Work and Pensions Committee
Oral evidence: Fraud and error in the benefits system, HC 1082
Monday 7 April 2014
Ordered by the House of Commons to be published on 7 April 2014.
Written evidence from witnesses:
– Department for Work and Pensions
Members present: Dame Anne Begg (Chair), Graham Evans, Sheila Gilmore, Glenda Jackson, Kwasi Kwarteng, Nigel Mills, Anne Marie Morris, Mr Michael Thornton
Questions 155-252
Witnesses: Lord Freud, Minister for Welfare Reform, Mike Driver, Director General, Finance, Department for Work and Pensions, Mr David Gauke MP, Exchequer Secretary to the Treasury, and departmental Minister for HM Revenue and Customs, and Nick Lodge, Director General, Benefits and Credits, HM Revenue and Customs gave evidence.
Q155 Chair: Can I welcome you along this afternoon to our third and last evidence session in our short inquiry into fraud and error in the benefit system? Lord Freud, if I can start with you, can you introduce yourself and your panel?
Lord Freud: Thank you very much. I am Minister for Welfare Reform at DWP, and I am joined by my colleague, Mike Driver, who is the Finance Director, who is I think well known to this Committee. I am very pleased that David Gauke, the Exchequer Secretary to the Treasury, can join me alongside his colleague, Nick Lodge, who is the Director General for Benefits and Credits. As the Committee is aware, I am on the whip in the Lords at the moment.
Chair: Right, so I have to look out for the bells ringing.
Lord Freud: So I may have to make short runs every now and then, but I will be as quick as possible.
Q156 Chair: It will not be the first time it has happened and it may not be the last. You do seem to be in front of the Committee rather often. Welcome to David Gauke; it is your first time in front of us, so thanks very much for coming.
Can I just look at the apparent emphasis there is on fraud? We always talk about fraud and error together, but official estimates indicate there is around twice as much error in the system than there is in terms of fraud. Yet in the 10 key initiatives highlighted in your 2010 joint fraud and error strategy, seven were entirely focused on fraud, one was related to fraud and error generally, and only two were specifically designed to address error. If the primary concern of the Exchequer is a financial one, should the Government’s top priority not be eliminating error rather than chasing the fraud?
Lord Freud: I will start off, from the perspective of DWP. Clearly, you are right about the figures. The most recent figures show that the fraud figure on the monetary value of fraud and error is running at £1.2 billion, which compares with claimant error of £1.6 billion and official error of £0.8 billion. It is clearly important for us to address all three components, and we do. We are making a number of steps, in both claimant error and official error, to make sure that we make some very substantial reductions. When you look at our figures, you can see that a lot of the reductions we are making, particularly in case-cleansing, are designed to get those figures down.
Let us just take official error first; I will just take some of the initiatives we have got. The first and the most important is simplification—the introduction of Universal Credit and PIP, much greater use of IT, data-sharing and the introduction of RTI and ATLAS. We have a new quality framework to make sure that we have complete information, technical accuracy and consistency of decisions, and we are going through a very substantial process of case-cleansing, which we will come on to, I should think, but there are very substantial savings from that.
Clearly the other area is claimant error; I will deal with that very rapidly. That is very much about making sure that our information, at the outset and further on, is very clear, using plain English, centralising our communication products, and getting awareness messages. We have also introduced a £50 civil penalty to make sure that people give as accurate information as they possibly can.
Q157 Chair: We have questions on almost all of that coming up, but is there anything from the Treasury’s point of view as to why, when two-thirds of the fraud and error is actually in the error side of the equation, there seems to be so much emphasis on the fraud?
Mr Gauke: Just to add to what Lord Freud has said, first of all, in terms of the numbers, of the overall proportion of tax credit expenditure in 2011-12, which was the last year that we have, fraud constituted 3%, official error was less than 0.1%, and claimant error was 4.2%. Again, the majority of it is claimant error. I will not go into all the details of the measures that we are taking at this point, but if you look at the various interventions that are made and the various steps taken to try to address this matter, they are as relevant to error as they are to fraud. As far as we are concerned, the priority is to address losses to the Exchequer, regardless of source—whether it is fraud or error.
Q158 Chair: The reason for asking this question is that in the public mind, fraud looms much larger than the error. The general public believes that losses to benefit fraud are more than 34 times higher than official estimates indicate. They believe that about 24% of all benefit payments are fraudulently claimed, whereas you have just said it is around 0.7%, depending on which benefit you are talking about. Does the emphasis on fraud not risk perpetuating the misconception that benefit fraud is rife in the system? You have admitted that fraud is actually a very small part of what both your Departments pay out?
Mr Gauke: I would reiterate that we are focused on both fraud and error, and there are steps to address that. The number overall, whether one is looking at fraud or error, is higher than we would want it to be, and so therefore we should seek to reduce it. Error itself covers a fairly broad spectrum in terms of the level of culpability on the part of the claimant, but we want to make sure that we address all of it. In terms of what we discuss as to whether we describe it as fraud and error or error and fraud, none the less we are going to reduce it.
Q159 Chair: As the two keep getting conflated, is there a case to separate them out and deal with the fraud on one side and the error on the other but never together? It is always fraud and error, or error and fraud.
Mr Gauke: A lot of the measures one would take to address fraud would be relevant for error and vice versa.
Q160 Glenda Jackson: The public perception, in my opinion promulgated by DWP and possibly your Department as well, is that by far the biggest take is via fraud. It is underlining what has already been criticised, regarding DWP: the misuse of stats, enhancing what, in my opinion, is their black propaganda, attempting to present everybody on benefits as though they are essentially “shirkers”, not “strivers”. Given that Lord Freud has already said that error is part of the Department’s problem by virtue of over‑complicated forms and, I presume, deliveries to claimants that can be easily ignored, surely the issue of fraud is quite minor in that sense? Unless Lord Freud is going to say at some point that there might be fraudulent processes actually within the Department, which I think is highly unlikely. The public image that is being presented is erroneous and surely that should be changed.
Lord Freud: Clearly, there are levels of fraud in different of the benefits that vary. You can look at some of them: you can see the level of fraud is running at 2.9% in Jobseeker’s Allowance, and 3.9% in Carer’s Allowance. There are particular benefits where it is higher. One of the important things to make sure of in a fraud strategy is to prevent fraud. It is important that people do not feel it is easy to do, other people are doing it, and that it is easy to get away with. Particularly when you see the statistics, they show that a lot of people drift into fraud. In other words, they start off correctly, have a change of circumstances and then do not adjust for it, and find themselves drifting into a position they should not be in. It is really important that people are conscious that they should not drift into that position.
Q161 Glenda Jackson: With respect, you completely ignored my question. My question was that surely it is the responsibility of the Department to make sure that the public understand where the greatest misuse of public money is. Clearly the statistics show it is not fraud; it is error in which the Department could be as culpable as the individual claimant, yet what comes out of DWP is the constant litany that the worst take of public money is coming by fraud. That clearly is not the case.
Lord Freud: I think we just have to agree to differ on that because we are very, very careful, in everything that we say, not to give that impression.
Q162 Chair: Is there a case, then, for publishing the statistics or the estimate for fraud separately from the estimates for error? Even have a different reporting time scale that might decouple the two.
Lord Freud: The basis of these figures is a concern driven not so much in historical terms from this Committee but from the PAC, which is worried about the correct expenditure of state funds. They have gathered that together, and we have got an elaborate measurement process of fraud and error that has been put together for these purposes in measurement terms. In this country, DWP—and I will let David talk for HMRC on this—have got one of the most elaborate measurement systems in this area of anywhere in the world. That is something that Parliament is extremely keen that we monitor and maintain. A pressure of the kind you are indicating would go completely counter to what we are responding to elsewhere.
Mr Gauke: To add to that, looking at tax credits, expenditure on fraud and error in total constitutes 7.3% of tax credit expenditure, and is higher than we would like it to be. It is lower than it has been in the past, but still higher than we would like it to be. In terms of the measures that we are taking, very often what applies to error also applies to fraud. In some cases, there is error that clearly occurs on the part of the claimant. HMRC is not always in a position to determine whether that error was reckless or had an intention to achieve a higher payment than they were entitled to, but it would be classified as error in those circumstances. Very often the issues and the problems that cause error also cause fraud, so therefore I think it makes sense to address to them together.
Q163 Chair: That is certainly true when we talk about what we might call the low‑level fraud, where people do not report a change of circumstances, or perhaps lie about the amount of money they have got in the bank or something along those lines. Do you have any sense as to what extent or what proportion of the fraud figures is that low-level fraud by individuals as opposed to the major criminality where you might actually have a fairly major fraud that is being carried out on a grander scale, organised by criminals but using individuals perhaps to put in false claims?
Lord Freud: We are reliant here on the survey approach, which gets our measures. Again, it is slightly different across the Departments, so I will talk for DWP. We are reliant on that way of measuring, which is a survey-based methodology where you go and talk to selected groups of people, check them out very thoroughly, and maybe put a fraud officer in to see it. There is therefore a breakdown of the types of fraud that we have, some of which may be highly organised and some may be individual. We have five main types: misrepresenting earnings; misrepresenting capital; living together; undeclared or under‑declared occupational pensions; and fraud from people abroad. Those are the five biggest categories of fraud that we have.[1]
Mr Gauke: In terms of tax credit, Nick might want to say something in terms of the breakdown, because as you say, there is a difference.
Chair: There is a difference between a fraudster running organised crime as opposed to the individual getting a few pounds extra.
Mr Gauke: And tax credit is subject to both of those.
Nick Lodge: It is. The figure that was quoted earlier of 7.3% in total for error and fraud is predominantly what you described as individual error and fraud. Organised criminality, of course, is something we do see in tax credits, but by and large it is not counted in those figures. We, like DWP, have a similar methodology for measuring error and fraud where we are taking up individual cases on a sample basis at the end of the year and seeing how much error and fraud there is in the system. But clearly we do suffer from organised attacks on the system from time to time.
Q164 Chair: Have you got confidence in your systems that, if there was that kind of organised attack, you would be able to pick it up fairly quickly?
Nick Lodge: We do pick it up. Only fairly recently we prosecuted a gang from the Czech Republic for exactly that kind of organised fraud. I am confident that we are able to spot it and address it, but no doubt there are always improvements that we can make.
Q165 Chair: Do DWP feel confident in the same way?
Mike Driver: Yes, I do. If you take Universal Credit, we are putting even more time and effort into dealing with cyber-fraud. If I could just build on the point that you make, though, because it is quite a difficult judgement between what is fraud and what is error in these circumstances. Take Universal Credit, which, as you know from previous discussions, is still ramping up: of 12 cases where individuals should have self-declared that they had an occupational pension at the point at which they made their claim, one did. So whether or not they had enough information to be able to do that is an issue, but I would suggest that what we are trying to do now is spend a lot more time trying to stop the fraud getting into the system itself at the gateway, rather than allowing it to perpetuate through the system.
Chair: We have got some questions on data sharing from the private sector coming up, so you will get a chance to say that then.
Kwasi Kwarteng: Can I just say one thing on that?
Lord Freud: I will just run.
Chair: Is it a question for Lord Freud or for the Treasury?
Q166 Kwasi Kwarteng: I just wanted to say, as a statement, clearly in the real world there is going to be a massive blurred line between error and fraud. To try to say that one is completely different from the other requires a window into people’s souls that we do not possess. You are quite right to say that it is very difficult to distinguish in many cases between error and fraud, and I completely appreciate that.
Q167 Chair: I have got some questions on funding the joint strategy, but I suspect people here can answer those questions even if Lord Freud is not here. The joint strategy indicated that £425 million of additional funding would be allocated to the range of measures over the course of the spending review period to April 2015. How much of the £425 million has been spent to date? Do you anticipate spending it all by April 2015?
Mr Gauke: That is more for DWP.
Mike Driver: I think it probably is. You are quite right. We originally estimated that the cost would be £424 million in the Spending Review 2010 period. Including new measures that we announced as part of the Autumn Statement, including more data matching with HMRC’s RTI database and some work that we are doing on abroad fraud, we now estimate that in that period our expenditure will be £286 million, so significantly less than was originally forecast. I guess the point behind the question is partly about value for money as well, so in terms of the relationship between our expenditure and the benefits we expect to accrue, our DEL spend will be £286 million, and we are currently forecasting Annually Managed Expenditure savings of £1.93 billion from that expenditure from the measures that are outlined in the strategy.
Q168 Chair: The joint strategy was also refreshed in a cross‑Government report published in February 2012. There was no mention in this update of some elements of this strategy, including the system of rewarding members of the public for information that leads to the recovery of significant public funds, and actions to educate and support claimants to get information relating to benefit claims right first time. Have these elements been dropped?
Mike Driver: No, they have not. If I could deal with them in the order that you said, as recently as last week our colleagues were meeting with Crimestoppers to understand what we could learn from what had happened in that situation. We are following up those discussions, because clearly if there are incentive arrangements that we can put in place that will help us to tackle fraud, in particular, we would be keen to pursue those. The work that we have done is slightly slower than we would have liked it to have been, but we are still pursuing that.
In terms of the whole issue of making people more aware of their rights, you may have seen that we have been running local campaigns at a number of locations, including Blackburn, for example, where we have been publicising the need for people to declare where they have seen fraud within their communities, and also reminding them of their obligations to tell us of changes of circumstances, the most common of which is changes in income.
Q169 Kwasi Kwarteng: I wanted to ask questions relating to the targets that you have set yourself. This is for Mr Gauke. HMRC missed its 2008 targets to reduce tax credit fraud and error to 5% of total finalised entitlements by 2011. You now have a target to reduce fraud and error to 5.5% by April 2015. Two years ago it was at 7.3%. How confident are you that you are actually going to meet your target for reducing the portion of fraud and error that you pay that is lost to the taxpayer?
Mr Gauke: There is a lot of work that is going on in this area, not least the announcement in last year’s Autumn Statement in terms of error and fraud additional capacity, which is much greater use of the private sector. Given the scale of the work that HMRC are undertaking here, we think that is an achievable target.
Kwasi Kwarteng: The other questions I have are to—
Chair: I think Mr Driver could answer it, if you want to ask.
Kwasi Kwarteng: From the point of view of the Department, what is your view?
Mike Driver: In a way, we have two targets. We have the business case that supports the fraud and error strategy itself. In terms of delivering the full benefits of that, it is fair to say that we are slightly behind schedule in terms of where we would like to be. We are continuing to look at opportunities to deliver more savings for the Exchequer from that strategy. In terms of reducing the monetary value of fraud and error to 1.7% by the end of 2014-15, it is fair to say that there has been a certain plateauing at 2.1% in terms of the monetary value of fraud and error over the last couple of years. Certainly it is lower than when the strategy was first developed. We are continuing to look at where we can do more in that space, and we are confident that we will reduce the monetary value of fraud and error further. A key point for us will be when the first set of data in relation to 2013-14 is published in June of this year. That will give us an indication of how successful some of our measures have been.
Q170 Glenda Jackson: When you detailed the five areas that you examine for the possibility of fraud, in every single incident, it seemed to me, setting aside the contribution from Mr Lodge of an external group of fraudsters, they would all be individuals rather more than any kind of organisation. In some instances, it might have been a genuine failure to realise that you had to report something. How are you splitting the money up between fraud and error? Are there separate organisations that are concentrating on those things, or is it all one?
Mike Driver: If you had asked me that question 12 months ago I would have been answering it on the basis that we were splitting that money up between fraud, on the one hand, and error reduction on the other. What we have created within the Department now is a single fraud and error service. We are looking at this far more holistically than had been the case in the past. We have also taken measures, as part of the fraud, error and debt strategy to introduce a Single Fraud Investigation Service, taking account of work that is already done within HMRC, and also local government in relation to housing benefit fraud, which is a significant issue that we need to do more on.
Glenda Jackson: We have got questions on that later.
Mike Driver: But, in terms of your question, we are trying to put our arms around all of these issues and fund it on a single basis.
Q171 Chair: In terms of how you compare with the private sector, which probably does not have anything quite as complex as DWP, do you compare favourably in terms of error and fraud in similar processes in the private sector? Do you do any benchmarking?
Mike Driver: It is quite difficult to benchmark against the private sector. If you take DWP as a financial services organisation, it is difficult to benchmark against what is happening within the banks, for example. What we have done, and I think the National Audit Office has said, is that if you take our organisation and compare that against standards that apply across the public service, certainly in other countries, we benchmark quite favourably.
Chair: You mentioned housing benefit and co-ordination with local authorities. Glenda, if you ask the questions.
Q172 Glenda Jackson: Thank you, Chair. Local authorities have complained to us that the automated transfers to local authority systems—ATLAS for short—gives them too much information. They have actually said that they have to wade through it to find what they actually need in relation to verifying housing benefit entitlement. What could be done, and are you doing it, to streamline the data received by local authorities via ATLAS?
Mike Driver: I must say, I have not had that direct feedback, so I will take that back to our team and make sure that we engage.
Chair: They were quite emphatic.
Glenda Jackson: And consistently emphatic.
Mike Driver: I will take that point away, provide that information to the team and make sure that we take that forward. One of the things that we have tried to do with ATLAS, because I accept there are a lot of data, is to create an environment or a system whereby, where there has been a change of circumstances, it is automated within the local authority system, so as far as possible it requires no human hands in order to take into account those changes. We have tried to automate as much as possible. If there is more to do, we will listen to what they have to say because it is in both the Department’s interest and the local authorities’ interest to do as much as possible to combat fraud and error in this space.
Q173 Glenda Jackson: Certainly what we heard in this Committee is that what is being lost, and what local authorities are finding very difficult, having to wade through all this information, is the local information.
Mike Driver: Yes. I will take that onboard and we will engage further with the local authorities to make sure we pick up their concerns and do something about it.
Q174 Glenda Jackson: And actually look at what they are saying; you could reduce the amount of automated information that they are having to receive.
Mike Driver: Of course, if that is the solution. I repeat the point that I made: it is in both our interest and the local authorities’ interest to get this data transfer and data‑matching operating as effectively as possible. Anything that we can do to improve what is still a relatively new approach, we will do.
Q175 Glenda Jackson: The estimated monetary value of housing benefit overpayments is more than twice that of any other social security benefit. Local authorities currently spend large sums of money uncovering incorrect claims including, to repeat, on the currently resource-intensive processing—that is another point—of ATLAS data. Yet the financial benefit of reducing housing benefit fraud and error goes entirely to the Treasury. Would it not be sensible to financially incentivise local authorities to reduce fraud and error while housing benefit administration remains their responsibility, and is going to remain their responsibility for an infinitely longer period than any of us originally thought?
Lord Freud: We do have an arrangement with local authorities where the subsidy that we pay on housing benefit is reduced to 40% when a fraud or claimant error is uncovered. The authority is able to pursue recovery of that fraudulently paid housing benefit.
Q176 Glenda Jackson: But that recovery of itself would be costly. It costs; it is not done for free.
Lord Freud: No, but that is the reason we have made those arrangements: to make it economic for—
Glenda Jackson: For whom?
Lord Freud: For local authorities to pursue fraud.
Mike Driver: Because local authorities, with their subsidy arrangements as they are at the moment, would receive a subsidy of 40%. If they identify the fraud and then recover it, they can receive subsidy, in effect, of 140%, so there are incentives within the system. Whether those incentives are sufficient is another matter.
Q177 Glenda Jackson: That could be linked to your examination of the ATLAS—them having to wade through that.
Mike Driver: We absolutely want to work closely with local authorities and local government in order to improve our performance in terms of fraud and error.
Lord Freud: Just on the ATLAS point, ATLAS notifications are designed to be machine-processed. Many local authorities have the ability to take the ATLAS feed on that automatic basis. Some of the concerns heard were from those local authorities that do not, and have to do it manually.
Chair: It is the very opposite. With the automatic feed they just get a huge amount of information that they have to wade through. They were very strong in their evidence to us about that.
Q178 Glenda Jackson: What has been lost, or could be lost, and which is the greatest benefit as far as local authorities are concerned, is that local knowledge and that working with groups of local authorities on these specific issues. I am still not clear, with all respect, why, if there is a reduction in housing benefit fraud, does it all go to the Treasury? Why would it not make more sense for some of it to go to local authorities?
Lord Freud: We have got this arrangement, which I and Mike have spelled out, that they do retain that 40% figure. That is designed to be the incentive to do it.
Q179 Chair: 40% of what? Is it 40% of what they recover that they retain? I do not know what else it would be.
Lord Freud: It equates to the money that they recover plus a 40% subsidy, so plus 40%.
Glenda Jackson: Of the money that they recover.
Lord Freud: Yes.
Q180 Glenda Jackson: We are seeing in London—I am sure you have seen it—increasingly adverts on the back of London buses, attempting to highlight the increasing use of sub-letting of social housing by the tenants. It is an illegal practice. Some of the evidence that we receive from local authorities is that pursuing those people through the courts is not only time-wasting but also expensive. So again I reiterate: would it not be better to give them more money than the 40%, if that is the benchmark?
Mike Driver: The point about the subsidy is that where a local authority finds that a fraud has been committed, the amount of money that we give them through the housing benefit subsidy arrangement is reduced from 100% to 40%.
Chair: I understand that.
Mike Driver: That is the way the funding works. What we then say to local authorities is that if they then pursue the recovery of that fraudulent claim, they are able to receive that money back into their bank account rather than it going to the Exchequer. So they have received 40% already; they know a fraud has been committed, so if they then recover more of that money they benefit financially, and that financial benefit could be higher than the 100% of subsidy that they could have received if it was a non‑fraudulent case. So there is an incentive within the system.
Chair: So the incentive is in the system, okay.
Q181 Glenda Jackson: Still on local authorities, they believe there is a strong risk that DWP will be unable to replicate the cross-checking of data currently undertaken within and between local authorities—this is the point that I have already made twice—when responsibility to support for housing costs transfers to DWP under Universal Credit. The Integrated Risk and Intelligence Service was intended to replicate these kinds of safeguards by linking UC and “a huge number of databases”. IRIS was not in place for the initial UC pathfinders, so has IRIS been redeveloped? To which databases does it link the UC system? Is it now in use in any of the UC pathfinder areas? Three for the price of one.
Lord Freud: That is a lot of questions. In the live service we have a relatively straightforward, data-matching process, which we call the counter-fraud capability. That is what works with the live service. We are then developing IRIS, effectively splitting it up and looking at it in four ways: information confidence, so it looks at every transaction and tests it; an analytical and intelligence hub, which was part of the original IRIS design, which will have analysts who will watch processes go through; a security decision service, the function of which is to take risk assessment from information confidence and reach a judgment about whether there should be an intervention; and then finally, just because we know who someone is because we have their ID, does not mean that their interface is not corrupted, so we need to maintain what we call session confidence, so that someone has not captured a particular computer in malware and made it do things we do not want it to. So that is the process by which we will be testing for fraud in the UC in the end state, and increasingly in the live.
Q182 Glenda Jackson: You have not got the data to work on at the moment, have you? So how is this going to come about?
Lord Freud: We will have 8 million different claims in the system covering 20 million people.
Q183 Glenda Jackson: Sorry to interrupt, but presently, how many UC systems is it linked to? How many of the pathfinders are actually employing this system?
Lord Freud: At the moment, we are using the first of these systems—the counter‑fraud capability—which is essentially a test against the other data systems that we have. Those are the DWP systems, and, on top of them, the other thing that we are getting an incredibly valuable feed from is the RTI feed—Real Time Information feed—out of HMRC, which gives us earnings information. That now covers the bulk of people who are employed under PAYE. As an example of that, we had one person declaring an occupational pension—self-declared; absolutely correct; that is the way you should do it—and we uncovered another 12 who had not declared that information in UC, where they had an occupational pension that meant that we should reduce their UC payment. That is the kind of flows of data that we simply did not have within the legacy systems.
Q184 Glenda Jackson: Correct me if I am wrong, but my understanding is that the only claims that are being processed in UC at the moment are Jobseeker’s Allowance. They are the simplest cases: they tend to be individuals; no family; no children; no complexity at all. It is highly unlikely that they would have an occupational pension because they seem to be, in the main, very young. So I am still asking the question: how is this system going to actually receive the information so that it can function when UC eventually rolls out?
Lord Freud: I have just given an example.
Chair: Can I clarify? The section of questions we are on is actually about housing benefit.
Glenda Jackson: And within UC.
Chair: The checks and balances that exist presently amongst local authorities—because they have the data not just of the people who are in receipt of housing benefit, but also the data around the properties. When that goes to UC, they will not have access to the data about the properties; they will just have the household claim about the individuals. In the new system, how will that data—the data that the local authorities use to do the cross-checking—work under UC, when they do not have access to the local authorities’ property register data in the same way as obviously local authorities do?
Lord Freud: To the extent that local authorities are using their own databases that effectively match a very similar process to what UC has got or will have—we will have people’s addresses and we will be able to match them across. One of the major issues that concerns councils is people who may have properties across different councils.
Q185 Chair: It is very difficult to put in a fraudulent claim where two people put in a claim for the same property, at present, if it is in the same local authority area. Under UC, that might be quite easy because you have two individuals who UC has not spotted are a household, so there only should be one claim for housing benefit, but they have both put in a claim for housing benefit on the same property. How will the systems under UC pick that up?
Lord Freud: To some extent, if it is doing the same thing as the IT of a local authority is doing, it will do it in the same way.
Chair: But it will not be. That is the local authorities’ point.
Glenda Jackson: That is the whole point.
Lord Freud: We will of course see when someone makes a claim in a property, which is the same as someone else making a claim in a property.
Q186 Chair: How?
Lord Freud: Because we have it in the Universal Credit claim. In that sense, there is no difference between a local authority system and the UC system.
Q187 Chair: Can I give you an example? In Scotland, we have got a lot of tenement blocks. Different local authorities use bizarre means of numbering them. I have a situation where every flat on the left‑hand side—there might be three or four flats— are all 26, and every flat on the right-hand side is 28. How will you pick out that these might be different flats or indeed they might be claims for the same flat? If it is different surnames, how will that data-matching happen? Local authorities have got local knowledge and they understand that if they have 26 Burns Road, there might be three different households or three different properties at 26 Burns Road.
Lord Freud: Let me try to have an uninterrupted go at this, so that I can spell it out. At one level you have the same kind of IT—it is actually enhanced IT in UC—as a local authority might have. At the next level you have the links between the UC organisation and our fraud investigators. We will have strong links with local authorities, and we do share a lot of data, so data-sharing will go on. We are clarifying the precise roles of local authorities and ourselves in housing, because local authorities maintain a very substantial position in the housing market, not least in the discretionary housing payments, where they have to make decisions in that area. As we build it out in detail, we will look to get the best of a systematised approach and some local knowledge.
Q188 Chair: Are you saying that under UC, DWP will have access to all of the property data that local authorities presently hold and use in order to cross-check housing benefit claims?
Lord Freud: We will look to get all the official—
Chair: “Look to get” is slightly different from “Will you have access?”
Lord Freud: I think I have to use the expression “look to get”. What the information confidence system will do is identify multiple claims from the same address across all the Department’s benefits. We will work with the local authority on the information that we can share for the fraud and error.
Q189 Chair: The local authorities were very concerned that a lot of the things that they have worked on over the years—such as that kind of data-sharing with them and their property—to cut down on fraud in housing benefit, which is one of the benefits that is most susceptible to fraud, was not going to happen. There is also the extra complication inasmuch as the rollout of UC has been slowed down. If the numbers of people who are on ESA do not go into UC initially at all, then there will still have to be housing benefit offices and officers in local authorities doing the work that they are doing now, because there will still be almost 1 million people who will not be in UC.
Lord Freud: There are that group and, of course, the current Pension Credit group. There will be quite a considerable housing infrastructure in local authorities for some time to come. We understand that, and will work with local authorities to make sure we optimise that position.
Chair: We are going to go on to some questions where there is another mismatch, and it is the Single Fraud Investigation Service.
Q190 Graham Evans: There is a lot of support for the Single Fraud Investigation Service. However, some local authorities have expressed concern over the potential of losing some experienced fraud investigators already located within the local authorities. There are opportunities for combining investigations, for example. Could you talk us through how you are not going to lose some very experienced members of the local authority investigation teams?
Lord Freud: The local authority concern is around some of what they call corporate fraud, i.e. the non-benefit fraud. They are keen not to lose that capability, and we are also keen that they do not lose it. What we are looking to do is consolidate the main anti‑fraud service, which covers the tax credits, UC and housing benefit, because that is all within the Universal Credit wrapper, or will be. We then want to make sure that the local authorities are not left without the cover that they would want for the corporate fraud. We are taking steps, first, to ensure there is funding for them to build that area up, and secondly, we have implemented a programme to help and support them build up that area so that they have greater capability in that area than they have ever had.
Mike Driver: Can I just add to that? The local authorities who have expressed the main concerns to us in this area have been some of the smaller district councils, where they might have one person who investigates a whole range of activities within corporate fraud, tenancy fraud, housing benefit fraud, and also potentially does some other things. They are very concerned that we might take that resource away from them so that they cannot do some of the functions that are left. There is still more discussion to take place, but what we have agreed with local government is that in those circumstances we will not try to take that person into the Single Fraud Investigation Service. We will potentially recruit separately, giving them the opportunity to retain their skills and expertise so that they can continue to deliver the business that they need to within their locality.
Q191 Graham Evans: The local authority union, UNISON, has commented that you have approached all of the councils individually—over 300 councils. They are asking why you did not look at a national framework. They feel that you are unnecessarily duplicating the approach.
Lord Freud: One of the issues there was very much that we would deal with each council’s needs on an individual basis. We are taking it pretty slow for that reason: to make sure we are able to deal with local authorities, because they all have different positions; some of them are in combination with other local authorities. We thought the right way to do it was to do it on an individual basis, which would get by far the best outcome for both sides.
Q192 Graham Evans: How many of the local authority fraud investigators will actually transfer into this new unit? Do you have any idea?
Lord Freud: We have got 380 local authorities, which is a little short of 800 full‑time equivalents. Our estimate is up to 790 are in scope for the transfer.
Q193 Graham Evans: Experience shows that when you make significant changes such as this, people naturally feel perhaps vulnerable or unsure of the transfer. What steps are you taking to make sure you do not lose the best talent from the local authorities?
Lord Freud: We have a local presence. It is not as if people are necessarily moving a large amount. We are keeping their terms and conditions so that, to that extent, there is not a massive change in their arrangements.
Mike Driver: What we found through the pilots that we have run—the pilot areas will be the first to join the Single Fraud Investigation Service—is that the reaction of staff is that working in this far more collaborative way across HMRC, DWP and previous local authority boundaries is a very positive step to take. I take the point of your question absolutely, because we need to have people who understand the business and have got the talent to drive down or to manage fraud in their locality. We need them in our organisation, so this is why we are engaging directly with those local authorities, and we will be with the people.
Picking up on your previous question, we have also been having discussions with UNISON more widely about the approach that we are taking. We do want this to be a success, because I think that the benefits associated with having this Single Fraud Investigation Service are significant.
Q194 Graham Evans: The concern is that you lose good people. It is okay Whitehall coming out with these fantastic policies, but at the end of the day it is the people, those middle leaders, that have to implement this to make it happen.
Mike Driver: I think it is worth saying that DWP—and HMRC, for that matter, as well—is not necessarily a traditional Whitehall Ministry of State function. It has that function, but a lot of what we do is in the locality through our Jobcentres and other provisions. We do understand, working closely with local authorities already on a number of issues, how we can manage this relationship effectively.
Q195 Graham Evans: In terms of the roll-out and the completion of the SFIS, it is 2016, I believe.
Mike Driver: We start the rollout in June at the pilot sites. We then take a break while we consider how that has worked. We will resume the rollout in October 2014, and we expect that to take about 18 months to complete.
Q196 Graham Evans: So by 2016, any glitches will be ironed out, and so on and so forth. Are you confident that it will be in place and working appropriately for the Universal Credit regime?
Lord Freud: It will be in place on this timetable, which we have slowed down for some of the issues that you have talked about. It will be in place and completed on that timetable by March 2016. That is the time at which we can start looking at very substantial volumes of Universal Credit movement. It should be working well across the country by the time we are looking at this new area. Do not forget, the investigation service picks up the leads, if you like, from where we think there is fraud. That is about the analytics hub in this example, as well as general concerns that we get alerted to, so there are obviously lags in that process.
Q197 Chair: It is largely in the process inasmuch as Universal Credit is going to be much slower, and so the SFIS, or whatever they call it—
Lord Freud: Single Fraud Investigation Service.
Chair: Whatever it is, they were meant to roll out together but, clearly, the fraud investigation is now ahead of the other. Because of the slowness of the rollout of UC, will that not leave the existing system more vulnerable to fraud and error simply because you have changed the investigation service?
Lord Freud: No, because in practice they are running out in fairly similar timescales.
Mike Driver: It is a service that has to operate in a Universal Credit world, but also a legacy world. We need to be careful how we implement, but we are cognisant of the risks associated with the point you make.
Q198 Chair: Graham also mentioned about the TUPE arrangement. My understanding is that they are not TUPE’d under the formal arrangements, but the principles of TUPE, according to DWP, should be followed as far as possible in accordance with business need. I had an email from UNISON this morning; they are very concerned that the draft SI with the TUPE regulations that have been laid actually has a number of TUPE regulations omitted, such as the effect of relevant transfer and collective agreements, insolvency, variation of contract and things like that. They are a bit concerned. While they understand that it is not the full TUPE that will be invoked, it would be as though TUPE did apply.
Lord Freud: That is what is called the COSOP principles. When you transfer within the public sector it is TUPE-like, although it is not formally TUPE.
Chair: They think it might be TUPE-too-light. That is their concern.
Lord Freud: “Like”.
Chair: Oh, “like”.
Lord Freud: Sorry; I shall enunciate with as great clarity as I can. That means that DWP steps into the shoes of the previous employer, and the contract of employment will be treated as having been made initially between DWP and the employee.
Chair: I think they accept that, Lord Freud. It is just the fact that the draft SI seems to have key elements missing that they would expect to see in a TUPE-like thing. Perhaps if you could maybe take that away and look at it?
Lord Freud: We are working through that right now, so we will have a look at it.
Q199 Sheila Gilmore: Having heard your response to Glenda’s questions in particular, at local authority level, in dealing with housing benefit over the last few years, there have been methods found for verifying things such as leases. People have to have that verified, and housing associations have sometimes been able to do this on behalf of councils. What is happening at the moment with Universal Credit claims that involve a housing element? How is that being handled? Who is verifying that?
Lord Freud: That is being done through Universal Credit. We do have a close relationship with the local authorities, so we are talking to local authorities as that process develops. We are looking at the detail of that right now.
Q200 Sheila Gilmore: So, basically, if you have got a Universal Credit claimant at the moment within the pathfinders or—
Lord Freud: The live service.
Sheila Gilmore: —is the housing element really dealt with as it is at the moment, by the local authority?
Lord Freud: No, it is dealt with under UC.
Q201 Sheila Gilmore: What does that mean?
Lord Freud: It means that the application and the information goes to Universal Credit and, depending on what the requirement is—to forward it or to show it when you come to your first interview—that information, where we think it is appropriate, we get, but we are able to check with the local authority. That is one of the things we are looking at very closely: what is the level of checking that we need to do? We are just looking at that area for the precise reasons that you are raising.
Q202 Sheila Gilmore: I recall you, in fact, addressing a briefing on this subject probably two or two and a half years ago, assuring everyone who was there that you thought it would be easier with Universal Credit. I am not really getting the impression it necessarily is, because the verification of things like documentation—at the time, for example, people were both signing up for tenancy and applying for housing benefit—was very useful and very important.
Lord Freud: As I say, we can require the same documentation. We have simplified housing benefit quite a lot within Universal Credit.
Q203 Sheila Gilmore: How?
Lord Freud: We have stripped down some of the areas to make it more simple, where we looked at some of the protections where the amounts offered are not very great in the context of how it works today, which is in the private sector, where a lot of the concern is, we have the LHA system. However, we have been left with a system that incorporated things that were historic protections that are no longer necessary, we think.
Q204 Sheila Gilmore: Housing associations, which are not the private rented sector, took on a lot of the verification, because it ensured there was a smooth transition in somebody coming into a new tenancy. Otherwise, they were signing people up for a tenancy and then it was taking quite a long time for the housing benefit to be sorted out. One of the elements within that was verifying things like leases. That was done at the housing association end, on behalf of the council, to speed the entire process up. Is that still going to be possible?
Lord Freud: Yes. We are going to have, and we do now have, strong data links between local authorities, ourselves and housing associations. Clearly, in Universal Credit one of the issues is how one appropriately shares data, which is an exercise that we are looking at very closely. However, our main concerns about error and fraud of the kind that you are talking about are not in the social rented sector. In fact, regarding the delays in the processing, at some stage, as you say, we ask: why are we going through these processes with social landlords that we trust?
Q205 Sheila Gilmore: That was precisely why they were authorised to do that.
Lord Freud: Exactly. We are working through this kind of detail now, but, when we look at our analytics centre about where the risks are here, chasing a reputable social landlord to check the detail of the leases is, I suspect, something that we will not spend a huge amount of time doing, whereas if you are process‑driven, you might have a process that insisted on it.
Q206 Sheila Gilmore: Identity assurance: I think you said there was a programme still to be developed for Universal Credit. How is that progressing?
Lord Freud: Identity assurance is something where, fundamentally, you want to be sure that the person who claims to be whoever is whoever, especially as you ramp up the IT relationship from the face‑to‑face and telephone relationship. What is happening is that there is now a Cabinet Office‑led programme to create an ID assurance process, which will be across all of Government. That is already underway. There are some beta sites; in fact, one of the beta sites is in HMRC, which David can take you through if you want to know the detail, and we are looking to use that system for Universal Credit.
Q207 Sheila Gilmore: We have had evidence from people suggesting you could use a biometric system, such as voice recognition. That is not speech recognition but voice recognition—the patterns of voices. Has that been looked at?
Lord Freud: We have done a pilot of that a few years ago. Interestingly, we found that for our purposes—and I am not going to say anything about whether a system might work for some—it did not work. We did not find it hugely helpful for us at this stage. The results of that pilot mean that we have not pursued that avenue at this stage.
Q208 Sheila Gilmore: What is the verification that you are developing?
Lord Freud: I went through those four stages and those four measures that we are building, which are about data, risk analysis, session confidence, and identity assurance. All those processes need to be put in place. What we are recreating in practice is something that is traditional in the security services more generally, where you have a group whose job is to monitor the information as it comes in, analyse its coherence, look at the threats and act on them. That is a 24/7 operation that we will be adopting and adapting in Universal Credit.
Q209 Sheila Gilmore: From the applicants’ point of view, what do they have to do? What are they being asked to supply if they are applying online, for example?
Lord Freud: What will happen is that people who go through on a green basis—i.e. it seems absolutely straightforward, no problems, information checks off against other information—will just go through. Where the red flashes, we need to look at the position more closely and we may look at asking, just as the bank would, the person to come in for a face‑to‑face to check that they are who they say they are and they are doing something they should be doing. That is something that the banking system will do if they are concerned about someone trying to take over your account.
Q210 Sheila Gilmore: If there was a failure to match address, for example, as the Chair was talking about, which we do find quite common, that might be a red flag occasion.
Lord Freud: If we suddenly see lots of families at one address, you see that almost instantly as a suspicious circumstance we need to sort out.
Q211 Sheila Gilmore: Is that automated?
Lord Freud: It will be automated. It is an automated system that kind of analytics and intelligence hub, yes.
Q212 Chair: What proportion of reds are you expecting? What proportion of the things, when they go through, will flag as red?
Lord Freud: You dial these things up and down. You can categorise, so you can move it up and down reflecting on what you think is the level of risk.
Q213 Chair: The reason I ask the question is that my own electoral registration office, like many, has just done a data match with the electoral register. In some of the poorer areas in my constituency, and I do not have that many of them, up to 40% of the people will drop off the register because there was not a match. That is the same system as you were just talking about there. How does somebody get into the system if they keep coming up as a red or the data around them is not there?
Mike Driver: Probably face-to-face.
Q214 Chair: That is why I am asking what proportion you are talking about. The people in the poorer communities will be the very client group that will be applying for Universal Credit.
Lord Freud: We have the capability of seeing, and we do see, on a regular basis, a large number of people anyway. Particularly the people who are in the conditionality groups, we do see them. We can see them fortnightly; we can see them more. So we have a process of seeing a large group anyway. What we are talking about is expanding that group, potentially, and clearly that is one of the immediate resorts. We have somewhere for people to go when we have a problem of reconciliation.
Q215 Chair: Without biometrics, though, how do you know they are who they say they are? It does not matter how often they come into your office if they have said they are Joe Bloggs when they are really someone else.
Lord Freud: That does come to training. The number of people who can carry off a face‑to‑face impersonation is a lot fewer than those who think they can make a claim for something to which they are not entitled. We are talking about a very small group and clearly we will need systems for that small group.
Mike Driver: You also talked about organised fraud earlier, and this is where the system that Lord Freud is talking about comes into its own as well, because an organised gang would, potentially, make 5,000 simultaneous claims for benefit through the system. What the resource that we are putting in place does already is look for those trends where you get multiple claims made at the same moment on the IT, so that we can investigate those more effectively and stop those claims entering the system even.
Lord Freud: One of the most attractive things in this area is what happens when you have fraud and the kind of semi‑criminal fraud is that somebody invents it somewhere, gets away with it and then it spreads across the country. By the time that we catch it under traditional methods, it has spread a long way and it can be very expensive to track down each of the groups that are doing it. When you have the analytics and intelligence hub, you can start seeing these patterns very early, and I am very keen that we spend effort in the SFIS environment in stopping new fraud patterns developing, so that they do not get replicated across the country and we stop the problem before it becomes a major one. We see examples of things being copied across the country today, which, under UC, we will be able to stop much earlier.
Q216 Sheila Gilmore: RTI is the other area where you expect to have savings. In 2010, you suggested savings of as much as £400 million a year across benefits and tax credits. Is that still the expectation and how far towards that are you at the moment?
Lord Freud: We have substantial savings in UC. We have a forecast here.
Mike Driver: The savings that you talked about we are still expecting to deliver.
Q217 Sheila Gilmore: That will presumably be once you have full rollout.
Mike Driver: Yes, and this links back to some issues that we were talking about slightly earlier in the discussion. What was announced in the Autumn Statement was a pilot to use the RTI data that we use in Universal Credit as a proof of concept to see whether or not we could apply the principles of Universal Credit into JSA and other benefits. We have also agreed with HMRC colleagues that we are going to use RTI data to compare the information that we already have in place for JSA, ESA, Income Support and Pension C redit, to look to see whether or not there are earnings that should have been declared that are being shown up by Universal Credit and the RTI system, and also changes in the amounts that are being declared in those benefits. In some ways, although we have not applied the RTI system quite at the pace that we would originally have expected in Universal Credit, we are looking to see whether or not we can use the approach in the legacy systems so that we can prepare the way for that. There is a question mark about exactly where we score the savings associated with this, but we still believe that RTI will offer us very significant savings in terms of both fraud and error.
Q218 Sheila Gilmore: Is that quite a manual operation or is there a method of linking those two things?
Lord Freud: It will be an automated system.
Sheila Gilmore: Into the legacy systems.
Lord Freud: Yes.
Mike Driver: It is a data‑matching between the Departments, so, effectively, a duplicate version of RTI exists that we can match against the live loads of the benefits that I outlined. We will then pass those cases, where we find the matches, to one of our Benefit Integrity Centres to follow up the detail and correct the case, and that could be error or it could be fraud.
Lord Freud: It might be worth talking about the RTI system. I will ask David to talk about it, because one of the things is that it has been enormously successful and smoothly put through and we are now getting these feeds. David, do you want to just take us through how all that works?
Mr Gauke: I was going to deal specifically with your point in terms of where RTI can provide savings in respect of fraud and error, because essentially, just looking at tax credit renewals from 2014‑15 to 2016‑17, we estimate that there are £820 million worth of benefits from RTI as a consequence of the information that HMRC will have. That is a significant benefit that we are seeing very early on with RTI. As Lord Freud has said, the introduction of RTI has gone very well. It is the biggest change to the PAYE system since 1944 and there will always be issues and consequences resulting from such a change, but that has gone well. In terms of being a very quick return on fraud and error in the context of tax credit renewals, we are going to see that very significantly.
Q219 Sheila Gilmore: Will that allow you to do what Universal Credit was going to do, which would be to update the difference, as I understand it, between Universal Credit and, say, tax credits, where often it is several months before there is a reconciliation between projected earnings and actual earnings. Would you be able to start using a system that did that more frequently?
Mr Gauke: That puts it well. The advantage of Real Time Information is that the information is real time and you can see it straightaway and that enables HMRC to move more quickly in terms of identifying a discrepancy and doing something about it.
Q220 Sheila Gilmore: That could be helpful, presumably, to applicants generally, because a lot of the problems with tax credit when people come to see us are often about the time lag. There is not necessarily fraud involved, but maybe the communication has been slower than it should have been, and things build up.
Mr Gauke: It is a good point. There is a customer-service element to this as well and so we are immediately seeing benefits from RTI.
Nick Lodge: In particular, tax credits, as you well know, is an annual system based on annual income, and at about this time of year, for the next few months, people will be renewing their claims. The information from RTI enables us to set their award up for the coming year with accurate income information from the start, and that is where many of the savings come from. It will also enable us to have a look in‑year to see whether people’s income is moving in ways that we would expect in line with the annual income information that they have given us. I absolutely agree with the point you are making about it being helpful to claimants as well, to keep their awards correct from the start.
Q221 Sheila Gilmore: What about people who are not paid on monthly salaries? Is that seen to be problematic?
Mr Gauke: RTI applies if it is a weekly salary or a regular payment. RTI is not based purely on a monthly payment basis.
Lord Freud: If you are paid weekly, there is quite a lot of technical detail about exactly how we use it, but essentially we capture the information as it is paid and we can work out what the payment over the month is, whether it is one employer or, indeed, several employers. That is not a problem to the system, and we can adjust for that. The bulk of people who are employed—I think the figure is about 94% or 95%—are on the PAYE system and will become, therefore, on the Real Time Information system. You have a small group who are employed who are not and who will need to self‑report. Likewise, we need a system for the people who are self‑employed, who will need to self‑report every month if they want to get their Universal Credit.
Q222 Sheila Gilmore: Have you made much progress on the self‑employed side?
Lord Freud: Yes. We are building a system so that they can report monthly what they have earned.
Q223 Nigel Mills: Just going back to the RTI system, how automated is that cross‑reference between the tax credits and the RTI feed? Is that a manual check where someone has to physically go and do it, or is it a system looking for things that are not consistent?
Mr Gauke: Essentially, it is a system point.
Nick Lodge: For the process that I described, at around this time of year, as people are renewing their tax credit claim, that is automated. We will pre‑populate with income information the renewals packs, which are going out to people from a few weeks’ time, and we will automatically set up people’s new awards using the income information that we have. We will also use it to verify income information that people gives us as they go through that renewal process.
Looking at people’s awards in‑year is different, because tax credit is an annual system. Therefore, the information that we get from employers three, four, five, six months into the year does not tell us what their annual income is going to be for the entirety of the forthcoming year, but it gives us an idea of whether their annual income looks likely to be broadly in line with what claimants have told us. That is not such an automated system. Although the information will come out to our staff in an automated way, we will have to then pick up with claimants to check whether the information that we are getting for part of a year matches what they think is going to happen for the entirety of the year, but the annual cycle is automated.
Q224 Nigel Mills: Does the system in‑year create some flags that say, “This person should have earned 10 grand by now; they seem to have earned 15. Someone should go and ask,” or do I physically have to go and look at the system and check that myself to spot something?
Nick Lodge: We are testing how best to use the system and the new information that it gives us in exactly the ways that you have just described. We are looking at whether we can do bulk data‑matching to throw out lists of people whose income looks to be falling out of line with the basis that we have used to calculate their award. Then we will be able to go to claimants to check and verify the information. That is exactly what we are in the process of testing now that we are getting the information from RTI.
Q225 Nigel Mills: Is that how you get to the £800 million saving over the three years?
Nick Lodge: No, the savings are almost all as a result of that annual process, setting people’s current‑year awards on the right basis rather than making provisional payments to them on the wrong basis, as can sometimes happen or has sometimes happened in previous years. We set somebody’s award up from the start on the basis of the correct income information and that is where many of the savings come from.
Q226 Nigel Mills: That is just because you have the RTI information a lot earlier than you have the year‑end data, so you do not have to wait for the P60s to start turning up.
Nick Lodge: Exactly.
Q227 Chair: What is the length of time between someone receiving their wages from their employer and DWP getting their RTI feed, so DWP knows how much has been paid? What is that timescale?
Lord Freud: It is very rapid.
Q228 Chair: Do employers now have the software that feeds that information automatically as they are doing the payroll?
Nick Lodge: Most payroll software is built to accommodate RTI and the information flows straight from that into HMRC and then virtually straight to DWP. There will, no doubt, be some small time lag, but it is quick.
Q229 Chair: Not the three days that the bank takes to clear an automated transfer.
Lord Freud: I think we are talking minutes, to be honest, are we not?
Nick Lodge: I think so. I do not have a precise timeline, but it is quick.
Q230 Chair: I still do not know why the banks take three days to clear a payment when you have done it automatically.
Lord Freud: It is one institution, effectively; that is the reason.
Q231 Nigel Mills: If we can get tax credits working so there is an automated flag-up if someone goes materially out of line, will UC work the same way? Is that a system feature you want to inherit?
Lord Freud: We have to integrate what is traditionally an annual system, as Nick said, into what is a monthly payment system. Clearly, we need to look at whether we just pay monthly or whether we have, which we are exploring in great detail currently, some carry‑forward systems. That is the issue. It is particularly in the area of the self‑employed, where you could have very volatile earnings month by month, so, yes, that is what we are looking at.
Q232 Nigel Mills: The holy grail is to try to spot a fraud or an error quickly, so the arrears have not built up over six months and it is a much easier amount to correct.
Lord Freud: This is less a fraud and error issue than a customer-service issue, to be absolutely honest—that we have a smooth payment system where we do not have what we have had in the past, which is these very large repayment obligations going back. That is one thing that we are determined not to have in Universal Credit. That is the other area that we are looking at: fraud, error and debt. We do not want to see large amounts of debt built up for that reason.
Mike Driver: To clarify your question, the point is that in Universal Credit, where a payment is being made and the person has earnings, their payments will adjust in the live service depending on what their income is. If they earn £20 this week and £30 next, for example, Universal Credit already takes that into account in the way in which the payment is made. In the same way that Nick was describing tax credits, that is happening in real time.
Q233 Nigel Mills: Are you looking at any other ways of accessing data you can match with what you have in your systems to check that people are, in error or fraudulently, getting more than they are supposed to be getting? Do you have any other ideas for how you could do that?
Lord Freud: Yes. As you roll out the system and develop it, you can clearly test against large numbers of data sources: whether it is a Government source like the DVLA or the Passport Office; we already check with the NINOs. You can see a large number of data sources, both publicly held and privately held and, indeed, we are looking at that as well. You can check for fraudulent behaviour against all these sources, and it might be quite easy today to make a claim to be someone you are not quite. Let us take living-together fraud. It is quite hard to maintain that position across a large number of databases and, where you have anomalies, you can pick those out and it raises a query.
Q234 Nigel Mills: Which privately held data are you thinking might be most useful?
Lord Freud: Both Departments have had an exercise with the credit reference agencies, checking our recipients with what is happening on the credit reference side to see if the expenditure pattern looks a little bit odd, given what their income should be. We found some extra fraud in DWP, but our data‑matching has been always pretty elaborate, so it was extra, but not a huge amount. What was very interesting was HMRC’s experience. David, would you like to take us through how you use that credit reference information?
Mr Gauke: This is something that HMRC has been exploring for some time, in terms of use of the data held by credit reference agencies. So far, something like £549 million has been generated as a consequence of the partnership with credit reference agencies, so it is quite a significant amount of work that has gone on here. We have had a somewhat different experience from DWP, and that is partly the nature of what we do, but in terms of identifying cases and so on, it is quite significant. Nick, is there anything you want to add to that?
Nick Lodge: We use it in two ways at the moment. First of all, we use it as part of our initial screening of every single new claimant we have, so we have credit reference agency data included upfront. We also use it to screen existing claims, predominantly to see whether there is a connection between people that might not be apparent from their claim—for example, somebody claiming as a single claimant when perhaps they should be claiming as a couple. That very often enables us to spot a financial link between people that we would not otherwise see, and that really has formed the basis of the success for us of using credit reference agency data.
Q235 Sheila Gilmore: I have had constituents who have had their tax credits stopped without any prior notice—just the money stops going into the bank account—on the back of, apparently, credit reference information that turned out to be wrong. In one case, it was information relating to a previous tenant at that address, and there had been some assumption that that was somebody still resident there, presumably a partner on another income. Is it normal to stop the benefit and then inquire rather than red flag it, as I think Lord Freud said?
Nick Lodge: No, it is not. Our process is that we always write to the claimant. We give them generally a good period of time to respond to us, because just establishing that there is or there could be a financial connection between people is not the same as saying they are living together as a couple. Our general approach is that we use the credit reference agency data as part of the information we have, and we write to claimants to explain that we have information that may show that they should have claimed as a couple rather than as a single person. People either respond to that with information to show, as in the case that you were mentioning there, that they are right to claim as a single person, or, if they do not respond, under those circumstances we may stop or reduce their payments. However, we write to claimants. We do not generally simply stop payments, so if you have an example where you think we have stopped payment without writing to the claimant first, which, as I say, is not the process that we follow, then I would be interested to see it.
Mike Driver: That is same process that we follow in DWP. If we are doing, for example, benefit integrity checks or we have information that would alert us to the fact that there is as problem, we will try first of all to contact that person on the telephone. If we cannot get hold of them on the telephone, we will write to them. We will give them 14 days to respond to us with the information we have requested before we suspend their payment.
Q236 Nigel Mills: It has been suggested to us that perhaps use could be made of bank data or payment sector data to try to match up accounts and check the behaviours are what you would expect. Is that something you have had any discussions about?
Lord Freud: We are talking to the Payments Council on quite a range of areas. One of the areas is to make sure that where we have known fraudsters in the system, whether they are in the insurance industry, the banking industry or tax credits, we can identify them and wherever they have come in on the system we can have an alert. The Payments Council is one of the groups with whom we are having those discussions to look at the multiple, organised frauds, where, if somebody tries to defraud one area, they are likely to have a go at another area.
Mike Driver: The point you make, though, is an important one, because if you take DWP and HMRC together, we now account for 49% of all volume on the BACS system. Using that approach, with that weight of business transaction, I think that there is more that we can do with banks in order to deal with fraud within the system. As Lord Freud has said, we are in discussion with the Payments Council, with VocaLink and with the banks themselves to look at what the opportunities are.
Mr Gauke: Just to add to that, HMRC, again, is also working with various organisations in this area. There was a pilot in 2011, for example, which looked at cases where tax credits or child benefit was paid into an account with no other UK receipts and where ATM withdrawals were, over a 12‑week period, entirely in one other foreign country. In that case, there was an 88% success rate. It was a pilot scheme, so one should not necessarily judge it by the monies, but the average savings were around £6,300. As a consequence of that, there is continued work with that particular bank to see what more work could be done in this area. That is an example of some of the work that is happening.
Lord Freud: The formal position is that the Cabinet Office is leading on a programme called the counter-fraud checking service, and that is what is corralling these different groups to work together for this initiative, with a series of pilots at different stages going on to make sure that we get on top of these particular problems.
Q237 Nigel Mills: It sounds like you have found a way of getting happy with the data protection implications of all this; is that right?
Lord Freud: Yes, we are very, very conscious of data protection. That is why we are going carefully and piloting as we go, to make sure that we do things in the appropriate way.
Mr Gauke: There is a particular issue as well with HMRC, because in addition to all the data protection issues, HMRC is also constrained by the Commissioners for Revenue and Customs Act, which means that it is very difficult for HMRC to share its data. Very often, organisations, to the extent that they are prepared to share their data, want something in return, and HMRC are constrained in what they can do in terms of data‑sharing themselves.
Q238 Nigel Mills: Do you think there is scope here to be a little more aggressive to push this forward quicker? You are not being too cautious.
Lord Freud: It is a real debate. It is clearly something where the data protection issues are very important, so we work with them. At this stage, what we are doing is making sure that we understand exactly what the most appropriate data is and then look at that in the context of what the data protection issues are.
Q239 Nigel Mills: Using more acronyms, I am afraid, CIFAS gave us some evidence where they suggested that they hold confirmed fraud data and share it around the private sector. They had offered that to the Department on a pro bono basis, but were turned down. Is that something you are familiar with? Presumably, confirmed fraud data would be something quite useful.
Lord Freud: Yes, it is, and clearly we have talked to CIFAS. Our approach now is to go with this counter-fraud checking service and to find the fraud in that way. We are talking to CIFAS; in fact, I think we are meeting them in a week or so to go on having that discussion about the appropriateness of sharing data.
Q240 Nigel Mills: My final question on this is: how critical is better use of data in trying to spot problems in the strategy of reducing fraud and error? Is this the only way of getting a step change or can you just get better on getting claims right in the first place and other, perhaps more traditional, fraud inquiries, or is this perhaps where it all needs to be?
Lord Freud: Using data smartly is really valuable to stop fraud getting in in the first place and to stop changes of circumstances drifting into fraud, which is how a lot of fraud develops, so it is really important. Clear and regular communication, being quick on data, responsive, talking to people, saying, “Look, something seems to have changed here. Do you want to tell us something different?” or whatever—getting all of that right is absolutely vital, because we have a large fraud investigation service: 3,000 people plus the people who are coming on from the local authorities now. It is very expensive to track down every individual case of fraud and prosecute. Prevention is infinitely cheaper as a route to go. Deterrence and prevention are absolutely critical in a counter‑fraud strategy.
Mr Gauke: You are absolutely right to highlight the issue of data, whether it is in prevention in the first place or subsequent detection and whether it be within benefits and credits, or whether it be in tax fraud as well. Certainly one of the things that struck me in my time in the role that I perform in looking at where HMRC has been very successful in dealing with, say, tax evasion is the use of the data. Data analytics are absolutely key in targeting effort in the right places. It is one of the reasons why HMRC has been able to make as much progress as it has in dealing with tax evasion, for example, and the same point applies with welfare issues as well.
Q241 Mike Thornton: Back in 2010, the 2010 joint strategy included a pledge to sanction a further 10,000 fraudsters a year. Do you know what the baseline for this pledge was and are you getting close to meeting that target?
Lord Freud: I do not have absolutely to hand the 2008-09 figure it is probably based on. Prosecutions were running somewhere below the 9,000 or 10,000 figure then. They have not changed very substantially to date. Our latest year, 2013-14, which is a part year—nine months—it is over 10,000, so we are running at a higher rate proportionately than we have been traditionally, but we are pushing the value of what we are chasing. We are consciously going for the higher-value prosecutions, if you like, and the higher‑value frauds within that. There is quite a substantial increase in the volumes involved when we do chase fraud, in the latest year or two.
Q242 Mike Thornton: Just on that point, you were saying earlier that it is much cheaper to prevent than to chase. I do not know how much you have the broken‑window syndrome on this, but surely if someone gets away with the small fraud, they might start increasing it, so if you ignore the small fraud you are going to end up with big fraud. If you catch the small fraud, they will not do the big fraud and you will save more money. Do you not think that could be relevant here, perhaps?
Lord Freud: There is relevance in the build‑up and how we handle people who start with a small fraud, particularly the drifting into fraud. We are talking about the individuals now, not the criminals or the criminal gangs, if we can make the distinction that Anne Begg made at the beginning, which I think was quite useful. There is a very substantial amount of drift in this fraud. It is about how one handles it through deterrence, a mixture of maybe civil penalties, which we have ramped up dramatically—they are now running at 50,000 a year—exactly to make the point that, “You cannot negligently provide us information”. There is a very swift penalty to stop it. That is the thinking behind that, but you are right in that it is not simply a question of ignoring small‑scale fraud, because sometimes that can grow. One of the things we will find out as we have the analytic hub is where you see patterns of small fraud develop, get on those very fast indeed, because they have the potential to replicate across the country as word of mouth gets out that this is an easy thing to do.
Q243 Mike Thornton: You were looking for a 10,000 increase. If you said, “We are not going to go over numbers, so we abandon the 10,000 extra target, because what we are looking for is amounts of cash rather than numbers of people”, is that something that you have probably switched over to?
Lord Freud: Yes, I think we have switched over. We are looking very closely at what is the right balance between administrative penalties, civil penalties, prosecutions, and so on. We have dropped cautions as part of our suite now, switching again to the civil penalties. We are looking very closely at the change of pattern that we have had in the last year or two and assessing whether we have the right balance by looking very closely at outcomes.
Q244 Mike Thornton: Is the switch to civil penalties partly to take away the risk of a criminal record for someone who has really drifted into it almost accidentally?
Lord Freud: Yes. The idea of the civil penalty is to go very fast and early and give someone a salutary warning that, “Look, you must put in correct information and casual information is not appropriate”. It is particularly designed for UC, but it also applies for the legacy benefits.
Q245 Chair: We started the session by saying that fraud and error get lumped together and that fraud only represents a third of the total, and of course the whole session has been dominated by fraud and we have hardly mentioned error at all. As a final question, I will ask what you might be doing to reduce both claimant and official error. It appears to us, across a number of inquiries that we have undertaken as a Select Committee, that communications is one of the areas that perhaps could be improved and might help to cut down on both claimant and official error. I wondered whether you have used any of the behavioural science findings about how you might address claimants and how you might communicate with them. Is the Cabinet Office nudge unit doing anything on this to help make sure that these errors do not creep into the system?
Lord Freud: I touched on some of this. I made a slightly large response at the beginning to your point, for the reason I was a bit concerned we might not discuss it anymore. It is something that is about communication, and we recognise at DWP that too much of our communication is too complex; there is too much jargon.
Chair: Nonsensical, actually.
Lord Freud: That is a hard word. I was going to say “lacking in clarity sometimes”. We are committed to really raising the quality in this area, and I think that will make a big step. I would make the point—and I can put that, if you like, standing outside the Department, and I do not think it is said enough—that DWP is managing a suite of horrifically complex benefits, which have grown up over the years. I personally find it remarkable that they run that suite of somewhat chaotic benefits with as little error as they do, because these individual benefits are really complicated with incredible rules. The level of error is something that is remarkably small in that context, but I do hope that once UC is in and we have a clean, straightforward system, which both the claimant and the operator can understand, and all of the rules are consistent across the elements, we will drive down error on both sides as a direct result.
Q246 Sheila Gilmore: Does that explain letters that appear to be disjointed and disconnected and include information that is not relevant? The answer that has been given before is, “It will all be fine when we have UC, so we do not need to look at the content of letters now”. However, if that is, for some people, going to be two, three or four years ahead, at least, then is it not possible to revise? Many of these letters are standard letters, but they do not make a lot of sense.
Lord Freud: As I say, we are now going through an exercise to raise the quality of our communications, recognising some of the criticisms that have been made and have been made again today in somewhat unkind language.
Q247 Chair: I am laughing because I have been on this Committee since 2001 and every single DWP Minister who has sat before this Committee in that time has said, “Oh yes, we are going to look at and improve the communications”, and, yes, they would look at the nonsensical letters and the computer‑generated letters.
Lord Freud: It is very, very difficult for an organisation, which is organised on the traditional lines of expertise in particular benefits. When claimants receive a multiplicity of those benefits, it is extraordinarily difficult to get coherent language and I think it is not an accident that you have been hearing this for 15 years or whatever it is.
Chair: And I suspect the 15 years before that as well.
Lord Freud: I am sure, and frankly that is one of the core motivations behind getting a clean, new system, which is consistent across the points. It is very hard to explain to someone, “On that you have to wait three weeks, but on this you have to wait seven weeks and on that the taper is so-and-so”. It is almost impossible to explain, so I do not think it is an accident and I hope that we will steadily, with the new system, get clarity to claimants, drive down their errors, and drive down the Department’s errors as a direct result.
Q248 Sheila Gilmore: One of the things that people find difficult is they are not necessarily claiming a whole lot of benefits; it not necessarily about those sort of comparators. When they get a letter that says, “Because of a change in your circumstances we are doing such-and-such”, and they are well aware there has not been a change in their circumstances, often what is being referred to is, perhaps, an annual up‑rating. Why not just describe it as it is, because that gets people thinking, “I have not had a change in my circumstances”.
Lord Freud: We are spending an enormous amount of time now in Universal Credit on how to communicate clearly. Each of these screens is tested. We are getting advice from the nudge unit. We have got our own teams working on it. With these communications, it is word‑by‑word clarity and it is an enormous amount of effort, even with a relatively simplified system. I am not surprised that there are odd complaints. I am not surprised that there are odd letters. There is an enormous amount of information to go out to people. It is a very, very complicated system, as we have seen. I do not think anyone—maybe Anne Begg excepted—understands the whole system. I do not think I have met anyone who does. It is too complicated for one person or even one organisation to understand, and we need a new system.
Q249 Sheila Gilmore: One individual who is on one benefit does not need to know all of that. They just want to get something that explains, for example, why they are getting a letter at all and why they sometimes get two or three letters over a short period of time that seem to be saying much the same thing but a little bit different.
Lord Freud: Yes, that is the issue. The issue is when you are organised on a series of service lines, which is what the Department is, it does not appear to the recipient in the holistic way that they think of themselves, i.e. “I am a person with these requirements”. That is the difficulty. That is what we did. That is why for 15 years, and probably 30 years, there has been lack of clarity.
Q250 Chair: It is still happening even with new benefits. The letters that go out to someone who has just been either awarded or not awarded ESA, whether they are in the WRAG or the support group, are quite confusing. I cannot remember the exact phrase, but one says that they are fit for work‑related activity and the other one uses the same phrase as “work‑related activity”. I have forgotten what the phrases are, but it is very confusing. The ESA is going to go on in payment beyond the rollout of Universal Credit and what Sheila is saying is that there are still a number of years that these poor communications will be in the system before—
Lord Freud: The ESA will become absorbed as two of the elements within Universal Credit, so—
Chair: Yes, but not until well after the roll-out of everything else that has been delayed.
Lord Freud: As I say, we have a group of ESA recipients, which we are going to be very careful with as we transfer, for obvious reasons, and I am sure you would welcome that, but we will be taking people in who are on ESA much earlier than that. In fact, because of the way we are rolling out, we have people today on those elements. I cannot pull them out the system, but I would have thought we already have people, in practice, on those elements. We will increasingly have those numbers and we will be finding out how best to communicate with them. I cannot promise you that every piece of legacy communication will be dramatically transformed in two years or even in the next 14 years, but I can tell you that we will roll out a system where it is possible that, because the system is one system, it is more likely that we are able to communicate direct to the person affected, because we are talking about all of their benefit in one.
Mr Gauke: Coming back to your question that we have talked mainly about fraud, the first point to make is, as both David and I made earlier, that there is not a clear dichotomy between fraud and error. Most of the measures we have been talking about today will apply to error very strongly, so the RTI introduction is very helpful in terms of sorting out error.
In terms of communications, you are absolutely right to raise that issue. There is a lot of work the HMRC is doing with the Benefits and Credits Consultation Group to try to improve communications, because we accept that tax credits is an inherently complex arrangement, if you like.
In terms of official error, of course we must always be trying to address that, but it is worth pointing out that that is a very small part of the overall amount of error within the tax credit system. It is very small in terms of total cost to the Exchequer, more in the region of £30 million in an overall fraud and error number of over £2 billion. That does not mean we should be complacent, and of course we must always be looking to improve customer service and ensure that there is greater accuracy.
Q251 Chair: Prompted by the Clerk here, I have just remembered what the phrase is in the ESA letter. If someone is getting ESA, one letter will tell them that they have “limited capability for work‑related activity” and the other one tells them they have “limited capability for work”. You would think that the one mentioning “work‑related activity” is the WRAG group, but it is not; that is the support group. The one that says “limited capability for work” is the WRAG group. That is incredibly confusing. Why on earth do those phrases have to be used? Why can they not say, “You are in the support group for ESA” and “You are in the work‑related activity group for ESA”?
Lord Freud: I must say we rather inherited those phrases. The trouble is once they are established it is quite hard to change them. I am putting my thoughts into what the last Government thought they were doing with the two phrases.
Q252 Chair: I can understand why the legislation might say that, but not the letter that goes out to claimants. That is the problem.
Lord Freud: I am not sure what the UC phraseology is. I am not sure we have even locked it down, but I will take your point very much to heart that something more straightforward would be appreciated.
Chair: On that note, can I thank you very much for coming along this afternoon? We really appreciate the time you have given. This is, as I said at the beginning, the last oral evidence session on this inquiry, so we will be going off to write our report, which will be with you in due course. Thank you very much.
Oral evidence: Fraud and error in the benefits system, HC 1082 30
[1] Note from Witness: In 2013/14: A total of 10,530 cases were referred to prosecution of which 863 were organised fraud (8.2%). There was a total of £159.1m prosecution overpayments, of which, £7.6m were organised fraud (4.8%).