Welcome, everyone, to this afternoon’s session of the Public Accounts Committee. First of all, apologies from the Chair, Sir Geoffrey Clifton-Brown, who has an appointment with the Prime Minister this afternoon that has diverted him elsewhere within our building. I am Clive Betts MP. I am deputy Chair of the Committee, and I will be chairing this afternoon’s session, which is about the Government’s use of external consultants. We want to look at how much the Government spend. There are different calculations of that, as we will explore as we go along, but one estimate is that about £1.36 billion a year is spent on consultants for particular purposes and projects, and on bringing in particular skills. The Government have a stated aim of reducing their spending on consultants by half. That was set out by the Chancellor in a previous Budget speech, so we want to examine whether it is being achieved. We will also look at how oversight of consultants is carried out, both individually by Departments and then, of course, by the Cabinet Office looking overall at what Departments are doing individually and collectively. We will look at how the Government plan their workforce to decide what consultants they should use in the future and whether they can make better use of in-house staff. All those issues are for us to explore today. Coming over to our witnesses, Cat Little is chief operating officer of the civil service and permanent secretary at the Cabinet Office—the ideal person to talk about this. Perhaps you could introduce yourself, Cat, and then your colleagues who are with you on the panel as well.

Thank you for inviting us here today on this important subject. As you say, I am responsible for the day-to-day running of the civil service. I am also permanent secretary at the Cabinet Office. To my right, we have Rebecca from His Majesty’s Treasury.

I am Rebecca Molyneux. I am the deputy director for workforce, public sector pay and public sector pensions in the Treasury, with oversight for the Treasury perspective of consultancy spend.

To my left are Andrew and Mark.

Good afternoon. I am Andrew Forzani. I am the Government chief commercial officer, so I am responsible for the commercial function across Government, which includes the consultancy playbook and the standards that we set for the way that we procure and contract with consultants.

Good afternoon. I am Mark Leigh. I am the commercial director at the Department for Energy Security and Net Zero. I work for Andrew and am responsible for the commercial activity of the Department.

Cat and Andrew, you have been before the Committee before. Rebecca, have you?

It is my first time.

What about you, Mark?

I have been once.

Rebecca, a particular welcome to you. I hope you enjoy the experience and decide to come back to see us again. Thank you very much for joining us this afternoon. We will begin with questions about the use and cost of consultants.

Good afternoon, Ms Little. I just want to start by asking about workforce planning, and what the Cabinet Office is doing to ensure that Departments and arm’s length bodies are properly carrying out their strategic workforce planning so that they can identify potential skills gaps in the pipeline.

This is one of the most important things that we are currently doing. For the first time, we are producing multi-year strategic workforce plans for the civil service. That is about right-sizing, right-shaping and making sure that we have the skills we need for the future. We started the process in the summer to run alongside the spending review, and we provided guidance to Departments on what to include. We asked for detailed forecasts for the next five years of the skills, the numbers and the shape of how every Department, including their ALBs, is planning to develop its people. For the first time, we have included contingent labour assumptions within the workforce plan as well. That process, now that we have concluded the spending review, is in refinement, because we are now doing business planning, which is where Government Departments take the strategic spending review and translate that into annual budgets. Once we have gone through the business planning round, which normally concludes in January, we should have, for the very first time, a five-year forecast of our people and their skills, and how we expect the shape of the civil service to change over the next five years. I do not know if it is helpful, but I should probably set out what the three big strategic workforce objectives are. This is current Government policy. We do not have a size cap. What we have are three policy objectives. The first is to reduce the administrative costs of the civil service. In the spending review, we announced that we would reduce administrative costs by 16% by 2030. That equates to just under £2 billion. The second policy is to halve consultancy spend. In 2024-25, we are on track to have met our halving of consultancy by £550 million, and have also set out the annual targets until the end of the spending review. The third objective is to become less back office and more front office. That is a bit jargony but, basically, we want to take more of the resources, capability and capacity of the civil service and put it on the frontline to deliver value for citizens. That means more shared services, more right-sizing and benchmarking of our big corporate functions, and trying to make sure that the civil service is growing only where we are adding value to citizens on the frontline.

I have two questions about that. First, in terms of halving consultancy spend, what we have seen in this Report is that the Government find it difficult to say how much they are currently spending on consultants. First, why is it so hard to have an accurate estimate of current consultancy spend? Secondly, if you do not know what you are starting with, how will you know when you have halved it?

Others will want to add to this, and especially Rebecca at the Treasury. The first thing to say is that we do know how much we are spending on consultancy. The annual report and whole of Government annual report and accounts, which this Committee knows very well, alongside the departmental annual report and accounts, set out in detail how much we spend on consultancy. We also report on services that are classed as managed services, which is not the same as consultancy. You have probably seen the definition set out in the NAO Report. We also know how much we are spending on contingent labour. The confusion comes because there are lots of other external measures of consultancy. For example, the Oxygen and Tussell measures are external measures that do not fully take into account our definitions, the accounting, or the way in which we contract within Government. Our view is that they generally tend to overstate the level that is being spent on consultancy.

Do you think they are misleading?

I do not think they are misleading. They are legitimate reports that help in understanding transparently how we spend our money, but they all have a slightly different basis. For example, Tussell is looking mainly at how we procure. If I take a big skills contract such as the one we have for civil service learning and development, some of that is consultancy and some is direct services for learning and development, but that is likely to be classified fully as consultancy. That is overstating the level of consultancy in accordance with Government definitions, but I am sure Rebecca would want to add to that.

Just building on what Cat said, Tussell uses what are called SIC codes, which are the standard industrial classification codes that companies register with Companies House about what type of firm they are. They might say, “I am a management consultant”, but what that does not get to in the external data is the individual things that a company that does management consultancy but also other things may do. Because of external companies categorising things in different ways and using different data sources, we really rely on the annual reports and accounts, which are, as you well know, the audited measure of what Government spend.

Just very quickly to ask on that, is it consistent between Departments? Are they all classifying consultancy spend in the same way so that you can take, say, the DESNZ accounts and know that consultancy spend is classified in the same way as in any other Department?

They all should be using, for the annual reports and accounts, the financial reporting manual’s definition of consultancy. All Departments should be using that. We know that Departments might find it difficult in some cases to classify their spend, and this report is really helpful in highlighting where some confusion may lie. As a result of that, we want to do a working group with Treasury and Cabinet Office officials, and then work with Departments to ensure that everyone is applying those definitions consistently going forwards.

Are you confident that you will not see too much difference, in either individual accounts or across Government, in spending on consultants once you have done that exercise?

Our National Audit Office colleagues audit our accounts to make sure they are materially true and fair, and that is an important judgment. It is independent and external. If an annual report and accounts is, basically, presented as true and fair, then I expect them to be substantively, materially correct. Having said that, as Rebecca said, we know that the devil is in the detail, so we constantly need to improve data. Being materially correct is one thing, but we need to make sure it is as accurate as it can be. Especially when you have large contracts or large service provisions, it is important that we try to disaggregate between the three types of expenditure that we are trying to get to the bottom of. It is only then that we can really get after value for money.

Just to quickly follow up, you talked about your three objectives—cutting overall costs by 16%, halving consultancy spend, and a shift away from back office towards front office. Do you see any conflict between the commitment to halve consultancy spend and the other two objectives? Does that make it harder to achieve either of the other two?

They are definitely related. When we look at consultancy and contingent labour, our main objective there is to make sure that we are spending less and upskilling civil servants, particularly in the areas of technology and project management, which are the two largest categories where we bring in consultants and contingent labour. We have to be able to make sure that the costs of doing that are cheaper within the civil service. As the NAO Report points out, it is quite often significantly more expensive for us to bring in contingent labour or consultants. They should be very much complementary for us to be upskilling, growing our skills within the civil service, and reducing cost overall, but it is something that we are watching very closely. What would be terrible is if we were reducing cost on one side while not upskilling and continuing to use high levels of contingent labour. Everything that I have seen in our draft strategic workforce plan shows that contingent labour and consultancy are coming down, and the proportion of people with technical specialist skills is going up.

Mr Leigh, how much does your Department, DESNZ, spend on consultants, and what do you use them for?

In the annual accounts, we spent £160 million on consultants in the last financial year. If you think about the Department as a whole, we are operating in largely emerging or nascent markets, often with a high technological content, resulting in a need for specialist design and financing advice in order to provide the right advice on large-scale programmes such as carbon capture, our work in the hydrogen market, or financing Sizewell C. Those are just a number of those topics. We are really focusing on using external advice and capability to help us deliver those programmes, and that is really where our focus is within DESNZ.

What processes do you use to ensure that you are getting the best value for money when you are procuring consultancy services?

I am sure we will come on to this in more detail through the rest of the Committee, but we are a framework first organisation, effectively. When we are looking to go out to market, assuming that we have already gone through the position that says we need to go to market, we look to use the CCS frameworks first, or our own frameworks, which are more niche, and look at using those frameworks first and foremost in an effort to drive value for money.

How successful is DESNZ in terms of strategic workforce planning? How much of a forward look do you feel able to do to get a better view of what your skills needs are going to be before you get to the point where you are having to hire a consultant?

In parallel with the exercise that Cat outlined, we have a strategic workforce plan within our organisation, brought off by our executive committee, supporting the Cabinet Office. That outlines how the Department will change over the next few years to become more agile, focusing on a number of key professions to build increased capability, skills and the ability to make a difference in those professions, one of them being my own as the commercial head of profession. Alongside that, I am also the head of the corporate finance profession, so I am working within that plan to bring those two groups together to share knowledge and experience, so that we are better able to respond to the needs of our departmental SROs so that they can execute the policy in the Departments. Overall, we have a plan and are working through that. It is clear that there are still areas where we will need to continue to use that technical capability and skill. It is also fair to say that, on occasions, there are going to be skills that we need, where investing and building that capability permanently inside Government would not be the right value for management judgment. In those cases, it is all about me and the team working really hard with our colleagues in policy to make sure that we are out there in the marketplace, buying in the right way and delivering that value for money. There is always a balance in our workforce plan between driving to build core capability, improving what we do and driving that forward, and other things in my commercial space where we have invested in different skills and capability taking us outside just core procurement to improve that capability, again so that we can be responsive to the Department and scenarios. There is a balance between that activity and not investing internally in those things because they are better sourced from the marketplace.

My question is for Cat. I am a little confused, so I would be grateful if you could just clarify. The NAO Report says, “As of 2022-23, central Government spend on consultants was estimated by HM Treasury to be approximately £1.36 billion, but other sources suggest the figure could be significantly higher”. Was it significantly higher, and why are we looking at 2022-23 and then comparing it with 2025-26?

I can give you the latest estimates and forecasts for subsequent years if that is helpful, but your question relates to the answer we just gave to your colleague, Ms Olney. Ultimately, the four different types of expenditure set out in the NAO Report all have a different basis. The one that we use, and that we think is most relevant and accurate because it is based on our definitions and what we are trying to measure, is the annual report and accounts figure. You can measure it in lots of ways, but the number we are most interested in is the £1.3 billion, and that is what we have used to measure the targets and the Government’s policy.

Is that accurate? Other sources say that it could be significantly higher. Is it significantly higher?

I do not think it is, based on the Government definitions.

You do not know.

I am saying that I do not think it is significantly higher. I am saying that I absolutely put faith in the annual report and accounts process and the external assurance process, which is undertaken by the National Audit Office.

Do you have a dashboard of what you have spent with each of these consultants? Can you tell me how much you spent with Deloitte or with each of these consultants in a concise spreadsheet? In my businesses, we look at how much we spend each year with each of our suppliers, and we then beat them up to get a discount. Are you able to tell the Committee exactly how many wheels you put on the Deloitte partners’ Bentleys?

For our strategic suppliers, which are monitored across Government, of which there are 39, yes. Otherwise, the accountability for expenditure is with Departments. I would expect every Department to know exactly what they have spent, and they will have their own processes and procedures for doing exactly as you say.

You might expect them to, but do you have that detail?

I could ask for it if I wanted to.

Is that not what we need to see in order to be able to decide whether you have halved it? Your stated aim is to halve. In order to halve something, you need to know what the figure you are halving is, do you not?

Yes, and the figure I am halving is £1.3 billion, which is our latest estimate.

It is an estimate; it is not actual.

We set a budget and a forecast of what we expect to spend on consultancy. We then have an actual at the year end. For the year 2024-25, we will have the whole of Government accounts available next year. At that point, I will be able to tell you exactly whether it was accurate.

It is extraordinary that you do not know the exact figures. If you are running a business, you know the exact figures. You have it and you can look at it, but it sounds to me like they are all estimates and are all up in the air.

That is because the financial year has not yet reported. As this Committee will know from the work on the whole of Government accounts, there is a time delay between the numbers being consolidated and audited.

Could I come in just on your questions around understanding what we spend with particular suppliers and consultancy firms?

It is consultants, really.

We have really good information about what we spend with the big consultancy firms. Another one of the reasons why there might be a difference between those total spends and what we are reporting under the target is because they do a really wide range of work for us. Not all of it classifies as consultancy, so that is another reason. We know what we spend with Deloitte, KPMG, PwC, et cetera, across Government and then across the wider public sector, but the work often sits within different definitions. It does not all sit within consultancy, which, for the definition of the target, is very much around advisory work. There are professional services and other types of work that they do for us.

So we can assume that these other sources are wrong. Are they wrong that spending is significantly higher?

Some of those sources are very much based on a supplier spend. They will take one of those large consulting firms and assume that all of that is consultancy spend and report against that. There absolutely is Government spend with that supplier, but it is not all classified as consultancy.

If the Chair will indulge me, I note that the civil service now amounts to a headcount of about 551,000, which is up 35% on 2016. When you are hiring all these new people, are you hiring people with the skills that you are spending money on with consultants? You say that you are now doing an analysis of what skills they have. It is not about the skills we have; it is the skills that you are hiring for. You have added a lot of people here. Is it logical to think that consultancy spend should naturally drop? The more people you have working full-time within the civil service, surely the less need there should be for consultants, or am I wrong?

That is entirely right, and the same is true of contingent labour. It is much cheaper to recruit permanent full-time civil servants with the right skills and expertise that we need, rather than to buy it in through expensive consultancy or contingent labour. As I said, what we are seeing in the draft workforce plans from Departments is exactly that trend, where they are recruiting more digital people and project managers, and reducing their costs on consultancy and on contingent labour.

Was there a reason why you shut this Government consulting hub that people seem to think worked quite well and seemed quite logical to me?

This decision was taken back in 2023, so it was before my time. There are lots of really good lessons learned from it. At one point, the hub had about 40 people in it. It did several things: it triaged consultancy; it helped to improve the intelligent customer use of consultants; and it also delivered projects and programmes. We estimate that it saved around £4 million compared with what we would have spent with consultants. The previous Government made the decision to close it down because, at the time, the policy was to reduce the headcount of the civil service, and it was felt that the headcount of the civil service and of that group should be reduced.

One would think that, with so many new people working for the civil service, an administrative burden would not be a problem, but logic probably does not play too big a part in it.

Just on a point of clarification, Cat, when Sarah Olney asked whether you knew how much you are spending, you said yes and disagreed with the NAO Report. Paragraph 1.33 of the NAO Report says, “The data available on Departments’ consultancy spending are inconsistent and vary from one source to another… This means the Government do not have a clear picture of how much is spent or how this spending has changed through time”. I take that to mean that, when the NAO was doing the research and speaking to relevant staff, the staff could not provide the NAO with information in a way that would allow it to write in the Report anything other than what they just did. When Sarah asked you the question, you said that you were confident that it was there, but that you had not asked for the evidence. There is an inconsistency here between what the Report is saying and what you have confidence in. Can you speak to that briefly? It is not my question but just a point of clarification.

Rebecca should add to this. Materially, in terms of truth and fairness, we are comfortable with the overall reporting of the numbers when we add them up. We know that, when we get into the detail, because we get the reports from Government Departments and are grateful for the work of the NAO, there is always going to be a certain level of judgment about how expenditure is categorised. This is about levels of assurance at central level as opposed to at departmental level.

If I may, one of the things that the NAO Report talks about is the difference between OSCAR data and the annual report and accounts data. As I am sure many colleagues on the Committee are familiar with, OSCAR is the internal, month-by-month data that Departments send in to the centre on what they are spending each month. They say what they spent the last month and project for the rest of the year. That data has not been exactly the same as the annual report and accounts data. That is because, when Departments compile their annual report and accounts at the end of the year, they start off with OSCAR data, but they apply additional scrutiny because it is the published data. They do additional checks, as does the NAO, as the auditors, to make sure that they are fully meeting the categorisation in the annual report and accounts data at the year end. That is the data that Government will use to check whether we are meeting our targets. In-year, we will use OSCAR to see how we are getting on towards those targets. The differentiation is because you have many people uploading to OSCAR and not doing as much assurance on the fine categorisation of spend in OSCAR as they are at the end-year accounts. That is where there is some of the discrepancy that the NAO report talks about.

I know that my colleague Blake was going to ask you about that, so I will hand over to him now.

It was only to ask what work you are doing to ensure that Departments submit better-quality data through OSCAR. If we look at figure 10, on page 27 of the NAO Report, we can see that consistently, each year, the OSCAR data is much higher than the ARA data that you are referring to. You are saying that OSCAR is incorrect, because you are saying that the ARA data is accurate. What are you doing to make sure that the data submitted through OSCAR is correct?

There are some inaccuracies in the OSCAR data when we are talking to Departments, particularly in this classification of spend between the categories that Cat and others have outlined—professional services, consultancy and contingent labour. The additional scrutiny that this Government are placing on consultancy spending through the target to halve it means that, in my team and other bits of the Treasury, we are spending additional time talking to Departments on a monthly basis to check their spending and to ask them, when they are uploading their OSCAR figures, “Are you confident that this is all consultancy?” and asking them to update it when they are not. We are working on it that way, and the additional scrutiny itself should improve the accuracy. Also, like I said, this report has been a really good spur to set up an additional working group, bringing together colleagues from Cabinet Office and Treasury, and then working with the various Departments to make sure that the definitions are clear and that people are applying them consistently. We are also talking to the GIAA, the Government Internal Audit Agency, to see what additional support it can provide, so that we are all, in-year, using the right categorisation as much as possible.

Do you know why the difference is so big each year at the moment, or is it work that you need to do through this taskforce?

We broadly understand that it is this classification issue. We know, for example, that MoD has a big contract that is largely professional services rather than consultancy, but earlier this year, it uploaded that mainly as consultancy. That is one example and was quite a large amount, and there will be smaller examples of a very similar ilk in other Departments. What is important is that you might take away from that that we are fudging the figures by just reclassifying things from consultancy to professional services. What we are also doing to check against that is looking at professional services spend and these other types of contingent labour. They are also on a downward trajectory in the forecasts. We are not seeing a massive reclassification out of consultancy, so we can be confident so far that the reduction we have seen in consultancy is a genuine one.

Just to bring all that together, the impression I am getting, not being a data specialist, is that the data is not as good as it needs to be, but you have put processes in place to ensure that it is changing. It seems like you are doing it in a very labour-intensive way if you are having all of these check-ins. What are you doing to ensure that, going forward, there are better processes to stop you having to do that?

If I might approach this through a slightly different lens, part of what we are doing in all of our corporate functions is investing in new technology. You will be aware of our shared services programme, which is bringing cloud-based enterprise risk systems into how we manage our finances and our people. Quite a lot of those cloud-based programmes allow us to use AI and to sort data in a much more interactive way, but that programme is currently under way. Every Department will cleanse and audit data in slightly different ways, but I would expect most functions to be using technology to better categorise and use information. It does not have to be labour-intensive.

That sounds like the future, though, rather than the present. Getting from the current labour-intensive, manual process to the wonderful new world of AI doing it for us requires a journey, so where are we on that journey?

In practice, it is a bit of both. We know that about 15% of the civil service is using AI on a very regular basis for all sorts of reasons. Part of it will be data cleansing and data integrity. I do not have a full audit of how much of this is manual, but I would expect for technology to be used in at least some of the departmental work that is under way.

I am going to push this a little more, Cat, in that we have had this conversation before about what technology will do in the future, but we are trying to ensure that we can measure your progress against targets with what you have now. From what we are hearing from Rebecca, it is a bit of a laborious get-together with people to make sure that the data is entered in a consistent way. It sounds to me that you have decided that, at the moment, AI cannot deal with the issue that you are dealing with now, and that, consequently, you have Rebecca, or her teams, doing this manually. While I get that AI is the future, it sounds like someone has looked at this and said, “AI and technology cannot do this right now. We are going to have to do this the old-fashioned, manual way”. I wonder if you could just talk about the journey to get from one to the other. Normally, when you are looking at this manually, it seems like you have a problem with data and data consistency, and you have decided that the best way to deal with this is knocking heads together, if I can put it that way. Am I correct in that assumption, and how is it going?

Others should add to this, but it is just the reality that every Department is in a slightly different place. I know that I have given evidence to this Committee before. Some of our Departments are on very mature cloud-based systems already, and some are still working off spreadsheets, so it really does depend on how data is held, the size and scale of that data, and how much technology you already have integrated into your ERP system. Inevitably, a lot of it comes down to finance directors and commercial directors knocking heads together, making sure that processes are appropriate, and doing the reconciliation controls between management information and financial reporting, which is fundamentally what Rebecca is talking about. That will always have an element of leadership, culture change, and senior people sending the right direction to staff. Over time, as we implement new shared service technology, we will get more and more tech-enabled.

I imagine that one of the problems is that there is not a consistent definition of “consultancy” across Government. I know you have said that you are working on a revised definition of “consultancy”. What is stopping that happening? What is stopping one definition of “consultancy” being used across Government?

We do have that. The guidance that goes out across Government to all Departments is called the consultancy playbook. That has a very detailed, clear definition about what consultancy services cover. It talks about it having to be advisory. It is predominantly around, quite often, the management of change. It should be for skills that Departments and the Government do not have. The people who supply those services should not be sitting within the organisation. It is quite a detailed definition about what consultancy is, and then it gives a clear definition of professional services and contingent labour. It is all there for Departments to make those choices. The confusion is sometimes that you can have a contract with a supplier, and it covers all three of those within one contract, so you can contract for consultancy services and professional services all in one bundle. Some of the challenges of coding come from that, but that is our best practice guidance that has been repeated and updated. All Departments have been reminded of that. As we have chased Departments around their compliance with and progress against the target, we have clearly shared the definition again.

Is that the target to reduce the spending, as opposed to the target to adopt a consistent definition? What target are you referring to?

The permanent secretary has written out across Departments through the year to remind them of the target to halve consultancy spending. As we track progress against that, we remind them all about the consultancy playbook and include in that the definitions.

Paragraph 1.33 of the NAO Report says, “Departmental annual reports use different definitions of what constitutes consultancy spending, which may or may not be based on the Cabinet Office definition for consultancy”, so the NAO is not assured that this is happening. You have something that you are telling Departments to do, but the evidence seems to be showing that they are not all doing it. Can you talk about what progress you have made in getting Departments to adopt this consistent definition? Are you just telling them that they have to? The evidence is showing that they are not.

We are back to the financial reporting issue. What I am saying is that there is a cross-Government standard definition. The Report, quite rightly, has picked up that that does not always follow through into the reporting, which is what we have been talking about. We are continuing to remind Departments of what the definitions are, and putting quite a lot of focus and effort into bringing Departments together to improve the situation, as Rebecca talked about.

Can you expand on the focus and effort part of that? Reminding Departments to do it is one thing, but it is the focus and effort, and the plan. You have been reminding them for a while, but that is not resulting in action that the NAO can evidence. I am more interested in drilling down into what you are doing and how you are measuring the success of these actions.

It comes back to the working group that we have established. We have asked the Government Internal Audit Agency for advice on how we can go about getting a much better grip on which Departments are doing what, and where the divergence from our definitions is. The NAO has given us a good indication of that, but we now need to look at wider assurance checks. Ultimately, we send directions in the annual accounts direction to Departments, and expect them to comply with the definitions that the Cabinet Office set in August. If we find that people are not complying, we will go back to those specific Departments to tell them to be compliant. There is quite a lot of follow-up going on. We are always very interested in other ways that we can go about doing this, but the combination of audit, data cleansing, improving the technology and also improving the mandate from the Cabinet Office is the main way that we are going about fixing this.

When will you have it fixed? How long will that take? I am pushing this a little because it is unusual for me, when I hear these discussions, to feel like I am not getting to an answer. I am getting that you are aware of the problem. You have raised it repeatedly with different Departments, and they are either not getting it or not agreeing with you, but it is not resulting in the action that you wish to happen. If this went on consistently in the past, and your approach is to keep doing the same in the future, what confidence do you have that you are going to get a different result when you ask them to do it next month?

I understand why you are pushing on this. It matters greatly to us as well. We want to get to having the very best information. We owe it to both the public and Parliament to make sure that that is the case. My personal view is that the combination of those actions will certainly improve it. I do not think the job is ever done. We all know that data quality and integrity are the hardest things to fix, and especially when you are spending this amount of money with such a large number of data sources sat on multiple different datasets. This is one of the hardest things in Government to fix. We keep going at it. We keep improving it. I really wish that I could give you a date for when it will be totally fixed. My experience of this over many years—and I am sure the NAO will look back over its Reports over many years—is that it is getting better, but it is never going to be a once-done and fixed job.

As a final point from me, considering all of the variances between all of this data and this approach, you have indicated that you are confident that you will know when you are going to meet your targets to halve the spend. Are you confident that the NAO is going to agree with you? Are you confident that this Committee is going to agree with you? You have to convince us as well as yourself.

Of course, that will be for our auditors to look at. We have looked at our internal monitoring. We have been tracing and testing the data that we have had from Departments. We can see that, based on that data, we are on track. Of course, now the proof is in the actuals, once the annual report and accounts are produced and have gone through a rigorous audit process. I am as confident as I can be, but I would prefer to have an external audit of the data.

I would probably leave you with the observation that I absolutely get that you and others will be confident within the Department. I am less confident that the NAO is going to be, based on the language that is being used and also the line of questioning. It is more of an observation that it is important that we all have a shared understanding and a shared confidence level.

What assurance can you provide taxpayers that any savings reported will reflect genuine reductions in spend and not simply reclassifications or continuing inconsistencies with the data across Departments?

It comes down to the data checks and the integrity. Rebecca should add to this. The Treasury is looking at the overall reporting of public expenditure. Whether it is professional services, management consultancy or contingent labour, we are monitoring all categories to see that the cost reduction is happening across the board.

But you are still using the manual approach.

We are looking at the annual report and accounts, which are published and audited. The bit I was talking about is the in-year data that we have internally—the OSCAR data—which is not published; it has never been published. We can see how we are performing in-year. That is what we use internal to Government to check that people are on track, and that is where we have more of these adjustments. The published data is externally audited, and we have more confidence in it. It is always slightly different in lots of categories from the OSCAR data, because the OSCAR data is then adjusted once Departments have done further checks. The OSCAR data does not go through any external assurance. It is literally just the flow into the Treasury, mainly. Once it is then published, that is when we will check it and judge ourselves on that final published data. Of course, we want the monthly data that we are looking at internally to be better, and we are working with Departments on that, but the NAO was particularly talking about that discrepancy between those sources and the external sources that are not Government sources. In terms of Government sources, we are clear that it is the ARAs that we will judge ourselves on. It is the ARAs that are audited. That is our public record, but we do not have that until quite long after the financial year end, and so, to judge whether we are on track, we have to use the in-year data. That is where there is more work to do on improving the quality.

If I can come in on this, I wanted to clarify, just for the Committee’s understanding, and to give an example of the type of data that you are seeing from the annual report and accounts that we are talking about. Just to give you an indication of the types of conversations that we were having with Departments when we were talking about spend, some Departments were saying that they used a Treasury definition. Others said that they found it really difficult to state clearly their spend, for reasons that colleagues have just been talking about in terms of definitions of what “consultants” constitutes. Some Departments said that definition changes had led to a year-on-year increase in spend because of definitions having changed over the years. Then you get into ALB spend, which also complicates the issue, with some Departments giving separate figures for core spending and group spending. It is just worth noting that, overall, the consultancy figure is in the front end of the annual report and accounts, not in the back end of the report and accounts, which does make a difference in terms of the level of audit assurance that those figures get.

What is coming out of this is not a direct criticism. It is just a reflection of the reality that we cannot be absolutely certain how much has been spent, can we, because different Departments use different definitions of what “consultancy” is. Indeed, you have already said, Cat, that the MoD changed its definition on one big contract from one month to the next. That has a significant effect on the overall figures when they are added up, does it not? Is that not a concern to you: that you may not know precisely what is going on, let alone anybody else?

I do not want to repeat what I have said, but we have a high level of confidence. Is it 100% accurate? No. Clearly, there is more that we need to do, but if we needed to give assurance at a detailed level, I am confident that Departments will have more information, depending on the nature of the expenditure that we are looking at.

So there are sometimes differences in how they categorise expenditure, but there are also differences in controls, are there not? There used to be central control over consultancy spending. That was dropped in 2023. Was that a retrograde step, in terms not of policy but of procedure?

That was a choice made at the time to improve the pace at which decisions were made in Government. When we look back at the data of how much time was taken up manually looking at consultancy controls in the centre of Government, what we find is that nearly every approval came very late on in the process and was approved largely because it was too late to do anything about it. Do I think that back-ended controls on consultancy, where the centre is double-checking that work, is the best way of us to go about assuring consultancy? No. We need to be much more front-loaded. There needs to be clearer accountability that it is the responsibility of the permanent secretary and the accounting officer to ensure that consultancy spend is well administered. Ultimately, what we did before weakened the accountability and the transparency of how this expenditure was used.

Do you have any responsibility, therefore, for ensuring that Departments have relatively similar controls over how they are looking at their consultancy spending?

This is one of the things that we are getting advice on from the Government Internal Audit Agency. What is very clear is that different Departments have different local controls, with varying degrees of accuracy and confidence in what is being identified. We need to have more assurance over how controls are being applied at departmental level.

How will you do that?

That is why we are getting internal audit to give us some advice on how best to do it. This is an independent assurance process that should be undertaken away from management.

So it is not about taking responsibility away from the accounting officers in Departments, but making sure that they are doing the job that they should be doing.

Yes, and improving the consistency of controls so that we can demonstrate locally that the control over expenditure is as good as it can be.

What about arm’s length bodies?

Arm’s length bodies are also subject to the same definitions and, again, we will get the internal audit team to give us advice on how best to improve controls in arm’s length bodies.

I could talk about the controls in DESNZ, if it would be helpful. We followed the Cabinet Office advice on setting controls that are executed at the permanent secretary level, at around £100,000 in three months, and controls at £600,000 and a contract of longer than nine months. What generally happens in the Department is that, if an SRO is looking to buy a consultancy service or, indeed, professional services or something such as that, given that we do not differentiate in the controls, because we recognise some of the challenges that have been laid out, we start that conversation early with my commercial team and the SRO looking at routes to market and how we might go to market. The SRO will go to either the permanent secretary or the Minister and lay out their case and why we think it is appropriate to go external for that support as opposed to meeting that need internally or reprioritising other resource inside our organisation. Then we follow through that process, coming out of a ministerial approval or a permanent secretary approval, into the value for money process, where we procure the service, as I laid out to Sarah Olney earlier.

I was going to ask whether you see much variance between what is going on with in-year reporting through OSCAR and when you get to the NAO at the end of the year.

Yes, there is some variance, and the NAO Report outlines some of that. It is for the reasons that I have set out in terms of the OSCAR data not being audited. It is uploaded by many different people within Departments. They may be making some assumptions that any spend with a big-name company such as KPMG is consultancy, whereas the contract is for consultancy and professional services. There is then a cleansing of that data for the annual report and accounts, which picks up some of these issues. What we are doing in my team and across other teams in the Treasury to assure ourselves that the Government are on track towards their target to halve consultancy is doing some of those checks early, in effect, just by talking to Departments and, through the spending team model that the Treasury has, saying, “MoD, you are looking at spending this much. Are you sure that is all consultancy? Have you checked what contracts it is?” We are doing a bit of that manual work to assure it in-year, so that we can get advanced indications about whether we are on track to meet those targets.

Mr Leigh, I wanted to ask about what controls DESNZ has in place for the use of consultants and temporary labour.

I outlined the controls that we have for consultants. In that process, we also run a commercial assurance board for expenditure over £10 million of contract value. That comes to an independent board, chaired by one of my colleagues, where we assess that case, determine whether it meets the six value for money tests, and whether that goes forward. On contingent labour, that is run outside the commercial process. That is run through a financial context. We operate to make sure that the CCS contract is the primary route for that, but it does not come through a formal set of commercial controls in the same way.

Can you explain how you ensure that Department teams do not use consultants unnecessarily?

There are a range of things that we are doing to try to avoid that, some of which I outlined earlier when we talked about the workforce plan. Essentially, what we see is that the controls act as a deterrent to people putting forward a case that is not as robust as it should be. What we see is that the controls force people to look really carefully at what they are trying to do, work with colleagues to assess where those gaps and issues are, and then move forward if they can make that case. In our workforce plan, we outlined earlier that we have put in place some activity in terms of introducing new capability to reduce that level of consultancy at the front end. We also work alongside Andrew’s teams in the Cabinet Office with the complex transactions team. For example, we have had 13 engagements using the complex transactions team in lieu of what probably would have been consultants in the past.

Are you noticing that you are implementing any changes in behaviour as a result of the reviewing of that?

Colleagues recognise the challenge in general. Coming to my team and using them for some skills and capability, or working with CTT, can often get resource there quicker. For the right things, we are pushing people into those routes. We are providing advice. As Cat outlined, it is better if the conversation happens earlier rather than at the end of the process. That does not necessarily result in not using consultancy, but it definitely results in activity that is much more fundamentally redesigned, more value for money-friendly, and avoids creating a capability trap for ourselves. Equally, there are definitely things that we sometimes need to buy, such as legal services, professional services or insight and capability in an engineering context, to make the enormously important programmes that we are delivering be successful. We are absolutely focused on making sure that they are bought properly in that context.

Thank you. We will now suspend for five minutes, until 4.25 pm. Sitting suspended. On resuming—

First off, in the consultancy playbook, we set out quite a lot of lessons learned about how to get the best out of the use of consultancy. A good example of that is knowledge transfer. A few years ago you might have seen multiple Departments buying similar things. We have really tried to crack down on how we share best practice, transfer knowledge and ensure that we are passing information on to permanent civil servants so that the value of the consultancy is delivered in future. That is something that we have to keep working at. The other thing is about ways of working. The NAO Report touches on this. We give guidance about having very clear scope and trying to blend teams so that we are not just bringing in consultants to operate in isolation. They should be working embedded within civil service teams with clear accountability, clear outcomes and clear ways of working so that we are operating as one organisation. The lessons learned that the NAO has produced, particularly figure 1, are incredibly helpful in reinforcing what we say within the consultancy playbook about how you go about getting the best out of consultancy in practice.

I have only been able to go back and find a certain amount of evaluation of what was undertaken. It is a bit mixed. Certainly, in the time that it was open, which was just shy of two years, it clearly demonstrated that you could provide expert advice and support to Departments in a cheaper, cost-effective way. There is no doubt about that. The one thing I would say is that, even though the hub shut, we have similar resourcing specialist arrangements. Both Mark and Andrew have talked about the complex transactions team, which in effect is a very specialised consulting unit within the Government commercial teams held within the Cabinet Office. The same is true of NISTA, the National Infrastructure and Service Transformation Authority. The same is true of many of our functions. The Public Sector Fraud Authority has specialists who can go and support Departments. Even though that hub has shut, we have taken quite a few of the lessons learned. There is clearly a demand for specialist in-house resources in the civil service to deploy support to Departments where consultants might have been used historically.

A couple of the remits of the consulting hub have been transferred and are being delivered by other parts of Government. Not all the services have stopped. The Crown Commercial Service offers an advisory service called Prosper, which is very much about supporting Departments. There is a small team that will support a Department by giving them advice about the best way to go to market, the best way to put their requirements together and how to contract. It is like a consulting arm. The best practice guidance, which was owned by the hub, has been transitioned to me in the Government Commercial Function. That is the playbook. We are still active. We are still pushing some of those services from different parts of Government.

We have done a review of the core advice, which talks about the definitions, the approach to market, the best way to contract, trying to contract for outcomes and how to manage consultancy suppliers. It is pretty much up to date. It still has the best practice guidance. There is some terminology that needs to change. It is going to be reissued in early 2026, probably by the end of the first quarter.

The Crown Commercial Service runs the largest frameworks. The whole public sector calls on many of these services. They run communities of practice. They run special advisory groups that are open to all public sector contracting authorities. There are a number of services there, not just for Government, that are available to local authorities. Those services bring together users of consultancy services and contracts. There are services out there much wider than for Government Departments.

Mr Forzani, can I just ask a little about the consultancy playbook? What work are you doing to ensure that Departments use it and are complying with its requirements?

Mark was talking about the controls that operate at Department level and the thresholds. Those numbers are the financial controls that have been set. Up to £100,000, it should be going through some formal governance process at Department level, and it should be signed off by the accounting officer. Over £600,000, it should go to a Minister for sign-off. As part of the guidance to support Departments, the consultancy playbook is the gold standard for how we expect Departments to define their consultancy requirements, go to market and contract. We also say to them that CCS should be the default route to market. It is really through compliance and the use of the playbook that we exercise the local controls. The consultancy spend flows through a Department’s commercial director. That is where the contracting is done. We are very focused on the spend that each Department is doing and the target that each Department has. We monitor that on a regular basis through the commercial directors and the commercial function.

The corollary question is whether you have a naughty list of Departments that are not adhering to it.

We get to hear about poor practice. The market will often tell you about poor practice, if Departments and end users are contracting in a way that is outside. One of the things we are trying to do is shift away from time and materials contracting only. We are trying to move much more to outcome-based contracting using payment by results or milestone payments, et cetera. It is very common for the market and suppliers to tell us in the centre where they think Departments are not contracting in compliance with the playbook. We then follow that through.

Thank you for the blush-sparing that you have done on that. Where does overall responsibility for the Government’s IT procurement sit?

The majority sits at departmental level, although we have just established a joint procurement team between the Cabinet Office and DSIT to look at where we can build guidance and more common standards for how we go about procuring IT consumables and services. We are now starting to experiment with cross-Government procurement in this space. Andrew can add more.

We also have a digital playbook, which sets the best practice in terms of how we should be approaching the market for those digital spend categories. There is this mix between what we hold at the centre and the Departments ultimately delivering the programmes themselves. We set some standards around the contracting. We also manage centrally the relationships for many of the large digital suppliers through the Cabinet Office. Many of them are on our strategic partnering programme. We will review the overall relationship and the overall spend. We will try to monitor performance for those big suppliers across a portfolio of contracts, but the vast majority of spend will ultimately go through one of the Departments. They are accountable for letting those contracts and ensuring that they deliver against those outcomes.

Is that working well at the moment?

As always, there is a mixed bag of success. That portfolio will always have some of the most challenging and difficult programmes because most of the Government’s transformation sits in that portfolio.

The hub used to do quite a lot of triaging to help direct where consultancy could be undertaken differently within Government. That does not happen in quite the same co-ordinated way, although CCS, through Prosper and the Government Commercial Function, will do a lot of that at local level. The other thing is delivery. There were only 40 people within the Government consulting hub at its peak. Some of that delivery will happen in different ways now. That is clear.

It is quite mixed. Again, this is a bit anecdotal from talking to Departments about it. It is very hard for a group of 40 people in the Cabinet Office to compete with the skills, expertise and breadth of the very large global consultancy firms. It had its place, but it was not in operation long enough to answer your question fully, if I am frank.

At the beginning you gave us some what I think is quite good news, if it comes off, about workforce planning. When will you be able to publish the strategic workforce plan for the civil service? I appreciate that is an absolutely enormous job. Given the amount of money involved on the consultant side and the amount of spend that you could also make if you identified skills through a skills audit, it is something that should be done sooner rather than later.

I totally agree with you. The Government recommitted their desire to publish this in the spending review. They mentioned it again at the Budget. Once we are through the business planning round, which in most Departments concludes in January, I would expect us to collate that together into something that can be published, but the Government will have to decide the timing of it. I would sincerely hope it is in the next three to four months.

That is good. Presumably, it will also be able to identify where the skills gap is weakest and where you are using consultants, and therefore where you can up the internal skills and reduce the external expenditure on these very expensive consultants.

I can give you some of that information now. The strategic workforce plan will set out precisely in which Departments and how we are tackling all the skills gaps. The two biggest gaps are in digital and technology skills—that will not be a surprise—and project delivery and project management. There is quite a range of different skills within that. Mark talked earlier about some of the very complex and quite sophisticated infrastructure consultancy and contingent labour that we use. That is definitely an area where we tend to see consultancy adding value because it would not necessarily make sense for us to rebuild some of the very specialist infrastructure support that is utilised particularly by DESNZ, MoD and the Department for Transport.

It is interesting that Blaise Metreweli, the new head of MI6, said almost exactly the same thing today about computer skills and digital skills. It is obviously not just the civil service.

It is a function of the labour market. When you talk to private sector colleagues, the demand is so high. It is not just the civil service that struggles to recruit the capacity and capability that we need.

That is good for the next generation, is it not? They are the ones who have it.

Can I ask a quick follow-up on contingent labour? I am recalling a question that we asked of the BBC director general when it came to using freelancers and other things. When you are looking at reducing the workforce because of AI, you are going to be able to look at this across consultants, permanent staff and freelance staff. You are also going to be able to change the mix slightly, in that you may be able to use more contingent labour. Do you speak to those who are part of that contingent labour to ensure they are happy being contingent labour? Going back to the BBC, there is a thing in broadcasting where often the freelancers do not want to be freelancers. Often they do, but often they do not. One of the easiest ways to check how the contingent labour are feeling about being contingent labour is to talk to them in the same way that you would talk to your staff in staff surveys. Do you talk to contingent labour? Do they give you a sense of how they feel?

We do. You will be aware that there are lots of different types of contingent labour. Quite often people choose it because of a lifestyle choice or for flexibility in ways of working. They do not want to be permanent, substantive civil servants, but they do want to come and work on exciting projects and use their skills in different ways. It is quite hard to generalise. For those Departments and functions that use very high levels of contingent labour, I am aware that permanent secretaries and colleagues regularly engage with them to understand how they feel. You will also be aware that there are quite complicated tax arrangements for how staff are treated. It really matters to us that we stick to the IR35 taxation rules and that we treat those individuals not as substantive employees if they are within scope of IR35.

You are talking to them, so if someone has come in on a freelance basis, you would give them the same opportunities to feed back on how they are feeling as you would to any other staff member?

Most contingent labour are embedded within teams. You would not be able to tell that they were any different from any other civil servant in most cases, in my experience.

I want to talk about AI for a minute. We have talked a huge amount with you about this gap in many different ways. The NAO Report, Use of artificial intelligence in Government, from autumn 2023, said that the Central Data and Digital Office “estimated the opportunity for productivity gains by analysing the types of tasks civil servants spend their time on and quantifying the efficiencies if routine tasks were to be automated by AI” and found that the potential productivity gains would be “worth billions”. When we talk about the AI productivity gain, we tend to talk about it within staff functions. Given that you have elements of staff, consultancy and contingent labour, are you ensuring that you can deploy AI across the three different types of workforce that you can bring in as required?

It is a really good question. Yes, certainly for the whole of the civil service and certainly by function. We are currently doing a lot of work to get the whole of the civil service upskilled, regardless of what category our staff are in, whether it is contingent labour, fixed-term contracts or substantive civil servants. As I have mentioned to you before, we are training the whole of the civil service on AI skills for all. That will include everyone. We are right in the middle of that training exercise. When it comes to consultancy, it depends on the type of consultancy, if I am honest. Quite a lot of it will be reliant on knowledge transfer into the civil service. I do not know whether we have done a more comprehensive look at it. It is something that I am very happy to take away and give you a fuller answer on.

It is more about whether you are evaluating the potential productivity gains from AI in staff, consultants and contingent labour. If AI is not being used to make gains across the board, one part of the workforce could bear a bigger brunt than another as we go forward. You might have to get back to us on this one. I would imagine that consultants themselves are very aware of data in the context of AI. When you are procuring consultancy services, are you finding that they are starting to put elements of AI into the contracts and ensuring what they are doing with data? Is data-sharing changing or evolving because of AI in a way it would not have done five years ago?

The whole of the professional services and consultancy model is completely changing because of AI. We experience that and feel it in the way in which we are working with that market every single day. Gone are the days where you had large pyramids within professional services. AI and agentic AI is fundamentally changing the way that they work and the services that they provide to us. We are absolutely seeing that in pretty much every single part of the work we do with that sector. Andrew and Mark might have some specific examples.

I am just trying to get to the bones of it. If I was one of your big consultancy firms, I would be doing everything I could to put into my contract that you are not able to use my data to get rid of me in the future by automating it through AI. There is potentially an in-built protection for consultants in a way that is not necessarily there for the workforce.

It depends. Mark might have some examples.

If we think about the work that we do and how we contract, we are using very standard terms. Because of the way we are contracting and using frameworks, the opportunity for other people to renegotiate those terms is minimal. At this moment in time, the standard protections about protecting our data and continuing data-sharing in the way that we have done traditionally still apply. In the past, not so much recently, you used to get people doing market surveys, data analysis or literature analysis. Clearly, our colleagues internally are using their own growing AI skills to develop that skill capability. We are seeing a shift in terms of what people are looking for, particularly more and deeper into that professional service space that I talked about earlier. We will continue to see that impact on the marketplace.

Just as an example, the latest framework the Crown Commercial Service launched in the autumn has, to the very point that you made, really strengthened the intellectual property provisions to give Government a much stronger position to have that knowledge transfer so we can keep the intellectual property for lots of the work, training, guidance and materials that get delivered. We cannot do that across the board. Some of the services that these companies supply are their intellectual property. We have really strengthened those provisions so that, as much as possible, we can utilise them and have that transfer of knowledge.

Maybe we are going to build our own AI models or software platforms. If you are bringing in Microsoft or any of these other companies and buying in AI, is there a difference between buying in an AI service and buying in a consultant? Will AI become, in effect, a consultant, as we would see it expressed in your playbook?

The large, commoditised AI packages such as those of Microsoft and Google are augmentative. They drive productivity. You are always going to need humans to apply judgment in the use of these AI tools. The sorts of things that we are seeing in the market, though, are around people developing quite specialist AI to go after fraud or the delivery of services. That is different. This is a fast-evolving market. It comes back to the same principles. You only want to be buying consultancy where it is genuinely adding value beyond the tools that you have, whether it is AI or people.

Is the concept of AI as a potential form of consultancy going to complicate your mission to have a uniform definition of consultancy services? We have talked about AI as a technological solution. We have not really talked about AI in any of these sessions as a consultancy solution. I suppose this is the place to explore that for a minute. Is it, “No, AI is never going to be a definition of consultancy” or, “Yes, maybe it is and maybe we will look at that”?

I personally do not think AI is ever going to displace consultancy because the market will reshape itself to make sure it is adding value over and above the technology. In my experience, the best value you get is when you put people, AI and traditional technology together. The blend and the shape of the way in which we use consultants and tech will change, but AI in itself is not going to become a different form of consultancy. It is going to shape the market.

Lastly, we have talked about the difficulties you have with your workforce, your consultants and even the definitions. For 50 years, workforce, consultants and contingent labour have been the three parts. AI is now becoming an extra part of the workforce. You are bringing in an incredibly disruptive and powerful fourth element to how workforce planning can be done. You have only had three in the past and, as we have been discovering in this session, even trying to rationalise how we talk about the three with a shared understanding is incredibly difficult. Is AI going to make your life, in terms of the definitions and planning, more difficult in the short term? It will realise, we hope, longer-term benefits. Is there a short-term additional complication and disruption even to workforce planning and these definitions?

The good news is that this question is being grappled by every single employer in the country in terms of how it shapes our workforces. We can only deal with the facts as we know them now and the way in which the civil service plans to use AI, which is to augment a lot of our work. We are embedding it and mainstreaming it into all the technology available for our staff. Where appropriate, we are using it to tackle specific problems such as going after fraud. We are using that information as best we can to project into our strategic workforce plans what we think will happen. If there is one thing I have learned about AI over the last couple of years, it is that it is very hard to forecast exactly how this technology is going to be used three, four or five years into the future. You are probably right that in the short term we will have to try to work out what it really means in practice, but it is not just us.

To give you my background, I was head of strategy at the Environment Agency and I represent what is almost certainly the most beautiful constituency in the country, Tiverton and Minehead, with its farmers and all their environmental problems, the sea and everything like that. I want to go back to your point about expertise. I want to ask Mr Leigh a question about that particularly in relation to DESNZ. For all your AI, Microsoft and this sort of thing, at the end of the day, for something as complex as climate change and net zero science, you are going to have to rely on expertise. Have you done a strategic workforce plan to inform that particular challenge? Have you been able to identify the skills gap there, recognising that you will still require external resources, such as very specialist consultants, which are jobs that just will not be able to be done by AI or anything else? It is really important, this climate change thing. I know that there are some people who do not think it is, but they are wrong.

From my perspective, it mirrors some of the things that I talked about earlier. We have a workforce plan. We have identified in the operation of the Department the key professions that we want to see improve and drive that capability. If we look at the technical and scientific end of that spectrum, we clearly have a chief scientific officer who gives us guidance. We will probably continue to require support in those areas from the marketplace, and we will be looking really carefully to make sure we do not get an AI swap solution—we have all heard that phrase—but instead we get a genuinely expert solution. That is the work that my team and I will do alongside my policy colleagues to look really sharply at that as the market evolves. That is what we would expect to do. I hope that answers the question. There will be some things that AI will be able to help us with. Andrew talks to me about those all the time. There will be some things that it will not be able to help us with, or it will not at this stage of its development. Let me put it that way.

That is very reassuring. Thank you very much.

It is not necessarily a blanket definition. We have definitions for how you measure efficiency, value for money and productivity. We expect Departments to apply those definitions to make their own assessments. It is not for us centrally to judge whether individual choices are value for money or not. We provide lots of guidance. In terms of outcomes, as Andrew mentioned earlier, we should only be using consultants where it is absolutely over and above the current skillset of the civil service, it is short term, it is something that you cannot develop the skills for quickly and nor is it desirable to do so. If it is absolutely necessary to achieve a Government outcome, it is perfectly legitimate that it could be seen to be value for money. I keep coming back to the point that we know that consultancy is very expensive. We know that sometimes it is used in ways that perhaps not everyone would judge to be value for money. We have to keep going at the definitions and make sure that Departments are taking decisions with all those factors in mind.

We embed international best practice through both the functions, and the Cabinet Office collects international best practice, which we disseminate. I know the Treasury does the same for financial management. A good example is that, every year, we take part in the global surveys of best practice, which look at financial management, digital and civil service administration. That is co-ordinated through the Blavatnik School of Government. We also participate in something called the Global Government Forum, which allows us to extract best practice and share knowledge and expertise. There is a huge amount of that going on both professionally and through central work that we co-ordinate.

Quite often, we will apply it within functions. Andrew, as head of the Government Commercial Function, will liaise with counterparts around the world and we will adopt best practice where we think it is better than ours. The NAO Report sets out a number of different jurisdictions where there are things we need to learn from and adopt as best practice. It is not unusual for us to update our guidance based on international best practice. In fact, we are doing it right now for some of the digital work that we do across Government, where other countries such as Singapore are clearly further advanced than us.

It is a really difficult thing for us to assure in the centre of Government. We do everything that we can to support Departments to adopt best practice, but we do not necessarily direct it and then check it.

It is the individual Departments’ responsibility. We run a federated accountability model. It is our job in the centre of Government to provide best practice guidance and tools to facilitate and check, where it is a Government priority. We do not do that for absolutely everything. We do it for the things that are most important to delivering Government outcomes. Ultimately, it is down to my permanent secretary colleagues across the 20-odd Departments.

Just picking up on that point, earlier we were talking about the definition of a consultant and the best way to feed data into the system. It strikes me that your Department within the centre, Cat, must occasionally be quietly screaming into the void when you put this best practice out and permanent secretaries are not adopting it. You have said that you have a model where you suggest they should do it but you do not force them to do it. It begs the question: if you are putting a lot of time and effort into producing best practice guidance and a significant proportion of people are not following it, what is the point? What is the point of doing it, if you cannot in some way insist on a change once you have identified the best practice and you know that certain Departments are not following that best practice?

This comes down to incentives and repercussions, does it not?

It is carrots and sticks, yes.

There are repercussions quite often for things that we see as absolutely mandatory and critical to the effective running of the state. There are certain priorities that are within permanent secretary objectives. For example, the 50% reduction in consultancy is something that we take very seriously. To avoid any doubt, I should really put on the record that we are very confident we will deliver our £550 million reduction in consultancy because we have put a huge amount of effort into double-checking those figures. We hold Departments to account. We have personally put it into senior colleagues’ objectives. There are certain things where we absolutely mandate and we expect people to comply, but you have to have a boundary over how much the centre is checking and controlling the system. It has to be proportionate to risk and what you are trying to achieve.

Why?

The centre has limited resources, time and capacity. We are not there to double-check everything. By the way, best practice often comes from Departments. As the NAO Report highlights, there are some really good case studies that come from Departments as well as from the centre.

I have two questions. In responding to my colleague’s question earlier, you mentioned something like a naughty list. I would be very keen to see the detail of which Departments are currently on course to reduce consultancy fees versus those that you perhaps have challenged. Could you provide this Committee with more information around that? I appreciate that we are not at the full end of the year, but certainly we are into Q3. You will get a sentiment on that. My second question is linked to that. I appreciate that Ernst & Young and the big four will have audit arms and consultancy arms, and the two are entirely separate. I would be keen to understand whether you have a total figure for the amount that one Department has given to Ernst & Young, for instance, for everything and then a total figure for consultancy and audit. I want to understand whether or not the overall fee that is going to a consultancy firm or audit firm has gone down or whether it is just the pure consultancy element. I am just wondering whether there has been further analysis of that.

On your first question, yes. Every single published annual report and accounts will set out in the front section their consultancy spend. We can provide further detail to brigade all that together for you. That is fine. On the second question, I would have to ask Andrew whether that is possible.

We certainly have it. We have the annual spend. We then have it broken down by Department. I do not know whether we have it broken down by type of service, but we could get that easily.

The reason I ask is because, if you do not have that information, how do you know what is consultancy versus audit? I will give you an example. Take forensic accounting, for instance. They are providing accounting services, but they could give advice in terms of how to restructure a business to ensure that you do not have the same issues going forward. My query from the answer is about whether you have that granular-level breakdown in terms of the total spend that each Department has made with, say, Ernst & Young? How is that broken down component-wise. I want to understand how you are calculating the figure that this has gone down by half, if that makes sense.

You are asking for two things, if I may. I can give you the calculation of how we have come to the £550 million by Department. That is absolutely fine. I can set that out for you. Your second question about how we distinguish between professional services and audit arrangements is a much more granular exercise. I am confident that, if I asked every single Department to give me that information, they could do it, but I would question to what end you want it. Most of our audit services are not provided by external organisations. They are where they are not under the remit of the National Audit Office, but the vast majority are.

To come back on that last point about why I ask the question, let me give you an example. If £1 billion was given by DWP to Ernst & Young over the course of the year, but £800 million of that is for audit and only £200 million is consultancy, the actual amount of money going out the door to an external provider to engage in services has gone down by only 10% rather than 50%. The reason I ask is because the public want to see real scrutiny of the amount of money that is shifting from the public sector to the private sector, to a company that many perceive to be consultancy. I appreciate that they do advisory and audit services, but I want to make sure that the public are getting the full provision around that.

The best way to do that is how we have assured ourselves, which is to look at every other category of expenditure and prove that it is coming down. I can show you that professional services, management consultancy and contingent labour costs are all coming down across all parts of Government. Rather than doing it by a specific supplier, it would be much more value for money for us to provide that information at a global level, if that is satisfactory.

As a liberal, I am not naturally draconian, but I am prepared to make an exception in this case. Going back to the point that my colleague Chris made about shouting into the void and your response about being federated, could we also have a list of the Departments that you have been in touch with to see their workforce planning that have not got back to you? That will undermine your ability to get everything finished, as you said, within the next three or four months. Could we also have an undertaking that, if this is a permanent secretary’s KPI and it is not met, there will be appropriate HR sanctions? If there are not, this could go on ad infinitum and you will not be able to do your job, which is unfortunate.

First, thank you for backing our processes. That is very important. I would say two things. On the strategic workforce plans, every Department is fully participating. That has been a co-ordinated and concerted effort between the Treasury and the Cabinet Office. All Departments want to take part. This is important information for how we manage the civil service and our staff. We care about it. Secondly, on your other point, permanent secretary performance is overseen by an independent remuneration committee, which is politically overseen and personally signed off by the Prime Minister. I can assure you that all our KPIs and performance objectives are robustly tested.

Let us hope he is watching this, shall we?

Thank you very much indeed, permanent secretary and colleagues, for attending today to give evidence to us. An uncorrected transcript of this hearing will be published on the Committee’s website in the coming days. The Committee will then consider the evidence that you have provided and will produce a report in due course. I will just take this opportunity to wish the panel and, indeed, everybody working in your Departments all the best for Christmas and the new year. Thank you for coming today. Hopefully, we will provide that report for you in due course.