and still get the right decisions and avoid people missing out on work. I said this at the beginning: at the heart of this is the fact that we recognise that work makes a difference; work is massively important to people’s lives, and we want people to be supported into work. That is why we are putting so much resource into Access to Work. It is about trying to address it in a segmented way so that we identify where we need to prioritise according to where the cases are most time critical. Hopefully, that gives some degree of reassurance to customers who might have got that initial voicemail when they made their original claim. Giving them the sense that there are ways to get to the front of the queue if your case is time critical because you are waiting to start work is a really important element to what we need to do to reassure customers.

When you want to start a claim and you would like to start work within 28 days—four weeks is pretty normal—I suppose that voicemail must be really hard to take. That is not a great start to the relationship with DWP and an Access to Work claim. I would really plead that you change that opening gambit with a member of the public.

That is a fair point. Helga, you might want to explain where the 37 weeks comes in, but the first thing you do when you make an application—it is on the online thing—is answer, “Are you starting work in the next 28 days?”. If you are, that automatically accelerates you to the front of the queue. How is that expressed in your telephony?

You have seen the telephone is not the only way to contact the DWP Access to Work teams; it can be done online, and 99% of our customers do online applications. As Peter was explaining, there is a box on the online application that identifies when you are about to start work, and if that is within four weeks, it will be fast-tracked into my teams that deal with people who are new starters. Although the voice recording is one way of contacting us, there is another way, and the vast majority of our customers contact us through the online portal.

We should check: does the voice recording also mention whether you are starting work in the next four weeks?

Yes.

Okay. Then it goes to the front of the queue. Hopefully, that gives some reassurance, Ms Hack.

The backlog has not appeared overnight; it has been coming for some time. Again, as an employer before I came here, the performance was not great in previous years. When was the last time that the 25-day target was hit?

Pre-covid.

I do not think it has been hit since we have seen the doubling in cases, so since covid.

Since covid. Blake was talking about productivity; if we recognise that the 25-day target has not been hit since covid—it has been five years, and we are only expecting to see improvements by the summer—what work has been done before now to start that improvement? One of the frustrating things, and something we have talked about a lot on the Work and Pensions Committee, is understanding the people we are dealing with. The need to safeguard them—there are lots of people with mental health difficulties—and to support them more quickly has to be something that we prioritise. What are you doing to try to mitigate the impact of delays, considering this has been something that has not appeared overnight, but been years in the making?

There are two questions there: how did the backlog emerge? What were we doing over the last five years, and what are we doing now? The backlog emerged because of the doubling of claims and the complexity of claims. What have we been doing? I can assure you that we have not been sitting back. The first thing we did was to double the number of folk working on Access to Work. Then we streamlined the processes. That is all set out in part 3 of the Report. We then sought to introduce more standardisation, but that has proved not to be enough. That is why we have had to go back and do things in a different way. By focusing on throughput, we were making the wrong decisions. We were not focusing on consistency and decision making, so we did not find ourselves working through the backlog in the way that we needed to. The things that we would often do when we have a backlog—put more people on to the work and drive capacity—have not been sufficient. The next thing we often do is streamline the process, but that has not been sufficient either. From 2023 onwards, listed in the Report, a whole load of actions that we tried to do did not work. What we did not do—which is what we have been doing over the last two years—is fundamentally to recognise that the caseload and complexity had changed, and that we needed to go about this in a different way, with different training for our people and a different operating process, so we could get the fundamentals right. That is why we have confidence today, but that is not to say that we have not been trying huge amounts of work. Huge amounts of resource have gone into this. I doubled the budget on Access to Work over this period. There has been a lot of attention, but we have not done it the right way, because we have not got to the heart of what had actually happened. I am not sure we recognised immediately that the caseload had changed. This was not just a volume problem; it was that the world of work and the nature of disability have changed. With many more people with health conditions that relate to mental health and neurodiversity, some of the standard types of support that Access of Work was used to providing for the 20 years prior to that were not applicable any more. We needed to get a better assessment of what effect the changes in the nature of disability had on the need for adjustments in the workplace—and the nature of the adjustments that were needed. It is not an easy story to tell, and I appreciate where the Committee is coming from—I feel that too—but we needed to get the fundamentals right. All along, we needed to prioritise those people who were most time critical, because those people who are waiting to start work in the next four weeks get prioritised. As figure 10 shows, they get addressed within a matter of days rather than a matter of weeks.

Thank you, Amanda—very helpful.

I will start, and then Bill might want to come in. The criteria and the boundaries of the scheme are well set out on page 14 of the Report. What you see there are the eligibility criteria, but in particular, the five principles of Access to Work. When we talk about more consistent decision making, we are seeking to not change those principles but apply them. We can have a debate about the ongoing policy, but were we actually applying the five principles of the scheme appropriately? Something that we identified through our management investigation and an internal audit report a couple of years ago, with Neil overseeing this whole area, was that we were not. The Report sets that out in two particular areas. The first was around employers not fulfilling their responsibilities under the Equality Act; they were expecting DWP to do the reasonable adjustments that would be reasonable for them to pick up. We can give you examples of big employers who put in applications to Access to Work for an ergonomic chair, which is something that I would expect an employer to provide—it is certainly something that I provide to my colleagues at the DWP. Likewise, with people claiming for jobs, the support worker plays a massively important role for so many customers, but we were seeing job aides whose role was not to help level the playing field for customers with disabilities, but more to do a task that an employer would normally take on an additional employee to do. It was misusing the scheme in a way that was inappropriate. When I talk about consistency, it is actually about getting us back to the principles set out here. The reason why we were not there goes back to Ms Hack’s “What were you doing?” question. We were telling people to prioritise throughput. I was saying, “I want to get this backlog down”. The danger of prioritising throughput at the expense of consistent decision making is that you make the wrong decisions, and with hindsight that was the wrong thing to do. It was done for the right reasons, because we all wanted to get the backlog down, but it had the wrong outcomes in terms of what we saw coming through. Figure 7, the spend on job aides, is one example of that—it went up tenfold in the period. We have seen growth across the board, but nothing tenfold, implying that we were not applying the five principles of Access to Work in an effective way. Bill, do you want to say a bit more about this?

The scheme is incredibly personalised, and people really value that. We could make it more prescriptive and say, “We are going to pay only for these 50 pieces of equipment”, but people would not like that. Last year, we had 30,000 responses on the Access to Work bit of the pathways to work consultation. The overwhelming response was that we needed to retain a flexible and personal form of support. That means that you need fairly loose guidance because, as Peter has said, you need to understand the person’s impairment, the specific job role that that person is doing, the thousands of different types of jobs in the economy and the size of the employer, which influences what it would be reasonable for the employer to provide. That is why the guidance itself is quite loose. The case manager has a very difficult job to interpret that and make a decision, on the balance of probabilities, about whether a person is eligible for support. We could be more prescriptive and have bands of support, and say, “We will have bands of £0 to £2,000, £5,000 or whatever”, and we could run the scheme much more quickly and cleanly, but it would not help people or deliver the outcomes we are seeing at the moment.

No, what we are saying is that the scheme has an element of prescription, but the five principles that I described were not being properly applied. Our concern was that we were not properly applying them in a consistent way, in the way that we should. It created the outcomes that I described, particularly around things such as employers’ reasonable adjustments and things like job aides, but it is legitimate to ask how reform should move forward in terms of Access to Work. As Bill described, that relates to the responses to the Access to Work Green Paper, where we had a number of options, one of which was that we should be more prescriptive. It was an option that was put in there just to see what people thought, because in theory it could be an easier way—we could process the application more quickly if you just said, “This is this situation; this is that situation”. I think I am fair in saying this, Bill, but the sense of the responses is that people like the fact that it is tailored; they would just like the decision made more quickly. That is what we are working on. We have not got it right up to the current plan, but I am confident about the current plan being the way to deliver that.

There are two things there. One is asking employers to do more. Secondly, there is the specific point that came up in the Report, which I might get Neil to talk about in a second, on third-party organisations that have a vested interest in identifying a particular need, because they make provision to meet the need and make a profit out of that. Regarding the employers, the crucial thing—I will talk specifically about Access to Work but I will briefly place it in a wider context, because this really matters—is getting into a process where you ask the customer what conversation they have had with their employer about the reasonable adjustment, and you follow that up. It is about giving our caseworkers confidence, and that research and challenge are an important part of the process. It is time consuming and it can be difficult. I have sat with colleagues and seen applications coming from big employers that literally have Access to Work departments whose job is funnelling claims to DWP. On one level, I do not mind that if they also have a bigger reasonable adjustments team that is looking at how they, in their own department, can do what they should be doing already under the Equality Act. I place that in the context of a wider debate. I talked at the beginning about the massively important role that work has, not just in terms of the economy but in terms of personal wellbeing, sense of purpose and confidence. Employers have a role in making that a reality. I link this conversation—I will do so only briefly—to Charlie Mayfield’s work for the Department on “Keep Britain Working”. There is a real debate about what the best employers are doing. Many people who are out of work because of health conditions developed their health condition when they were in work, so what can employers do, not just about reasonable adjustments but about identifying and helping to support people when that health condition first starts to develop? Often, it is a reasonable adjustment or a type of support—Charlie is looking at workplace health provision in particular—but how can we make sure that interventions in the workplace can help to avoid people developing health conditions in the future? Charlie is leading vanguard work with 100 employers and 10 mayoral combined authorities, as well as with small employers, to help us to understand what best practice is. I just wanted to place reasonable adjustments and Access to Work in the wider context of employers’ role in helping to improve the health of the nation and the workplace, and helping people stay in work.

The Report has a survey of cases. We focused on the consistent decision-making process and asked how many awards were probably inappropriate. There are some statistics in there; I cannot remember which page.

The independent consultancy looked at this and thought that 69% of the funding to employers was correct, and that employers should be meeting about 31% of the costs.

That gives you a sense of the magnitude.

Figure 7, on job aides, which we have alluded to before, suggests that we were making mistakes on cases in 2023-24 as we attempted to clear the backlog in too much of a hurry, as Peter suggested. That is probably what some of you are seeing in your postbags. People did not know that we had made a mistake on their case, but those cases are now coming up for renewal and are producing lower awards, so they are asking, “Why have I got a lower award? Nothing has changed in my life.” The answer is that we had wrongly given them a job aide for, normally, 100% of the time, when we should have given is for about 20% of the time. The job aides are not designed to do the work; they are meant to lift the disabled person to the same level as another employee. There is evidence in the data that Access to Work, instead of being based on the skill level of the task, is based on the skill level of the person. For example, a job aide should be doing admin tasks, but not actually doing the work that the person is employed to do. Everything in the job aide data suggests that some of the money will be correct, but some of it will be incorrect. Over the next three years, as we work through the cases that were made without the consistency that Peter has been talking about, you will see those awards fall—not to zero, but to what the scheme actually provides for.

If you have page 27 open, it is that light blue line. That figure is all spending on support workers of different types, and the lines rightly go up, because we are seeing growth in demand and the demography, but the light blue line—which is job aides—makes that dramatic tenfold increase. You might have thought that the gradient of growth would be more in line with the other types of support worker. Obviously what we will do over time is apply consistent decision making to all the different types of applications coming in. We will keep this data up to date, and it will enable us to understand the impact of the errors that we probably made in the past.

We will take a break in a minute, but Anna Dixon has a supplementary question.

Apologies to everyone for my late arrival. Further to the figures relating to support workers, particularly for deaf people, the support of a BSL interpreter is one of the categories that has increased most proportionately, yet the feedback coming from the deaf community is that they are experiencing reductions in the number of hours that they are able to secure for a BSL interpreter. In fact, some deaf people are having to give up employment altogether. I wondered whether any of the witnesses might comment on how to ensure that people who need BSL interpretation basically for the whole of their working week are to be given sufficient funds to secure that support and so remain in employment.

It is an odd one, because if you look at the principles of the scheme, BSL interpretation—if that is the sole role, as you describe—is something that absolutely helps that customer to level up to where they need to be. In those situations, where on renewal there has been a significant reduction in the award, they have to go through a second pair of eyes within the team, to help us to understand that and the quality control, and to make sure that the right decision has been made. I cannot comment on any individual cases, Ms Dixon, but I am happy to look at any particular cases that you want to draw to my attention. They can also go to reconsideration.

There may be a difference here between employed and self-employed people. Those who are employed are restricted to a standard working week—there is a cap on the number of hours that can be provided—but the self-employed are not. Some self-employed customers have been putting in very large claims for the number of hours they are working. We know that, in general, self-employed people work longer per week than employed people. The cases you are seeing could be caught up in the challenge the teams have been putting to people about the hours they are actually working.

Yes, that may be the case. One of the examples is a business owner. If she does not get the BSL support and she has to give up her business, that will have consequences for the other 24 people she hires. It would be possible to get some examples from the RNID.

Yes, please do.

We could follow those up specifically to make sure that reducing job aides to 20% is not inadvertently reducing support for people with BSL needs, which should be at 100%.

That should not be a job aide category; that is a different support worker category. Please do ask the RNID.

Indeed. Thank you, Chair, for allowing me to raise that.

I thank our witnesses for the information they have given us so far. I apologise to everybody for starting a little late; we had a lot of business to transact before we started the formal hearing. It is almost half-past 11. Let’s try to be back by 25 minutes to 12, because we still have quite a bit to cover. Sitting suspended. On resuming—

Welcome back to the Committee. We will now start with questions from my very able deputy, Clive Betts.

Thanks, Chair. I want to pick up the themes from before about support workers and the big increase—this exponential increase—shown in figure 7, which has already been referred to. I do not know whether you, permanent secretary, or Neil Couling want to respond. This is a real difficulty in terms of how the scheme operates; it is not too difficult to look at a physical adaptation and say, “Is that something employers should be doing, or isn’t it? Is it justified?” but when you get to support workers, it is a bit more abstract and difficult to pin down.

That is a fair challenge, Mr Betts. It relates to the changing nature of the applications that come into the scheme. Support workers can play a really important role. Just before coming in, I was reading a case study of someone with a support worker. The customer had ADHD, and they had written in to say, “Just writing the client brief has triggered an emotional reaction in me. The questions on the pro forma made my stomach clench. Without my supporter being there, I would have given up on it. My support worker was able to understand my response, calm me down, tease the information from me in the form of posing questions and eliciting the answers conversationally, and using an AI tool to help construct my responses into the pro forma.” She ends up saying that she could cry because of the difference the support worker has made to her life. Support workers can play a massively important role, but the challenge—particularly in the case of job aides—is how you identify when someone’s role as a support worker is to level the playing field and help the customer manage in the workplace, compensating for their health condition or disability. There is sometimes a grey area between that role and asking them, while you have this person around in the workplace, to do some photocopying or some admin support that, in the ordinary way, you would have employed someone else to do. Those are different categories: one is an enabling role, and the other is a replacement role. As Neil says, the caps for the payment allowance are different for both. You can just imagine that happening in a busy office environment. What exactly is your role? Are you doing this really important support work that was described by that customer in that case study, or are you doing something that is enabling the employer to avoid having to employ an extra person at the taxpayer’s expense? It is difficult. Support workers have been part of the scheme for many years. Ms Dixon talked about BSL support for people with a hearing impairment, for example. But it is when you get into the role around mental health and neurodiversity, in particular, that it is more difficult.

Is that the big growth area that you are having to tackle?

Yes.

Yes, it is. There is quite a lot of data about that in the Report. The Report rightly challenges us to have a greater breakdown of the type of mental health condition and more identification of neurodiversity within that. But I think it is fair to say that it has absolutely been the area of most growth.

I may be the only one in the room who has administered this scheme; I used to do it back in around 2003. You would never see an application then from somebody with a mental health problem. It was very common, in what was very small scheme then, to have people with visual issues and auditory problems; that was the predominant nature of the claims. To judge a support worker now inside the scheme—particularly to the earlier question about whether there is more guidance we could offer—would be perceived externally as a change in policy, but it might also help you administer the scheme more effectively. As Ministers decide on the future shape of Access to Work, that is one of the questions they will need to wrestle with. You are right, Mr Betts, that it is trickier to assess support worker needs. The exact definition for an enabling—that is, 100%—support worker is that that person is instructed on what to do. In some of the mental health categories, that instruction is not happening, so a replacement job aide would be more appropriate there. That is limited, and as I said earlier, some of you will be seeing in your postbags those whose renewals have changed, and that is the predominant reason.

This is probably a slightly wider question, but is the issue of claims based on mental health or neurodiversity a wider challenge in your systems? It goes back to the discussions we have had with colleagues in other Departments about SEND—special educational needs and disabilities. It goes on to NEETs and the applications there for universal credit. It even goes on to PIP and the changing balance of applications there over the years. Is this something you are looking at in a wider sense, as opposed to just focusing on mental health and neurodiversity with regard to Access to Work?

You are absolutely right. We are seeing that across the board. Bill might want to come in a bit more on this, but on child DLA and young people’s PIP, about 80% of young people claiming PIP are doing so for mental health conditions.

Actually, it is 89%.

It is 89%. That is something that Alan Milburn is looking at in his wider review on young people.

It is absolutely a societal phenomenon and is very challenging. The Department of Health and Social Care is looking at this in its review of prevalence and what the best approach is to support people. Alan Milburn is looking at the issue of NEETs, where that will come up. I am sure part of the Timm’s review into personal independence payments will look at the changing nature of the caseload and the pressure on the SEND system—the SEND reforms that have just been announced are absolutely key to that. We are doing our best to ensure that there is joined-up thinking between those reviews, and they will be complementary once they are published throughout this year. The Department of Health and Social care review on prevalence will be quite important regarding what is actually happening out there, what is driving it and what is the best way to support people. We will look to go off the back of that.

Neil, what you said almost has me feeling great sympathy with your decision makers. They are having to make really quite complex medical decisions. Is there a medical team of whom they can inquire, “This person has this condition. What should they be expected to be able to do?” Do you have that sort of medical back-up?

No, we do not do it like that, but you touch on a really important point I was going to bring out earlier. This is not a benefit, and it is not like assessing a benefit—“Are you entitled, or are you not?” This is very much an assessment, which is why it is not easy to flood my teams with people to clear it, as I did, for example, with the child DLA backlog. Fantastically, when we come to look at the annual report and accounts—without Peter—next year, I will be able to show you some really marked improvements.

I am going to sign off the annual report and accounts before I go—let me just get that point in.

Good.

This is much more of an assessment. You need to interrogate the situation. We send people out to visit the employer to talk to them and see what they are providing. We assess the workplace. That is why it takes longer than a standard benefit claim. Decision makers have said to me that they would like more guidance, but as I said, providing more guidance would be interpreted by many people outside as a change in policy, and we have not changed the policy on Access to Work. We are trying to use the five principles to anchor decision making.

There are two different things here. Guidance is one thing, but being able to make an informed decision on the information you are going to be given is another. As an ordinary decision maker, you are not a trained medic. I do not know quite how, with complex conditions, those decision makers make a decision in the absence of medical advice.

I may have inadvertently misled you. They can ask our own medical team for advice—we have doctors in the Department. What I was saying was that it is not part of the process; you do not have to pass a medical assessment to get Access to Work.

I get the difference. You have answered my question. Thank you. Steve, you are very welcome to the Committee.

It is a pleasure. Thank you for inviting me. I would like to understand value for money, how that pans out and what the perception of value for money is. I would like to look at it through two different lenses. The average award for Access to Work is £4,000, yet institutes have looked at the value for money of other Government schemes and it can cost as much as £25,000 to get people into employment. How does that value for money work out? The other lens I want to look at value for money through is, how do we measure ourselves against other nations that are much more progressive on getting people into employment? In the United Kingdom, the difference in employment rates for those who do and do not have a disability is 30%, whereas that shrinks massively for Denmark and Sweden, down to 20%. How do we see the value for money of the scheme playing out through those two different lenses?

I might get Bill to say a little more about that, but I will give some of the headlines, and this comes out a bit in the NAO Report as well, of course. There are two headings for value for money. The NAO says we have not done a full value for money test, by which we are referring to a quantitative test. Last year, we spent £320 million on Access to Work supporting 74,000 people. That was an average of about £7,200 each, I think. The difficulty is that to do a formal value for money test, you basically need to have a control group to which you effectively deny Access to Work so that you can create a comparison and see what effect the Access to Work support has. What that means in practice, of course, is my identifying people for whom Access to Work would be appropriate and then saying to them, “For the purpose of my value for money test, I am going to deny it to you.” The work we did in 2018 with the NatCen research group indicated there was effectively no ethical way of doing this. From a quantitative point of view, we have not been able to do it, so we have looked at a qualitative point of view. It goes back to the heart of what I was saying at the very beginning, in the sense that work is not just about the fiscal, the economic and the wage per se, although that all matters, of course. Work is about value. It is about people’s self-confidence. It is about a sense of purpose. It is about feeling that they are part of society and making a contribution. It gives a sense of wellbeing as well. It is important for health, as well as the other way round. One of the things we did in our qualitative research in 2009 and again in 2018—we are doing further work now—is talking to participants and employers. Participants have talked about how Access to Work creates a level playing field in the workplace, reduces sickness absence and improves confidence and wellbeing. That chimes with me and what I hear when I talk to customers who have really benefited from Access to Work. There is that sense of what having a job means for them in terms of confidence and wellbeing. That is part of the value story. The other part of the value story is with employers, who talk about how it has enabled them to recruit more disabled people and to better understand needs and provide effective support. One of the additional benefits of Access to Work is that it has encouraged employers to look beyond it at the other aspects of reasonable adjustments that they should be doing anyway, and to provide broader support; it has encouraged them with their work as well. So it goes beyond the numbers, in terms of some of the qualitative impacts. As the accounting officer, those are the sorts of things I look at when thinking about the value for money of the scheme. In terms of what other countries do and how we can do more, do you want to say a bit more on that, Bill?

Across the OECD, there are large disability employment gaps. The UK is broadly middle of the road, but you mentioned a couple of countries where it is slightly smaller. I think you need a Government system approach to it. Access to Work is obviously quite a small scheme; we are talking about 5 million disabled people in work, and if we want to close that disability employment gap, Access to Work is probably not going to be the main lever to do that, given the size of the scheme. You need to look at the broader employment support that we offer to people out of work and the Pathways to Work roll-out that is going on at the moment. We need to look more at what employers can do and at the work Charlie Mayfield is leading on that, and at how we can retain people. There is a whole issue of people coming out of employment, the fit note system and how we look at that, which I think the Government will have more to say on shortly. We also need to look at the role of the benefits system and how that supports people into work at all times and how enabling that is. If you look at all those different factors, I think you could make movement. One thing I would say is that you often see the cultural attitudes to disability in disability employment measures. The number of people who will say in a survey that they are disabled will also move the figures quite a lot. Historically, the disability employment gap has closed when more people have said they are disabled. We are having a slight reverse of that at the moment.

I just want to unpick a bit further the effectiveness of spending on Access to Work, as compared with other schemes, where it might be £25,000 to get somebody into work. I accept there are cultural changes and so on in the bigger picture, but does the Department not have an opportunity to say, “Actually, this is a lever that could have a greater impact on getting people with disabilities into work”?

As Peter said, we do not have a quantitative analysis of the value for money of the scheme, because of the impossibility of getting a counterfactual, but we do know quite a few things. We know the unit cost of the scheme, which you mentioned was £4,000 for any provision. For people who actually get a grant, the average unit cost is £7,000. For people who get a support worker, it is £10,500. We therefore know what would need to be true about the impact of the scheme for it to break even. It would have to have quite a large impact to break even. We would need additionality of about 23%—that is, 23% more people being in work than otherwise would be the case. We know the effectiveness of other DWP labour market programmes; would that impact be reasonable or feasible compared with other DWP labour market programmes of the last 20 years? That would require Access to Work to be twice as effective as any DWP labour market programme, if we were going to break even on that basis. We also know the employment history of people on Access to Work, so we track their previous employment data. People applying to Access to Work have a very strong employment history compared with other DWP employment programmes. Over 80% have been in full-time employment for the previous six years, so we get that additionality. We also know that over half of the people applying for Access to Work are employed by a large employer, mainly in the public sector, but also in the private sector, so the extent to which the employer would provide that adjustment, or not, is in question. You triangulate that with a very strong qualitative analysis that we have about how effective people perceive the scheme to be in being in work. Probably it is quite unlikely that the scheme pays for itself fiscally, but we do not know that for certain. That is not unusual when we talk about schemes that support disabled people, but there is a large societal benefit to it. That is why we run the scheme, because there is a need to provide these above-reasonable adjustments. It is a good thing that the state provides that support.

It just gives you a sense of the work that is being done. But as I say, from an accounting officer point of view, it is over and above the fiscal and over and above the economic. It is about the societal benefit and the benefit to the individual. That is the thing that makes it appropriate to do, and to do at the scale we are doing it now.

That has been true since 1994. When I looked at it, as I said, when I used to administer it, my first thought was, “Crikey, some of this is rather expensive”, but when you saw the lives that it was changing and the impact it was having for the individuals, you could justify the scheme then, and I think you can justify the scheme now.

The other area I wanted to explore is managing budgets and where you are managing this budget from. Clearly, it is a grant budget. I want to unpick how you are managing that as far as finding other moneys to assist with that. I understand there are some schemes that are annually managed expenditure, whereas this is a grant pot, and there are a lot of benefits that are annually managed expenditure. Can you explain why this is a grant pot and not annually managed expenditure? A number of you have said that it has a significant societal positive impact, so why is it managed purely from a grant pot rather than other benefits that we would say have a societal positive impact? Why not the annual approach? I would welcome reflections around that area and how you are filling the gaps by having the grant.

As you say, it is a grant scheme set in our departmental expenditure limit, whereas benefit expenditure fits in our annually managed expenditure. It is the nature of the scheme that makes it a grant scheme. Obviously, the Government could decide to change the policy and change it into a benefit entitlement, in which case the Government might decide to put it in AME. We are working within the parameters of the policy, which is to treat it as a grant scheme and to treat it in the way that I do in DEL. That means—exactly as you say, Mr Darling—I then effectively have a budget for it. We touched on this a little earlier. It means I have to make a forecast each year for the number of claims that are likely to come in and be successful and the grants that we would pay out as a result, and make provision not only within the departmental expenditure limit for the running costs of the scheme—the number of people I am hiring to run it and colleagues working on it—but the spend of the grants. They all come under the DEL and I need to budget for them each year, and each year I have budgeted often a bit more than the year before, but those are conversations we have in the context of the spending review with the Treasury as one line in an overall conversation about our overall departmental expenditure limit.

Well done. Thank you. Anna is next.

I have a quick follow-up on this point. Steve asked a really excellent question about whether it would be possible to come up with a cost-benefit analysis of the value for money relative to other schemes that might try to close the disability gap, and indeed, as we reform PIP, where that personal independence payment is used by people to get to work and so on. While as a former chief analyst I am delighted to hear that you would aim for the most robust case-controlled study, shy of that being available, have you not at least done some sort of modelling? If Access to Work were not available to some individuals, one assumes that they would not be in work. For the constituents who approach me who have applied for Access to Work and been refused, it effectively means that they are not able to work. If the counter is that they are not working—they are on benefits, they are not paying taxes, there is a detriment to their health, and they potentially require additional care, have you not been able to model it? I think it would be beneficial. I think it probably would be very good value for money. I do not think that we are suggesting that this is poor value for money; if anything, the modelling would demonstrate that it is very good value for money and effectively give you the ammunition to argue for bigger budgets.

I would say two things. One is that we asked ourselves exactly that question, which is why we got NatCen in to look at what we could do. It came up with the conclusion that I described. The modelling that we have done is the modelling that Bill went through in detail earlier, in answer to Mr Darling’s question. It would show an additionality requirement—

It would be 23 percentage points. Sir Peter Schofield—of 23 percentage points. That is a lot compared with other employment schemes, so it would need to be much more successful than those schemes to be value for money. That is purely from a fiscal perspective, before you get on to a societal perspective. The broader point is that it plays a role in a much wider set of schemes and I think it does its job, which is all about levelling the playing field—that is the point of it. You have to look at it alongside what employers are doing and other, different types of employment support such as Connect to Work, which we have now fully rolled out across the whole of England and Wales. Connect to Work provides employment support to help people to get a job and to stay in work. That complements Access to Work, which helps people to manage a disability in the context of the adjustments that need to be made to level the playing field for them in the workplace. For many customers, you can imagine a whole suite of different things coming together to help them make the journey from out of work to in work. This is only one of them.

I cannot see how the counterfactual could be that 0% people in employment. No employment programme would ever assume that. I do not agree that 0% would be in work, when 80% have had a work history of full-time employment for six years and 50% are with a large employer. The NAO Report states in paragraph 116 that in a review of 200 cases, it was felt that 40% should have been funded by the employer. I cannot see how the counterfactual would be that low. We know that most people are not applying for a new job when they go to Access to Work; 90% are already in employment. We have done some qualitative research on why people apply for Access to Work, and there are three reasons: one is that a health condition or impairment has got worse; the second is that they are changing role within their current employer or taking on a new role; the third is that they just found out about the scheme, and the employer says, “Okay, I can get some funding here.” That is why I do not think that would be a reasonable counterfactual. Other studies, in the early 2000s, used a similar approach to the one you suggested, but we do not think that that is robust.

Sorry—just to clarify, what is your robust modelling? You keep talking about this 23%; what is that?

That is our analysis of what would need to be true for the scheme to break even, based on the unit cost of the scheme and how that compares with other labour market programmes over the last 20 years that the DWP have done. The two schemes with the largest additionality were the Future Jobs Fund and Work Choice, at around 11%.

We could go on discussing this all morning. We will look at the evidence and we may well come back to you, if we may, Sir Peter, on this aspect.

Yes, of course.

I want to put this in the wider context that we explored before. The scheme has been around for a long time. It has not fundamentally changed, although perhaps the demands on it have. We have had the Green Paper, which was published in March ’25 and looked at proposals to reform incapacity and disability benefits. We then had the Mayfield report and a Government response to the Green Paper. What is your thinking now about how you are going to pull that together? Is it going to be a significant change in the Access to Work scheme? Will it fit into a significant change in the wider benefits scheme, trying to encourage more people back to work? What is your thinking at this stage?

That is getting into areas of policy, which are issues for the Government. I will tell you about some things that are going on that helped to pull that together, which the Government will then take a view on. I think that Charlie Mayfield’s report is important for that, because it is trying to get a body of evidence about what is going on among employers and what good looks like. That, alongside the effectiveness of things such as WorkWell, Connect to Work and other interventions, helps us to understand the nature of what needs to be done when you bring together the roles of Government—within Government you have employment and skills support on the one hand and benefit expenditure on the other—of employers and, particularly, of the health system and health professionals. For example, a BBC report I listened to as I was coming in this morning suggested that many GPs were simply waving through fit notes relating to mental health. The key thing for any Government response is to look at the full partnership across each of those systems, because they all need to come together in a way that helps people to get into, stay in and succeed in work. Those are all the elements, and I can go through each of them in turn, but the main point is that—while it is for the Government to decide what to do in terms of how to take the policy agenda forward—that has to be seen in the context not just of trying to fix one bit of one part of the system, but of looking at the role of the health system, employers and Government operating together, because all those together impact on people’s lives.

You are beginning to answer the question that I was going to follow up with. Obviously, you do not do policy, Sir Peter—maybe you will be invited back to do it at some point, when you have retired, but at this stage you are not doing it. To inform the Government’s policy, what are the key issues that you think need addressing in terms of your concerns about how the system is currently working?

One is what interventions are most effective. The thing that was not explored until Charlie’s work was interventions in employers. Employers have the resources and the day in, day out contact with people before a health condition develops in a way that causes them to fall out of work and need medical support and support from the benefits system. What is the role of employers? The first thing is the work of the vanguards: the 100 employers that are developing workplace health provision and other types of interventions. What comes out of that, and should the Government do more to incentivise, encourage or mandate it? Those are all questions that would need to be resolved following that work. The second thing is the way that the systems work across both health and employment. I will give you examples of what we are doing with the health system. One is the fact that we have disability employment advisers, DWP staff, going into many GP surgeries. I have seen that in places such as Leeds, Ilford and Poplar: people go to see the GP to talk about having a fit note and as part of the conversation the medical professional can say, “Look, actually, we think work could be a help for you. Can I refer you on? It is a voluntary thing, but can I refer you on to this colleague in the same building who can talk to you about work and support?” That is already happening, and we have a new pathway called WorkWell that is putting money from DWP into the health system to develop those referral routes into pathways of employment support. The Secretary of State recently announced a further expansion of WorkWell to do just that. The third element is how the benefits system works to encourage and incentivise work, and how it encourages people to have a conversation about what they can do rather than what they cannot. Of course, Stephen Timms’s review will play into that effectively, and Alan Milburn’s review will probably get into the same space as well, although that has a broader focus on young people. You can see that these different elements are all happening at the same time. They will come together, I imagine, around the autumn or later on in the year. That will give an evidence base for policymakers to make decisions.

That is helpful, thank you.

This question moves smoothly on from Clive’s. Understanding all the changes and reviews that are coming forward—we are undertaking an inquiry on supporting disabled people into employment at this moment—how do you envisage the Access to Work scheme fitting in to that new changing environment? How do you see the role changing?

I think that it is about developing what we have been discussing already this morning, and thinking about how the world of adjustments needs to fit into a wider type of support. I would particularly look at the workplace, because that is where Access to Work operates, and the work of employers and reasonable adjustments alongside that. If you look at a workplace situation, where does workplace health provision sit? Many employers offer health support in the workplace, to help someone to see a practitioner quickly if they develop a health condition, and to get that addressed more quickly than having to wait to see a GP or go into the NHS system. We are working through what the best practice is on that. What about helping to prevent illness in the first place? What can be done in terms of prevention in the workplace? Many employers do flu jabs and things like that, but how do you develop that support onwards? There might be access to blood pressure monitoring and, alongside that, other types of mental health support for someone with a low-level mental health issue, to address that sooner. Reasonable adjustments are a part of that whole story. Access to Work fits in by starting with the employer and what they should do. Raising the expectations of what employers should do is where Charlie’s challenge comes through. Access to Work covers those adjustments that are necessary but not reasonable for the employer to do. Of course, that varies according to the nature of the employer. We talked earlier about large companies that come to Access to Work for an ergonomic chair, but you can also imagine a very small business for whom some of the adjustments that a big company could easily afford are just not affordable, so it would be reasonable for Access to Work to come in and provide that type of support. I see Access to Work doing the topping up, but you have to start with what else is going on in the workplace, particularly with the role of employers to do what they need to do. When you talk to enlightened employers about this, it is a win-win: you can keep talented people in the workplace, and you can bring new talented people in, just by making small adjustments. I still have in my mind a very poor situation. A number of years ago, I was visiting a firm in Bristol that had been set up by a group of parents of children with Down’s syndrome. Their thing was, “What is going to happen to our children once they leave the education system? What are the employment prospects for them?” They created a business making flatbreads—quite wonderful flatbreads, I have to say, although I am gluten intolerant so I could not eat many of them. The young people had a period of time working there to demonstrate their employability, and then the parents went to other big firms, some of which said they were disability confident, and said, “Can you take these people on?” I remember one story of a big firm—I won’t say who they were—who said, “Well, our shifts are eight hours.” The parents said, “These young people have proved that they can work, but the nature of their conditions means that they can really only do four-hour shifts. Can you split your eight-hour shifts into two four-hour shifts?”, and they said no. It is heartbreaking. There is much more that employers need to do in many cases to raise up to the standard of the best. Access to Work fits on top of that, to do the things that it is unreasonable to ask an employer to do. Does that help to answer the question?

Yes.

I have one or two rounding-up questions after listening to what we have heard this morning. The more I have listened, the more I feel sorry for your decision makers, because this is a very complicated scheme. You are going to help by giving them better training, training new people and getting better IT systems, and so on and so forth, but you are about to do some reforms on the programme from the Green Paper. My first question is: what is the timing of that? When are we likely to see what the new shape of the grant is likely to be?

We had some collaboration committees last year, which ran from July to October, to help us develop our proposals. Ministers would like us to continue that work with the independent disability advisory panel chaired by Zara Todd, which formed at the end of last year and the start of this year. That is the first major programme that they are going to support the Department on. We will be working with them over the next three to four months. I imagine, depending on what Ministers want to do, it will be later in the year, maybe the second half of the year, that we will be able to set out any changes that we want to make to the scheme. We have heard loud and clear that we need to retain the scheme as a flexible and personal support to people, which is focused on that individual person’s needs. That will still be a central part of any reforms.

To be clear, because we do not want people out there worrying, the basic scheme will still remain. Any reforms are to make it clearer. We should bear in mind the summary in paragraph 9 on page 8 of the Report, which says, “awards have not always aligned with the policy intent and funding decisions have been inconsistent.” The Report goes on to give the examples of “Reasonable adjustments” and “Funding for support workers”. What we want to do is try and help your decision makers, and indeed, all those who are applying for the grant, understand what it is clearly supposed to achieve. Is that where we are all trying to get to?

Yes, I think that is a fair summation of what Ministers are trying to do.

I was with some decision makers last week up in Halifax. I said, “What is the one thing I can try and get across for you?”. Your response, Chair, suggests I have been successful, because the point that they would like you to understand is just how difficult this is.

On that note of agreement, I thank you all very much. Sir Peter, I particularly thank you and I wish you well for the future. I was mulling it over in my mind when you said you were going to sign off the accounts, whether we might not have you back from whatever you are doing next to answer for your policies and the results in those accounts. But that is for another day.

I am totally at your disposal, Chair.

That is very kind.

What a kind retirement gift.

Whatever happens, we will this hearing with a huge thank you to you and, of course, the other witnesses. I wish you very well in whatever you are going to go on to do in the future.

Thank you so much; I really appreciate that.

The uncorrected version of the transcript will be available in the coming days, following which we will produce a Report with recommendations. We hope you, Sir Peter, and your assistants will look at that carefully and see whether you agree with our recommendations.