Public Authorities (Fraud, Error and Recovery) Bill (Second sitting)

Rebecca Smith Excerpts

Committee stage
Tuesday 25th February 2025 Public Bill Committees

[The Chair]

We will now hear evidence from Joshua Reddaway from the National Audit Office, and Richard Las from His Majesty’s Revenue and Customs. We have until 2.30 pm.

[Rebecca Smith (South West Devon) (Con)]

Q51 This question is for you, Joshua. As you no doubt know, the Bill will allow the Public Sector Fraud Authority to be established as a separate body from the Cabinet Office. Do you think that is the right approach, and what benefits will it bring?

Joshua Reddaway: I guess I am agnostic as to what is done, but the benefits would be an opportunity for governance and an opportunity for accountability, clarity and transparency. I am sure that we would be absolutely delighted to audit the accounts for the PSFA and help to provide some of that transparency. Of course it is currently incorporated with the Cabinet Office, so it is about a clear line of sight. You have to offset that against the fact that there is an administrative burden for producing things like sets of accounts, and having governance and so on. The bigger question, and the one for the Committee, is whether it will enable better oversight of the powers in the Bill.

[Rebecca Smith]

Having a separate organisation?

Joshua Reddaway: Having a separate body.

[Rebecca Smith]

Q Do you have an opinion on that?

Joshua Reddaway: I do not have a major opinion. I would ask whether you are comfortable with the oversight arrangements. One thing to point out is that this will be the closest thing to an anti-corruption unit that the Government have, with search powers. Are you happy with that being constitutionally in the Cabinet Office or not? I am also interested in what the criteria are. The explanatory notes have set out that there will be an efficiency assessment for the powers in the impact assessment. I am not sure the Cabinet Office is clear on exactly what that means. It is interesting to think about what would actually trigger it to exercise that power under the Bill to create it as a separate body.

[Rebecca Smith]

Q How far do you think the increased time limit that would be in place for investigating fraud related to the covid-19 schemes will improve recovery rates? That is a slightly different topic, but do you think those longer time limits mean that we are more able to recover that money?

Joshua Reddaway: In short, yes. Would you like a longer answer?

[Rebecca Smith]

That would be great.

Joshua Reddaway: First, we should be clear: with most fraud, once the payment has gone, you are not going to get it back. I have a professional next to me who can talk to you about the challenges and the pursuit, but if you ask how much fraud is out there, the answer is a lot. If you add up all the official estimates from the different schemes during covid, it is £10.5 billion-worth of fraud. The Government have so far recovered £1 billion of that, mostly from HMRC and less from others. Of course, HMRC stopped collecting it because it knew that its resources would have a higher return of investment if they were re-diverted back to tax rather than fraud recovery. I am afraid you are always on to a losing game if you enter the recovery phase, but every million counts. It is always nice to get something back.

The covid counter-fraud commissioner has only just been appointed. Their role is to review these schemes and see whether there is a way to get the money back. My understanding is that the six-year time limit under the Fraud Act 2006 will be expiring next spring so, with that timetable alone, if the fraud commissioner is going to bring forward anything that has a chance of working, it makes a lot of sense to give them a bit more time. Like I say, we are really sceptical that it is possible to get the majority of that £10.5 billion back. Some of it will come back from the bounce back loans anyway, but the vast majority of it has gone. But every million counts.

[The Chair]

Do you want to add anything, Mr Las?

Richard Las: On the covid side of things, we have not stopped our efforts, but we have recognised that we are not going to pursue it as a lead subject. However, we are conducting other inquiries and looking to other taxes. We will be looking at whether there was fraud under the covid schemes, and we will still be pursuing that. I still have a large number of cases going through the courts or heading towards prosecution in relation to the scheme. A bit like Joshua, I am certainly not giving up on it—we will keep pursuing it—but, in a decision on how we deploy our resources, we are saying, “We’ll look to what we think are the higher risks, and we will pick up the covid risks as and when we come across them at the same time.”

--- Later in debate ---

[Andrew Western]

Q I have a very general question to open with. You will be more aware than I am of the changing nature of fraud and the increasing sophistication that we see from perpetrators. Do you agree, in general, that the DWP’s powers would need to be modernised in order to cope with that shift? Also, what are your views on the general provisions within the Bill, where it pertains to the DWP, to detect and prevent fraud?

John Smart: At the risk of echoing what has been said before, I think it is critical that we modernise the approach to fraud, and the Bill is a good step towards that modernisation. The critical part of a lot of investigations now—and of identifying, preventing and detecting fraud—is the use of data. Getting that data and information quickly and effectively is critical. I think the Bill will go a long way towards speeding up and broadening the available information that can be used to prevent, detect and prosecute fraud. That is a really valuable thing that we should be pushing for, because relying on pieces of paper to seek information from organisations is crazy in this day and age, when you can do it electronically and get an answer relatively quickly. If you are turning up with a piece of paper, it can take weeks or months.

[Rebecca Smith]

Q Based on your experience and the work that you are already doing with the PSFA, are there any gaps in the Bill? Is there anything that you think is not there that would help us to tackle fraud against public authorities, or do you think that as it stands, it is about as complete as you would want it to be?

John Smart: Having worried about this for a number of years, I think there are a lot of steps that the Government—the PSFA—can take over time, but we are on a ladder to get to a position that is constantly moving because the fraudsters are developing all the time. One critical thing that I have been concerned about for a number of years is the use and sharing of data across Government. Government have so much data available to them, and third parties have a lot of data available to them. There is clearly a privacy question that rapidly comes into play, but from my perspective, if the data is available to Government, they should use it. They should use it proportionately: they should not exploit those powers to use that data on some sort of phishing trip, but if there is evidence that fraud is being or has been committed, getting that evidence in the hands of investigators quickly is critical to preventing the fraud from continuing and to identifying and recovering any money that has been lost. To my mind, there is quite a lot of work still to be done on data sharing across Government.

[Rebecca Smith]

Q Given what you just said about the nature of the data that is already collected, could we avoid the fraud in the first place by sharing a lot of that information? If cross-departmental working is tightened up, might there be opportunities to flag fraud as soon as it starts to happen?

John Smart: Absolutely. There are two points to make. The first is that that frauds that are already happening would be identified if the data was shared more effectively and quickly. Secondly, by joining up data that is sitting in Companies House, the licensing authority, or wherever, you can find evidence that a fraud is being carried out and prevent frauds from happening in the first place.

[Rebecca Smith]

Q Interestingly, one of the previous witnesses talked about powers that other countries have but we do not, which potentially would do exactly what you have just described. Are there any countries already doing something along the lines of what you just suggested it would be helpful for us to look at—perhaps not for this Bill, but in the future?

John Smart: An obvious example is the United States; there is an interesting case in point at the moment, which I have dealt with quite a lot. The US has whistleblower reward legislation in place, which is very effective at flushing out issues affecting payments made by Government. Their qui tam legislation, as it is called, flushes out frauds by incentivising whistleblowers to blow the whistle. It creates a lot of work for various organisations, but it encourages people to think about whether fraud is being committed against the Government in the US. That is an obvious piece of legislation that might be worth considering in this country.

[Georgia Gould]

Q As an adviser to the PSFA, you have seen the organisation grow. What is your assessment of where the PSFA is in tackling public sector frauds, and how the Bill moves us on?

John Smart: That is a big question. I have been involved with the Cabinet Office for over 12 years, so the inception of the PSFA came about while I was working there. In the 18 months since it was formed, the PSFA has gone a long way to reach a better understanding of where the issues sit across Government. Clearly, it plays best outside the DWP and HMRC. My passion has been identifying where fraud is taking place, which I have worked on for the past 10 years, and trying to quantify the fraud occurring within Government. As you all know, that is very hard to quantify because it is hidden and therefore unknown. The PSFA has gone a long way and is continuing to flush out where resources should be committed to preventing, investigating and deterring fraud across Government outside HMRC and the DWP. That is critical. When I first started asking Departments where frauds were within the Departments, they replied, “There’s nothing to see here.” At least now, particularly because of the work the PSFA has been doing, there is recognition that there is a real issue to be addressed, and that it is not just expenses fraud, or whatever they used to think it was.

--- Later in debate ---

[Georgia Gould]

Thank you very much.

[Rebecca Smith]

Q We see that there has been a lack of robust assessment of business costs so far. The Government have been unable to say what the cost to the banks will be. Do you have any concerns about the costs or the other impacts on businesses in the sectors? I suppose the flip question of that is: do you think there should be incentives to get them to actually work with it? It feels to me like there is a huge amount of good faith here, in that banks are expected to bear the costs of providing this information, but are there unforeseen costs that we need to be concerned about as well?

Daniel Cichocki: In terms of broad principles, obviously wherever there is additional legislation and regulation on the sector, we would hope that that is proportionate. We anticipate doing further work with the Government to help to support the impact assessment as a result of the more detailed work when we see the draft code of practice, when we are better able to understand the methods through which this information will be shared, the practicalities of how it works, and the scale at which the powers will be used. We therefore anticipate more work being done around the impact assessment.

We would hope and anticipate that the Government would recognise that the impact on the private sector needs to be proportionate. As well as the cost implications around resource, this is also around prioritisation. To my earlier point, many of the teams that will be complying with this legislation will currently be complying with the broader legislation and regulation that we have in place, sharing information with the Government and law enforcement, and ensuring proportionality of how that resource is deployed. Certainly from an industry perspective, as a broad principle, we would see it as appropriate and desirable for much of that resource to be focused on serious and organised crime in the round.

Eric Leenders: I have a couple of brief points. First, one consideration is congestion. There is quite a crowded mandatory change stack, as we call it. There is a sequence of changes in train that firms are already implementing. Secondly, to your specific point about the cost-benefit analysis, we recognise the challenge that the cost will be direct, as in the build costs that we have just summarised. The benefits—reducing and deterring criminality generally, and perhaps even preventing it—are perhaps more indirect. I suppose that leads to another point: the extent to which we need to be thoughtful about circumvention and how to ensure that the legislation is suitably agile, so that bad actors cannot game the system no sooner than it has been introduced.

[The Chair]

There being no further questions, I thank our witnesses for their evidence. We will move on to the next panel.

Examination of Witness

Ellen Lefley gave evidence.

[The Chair]

We will now hear from Ellen Lefley, senior lawyer at Justice. We have until half-past 3 o’clock.

[Rebecca Smith]

Q You will have heard our previous questions. We talked about the cost-benefit analysis for businesses in the private sector. In your Justice work, what is your assessment of the proportionality of that measures in the Bill that, in essence, tell financial institutions that they have to share that information with the DWP? What do you perceive as the cost-benefit of that?

Ellen Lefley: It is right that the bank power, which is the eligibility verification measure, is separated out in terms of proportionality because, just to clarify, it is important that the other powers of information, search, entry and seizure, which are extended by the Bill to the PSFA and to DWP, all contain that threshold form of words of needing “reasonable grounds” of suspicion or belief. That threshold for the exercise of state power requires reasonableness and objectivity—for there to be something there. That rule-of-law barrier prevents fishing expeditions and state intervention in people’s lives when there is simply nothing to it.

Any such form of words, however, is missing from the eligibility verification measure, which is why the privacy concerns and the concerns about the proportionality of the measure have been so concentrated. Justice is concerned about the proportionality of the measure precisely because it does not have that threshold of reasonable suspicion and because of the vast numbers that could be subject to it, albeit that the state pension has been taken out of scope—it was in scope before, under the almost-equivalent measure in the Data Protection and Digital Information Bill last year.

The concern is with the broadness of that power, the lack of a threshold and the fact that the fundamental right to privacy is involved. We all have a right to privacy, and we all have a right to enjoy our privacy in a non-discriminatory way, and that is the further issue that I would raise. I am sure that others will raise this today, too: the almost inevitable disproportionate impact that those financial surveillance powers will have on people who are disabled. There has been no equality impact assessment for this Bill, but there was for the previous one—not that it was released, I think, but it was the subject of a freedom of information request and I had sight of it. It revealed that, even though about 23% of the population at large are disabled, that figure is about 50% for the benefits-receiving population. There is that prima facie disparity. The financial privacy that is enjoyed by citizens of this country and people who reside here is less protected for disabled people than for others. That very much needs to be proportionate and justifiable, given the fundamental rights that are engaged.

[Rebecca Smith]

Q On eligibility verification, the Government have said that the final decision will always be made by a human. There is an aim to automate some things, which is not necessarily a bad thing, but do you feel reassured by the human element at the end of the process, so that people are not adversely affected by automation, or are you concerned that that will still not be far enough or good enough?

Ellen Lefley: Reassurance cannot be the word, unfortunately, given the moment we are in, which is one of increasing automation and increasing investment in data analytics and machine learning across government. Last month, I think, we had a Government statement about mainlining AI into the veins of the nation—that includes the public sector. Knowing that that is coming and having a clear focus on how the functions in the Bill will be operationalised need to be a key concern.

The preservation of human intervention in decision making might have been a statement that has been made, but it is not on the face of the Bill. Indeed, we need to remember that the Data (Use and Access) Bill, which is also before Parliament, is removing the prohibition on fully automated decision making and profiling. That is happening concurrently with these powers. In addition, over the years, there have been numerous Horizon-like scandals that have happened in the benefits area. One, quite close to home in the Netherlands, was a childcare benefit scandal, which Committee members will know of. In that scandal, recipients of childcare benefit allowance in the Netherlands were subject to machine-learning algorithms that learnt to flag a fraud risk simply because of their dual nationality. So there is a problem here. Even with the powers that are subject to reasonable grounds, we need to have a wider discussion as to what reasonable means and what it definitely does not mean when we talk about reasonable grounds of suspicion, when suspicion is an exercise that is informed in a tech-assisted and technosocial decision-making environment.

Justice has some suggestions as to how reasonable grounds can be better glossed in the Bill in relation to generalisations and stereotypes that a certain type of person, simply because of their characteristics, is more likely to commit fraud than others. Perhaps it could be recorded in the Bill that that definitely is not reasonable.

Some useful wording from the Police and Criminal Evidence Act code of practice A is not in the Bill because it relates to the power to stop and search, which is not being given to DWP officers, probably rightly and proportionately, but some explicit paragraphs in the code of practice for stop and search for police officers say that they cannot stop and search someone based on their protected characteristics. Under the Equality Act 2010, they cannot exercise their discretion to stop and search someone due to generalisations and stereotypes about a certain type of person’s propensity to commit criminal activity. Amendments like those could strengthen the Bill against unreasonable, but perhaps not always detectable suspicions being imbued by machine-learning algorithms. Of course, if there will always be a human intervention in the decision-making process, perhaps that could be explicitly recorded in the Bill as well.

[Andrew Western]

Q Your last point about stop and search and decisions being made purely based on protected characteristics speaks to what you said earlier about the perceived impact on disabled people. Are you suggesting that the eligibility-verification measure would directly discriminate against disabled people, or is it merely that disabled people make up a larger number of the cohort?

Ellen Lefley: They make up a larger number of the cohort, so we would analyse a prima facie indirect discrimination potential risk there, which would then need to be justified as being necessary and proportionate. The proportionality assessment of course is for Parliament, but we consider that a significant amount of scrutiny is required not only because of the privacy impacts, but because there is that clear indirect discrimination aspect. I am not alleging direct—

--- Later in debate ---

[Georgia Gould]

Q Concerns were raised earlier that the PSFA would take its eye off the ball on prevention and the wider area of work, and become an agency focused only on enforcement. It would be helpful if you could address and answer those concerns.

Mark Cheeseman: The Public Sector Fraud Authority has two elements to it. One is overseeing Government and how individual Departments are doing in dealing with fraud and what they are doing on it; the Bill itself says that Departments would refer cases to the PSFA and ask for them to be dealt with under it. The second is providing some of the services that support Departments around taking action on fraud where it happens.

The biggest difference we will make, alongside that, is through prevention. We heard from witnesses earlier about the use of data and analytics. We have a data and analytics service that works with public bodies to use that to find and prevent fraud up-front. We also have a risk service that works with other parts of the public sector to understand the risks they face, in order again to prevent those risks by putting in controls.

While there will always be that balance, there will also always be some element of fraud that is still committed. We will not be able to design a system where there is no fraud risk or design out fraud. There will always be cause for an efficient, effective and proportionate part of the machine to take action on those instances of fraud and to investigate them thoroughly and properly.

[Rebecca Smith]

Q This is possibly a slightly nuanced question, but there has been a lot of talk about the authorised officer role. To my mind, when you go from the Minister to the authorised officer, that feels like quite a big jump in title. It might be quite a small thing but, from your experience at the PSFA, does that terminology work when looking to expand beyond that, in terms of people’s understanding of how senior the people doing these investigations are? The word “officer” is often used, particularly in the public sector, to mean quite a junior role. How do we ensure that the public and people across the public sector understand how senior these people are? Does that need to be on the face of the Bill or in the code of practice? How would you look at that?

Mark Cheeseman: Again, I do not know whether it would need to be in the Bill; that would be for you to debate. As it gets past the authorised officer, there is a structure: there are senior leaders with deep experience in investigating fraud who are overseeing them. We have structures of senior investigation officers overseeing your investigators and the individual authorised officers. While it may feel like a big jump, there is a structure to ensure quality, to ensure the right practices, and so on. That directly compares with what happens elsewhere.

I am pretty comfortable that “authorised officers” is a term used elsewhere. I recognise what you say about the seniority of grade; I had to have a wry smile, because it took me a while to get to HEO and SEO—higher executive officer and senior executive officer—but those are still senior, experienced roles. They are experienced administrators with a high level of skillset and expertise doing those roles. Part of the reason for creating the counter-fraud profession is to show the expertise and capability that those experienced counter-fraud experts have in taking action on fraud.

[Rebecca Smith]

Thank you.

[The Chair]

There being no further questions, I thank the witness for his evidence and we will move on to the next panel.

Examination of Witness

Jasleen Chaggar gave evidence.

[The Chair]

We will now hear evidence from Jasleen Chaggar, the legal and policy officer at Big Brother Watch. We have until 4.10 pm.

[Rebecca Smith]

Q The Bill is sort of a move on from the Data Protection and Digital Information Bill, which was introduced by the last Government. I know that the Information Commissioner had some concerns about the Bill’s previous iteration, and he has said that those concerns have roughly been addressed in this new Bill. Are you similarly reassured or do you have outstanding concerns about this piece of legislation?

Jasleen Chaggar: We recognise that the Bill is different from the previous Conservative Government’s Bill and some changes have been made. However, we are still concerned that the purported safeguards in the Bill are really insufficient. One of the major safeguards that is pointed to as a reassurance is the fact that financial transaction information and special category data will not be handed over to the DWP from the banks. However, it is a circular safeguard in reality, because once the account number and name of the individual has been passed on to the DWP, it can very easily go back to the bank and request that granular financial information. That is incredibly privacy invasive, as you will know, so we are still concerned about the safeguards in the Bill.

A similar safeguard is the provision for an independent person, but there are no safeguards about what qualifications that person should have. They are expected to provide an annual report to Parliament, but we are concerned that their oversight role is more to do with enforcement than accountability. There are provisions about the efficiency of the measures but no provisions about how they impact equality or the adverse consequences on benefits recipients, so we are not reassured by these safeguards.

[Rebecca Smith]

Q What would get you to a place of feeling more comfortable with those safeguards? A provision that you can use the data once and then you cannot use it again? Where would you need to see movement in order to get to a place where you were comfortable with the safeguards?

Jasleen Chaggar: Is that in relation to—

[Rebecca Smith]

Particularly in relation to bank account details and information on spending, and that sort of thing, which you just used as an example.

Jasleen Chaggar: On the eligibility verification measures—what we are calling the bank spying powers—we are recommending that they be removed in their entirety. They really are unprecedented financial surveillance powers. There are no other laws like this in this country. The powers would permit generalised mass surveillance of everybody’s bank accounts. It is not just benefits claimants who will be targeted; it is everyone’s accounts, including yours and mine. They will be scanned using algorithmic software to make sure that the eligibility indicators are not met. Even if you are a benefits recipient, you can appoint an individual—a parent, a guardian, an appointed person or your landlord—to receive the benefit on your behalf, so those people will also be pulled into the net of surveillance. We do not really see a way in which these measures could ever be proportionate.

[Andrew Western]

Q Thank you for joining us. I have a couple of initial questions. You have neatly set out your position that there is no circumstance in which you would support the eligibility verification measure. I was interested when you said there are no other laws like this in the country. We heard from HMRC today about its ability to receive bulk data with regard to every interest-bearing bank account in the country, and it does that on a regular basis. How do you consider this power to be different from that one?

Jasleen Chaggar: What is really important about the Bill is the conflation of fraud and error. It is not just people suspected of serious crime, or even low-level crime, who are pulled into the net of surveillance. It is also people who, while navigating the complexities of the benefits system, may have found themselves on the wrong side of making a benefits claim and made a mistake. It also involves DWP’s own errors, which make up one in 10 errors. What is critical when we are thinking about the Bill is that it is suspicionless surveillance that applies to everyone.

--- Later in debate ---

[The Chair]

We will now hear oral evidence from Geoff Fimister, of the Campaign for Disability Justice, and Rick Burgess, from the Greater Manchester Disabled People’s Panel, who joins us via video link. For this panel, we have until 4.40 pm. I have introduced the witnesses already, so we will go straight to Rebecca Smith.

[Rebecca Smith]

Q Thank you, Chair, and thank you, gentlemen. We have heard quite a lot of concerns already this afternoon about the potential disproportionate impact on disabled people. It would be good to hear from you as disability campaigners—because we have kind of heard people talking on your behalf so far—about how you think the measures will affect disabled people, and what could be done to address the impacts that may result from the Bill.

Geoff Fimister: I should say, first of all, that the Campaign for Disability Justice was launched relatively recently—a few months ago—by Inclusion Barnet. We now have a substantial number of individuals—several hundred—supporting us, as well as a substantial number of organisations, ranging from large charities to grassroots disabled people’s organisations, so we get quite a lot of feedback.

I suppose our concern with the Bill include a broad aspect, but also a very specific aspect as to how it may impact disabled people. The broad aspect is that, because it focuses very much on means-tested benefits, it will, by definition, disproportionately affect people on low incomes, and disproportionately affect disabled people, because they are more likely to be on low incomes than others.

The practical issue, which I think has attracted the most concern, from the conversations I have had, is false positives, as the previous witness, Jasleen Chaggar, mentioned. We are all familiar with a world in which we have problems with malfunctioning technology. Every few months, my internet provider locks my inbox because of “suspicious activities”, which have included sending an email to an MP’s researcher or one to Mencap. Every now and then, my bank freezes my wife’s and my bank accounts because of “suspicious activity”, such as, on one occasion, purchasing a sandwich from a Marks and Spencer in Deptford.

That might sound entertaining, but it is a serious business; this tech goes wrong, and I think the previous witness made the point that, if large numbers of people are embraced by this kind of trawl, it will go wrong for a percentage of them. We do not know whether that will be a large or a small percentage, but even a small percentage of a big number is a lot of people. People being left without any income if technology triggers the cessation of their benefit is a serious business. Not having any income can cause hardship, debt and stress. In extreme cases, there can be serious health and safety issues. Disabled people are concerned about that kind of eventuality.

As to what we can do about it, I understand the thrust of the Bill and where it is coming from. In parliamentary terms, it has widespread backing, although a number of reservations have been expressed. We would like to see some sort of safeguard whereby benefits could not be stopped unless and until it was established that there was an overpayment—not that the DWP thinks that there might have been because the tech spotted something. We do not want to see a “shoot first and ask questions” later approach. If we could have some protection along those lines, that would be helpful.

Rick Burgess: I stress that I am from the Greater Manchester Coalition of Disabled People. The panel is something we do, but I am not speaking in that role today.

There are particular worries about how this affects people living with mental distress, particularly those with diagnoses of paranoia, schizophrenia, depression or anxiety. This adds to the feeling of being monitored, followed and surveilled, because you literally are being surveilled by your bank on behalf of the Government. So it will necessarily reduce the wellbeing of disabled people who are claiming benefits that are monitored by the system. There is no getting away from that.

On the potential risks, when you enter a trawling operation, you are not targeting it in any way; you are simply looking at everyone. So the error rate becomes extremely important. We do not know exactly what the technology is. We have not seen the equality impact assessment, but even if it had a failure rate of 0.1%, which would be a quite respectable systemic failure rate—it is pretty acceptable in a lot of these areas—that is still 1,000 people per million scanned. If you are talking about even the means-tested benefits, that is going to run to thousands of people getting false positives. If you think about the entire DWP caseload, which is 22.6 million people, that is over 22,000 people. Bearing in mind that the Post Office scandal involved fewer than 1,000 people, you are at the inception stage of something that could be the greatest miscarriage of justice in British history, if you go ahead with this with untested technology that has not had proper impact assessments.

I stress, though, that we are against this measure in its totality because it treats disabled people as a separate population who should have lower privacy rights than the general population. In that respect, given that the United Nations has condemned the UK twice in a row for grave and systemic human rights abuses, this is going further in the wrong direction and failing to address the failures identified by the UN. It is further marking disabled people for additional state oppression and surveillance, which, as I said, will necessarily be harmful to a great many of the people under the surveillance regime.

[Andrew Western]

Q Could I come back on the last point you made, Rick, and the suggestion that this treats disabled people as a distinct part of the population under different rules and measures. The Bill targets the three benefits that have the highest levels of fraud and error at present: universal credit, pension credit and employment and support allowance. I would accept that there is a higher prevalence of people who are disabled in those cohorts, but this is not restricted exclusively to disabled people. Can you elaborate a little on how you feel that disabled people, in isolation, would be treated as a separate entity?

Rick Burgess: Because we are over-represented in those classes. If you choose to target it at those cohorts, you are accepting an additional level of targeting towards disabled people, which is discriminatory.

--- Later in debate ---

[Michael Payne]

Q There has been some discussion today about the use of technology and AI. As Ministers, what are you doing to ensure that humans still take the final decisions on whether the powers should lead to enforcement?

Andrew Western: As it relates to the DWP—I do not know whether you want to say anything about the PFSA powers later, Georgia—it is worth reflecting that the proposed eligibility verification power is in effect a data-push power. The banks will not make any decisions as to someone’s culpability, on what penalty they might receive, or on whether the overpayment flagged on the account is legitimate; all the banks will do is send back a marker against an account to suggest that someone is in breach of their eligibility requirements.

For example, that might include someone who has more than £16,000 in their account, but is in receipt of universal credit. It is important to say that the flag is then passed to a human investigator to analyse the information and look at what the reasons may be, because there can be very legitimate reasons why someone has more than £16,000 in their account and is still entitled to benefits, such as someone who has received a compensation payment that is out of scope of what would be considered capital for benefit-eligibility reasons.

In all the five principal measures on the DWP side of the Bill, a human is involved in the decision making: on eligibility verification, it is passed to an investigator; on information gathering, when we receive information, it is passed to an investigator to consider the next steps in a fraud investigation; on debt recovery, an individual—a person—would make a decision as to someone’s ability to repay a debt; and on penalties reform as proposed, a human will determine what actions will be taken against a person who received a penalty for fraud against a DWP grant scheme. That is entirely the way that it works with any other penalty that can already be applied. Finally, on the powers of search and seizure, as we would expect, a human judge will take a decision on whether to issue a warrant. At every stage, a human decision maker is baked in before any final decision on sanction or otherwise.

[Rebecca Smith]

Q I will ask one question in a second, but I wanted to come back on the point that my hon. Friend the Member for Kingswinford and South Staffordshire made about the code of practice. The Minister has said that each time in our scrutiny we get to a point in the Bill that relates to a bit of the code of practice, he will bring that to us then and we can discuss it. Will that be in writing? It strikes me that if something is there for us to discuss at a particular point, it would make much better sense to have discussed it all at the beginning, so that we can look at it as a whole. Otherwise, I am not quite sure how we will do it in debate.

Andrew Western: A draft code of practice will not be available at that stage, so I will speak in general terms about what we intend to include, but there will not be a written document at that stage.

[Rebecca Smith]

Q It would still be helpful to have something in writing, even in advance of each issue that we can scrutinise, because we are being asked to scrutinise something in its fullness without a level of detail. Anyway, I will leave that there, because it has been covered enough.

Now the question that I was coming to, if I may. The state pension has been explicitly excluded from the eligibility verification measure, and the three means-tested benefits are the initial focus. I wonder why the Government have left it open to include other non-means-tested benefits in future, and what data would the Government ask for in those cases?

Andrew Western: The state pension is excluded—because of the particular nature of the eligibility criteria for state pension and the consequently incredibly small amount of fraud that we see on it—considering the number of people we would have to bring into scope of the measure to go after what is a tiny amount of fraud. It is not considered proportionate to do that, as far as I am concerned.

The determination that we have made as to the three benefits that should initially be in scope is entirely predicated on current levels of fraud and error. We want to retain the ability, if necessary, to bring other benefits into scope, should there be a surge in fraud in those benefit areas. We do not anticipate this, but we want to future-proof the Bill as best as we can, should there be any material changes in the level of fraud in those areas. For instance, if we consider the tiny amount of fraud in the state pension versus the £1 in every £8 currently spent in universal credit that turns out to be fraud or error, it is clearly right to distinguish between benefits and consequently to have some in scope and others not.

[Sally Jameson]

Q What happens when a flag on an account is made under the EVM, and is that sufficient to find that someone has committed fraud?

Andrew Western: I answered this slightly in response to Mr Payne, but the flag in of itself does not mean that someone has been found guilty of fraud. A bank indicating to us that someone has above a certain amount of capital in their account does not mean, “Job done, box ticked”, or that person receives news that they have been found to have committed fraud, or that we then go through the penalty process with that individual. It would be referred to the most appropriate team for investigation—in the case of capital fraud, the team that looks at that particular type of fraud.

The principal other type of fraud that we think would be in scope is people who have been out of the country for longer than they are allowed to be as a condition of their benefit. Again, it is really important that we do not automatically penalise somebody for having done that, because it could be on grounds of a health emergency abroad. I had somebody in my advice surgery recently whose flights had been cancelled due to an environmental issue in the country that he was seeking to return from. It is really important that this is triaged to a human investigator to look into what the nature of the flag is, what the benefit eligibility criterion that we suspect may not have been satisfied is, and then take the appropriate steps needed to establish whether there is any legitimate reason for that.