Work and Pensions Committee
Oral evidence: Spending Review, HC 2169
Wednesday 10 July 2019
Ordered by the House of Commons to be published on 10 July 2019.
Members present: Frank Field (Chair); Neil Coyle; Ruth George; Steve McCabe; Chris Stephens; Derek Thomas.
Questions 1 - 57
Witnesses
I: Paul Johnson, Director, Institute for Fiscal Studies
II: Robert Chote, Chairman, Office for Budget Responsibility, and Andy King, Member of the Budget Responsibility Committee, Office for Budget Responsibility
Witness: Paul Johnson
Q1 Chair: Welcome. Might you just identify yourself for the sake of the record and then we will begin with Chris?
Paul Johnson: I am Paul Johnson, I am Director of the Institute for Fiscal Studies.
Q2 Chris Stephens: Good morning, Paul, and thanks for coming in. Just a light question to begin with, what do you think are the main questions that we as a Committee should be asking DWP as it thinks about its submission to the Treasury on the three-year spending review?
Paul Johnson: There are quite a lot of issues facing DWP. One question is what direction is policy going to go in. We will have to remain neutral on that. Is it going to become more or less generous?
Q3 Chair: Paul, when you say direction, what do you mean by direction?
Paul Johnson: More or less generous. Given where the current candidates for the Prime Ministership are, in terms of it being more generous across the board. We normally think of spending reviews as squeezing money out but maybe they are not looking to do that now.
There are a number of issues that it is facing. One clearly is the rollout of Universal Credit and what speed they are realistically looking to do that. There remains something to be said for making the transition even slower. It is not, obviously, going to be expensive for them, and potentially less risky. There are issues about levels of benefits. Working-age benefits have been frozen for the last four years and it would be good to quiz DWP about what they know about the hardship that that is causing.
There are issues about the impact of things like the two-child policy, which is particularly difficult for a small number of families. Again I think the DWP should be giving evidence on what they know about the impact of that. Within Universal Credit, there remains quite a lot of uncertainty about the impact of the requirement to see job coaches and so on for people who are in work. The conditionality regime for people in work is completely new. DWP is doing some trialling on that, but what will that look like in steady state and what will employers think about that? Those are serious issues.
There are big issues around some apparent policy changes that have happened within DWP. Sanctioning appears to have gone up a mountain and fallen off a cliff in recent years without any obvious announcement of policy change, but there clearly has been a policy change there so it would be good to get some sense of where that is going. Of course, if I were in the Treasury and looking to save money, I might well be challenging DWP on that. The figures are very dramatic.
Then there are the areas where spending continues to rise. For example, we have had in the last decade near enough a doubling of the number of children receiving disability living allowance and a doubling of spending there, obviously not intended. Is that something the DWP are comfortable with, are they looking to do anything about that?
If you are looking at the pensioner end of things, there remain the common issues around winter fuel allowances and whether they are going to continue. Is the triple lock going to continue and what is the benefit of doing that? There are a lot of individual areas where there are some quite significant and important questions to raise.
Q4 Chris Stephens: Thanks for that, Paul. You mentioned two things there that interest me. The first one is the level of benefits and the benefit freeze.
Chair: We do have questions coming up on that.
Chris Stephens: Okay. You mentioned pensions. Obviously pension credit might be a question we could ask because there is a lack of take-up of pension credit and obviously, with the licence fee issue and all of that, that is something that we could be asking as well.
Paul Johnson: Indeed. The licence fee is an interesting issue in itself. It might be interesting to talk to the BBC about some of that, because the BBC has said that it is putting aside £250 million to cover licence-fee payments for pensioners over the age of 75 on pensioner credit, but if you multiply the licence fee by the number of pensioners over 75 on pensioner credit you get to a number £100 million less than that. There are all sorts of reasons you might do that but it is not terribly clear, certainly in the BBC press release, why that is.
If you delve into some of what it has written, it says it is going to do a campaign to increase take-up of pension credit, which seems an odd thing for the national broadcaster to be doing. That might be, as it were, a risk for the DWP budget if it is significantly successful in that. On the other hand, the BBC might have several tens of millions of pounds spare if it does not do it or if it is not successful in doing that. You are right that the level of take-up of pensioner credit is relatively low.
A related issue that we have written on recently is the levels of pension credit are hugely more generous than the levels of working-age benefits for people out of work. Two things have happened. One is obviously the state pension age is rising so people are having to wait a bit longer before they access these much more generous credits. They are about 130% more generous compared with a much lower difference 20 years ago. But we have had this move to a world in which if one of the couple were under pension age and one over, they could access pension credit. Now they both have to be over pension age. This has a big effect on about 100,000 households, which is driven by these very different rates of minimum means-tested benefits above and below state pension age.
Q5 Chair: Have you done any calculations, Paul, on the cost to the BBC on finding new pensioners eligible for pension credit, with what the loss would be to the Department of that person registering for pension credit?
Paul Johnson: We have not done those calculations. It would not be per additional person who registered. It would not be terribly hard to find that number. I believe that Frontier Economics did a big piece of work for the BBC on this and made its own estimates of the costs of a take-up campaign and that kind of thing. It may be within its report that it has also looked at the cost to the public purse, but I do not know that number at the moment.
Chair: It is a good way for the BBC getting back at the Government, to say, “We are running a campaign that you might have done over the years. Although it is costing us the licence fee, it is going to cost you even more”.
Q6 Steve McCabe: Could you outline what you think the main trade-offs are that the Department will face in policy and spending terms over the coming year or so?
Paul Johnson: In a sense it is going to face the same trade-offs it always does. It has a huge budget in terms of the scale of Government spending and there is always a trade-off between savings and generosity within that system. There are trade-offs about the speed of rollout of Universal Credit, there are trade-offs around the way in which housing benefit works, for example. We have a system where we have moved that trade-off quite a lot over the last seven or eight years to a system that used to pretty much go up with rental costs. It no longer does. You have that gap between rental cost and housing benefit. In a fiscal incentive sense that looks quite sensible, but from the point of view of the recipient that can look pretty tough. There are those sorts of trade-offs.
There are the usual trade-offs in things like Universal Credit around do we want a bigger work allowance, because that would have more incentive for people to get into work but reduces incentives further up the earnings distribution because you have a higher work allowance, and of course it costs more money. You would think it would be thinking about trade-offs between people of pension age and people of working age. That is not quite how it works but if you thought there was a fixed budget across the two, do you think that it is all being used in the most appropriate manner at the moment and might you want to move some money down from those over pension age to those under pension age.
What I cannot speak to is some of the other trade-offs that it might have over the amount that it spends on the administration of the system to make sure that it is working effectively, and how significant those trade-offs are. For example, if you put more work into active labour-market policies, if you put more work into work coaches with Universal Credit and so on, what is the payoff to that? That is a big question for it, which I do not know the answer but it probably needs to think about that quite seriously.
This is one of the problems that I think the DWP can face in spending reviews, which is that often the spending review focuses on the DEL rather than the AME part of it. You can cut the DEL and that can lead to increases in AME spending because you are not controlling the AME spending so well. It is certainly a challenge for DWP to make sure that in its conversations with Treasury it is absolutely clear about those sorts of trade-offs and does not accept cuts to DEL if it thinks that is going to have a knock-on effect in terms of increased spending on benefits, which has happened at times past. It certainly happened during the 1980s in a very big way. I think some of those lessons have been learned but they remain an ever-present danger with this budget.
Q7 Steve McCabe: Do you have any sense that there is a distinct direction of travel in the Department that demonstrates that it knows where it is trying to go in terms of these trade-offs?
Paul Johnson: There has obviously been a real direction. A large part of the Department has clearly been very focused on delivering Universal Credit, for a long period of time. In that sense there is a clear direction, which is rolling that out and making it work. Whether it is within the Department or within Government more broadly, there is a clear sense that there is a pensioner pot and a working-age pot and those are treated very, very separately in terms of policy approach and in terms of not addressing the trade-offs. Effectively it is two completely separate Departments, so it is not internalising any trade-offs at all between those two parts of its remit.
I do think that over a long period that has been a focus on—certainly for those who are out of work on JSA and so on—a clear direction of travel in terms of active labour-market policies. I wonder whether the same is true for things like DLA PIP, where one’s impression is of a regime that has come in in ways that were wholly unexpected by the Department, and indeed by everyone else, cost a lot more and involved a lot more people claiming, and policy having to adapt very quickly to a much greater pressure than had previously been expected and now, one senses, just managing what has come out at the end rather than trying to push in any particular direction from where we are, it very much being a demand-led system because it is managing something that, for a period at least, if not in crisis was certainly putting more pressure on it in ways that it had not expected.
Q8 Steve McCabe: If you were trying to predict the biggest problem that you foresee confronting the Department in the months ahead, what would you say is your greatest concern?
Paul Johnson: There are certainly risks. Until it is fully rolled out there are going to be risks around Universal Credit and I am sure you are aware of most of those. There are some risks I have not seen much discussed there. For example, a lot of the tax-credit debts that people have will be rolled into the Universal Credit system. That in a sense is not specific to Universal Credit but it is yet another thing that might begin to undermine confidence in that system.
I think there are risks around work requirements for people who are in work. We do not know how employers are going to respond to the fact that someone else externally is maybe trying to encourage their employees to go and work somewhere else and that it is the Government doing that. I think there might be some concerns or risks around that.
The minimum-income floor for the self-employed. I completely understand why it is there, because it is so hard to see what the self-employed are earning, but that is one of a number of things where one might see serious hardship relative to where we are at the moment. The other big thing is the two-child policy, which I think is certainly at risk of doing that. It has seen a court case, that at the moment it has lost, over the benefit cap, so there are clearly some risks over that. Some of the risks—perhaps more than in recent years, perhaps rather like some of the bad publicity it has certainly had around access to disability and incapacity benefits—might start to push into other parts of the system where the increasing harshness of some bits of the system begin to bite on them.
Q9 Chris Stephens: Paul, is there anywhere in the Department that could see future savings or have we reached the end of the line in terms of that?
Paul Johnson: As I say, there are some obvious options, for example getting rid of the winter fuel allowance. I cannot remember the numbers but that is a couple of billion pounds so that is not a trivial amount of savings. We still have the triple lock on the state pension, at the same time as we have had nominal freezes for working-age benefits. You could change the indexation arrangements of the state pension. That would save a significant amount of money over time.
Q10 Chair: This side of the election, Paul?
Paul Johnson: I cannot speak for what the—you have a benefit system. It is very unpopular cutting pensioner benefits. There is not much in the way of working-age benefits that does not go to people who are on relatively low incomes. If you are looking for politically easy wins, I am not sure that there are any of those.
Q11 Chris Stephens: Of the savings that you have suggested there, all of those would have an economic impact, wouldn’t they? If there was a further freeze in benefit or if you were cutting anything to do with pensioners, that reduces their spending power to spend on goods in the economy, so there it may be a saving to the Department but there would be an impact.
Paul Johnson: Suppose you were to do the same to the winter fuel allowance that the BBC wants to do to TV licences: only make it available for people on pension credit. The saving would be small, it would be in the £1 billion mark, something like that. I think that would not show up in anything you could measure in the economy.
Q12 Chris Stephens: The uptake in pension credit could knock that saving out, couldn’t it?
Paul Johnson: The uptake in pension credit could certainly change that. My guess—and you can ask colleagues from the OBR later—is that even the £5 billion or £6 billion that has been saved from freezing working-age benefits over the last few years, I do not whether that feeds through into any of the growth forecasts. In terms of the overall economic growth, in a £2 trillion economy, a couple of billion is not going to—because this is spread across the country, I think the impact is going to be pretty small.
Q13 Chris Stephens: Perhaps to the overall economy but not to the individual.
Paul Johnson: Not to the individual, no. Absolutely for the individuals, of course, and that is the difference. Clearly for people of working age who have seen nearly a 10% cut in the real value of their benefits, that has been a really significant impact. That is why I say one of the things I think you might want to ask DWP is what it knows about what that has caused in terms of hardship, whether it still thinks this is an adequate or subsistence level of benefit.
I think the same is true for some of those hit by the benefit cap and some of those where the two-child limit starts to bite. These are significant. The two-child limit and the benefit cap are for small groups of people but have big impacts on their living standards. For the benefit freeze for large numbers of people it had a 10% or so impact on their living standards. I would really like to know what DWP knows about the impact that that has had on people’s behaviour, on their wellbeing and so on. I absolutely agree for individuals it really matters. In response to your earlier question, a measurable impact on the macro economy I think is less likely.
Q14 Chair: You reported recently the increase in the number of people work-poor and you say there are no easy answers. There are easy answers, aren’t there?
Paul Johnson: We are now in a world where something like two-thirds of people defined as poor are in work, and that has changed dramatically. It is a bit less than two-thirds but that has changed dramatically since 20 years ago. Partly that is good news because there are a lot fewer people out of work. We now have two-thirds of lone parents in work. Twenty years ago only one-third of lone parents were in work. That has pushed up in-work poverty because most of those or a lot of those lone parents are on relatively low earnings.
A second part of what has happened, because we measure this on an after-housing cost basis, and an important part that often gets neglected, is that in-work poverty has been pushed up by increasing housing costs for people towards the bottom of the income or the earnings distribution. I think one of the most striking things that we have found is that over the last 20 years housing costs for people of working age at the top of the income distribution really have not changed at all. That is because interest rates have come down so significantly. Homeowners and particularly older working-age homeowners have relatively low housing costs by historical standards but people in the bottom quintiles have seen big increases in housing costs because more of them are in private rented accommodation and both social and private rents have risen very substantially, particularly over the period between 2000 and 2010.
One of the big issues of in-work poverty is housing costs and we have been cutting the generosity of housing benefit. A housing benefit might be a short-term solution there. In the long run, there are big institutional issues in the way that the housing market works. In terms of low pay itself, again we have seen a number of—
Q15 Chair: Can I go back on that? The theory that Governments put forward for capping the size of taxpayers’ contributions to people’s rent was that landlords would respond by holding down the price of housing. That has not worked, has it?
Paul Johnson: We did a short-run evaluation of that when the policy was brought in and there was not much evidence of that working. What we do not have access to—another thing you might talk to DWP about—is longer-run data. The only people in the world who are able to tell you the answer to that is the DWP because they are the only people in the world with access to the data. It is a very good question, to which we do not know the answer.
It is worth saying that real rents have not risen very much in the last 10 years. Whether that has anything to do the housing benefit I have no idea, whereas they did rise very fast in the 10 years before that. Again it would be a very good question for DWP as to what it knows about that impact. There may be some published research that I am not aware of.
The key thing that you are getting at in terms of thinking about in-work poverty is earnings. There are several things going on there. One is obviously hourly wages. Hourly wages towards the bottom have done pretty well over the last seven or eight years because the minimum wage, the national living wage, has gone up quite significantly. Certainly if you look at the last period since 2010, hourly earnings at the bottom have gone up faster than hourly earnings further up the distribution.
If you look over a longer period, there is a bigger issue about number of hours worked. For the first time ever, we are seeing significant numbers, really significant numbers, of low-paid men working part time. We do not understand why that is. That might be associated with the benefit system and the fact that there are lots of 16-hour and 24-hour jobs out there in low-skill pay because that is what the benefit system directs. It might be to do with the expansion of the gig economy and zero-hours contracts and all those sorts of things.
If you want to solve low household incomes for people in work, it is not just hourly earnings that matter, it is also the number of hours that are worked. Of course, their overall income is impacted by the benefit system. The benefit system has played a very positive role in holding up incomes at that level. If you look over a long period you have seen a big increase in earnings inequality among households in work, you have seen a much smaller increase of income inequality because the tax-credit system has been quite effective at increasing those incomes. Where do you need to look? You need to look at hourly earnings, hours of work, the benefit system and housing costs to begin to address that issue.
Q16 Chair: If we had had you here before the Government’s policy substantially to increase the national living wage, would you not have reported that all economists would be saying, “If you do that, there will be significant increases in unemployment”?
Paul Johnson: No.
Q17 Chair: Hasn’t one of the consequences been that that has not occurred but those at the bottom who might have been made unemployed are now faced with nil-hour contracts, with self-employment? You answered immediately, “No”, but most economists would have said that these sorts of increases that we have seen in the national living wage would have caused unemployment. They are very significant jumps, aren’t they?
Paul Johnson: Yes.
Q18 Chair: But we have seen big changes in the nature of people’s employment within that living-wage structure.
Paul Johnson: I think what I would have said and what I still would say is the honest answer is we do not really know.
Q19 Chair: That would not stop economists telling us, though, that that would be the consequence of it, even though you say we would not know.
Paul Johnson: If you look at all the evidence on the minimum wage up to 2015, there is no evidence that it had any impact on unemployment levels. That is very clear. It is very clear there is no evidence that it had an impact. That is a rather different thing from saying it definitely did not have an impact, because proving a negative is rather more difficult and a lot of these studies have relatively low power in terms of the certainty around them.
The issues around a swiftly growing minimum wage are first what level of evidence and certainty you need in order to increase it. There is a difference between the Low Pay Commission’s previous remit, which was essentially do as much as you can where you are quite confident it is not going to increase or have an effect on unemployment, as opposed to the Government moving it up more quickly, I think taking a judgment that they are willing to take a bit of a risk around that.
Your point is apposite about the risks around different things, the risks around how this plays out in terms of the structure of the labour market. If you think, for example, of the swiftly increasing numbers of self-employed or company owner/managers, most of these, the vast, vast majority of these, are individuals working on their own account, often for just one employer, sometimes for more than one employer. Why? Everything in the employment system makes it easier or cheaper or more convenient for employers to contract with a self-employed person. The minimum wage is one of those things, because obviously the self-employed are not covered by that. Pensions auto-enrolment is another. The tax system very strongly is another and a whole series of employment legislation is another.
I do think we need to think about the extent to which we have created accumulation of incentives for hirers to prefer to contract with people who are formally self-employed. I do not know whether that accounts for the very, very fast rise in those numbers but it would seem plausible that it is a significant contributory factor.
In terms of how far can you push the minimum wage, the Low Pay Commission has done huge amounts of work on that. There are concerns that go well beyond unemployment, including the things that you have described but also the structure of wages within companies and the potential difficulty of persuading people to take very small increments to move the next level up. There is clearly now a big jump between age 24 and age 25 in terms of the level of the minimum wage. One might think that the risks are greater the further down the age distribution that you go.
Q20 Derek Thomas: You have said quite a lot about the freezes and caps that we have seen. Are there public policy justifications for using these? Do you think there would be an argument to continue the kinds of freezes and caps we have seen, in the spending review?
Paul Johnson: I think it is very hard to justify a freeze because the actual impact on both real spending and real living standards is unknown ex-ante because that depends on what happens to inflation. If you think that you want to save money by reducing the real value of benefits, I would much rather see a policy that says we will increase them by CPI-1 or CPI-2 rather than by zero, because that could turn out to be, as it did in one year, nearly CPI-4. That is a very uncertain impact on future spending but particularly on future living standards. In the end it is a question for Government whether they think the benefit levels are appropriate. If they do not, if they think they are too high, then probably something that leaves them not falling in nominal terms is probably the easiest way to achieve that.
In terms of the two-child limit, in terms of thinking about the appropriate policy going forward, clearly there is a public policy justification for having a limit, which is if you are on benefits and you know the state will pay for additional children there is an incentive issue—incentive might not be the right word, maybe a fairness issue—about the state picking up the tab if you have a third, fourth and fifth child. That does have to be weighed against the fact that a lot of people who move on to benefits after they have had a third child might not have expected to be moving on to those benefits and might have done that in the expectation that things would work out fine. As ever, I think there are tough trade-offs but you can find not daft public policy justifications for something like that.
Q21 Derek Thomas: Thank you for that. When it comes to the cap, are you aware of work that is being done to understand the extent of the impact? In other words, how many families are going to find that their spending power is reduced when inflation goes up under the cap? If the cap is set, is there any work that is done within DWP or elsewhere that looks at how many families would be affected if inflation, as you say, grew by 4%?
Paul Johnson: Yes. In a sense, that is a relatively easy calculation because we know the numbers of people who are affected by it. Essentially that is the number of people who will be affected by that cap. There is a slightly more complex calculation, which is that clearly people move on and off benefits so over a period more people will be affected than are affected in the cross-section. That certainly is something that is knowable.
Derek Thomas: Assuming that not everybody will be up to the cap.
Paul Johnson: Which cap are you talking about?
Derek Thomas: I am talking about the overall cap.
Paul Johnson: Sorry. Yes, I do not have the numbers in my head but it is known how many families are affected. It is one of these things where, in terms of the millions who are receiving benefits, it is a relatively small number but the impact on them can be very big indeed.
Q22 Derek Thomas: You mentioned the pension triple lock, the state pension triple lock. How sustainable is that? The Chairman made reference to the general election. I am assuming it was a manifesto pledge so I should point out the reference. Is it something that can carry on indefinitely or do you think that it has to be addressed, that it is not sustainable?
Paul Johnson: Relative to the underlying legislation, which is that pensions should go up in line with earnings, it adds something like 1% of GDP in expectation to pension spending a number of decades out. So 1% of GDP is a significant amount of money in public spending terms; it is £20 billion at the moment. Relative to some other pressures, like on health spending, that is a relatively small number. Given the scale of pressures from an increased number of old people, social care, healthcare and arguably a number of the other promises that are being made at the moment, the real question is is that the best way of spending money going forward. My instinct would be that there are probably better ways of spending that money across the piece in terms of either thinking about the impact that has been had within DWP on those of working age or, perhaps even more, impact of things like the justice system and social care and some aspects of local government funding that have been cut so much in recent years.
Chair: Thank you very much. We may be writing to you following up some of those answers. Thank you very much for coming.
Examination of witnesses
Witnesses: Robert Chote and Andy King
Q23 Chair: Robert, welcome. Might you introduce yourself for sake of the record?
Robert Chote: Yes, I am Robert Chote and I am Chairman of the OBR.
Andy King: Andy King on the Committee at the OBR.
Chair: Thank you very much.
Q24 Ruth George: Good morning. I will start off by asking you to give a bit of an overview of where we are going and in what sort of areas should we expect to see the biggest changes to the current trends in welfare spending over the next two years based on current existing policy.
Robert Chote: The forecast horizon that we generally look over is a five-year one. Typically, because we are focused, obviously, on the costs of the welfare system in terms of the payments that are made and its implication to the public finances, we normally look at that in the context of a share of GDP, relative to the size of the economy that has to finance all of this. If you look at our latest forecast over the next five years, through to 2023-24, welfare spending is a little above 10% of GDP and the projection is that it would fall slightly by 0.2% of GDP.
You have things moving in both directions. You have state pension spending rising, probably, as a share of GDP over that period. That is down to the triple lock that you have just been talking about. In previous years that effect would have been offset by rises in the state pension age, which you do not have so there is less of an offset there. We also assume that there will be a rise in disability benefit spending as a share of GDP, primarily as higher incidence feeds through to caseloads. On that, there has been a rise in survey prevalence of disability. The rise in spending has been faster than that so the relationship between the two is not one for one and entirely predictable but the expectation is it goes in the same direction.
The offsetting, those two things are more than offset, essentially, by lower spending on working-age benefits. There you have the policy cuts working their way through, plus an expectation that you will have earnings picking up as we go through the forecast. That means that given the way in which the thresholds are set, you have a downward pressure on spending as a share of GDP from the proportion of people who are not eligible.
Q25 Ruth George: Obviously being a proportion of GDP, it is quite dependent on what GDP is and, as you say, on working age the assumptions have been that we will continue to have rising wages and rising employment. What is going to happen to those predictions if we end up with one of the scenarios that has been predicted due to Brexit or any other impacts on the economy?
Robert Chote: If you think generally about what would happen if there was a recession, from whatever source that came from, the short answer is it would depend on what sort of recession it was. It is quite striking, if you look back at the way in which welfare spending evolved over the last two recessions, it evolved in different ways, partly because of the characteristics of those. A key issue obviously is whether the downturn manifests itself more as a rise in unemployment, as was the case in the 1990s, or as weaker earnings growth, as was the case in the more recent recession.
There is also the question of what happens to inflation over this period. The relationship between inflation and recessions is partly down to what caused the recession. Sometimes you have a recession because inflation is higher than you want it to be, and that has to be choked out and it is brought down, but it is quite high to begin with. You can also have situations when there are external upward pressures from oil prices or the exchange rate or whatever that might be that pushed up. In the last recession, for example, you had retail price and consumer price inflation stronger than was expected before you thought there was going to be a recession, despite the fact that the economy was weaker. In the last recession you had upward pressure more on the pension side because of the way in which the uprating works. Go back to the previous one and it was showing up more from unemployment, housing benefit and so on.
There are those direct effects from the economic determinants, but the other issue of course is how the Government respond to this and how the particular bits of the welfare system behave, tax credits operating as something of a stabiliser in the past. Will Universal Credit—or the sort of hybrid as we are halfway across the tightrope getting from one system to the other—perform as an automatic stabiliser and as a source of different spending in a different way because of the differences in conditionality, for example?
There will also be choices the Government will have to make about—this brings you to some of the questions that DWP will be thinking about when it is thinking about what DEL to ask for, which is what would you want to do in terms of injecting more money into the operation of the system in that event. One of the sensible things that was done ahead of the previous recession was to make sure that jobcentres are adequately funded for this when it hits, having learned the experience from previous times.
Q26 Ruth George: Obviously a lot of the impact over the next few years and the cuts in welfare is going to fall on people of working age. That is dependent not just upon monetary and financial factors, such as the level of the minimum wage, but on employment regulation as well. Obviously the last recession happened in an era where we have seen much more flexibility within the labour market, much more of a gig economy and that has grown since then. How far is any regulation on employment factored into what you can predict?
Robert Chote: I do not think, trying to understand the previous behaviour, that we have extracted a financial element that you would attribute to what was going on at the regulatory level. Is that fair?
Andy King: Thinking back to the last recession as it was unfolding, sat where I was at the time in the Treasury, we were slightly worried that flexibility would mean firing. That was what happened in the US, whereas flexibility manifested itself in terms of hours and hence showed up in tax credits rather than unemployment benefits. If I learned one thing from that, it is that it is difficult to predict as it is happening. It could be the same again or zero-hours contracts could mean that it manifests itself in a different way again.
Q27 Ruth George: The employment market is broadly similar in that those rights and the short-hour contracts and the zero hours and an increase in self-employment seem to be here to stay in terms of the structures of the labour market. It is how to eradicate in-work poverty and therefore a lot of child poverty while we have a labour market that is in that situation. That is why I was asking if you were looking at things to maybe increase hours of work as well as basic wages.
Robert Chote: The job you have given us, Parliament has given us, is very much to focus on what the amount that would be spent is. We have been instructed not to look at the distributional and poverty consequences. Obviously the IFS is all over that like a rash. The Government have produced their own analysis there. We have not looked at that, the effectiveness of different policy options from that perspective, important though that is. It is not our area.
Q28 Chair: The Government could easily change that, Robert, couldn’t they?
Robert Chote: Easily. They could legislate to change the remit if they wished to. This, as you know, was debated back when we were set up, and re-debated a few years later as to what the scope should be. It is interesting—partly because of the work that Paul is doing now and that I was doing when I was doing his job before—that lots of people have come up with things that they think that we ought to do. Few of them have said that there is a big gap in the quality of the distribution analysis that we need to fill. I think that is partly testament to what the IFS does anyway.
Q29 Ruth George: There is, but the Government do not have a duty and responsibility to listen to IFS, as we see in quite a few Ministerial statements, whereas they do have a duty and responsibility to listen to OBR. Who do you liaise with in terms of poverty within Government and what impact policy is having on poverty?
Robert Chote: As our remit is set out at the moment, that is not part of the—
Q30 Ruth George: Not part of yours, but you do not do any sort of liaison within Government?
Robert Chote: No. We are obviously looking at the way in which the interaction between what we think is going to go on in the economy and the Government’s stated policy at the time about the structure of the welfare system is going to affect expenditure over that period. Obviously future changes in policy will be affected by what the Government of the day’s objectives are for poverty and for inequality more broadly and how seriously they take the idea of pursuing those. It is not something that we have been tasked with checking that they are on course to achieve, although, as I say, some people have suggested that we should.
Q31 Ruth George: There is no other part of Government that you liaise with to say, “These are our predictions, here they are” to see what another part of Government is doing to map the impact on poverty, as we had when we had a statutory target?
Robert Chote: Not on the income distribution side. The Treasury has published distributional analysis. It stopped for a while and has restarted. I now cannot remember the reporting requirements under the legislation for the child poverty target. That was never relevant as a reporting function for us. I cannot remember what the former entitlements are under the child poverty legislation as it was. There was a requirement of a comply or explain variety but I cannot remember how it worked. I knew 10 years ago.
Q32 Ruth George: Fair enough. We may have to turn to those with longer years of experience on that. In terms of state pension spending in coming years, I think you said that was falling, but from the figures it looks like it is going to be rising substantially.
Robert Chote: It is rising. It is offset by other things that lead the overall decline. As will all of the numbers you are looking at, you are interested in the caseloads, the number of people who are going to be receiving and the generosity. In terms of long-term projections you are obviously looking at the population projections, life expectancy, you are looking at the state pension age and the link with longevity that is now put in here and that is all affecting your caseload side. On the generosity, we are down to the uncertainties around how the triple lock would work.
In our long-term projections, we calculate the average effect of the triple lock operating over a number of decades. I think we assume, roughly speaking, that the triple lock is equivalent, over the long term, to earnings plus 0.35 or 0.36-ish percentage points. Obviously that is quite lumpy. This depends on when the 2.5% element kicks in, so therefore will depend on what the long-term path is for productivity growth and for earnings growth as well. All we can really do in projecting that is to ask, “What do you think the average will be based on past behaviour?”
The key point, and this comes back to what Paul was saying before, is that the UK is an outlier in having a triple-lock arrangement of this sort. Built into it is a ratcheting up of expenditure as a share of GDP, all other things being equal, because you assume broadly speaking that earnings will rise in line with the growth of the economy. You are basically saying there will be circumstances in which we want to kick that path higher so that is an upward pressure and a source of risk to the public finances over the longer term.
Q33 Ruth George: What have you built in in terms of caseload? We have pretty much flat life expectancy now rather than it having been rising. We have seen full pension qualification under single-tier rise to 35 years rather than 30, more people starting work later in life, at 21 or 22 rather than 15 or 16. Are those impacting on caseload or when do they start to kick in?
Robert Chote: Broadly speaking, obviously we do not do a population projection of our own. We choose from among the available variants that are produced by the Office for National Statistics. Then we take the Government policy as it is stated and apply that through. Is there anything on the age-specific employment rates?
Andy King: Yes, in outturn, life expectancy improvements have paused, but in the projection that the ONS produces and that we use, it is still rising over time. On the back of that we assume the legislative path of state pension rises plus our own assumption where the terms of the longevity link would bite in the future, so the state pension age gets to 69, I think, in our projection. In terms of employment rates, we have focused more on what is happening at older ages where movements in the state pension age seem to generate higher employment rates for those close to it. In a sense, raising the state pension age boosts GDP because it boosts employment among those affected.
The net effect on the caseload side, the caseload does rise as a share of the population because of ageing but that is tempered quite a lot by the longevity link of the state pension age, and as a share of GDP it is tempered again because more people work. The big thing in the projection is the triple lock and there is not a lot to base that on, so that is quite finger in the air. The 1% of GDP figure that Paul mentioned is one of the largest assumptions in our long-term projections. The risks around it are quite great. It is one of the issues that we feel beholden upon us to look at those risks, especially when productivity growth risks are quite important.
Q34 Ruth George: Do you look at personal pension provision in any of your forecasts and make assumptions based on that or is it purely state spending?
Andy King: The DWP models that we use as the vehicle for these projections do contain assumptions about personal pension provision. They are not important to our long-term projections in a quantitative sense. They probably would be more important to the questions you were asking earlier about poverty, that are beyond our remit. They would be important on the revenue side if we thought about things in that way at the long-term level, but we simply think about the age profile of tax revenue, so again it does not feed into the numbers.
Q35 Ruth George: Presumably you are not looking at impact on average income per pension and how that might impact on pension poverty?
Andy King: No.
Q36 Ruth George: As far as you are aware, is there anyone within Government doing that analysis?
Robert Chote: It is DWP or the Treasury if anybody is doing it. The IFS would be looking at it as well.
Q37 Chris Stephens: How would you rate the Department for Work and Pensions’ use of evidence when it comes to making its spending decisions?
Robert Chote: I think if you look at DWP relative to other Departments, partly because of the nature of the business it is in, there is more detailed microdata that it has that it is working with than would be the case in some other Departments. If you look at the evidence that we see from them in the job that we are doing, it is important to make the point that we are not its only customer in terms of the material that it is producing. We are interested in what it can provide us with in order to help us with the task of forecasting. It obviously has a much wider set of decisions, for example about where it is deploying people geographically across the country and so. It needs to be done on that basis.
Q38 Chair: Are you frustrated by their data? Are you frustrated by the quality of its data when you begin your work?
Robert Chote: Frustration with the quality of data is a daily part of our lives across the whole thing. You would always like it to be better. There are some areas where obviously it matters more than in others. There are some areas of the forecasting activity where I think you have pretty good areas. It is always more problematic if you are an area where policies were forming and the system is changing. That is not unique to the welfare system. The same would be true if you were trying to forecast capital gains tax and the capital gains tax system keeps changing around.
Q39 Chris Stephens: Just to go back a second, in comparison to the other Government Departments, where is DWP in terms of using this evidence to make its spending associated with spend on clients? Is it the relegation zone, is it competing for a European place? Where would you say the DWP was in compared to other Government Departments?
Robert Chote: Obviously we have more experience of it than we do of the other ones. As I say, given the nature of the business it is in, there is more data that it has but there are more complicated tasks to be done with it. Clearly you can look at some examples. There is a particular challenge at the moment with the move from the legacy system to Universal Credit, where you are moving from—I would liken this to walking a tightrope from one relatively firm piece of ground to another, where when you have lots of information and lots of data about how an existing system has been operating for some time, you are on firm ground. You will be on firm ground when you are in the new system and that has been operating and running and you have time.
There are particular difficulties in the transition between the two, and knowing when you have new information about what is going on with expenditure in the current year about how much you can learn from that. Can you tell whether it is because the rollout is moving more or less quickly or because the cost of individuals—the mix is different. That is a challenge. It is an area where you could probably do better but it is intrinsic to the nature of a reform of that sort.
If you want an example of a less successful use of evidence, go back to the re-costing of the DLA to PIP transition in 2012, where there was a question about what was the available evidence base. We spoke about this a bit in our welfare trends report where we and it put too much weight, I think, on the limited evidence, which was not designed for that purpose and turned out not to give a good indication to how things would move in the longer term.
Andy King: I think the frustration often is that the thing that you really want to know about is the one that it is most difficult to know about—how you are walking across the tightrope, in Robert’s analogy. We ask an enormous number of questions of DWP analysts in the process of forecasts and our other reports and generally we get good answers, well thought through. People are willing to work hard to answer our questions quite quickly during forecast rounds, but the areas of frustration are material to our forecasts. If you look at the way the disability benefits forecast has been revised, that is a multi-billion pound a year revision over time and it is linked back to use of evidence.
The thing that as an institution we have discussed with the Committee in writing is about Universal Credit, which is such a large amount of spending that is forecast in a suboptimal way at the moment. It is the only way it can be forecast at the moment because of where we are in the rollout and what that means for the available data and the way you can use it. That is an enormous issue that I fear is going to last a number of years before it can be resolved and forecast and analysed in a normal way.
Robert Chote: Another relevant general point is often the data and the analysis is better and more focused on the caseloads than it is on the average amount. Obviously, if you are doing a forecast for the spending you are multiplying the two together. But you have more information and there is more focus on the number of people affected than how a particular bundle of features of the system adds up to the total amount of money that people get.
Often the forecast errors will not be the fact that you have the number of people getting it wrong but that the average amount and the makeup of what they are getting is different from what you would expect. That often takes longer to emerge and to understand than the former.
Q40 Steve McCabe: You mentioned your welfare trends report there. I think you said in it the Department need to look more deeply at the nature and interpretation of key pieces of evidence when it comes to costing policies. What did you have in mind, in particular, when you said that?
Robert Chote: Partly that would have been, as we have just been discussing, in 2012, the information there on what weight one could have put on it would be quite a good example of that.
Q41 Steve McCabe: The PIP?
Robert Chote: Exactly, that is right. You are often in a situation where you might have a formal sample or a trial or it is not designed as an experiment in that way but you have a small sample, working out how representative that is likely to be of the system of the whole when you scale it up is often a challenge. Sometimes the difficulty is that it is the best you have to go on, but we and others could have done more to probing whether it was merely uncertain or likely to be biased in one direction or another.
Q42 Steve McCabe: Is that quite a regular occurrence in the Department, in your experience?
Andy King: No, it relates to where things are changing and so the evidence of what has just happened is not the relevant piece of evidence.
Q43 Steve McCabe: The PIP example you quote relied heavily on a population that was very unlikely to be representative of the people who ended up being assessed for PIP; 900 people and self-selecting. It was hardly the sort of thing you would normally expect to give you good evidence.
Robert Chote: Indeed, and it was not designed to be used for that purpose therefore to have those sorts of more randomised, more representative features that you would ideally like, but often you are in the world where you are looking for the information that you can find. The question then is how much weight you put on that.
Another related issue is you always want to think about the behavioural effects that policy is likely to have and there is a natural tendency, if you are a Department that is invested in the success of a policy, to be very upfront about what you think the positive behavioural impacts of something are likely to be or less likely to come out and trump it and say: “Look, I think we have identified something that is not going to work terribly well on this”. That is in the nature of the beast and it is part of our job obviously to try to ensure that we correct for that.
Q44 Neil Coyle: The Government gave itself the power to extend PIP to children and over-65s, so if the Government decided to go ahead with that approach how would you do it differently?
Andy King: The OBR’s role is essentially to scrutinise the way DWP approaches the question. What we have in the report—and some of the things in hindsight do not look very good—is a much better list of questions that we ask. We ask them routinely now of costings that have this kind of nature. If that were to happen, as ever they would start with the evidence of what is happening in DLA, say. If there were a question about whether PIP would result in different proportions receiving the award when they claim, then maybe they would design a trial that was better. Maybe we would wait longer before we put a number on what went in the forecast. It would all depend of course on how concretely the policy was stipulated at the start.
Those are the kinds of things and obviously we have quite a long list of questions about, “You did a trial, did you use your best people to run it? Is that going to be what happens when you scale it up and the average person is doing it?” and those kinds of questions.
Q45 Steve McCabe: Universal Credit is obviously the big issue in the Department at the moment. A dominant feature of Universal Credit appears to be the use of advance payments now, which the Government regularly refers to as ensuring that people do not get into debt or do not end up in hardship. Was there any evidence produced at the outset to show that advance payments would be such a significant feature of Universal Credit? What impact do you think the acceleration of the use of advance payments has had on the overall management of the Universal Credit policy?
Robert Chote: I was wondering whether at the outset it was designed to be as important as it has ended up to be, which I cannot remember.
Q46 Steve McCabe: There has been a report just the other day querying how exposed advance payments are to fraud. They are running at a very high rate. Government Ministers regularly tell us that we do not need to worry about the five-week delay in payment because they are available and they do not charge interest on them. They are now a very significant feature of the system. What was done to identify that that was going to happen or has that just emerged as it has been going along?
Robert Chote: One issue has been that obviously, in terms of the net expected cost relative to the legacy system, Universal Credit has been through a journey. We started off with a vision of this as being deliberately designed to cost more than the legacy system that it was replacing and the aim there was obviously the “making work pay” arguments, and so on. That was a design feature of the system.
You then had an extended period of time in which this was whittled away and you got to a position in which the objective was that UC would save money relative to the system that it was replacing. What happened subsequent to that is gradually, having gone down the hill, we are now moving back up again and we are back into a world now in which the system is expected to cost more than the legacy system that it is replacing.
One of the reasons that that has happened is that there have been areas where difficult cases, losses, having been identified and policy, partly because of the political calculus at the time, has responded to that and those savings have been whittled away. Presumably this is part of, to some degree, an element to that. One of the risks we would highlight about UC going forward is that as it continues to be rolled out and more difficult cases and categories of people who do not do well emerge, it is quite possible of course that the pain will be alleviated for them and we will move further up the hill, having gone down it in the first instance. But I do not know to what extend the advance payments are specifically dealt.
Q47 Chair: Can you go back to the earlier part of your answer there? Are you saying now that at this point in time Universal Credit is costing more than what the bill would have been if we had just kept the legacy benefits?
Robert Chote: In the steady state when you get through it, it is slightly complicated by transitional payments, but, yes, you are in a world where it was originally £3 billion more, £3 billion less and now £1.5 billion more I think is roughly the—
Q48 Chair: But Ministers’ figures when they say that there are losers and gainers, they are talking in rough terms of half the population will be gainers, half losers. Are you saying that that proportion is going to change over time to more gainers?
Robert Chote: That is a slightly different question depending on what the average loss and the average gain is. I cannot remember whether the report that we did on that we had numbers of losers and gainers.
Andy King: I think we did in money, but the other thing is we did the report before a number of the recent giveaways that have tipped the balance again.
Q49 Chair: You say “giveaways”, Andy; most of them thought these were a restoration of cuts.
Andy King: Absolutely, reversal of a variety of things that were announced mainly in 2015 or at least in spending effect. They are not always direct return to what was there before. The IFS did a recent look at this from more of a families’ perspective rather than pounds and pence overall.
Robert Chote: It is one of the striking features that it is easy to forget about UC in terms of the risks around the eventual costs. At one level you look at it and you say the legacy system costs a bit more than £60 billion and at the end of the day it is going to cost a bit more than £60 billion and maybe another £1.5 billion on top of that. Therefore you think if you are a bit wrong about the £1.5 billion it is not going to make a huge difference.
What it is easy therefore to forget is that that move from a bit above £60 billion to a bit more above £60 billion comprises a large chunk of gross savings and a large chunk of gross additional expenditures. There are errors and uncertainties around both of those.
The apparent difference, in the steady state eventual cost of UC versus the legacy system, understates the risk and uncertainty around the size of that. Normally you would be quite worried about saying you have a change here and it could be twice as big or only half the size. You think that looks quite a big forecasting error, in this context it would not look at all surprising given the fact that you are looking at the net effect of large giveaways and takeaways.
Q50 Steve McCabe: I suppose the only thing I am left wondering from that is if it is very difficult to produce a good forecast for a big policy area—and we could be talking about a few billion pounds here or there—does that mean that the policy is determined by whatever the values of the Minister of the day are and it can shift? Universal Credit is there not to save money, then it is there to save money, it is there to make more pay, then it is not there to make more pay. Does it just mean that the numbers are not that important to the Department, it just depends what the Minister of the day thinks is important? Is that a reasonable way to view it?
Robert Chote: It is a combination of the two. Clearly you are always having the Treasury very focused on what the costs are and you may have the Department with a different set of focuses and a different set of objectives. Clearly, looking ahead, it will be interesting to see what the constellation of relationships between the centre of Government, the Treasury and DWP are, and how that all manifests itself in the way in which the direction of this policy goes. Are we going to be in a world like the Gordon Brown era where the Treasury is very interested in the precise design of the welfare system or is it a much more “all we care about is the overall numbers that you get out at the end of the day”?
One lesson, looking back at the way in which the numbers move around, is that obviously whether it is a good or a bad idea and what its implications are for poverty and income, and so on, if you simply cut the generosity of something, it is a not entirely predictable but a relatively predictable answer to how much that is going to save you in the future.
If you have a reform of the system, particularly a reform of the system that involves gateways around who is allowed to qualify for particular amounts and how likely are those decisions to be challenged, the lesson is that big reforms—and you can see it as across the incapacity and disability and you see it with UC—often take longer than you thought they were going to and end up being more expensive than you thought they were going to be as you had set out.
It is certainly not for us, but it is for you to say, “Does that mean you should basically say let’s not meddle around with this anymore than is absolutely necessary because of the danger of unintended consequences?” But on the other hand, people obviously are entitled to have a vision for how they think the welfare system ought to operate. A change in vision that involves a change in structure brings with it risks and uncertainties around cost.
Q51 Steve McCabe: Would that be a good argument to apply to the five-week waiting time for Universal Credit? There does not seem to be any evidence base to say why that was necessary but the consequence of it seems to be now that we are throwing money at advance payments. Is that the sort of thing you are trying to discourage?
Andy King: We are not forecasters of how policy will evolve but we can—
Q52 Steve McCabe: I cannot see why at any stage in the construction of Universal Credit there would be an evidence-based argument to say, yes, people must wait six—then subsequently five—weeks before they get any payment. Maybe you are aware of something that I am not aware of. Somebody must have artificially dreamt that up. The consequence of what they have done is they have now created an entirely different issue around advance payments. Is that exactly what has happened?
Andy King: It is that second one I was trying to get to. I think I remember the rationale being that that was how a full-time employee might experience their move into work, so you get paid in arrears.
Q53 Chair: It is an ideal benefit, isn’t it, for the middle class? But not for the people for who it was designed.
Andy King: I could not possibly say.
Q54 Chair: We would be all right in this room but not if you were depending on daily or weekly wages, would you?
Andy King: Exactly. The second point that I was getting to is when we look at how the system is evolving, you can see that one thing gets squeezed and it pops up somewhere else. The seven days added to the five weeks; they caused a problem, they left, that advances had been doubled. Those advances are taken back over a period.
Steve McCabe: Unless they are subject to this major fraud that has been described in—
Andy King: Unless they have absconded. That money is taken back over a period. The period over which it is taken back is then extended. From our perspective, looking at risks to the numbers that we are producing, that is the kind of thing we are thinking of and again that is more broadly with a system where there are losers from some elements of it and winners from others, and they are different people. We think about what your postbags might be looking like and where the next pressure point is.
Chair: Ruth, you are going to continue on that, aren’t you?
Ruth George: Steve has covered the Universal Credit fraud aspect, thank you.
Q55 Chair: Can I just ask one last question then? Are there examples in your experience when the Department has provided you with data about what they think is happening to expenditure and that Ministers and Secretaries of State are coming out with a different statement to the public?
Robert Chote:. I do not recall anything of that sort. The level of data that we are looking at is not normally the sort that they would be talking about in public in that context.
Chair: But if you have Ministers defending the move to PIP.
Robert Chote: How much that is saving relative to—
Q56 Chair: Wasn’t the case put forward that this was going to give the most disadvantaged people more money at the expense of the less disadvantaged, but there would be also a saving to the public purse? If so, when did you become aware that was not what was happening?
Robert Chote: On the relative giveaways and takeaways, that would not have been something that—I am trying to remember. We were looking, as part of the welfare trends report, at the issue of was it saving money relative to the legacy system, and we concluded that—rather than saving the one and a bit when fully rolled out—when it was two-thirds rolled out, it was costing somewhat more than that. I do not think the Government said the opposite of that—after we had made that assessment, whether they had on the basis of earlier evidence said that they thought, as was indeed the original intention, that this was going to save money. That may have been the case but I cannot remember the precise sequence.
Q57 Chair: In your role, your office—as a check to try to increase the amount of truth in public debate—if Ministers were saying things, which were not backed up by the data the Department had provided to you, would you think it your duty to make a public statement to that effect?
Robert Chote: We have in the past: when somebody has either deliberately or inadvertently misrepresented what we have said, we respond to that. That happened famously with the Prime Minister some years ago on the impact of austerity on growth. We try to use our reports to set out and to increase the—obviously the welfare trends report is a good example of providing a lot more data and a lot more information and transparency there. That is the most valuable service that we can produce rather than going back and fact-checking every statement that Ministers come out with.
Obviously when those statements are related to official statistics—that is something that the statistics authority has made statements on—they would do that more in real time. Our value is more setting out of well-argued detailed analysis of something and allowing people to learn from that and to draw from that.
Chair: Thank you for gliding past my question with great skill. Thank you very much, both of you, for coming along this morning. It has been very helpful. We will no doubt be writing to you though.