Contents
- Attendance
- Decision on Taking Business in Private
- Pre-budget Scrutiny 2020-21
Pre-budget Scrutiny 2020-21
Under agenda item 2, as part of the committee’s pre-budget scrutiny, we will take evidence from the Scottish Fiscal Commission on its role and its forecast for Scottish social security spend. I welcome Dame Susan Rice, the chair of the SFC; Professor Alasdair Smith, one of the commissioners; Claire Murdoch, the head of devolved taxes and social security; and Claire Mellor, senior fiscal analyst.
I thank the members of the panel for coming along this morning, and I invite Dame Susan to make an opening statement.
Good morning. As this is our first formal session in front of you, I ask you to bear with me while I give you a little bit of background—it will not take long.
Thank you for inviting us to this session of the Social Security Committee. We welcome the chance to discuss our role and our work on social security. I will say a few words about what we do, our role in the budget process and how we work with the Scottish Government, but first I will provide a bit of context.
This year will be the third in which we have produced the independent economic and fiscal forecasts that must be used by the Scottish Government in its budget. It will also be the first time that social security makes up a significant component of the Scottish budget. With the devolution of executive competence for new benefits in April 2020—next year—around £3.5 billion of the Scottish budget will be spent on social security. Our job is to estimate how much will be spent on devolved social security benefits and programmes in the current financial year and over the subsequent five years. These forecasts are of the amount of money that will be paid out to claimants.
Our forecasts determine how much of the Scottish budget is allocated to social security. Any differences between the allocations, based on our forecasts, and the actual amount that is spent on social security will have to be managed by the Scottish Government during the financial year.
While we are forecasting the spending in Scotland, our colleagues in the Office for Budget Responsibility are forecasting benefit spending in England and Wales. Those forecasts are used by the Treasury to estimate the block grant adjustments—that is, the funding that is received here from the United Kingdom Government for the benefits that are devolved to Scotland. Any variation in those block grant adjustments will also have to be managed as part of the Scottish budget.
That is a bird’s-eye view of what we do and how our forecasts and the OBR’s forecasts fit into the Scottish budget. I will now explain a bit about how we forecast and how we work with the Scottish Government.
Forecasting can seem like a dark art at times, but, in principle, what we do for social security benefits is relatively simple. The cost depends on three things: the number of people who are eligible to receive the benefit; the fraction of those who apply for and are awarded it, which we call the take-up rate; and the amount of payment awarded. We look at past trends and consider how well those will predict the future. If trends are stable or there are no policy changes, the past is usually a reasonable predictor of the future. However, when policy is changed or a new benefit is introduced, we have to make assumptions and judgments about what will happen.
With new benefits or new policies, we do not have past data to look at and use as the basis for our forecasts. Through the process, we work closely with the Scottish Government, and we ask it to provide all available information on new benefits, such as the eligibility criteria application process and its communication plans. We review that information along with data on any existing benefits, to consider how things such as benefit take-up might change. We then use those assumptions to produce the forecasts, which will have a greater level of uncertainty because they are new—indeed, our colleagues in the OBR have had a similar experience when forecasting the UK Government’s changes to, for instance, the disability living allowance and personal independence payments. The early forecasts are quite volatile until things settle.
What we can do is explain transparently how we have produced our forecasts and highlight the forecasts that are most uncertain. Inevitably, of course, our forecasts will be wrong—no prediction of the future will be precisely right. What is important is that, once we have new or outturn data, we look back, understand why the forecast was wrong and try to improve it in the future. Every September, we publish a forecast evaluation report, and this year we considered our social security forecasts for the first time.
Having just discussed the challenges of forecasting new benefits, I can point to two examples that illustrate that point. First, we underestimated spending on the new pregnancy and baby grant last December by £2.5 million, or 59 per cent, because of an unexpectedly large number of claims in the first few days after launch. The early data on claims for that grant, combined with a better appreciation of the Scottish Government’s approach to launching new benefits, led the commission to increase its forecasts of spending in May. Secondly, the forecast of carers allowance that we produced in December 2017 was based on a full year’s spending, as the exact date of devolution was not available at the time that we made our forecast. Although the forecast error for that benefit appears large, at 74 per cent, accounting for the actual date of devolution and using our model reduces the forecast error to 2.6 per cent, which is very much in line with the kind of errors that the OBR would have.
Our next forecasts will accompany the 2020-21 budget, and we see two areas of particular interest relating to social security. First, as I have already alluded, there is a significant expansion in social security spending in the budget, and the Scottish Government will have to prepare to manage in-year any variations from the forecasts—the fact that that must be done in-year is the most important point. Secondly, the Scottish Government is proposing significant reforms to the social security system.
All the policy changes and new benefits that are planned for 2020-21—the next fiscal year—will be included in our forecasts. As we get information on the Scottish Government’s plans, we will incorporate them in our forecasts. Where the Scottish Government proposes changes to be introduced after the start of the fiscal year, we do not anticipate including them in our forecasts, and the benefit will be forecasted on the basis of a continuation of existing policy.
Thank you for your patience. I hope that you found that overview helpful. We are happy to answer any questions that you have just now, or at any other time.
We expect that it might be useful for you to hear from us at this point each year, because we will have just published our forecast evaluation report, and also perhaps in the weeks or months following the publication of the Scottish budget, which, for the past few years, has been in December.
Thank you very much, Dame Susan. I suspect that we will be keen to have you back a number of times over the year, given the increasing complexity of the area that we are looking at.
One aspect of that complexity is that the Scottish Government has to set a budget that is based on forecasts from the OBR and the Scottish Fiscal Commission. For the 2020-21 budget, which is the one that we are about to look at, the Government will set a budget that is based on projections, and there will then be reconciliations. I apologise because, even having read the briefing paper, I do not have much technical knowledge about how those reconciliations go through block grant adjustments or whatever, but I understand that their impact will not kick in until the 2022-23 budget. There is a time lag, so what the Scottish Government does for the coming budget will not have a direct impact until the 2022-23 budget, when that reconciliation is done. The Government will have to use borrowing powers to manage in-year overspends and will take a consequential hit or benefit in 2022-23.
The example that we have been given is that, if the outturn for disability benefits is 3 per cent more than predicted, there could be an £83 million overspend, which is set against the Parliament’s overall borrowing powers of £300 million. My question—it is the only question that I have today, because I want to let my colleagues in after this—is whether it is realistic to expect the Government to manage those uncertainties within the constraint of borrowing powers of £300 million, given that that 3 per cent error would result in £83 million of the total being used.
That is a good question. We are not fully equipped to say whether the borrowing powers are appropriate, because they are agreed between the Scottish and UK Governments, but you have highlighted the dynamic nature of having devolved social security as well as devolved taxes. The Scottish Government’s borrowing powers would cover not only any shortfall in social security payments in-year—perhaps because more people were eligible and applied, and so forth—but any issue in relation to devolved taxes. Compared to what this country has been used to, that is a completely different approach, and it is dynamic. The Government is cognisant of the issue and will have to be very mindful of it. The Government has the ability to borrow and it has the Scottish reserve, which can be used for capital resources and can hold any spare money. It is possible for forecasts to go either way—we could underforecast or overforecast the spend.
It is complicated. Alasdair Smith might want to add something.
I will emphasise a couple of points. One is that those issues already arise in the area of income tax, where the difference between the Scottish Fiscal Commission’s forecasts of income tax and OBR forecasts of what would have been Scotland’s share of the UK tax take is running at £200 million for the coming year, with the figure likely to be even higher in the future. The income tax reconciliations are large relative to the borrowing powers that the convener has referred to. If we then turn to the £3.5 billion social security budget and assume quite reasonable forecasting errors where nothing has gone wrong with the system—just the standard level of forecasting errors—we are looking at adding to the existing income tax reconciliations further reconciliations or adjustments that are very large relative to the Scottish Government’s borrowing powers. However, as Susan Rice said, it is not for us to make judgments about whether the borrowing powers are right or wrong. We simply observe how the fiscal framework works and alert the Parliament and other bodies to the considerable fiscal risks that are associated with it.
09:45
Therefore, the Fiscal Commission would not have a view on what would be a reasonable divergence on the projections. If we were to say that 3 per cent was a reasonable variation and you were 3 per cent off forecast, that would not be bad going. Across the policy portfolios, where there is risk around divergence, there are financial liabilities on the Government because there are borrowing constraints. Would the Fiscal Commission not take a view on what the borrowing powers should be vis-à-vis the financial exposure that the Government now has with its new powers?
The limits are a matter for the Governments to agree. It is worth flagging that, although the resource borrowing powers have a £300 million limit, the Scottish Government has other tools to manage any differences between our forecast and the amount that is spent. It also has the Scotland reserve, which can be used to manage over and underspends but can also be drawn down for forecast errors. The Scotland reserve has a limit of £250 million, and there is £300 million of resource borrowing.
That is helpful.
My questions are on similar lines to those of the convener. Your letter drew our attention to the reconciliation figure of a reduction of £229 million. That is alarming because it is a large reduction. You are clear in your submission that that means that the Scottish Government might need to adjust its spending plans and that it could be in a large deficit—close to the upper limits of the borrowing powers.
As Alasdair Smith said to the convener, it is not for you to comment on whether the borrowing powers are right or wrong; you are here to comment on the forecasting. However, are you able to say whether the Smith commission principle has been adhered to? The principle is that the budgets should be no larger or smaller
“as a result of the initial transfer of tax and/or spending powers”.
The process that is in place reflects that principle, but it is a multiyear process and therein lies the challenge in any fiscal year. We do reconciliations of OBR forecasts, BGAs and our forecasts. We come up with the numbers and we look at outturn data, which is the data that tells us—in the case of benefits—how many people received a benefit and how much it was. We then compare that to what was forecast. Those elements are there, but they happen over several years or, at least, over a two-year period, whereas the spend must be made in a given year. That is the challenge. It is like managing our household budgets. Sometimes, we have to make a decision: “I won’t spend on this thing that I want, because I have to spend on something else.” There has to be that thought process.
It is important to be clear about the distinction between the reconciliations problem and the general scheme of the fiscal framework, which has devolved responsibility to Scotland. If, as a result of Government policies or other external events, Scottish expenditure on social security rises above the level that is provided for in the fiscal framework—that is, above the level that would have existed had the UK Government been funding it—under the principles of devolution, that bill has to be met from the Scottish Government’s finances. That is the system that the Smith commission set up. It is devolution in practice. When we consider a budget of £3.5 billion, we must be sensitive to the fact that policy changes can have unexpectedly large fiscal consequences.
Reconciliations happen when we—the forecasters—do not catch up with the policy effect fast enough. That is a balancing issue between years that gives rise to other issues, but it is important not to think that reconciliations are the only issue or even the most fundamental one to focus on. The most fundamental—
Yes, but you say in your letter to the committee that, as a consequence of the reconciliation, the Scottish Government
“may ... need to adjust its spending plans.”
I am trying to take all of this in. It is new to us in some respects and we are trying to get our heads around the fiscal framework. You believe that the Smith commission principle has been adhered to; it is just that there might be some years in which things do not catch up. However, you accept that, in those years, the Government may have to reduce its spending plans.
The Scottish Government could well have to adjust its spending plans not because of reconciliations but because the devolved social security budget has grown, which then becomes an issue to be managed within the Scottish Government’s budget. That is the implementation of the Smith commission’s set-up.
An example might be useful, although Claire Mellor and Claire Murdoch can tell me if it is not. If the Scottish Government widens the eligibility in Scotland for a given benefit or does something else to attract more recipients in Scotland by changing an aspect of it so that it will cost more than the same benefit would have cost or continues to cost in England and Wales, the only money that is transferred is the equivalent of what it would have cost or is costing in England and Wales. If the Government feels that that additional spend is important for the people of Scotland, it has to be managed from within the Scottish Government’s budget. Does that make sense?
The committee has latterly understood that point about changing eligibility—that is the policy overspill issue. What part of the £229 million reduction that you have forecast is down to social security, given that, so far, nothing has changed?
The £229 million is the income tax reconciliation for the 2017-18 financial year, and it will be applied to next year’s budget.
I will pick up on those points about the block grant adjustment. I am keen to hear a bit more about the divergence that you have alluded to between Scottish spend on social security and the spend in England and Wales. You have indicated how the system could or should work, but I am interested to know a bit more about what that divergence looks like and how we should account for it in the fiscal framework.
Forgive me, but I will answer the last part of your question first. The fiscal framework is not our design. We implement the aspects of it that relate to the commission. Any changes or thinking about the fiscal framework are matters for the UK and Scottish Governments to sort out. That is important.
What was the first part of the question?
I really just wanted to get more information on what the divergence looks like. Rather than asking you to manage it, I was just asking for a bit more information on what it looks like and how great it is.
In our letter, we refer to what happened with the introduction of the best start grant, which replaced a UK benefit for newborn children. The eligibility rules for the best start grant were a bit different and more generous than those for the UK benefit, and the take-up of the Scottish benefit was much higher than we and the Scottish Government had anticipated. The issue was not just with new recipients; there were people who could claim under the old system or the new system because the birth of their child fell within the windows for both. They quickly found out—through social networks, which worked well in that regard—that it was better to apply for the new Scottish grant rather than the previous UK grant.
The best start grant is a relatively small sum of money relative to the Scottish budget, so that is not a big problem, but it is an illustration of the fact that a not-huge difference in the eligibility rules for a Scottish benefit as compared with the UK benefit can lead to a big proportionate change in expenditure.
There is another aspect of divergence around new benefits, such as the new Scottish child benefit, which will come in next year and will provide a benefit for children who were not eligible for a UK benefit. That is another source of growth in the budget, because it is a new benefit; it is also a source of uncertainty, because it is hard to predict the take-up of a benefit that does not exist in the UK system.
There is a third issue, which is the general one of the Scottish Government’s broad stance on social security benefits, which involves an attempt to get away from a situation in which benefits claimants are stigmatised and instead move to a situation in which people are treated with respect and benefits are viewed as something that people should get. We have had interesting discussions with the Scottish Government about that. It is hard to imagine that, if that change in stance is effective on the ground, it will not have substantial budgetary effects.
Those three elements—eligibility rules for benefits as they get handed over, new benefits that come along and the Scottish Government’s general stance on social security policy—can have budgetary consequences that we all need to be aware of.
I will just add a footnote to that. It is fair to say that we cannot forecast the divergent number. Each benefit is different, and Scotland is taking administrative responsibility for the benefits at different times. The elements that matter to our forecasts for many of the benefits are data on past recipients—it is relevant that we have data that we can use.
On the communications plans, the Scottish Government has already seen the impact of communicating information about various benefits in different ways.
The process designs are important, and can include things such as having people use a plastic card, which is easy to use and means that they are not called out, as it were, when they use it in a supermarket, for example.
All those things will make a difference. It is fair to say that, as new benefits come to Scotland and are managed and administered here, our forecasts are likely to be quite volatile. The OBR has found the same thing. It is just the way that it is. As we get outturn data, the numbers—including the number that you are looking for—will probably shrink over time, although I cannot guarantee that. However, it is impossible to say up front what the amount will be.
Forgive me if my next question sounds like it is slightly outwith your remit, but there is a logic to it. You have described a situation that will be much more dynamic and divergent than the situation at present. Clearly, that has implications for how the public and Parliament scrutinise the process. Are there any implications for us that occur to you with regard to the way in which we scrutinise the process?
I could not begin to tell all of you how to do your jobs. The only thing that I would say is that engaging in the way that we are doing today is important. We want to share our methodology and thinking. By design, the Scottish Fiscal Commission is highly transparent. We publish on our website just about everything that we do. We are extremely keen to be as open as possible in order to have you understand the factors that might cause uncertainty. However, what we cannot do is advise that that uncertainty would diminish if you did X or Y—at least, I could not give you that advice.
I have some questions about how policy spillovers would affect the block grant adjustment. The impact that direct effects of a policy change would have on the block grant adjustment seems clear; what is not so clear is how behavioural effects of policy spillover would affect the block grant adjustment. When it comes to behavioural effects of policy spillover, how do you forecast changes to the block grant adjustment?
10:00
Let me try to parse that question, so that everybody understands.
A spillover is when a benefit is administered in England and Wales and, in Scotland, a similar benefit is administered for a similar purpose but has different features. Perhaps the Scottish Government has added an additional feature or increased eligibility. That might pull individuals into that Scottish benefit, making them eligible for a benefit that comes from the Westminster Government for which they might not otherwise have been eligible. When that happens, it is a matter for the two Governments to discuss and agree whether there is a spillover. There is a role for the Fiscal Commission and the OBR, but only if we are approached and asked, on a technical basis, to cost that. Claire Murdoch can probably speak about that much more effectively than I can.
The block grant adjustments depend on the OBR’s forecasts of spending in England and Wales—that is, what it thinks will be spent on the corresponding benefits in England and Wales. Changes in the OBR’s forecasts depend on how it thinks spending will change in England and Wales. Its forecasts are not affected if we change policy in Scotland; they are affected only by change in England and Wales. If there was a spillover, that would be a matter for the Governments to consider themselves. They would discuss it and decide whether they agreed. If they did not agree, they could raise a dispute. At that point, we would get involved in estimating the potential size of the spillover.
Mark Griffin asked specifically about behavioural differences.
The two Governments have agreed a process for what they call a direct effect. They have not yet agreed a process for a behavioural effect. We do not yet know what that would look like.
Have you been asked for your input on a resolution process for a behavioural effect?
The committee has a deep interest in benefit uptake. Let us say that the Scottish Government does something around uptake of devolved benefits, such as a benefits awareness campaign, and that that leads to a difference in uptake in Scotland of reserved benefits in comparison with England and Wales. If that is identified as a behavioural effect and as a policy spillover, how will the situation be agreed and resolved?
Again, that is a matter for the two Governments to work through. We have not been consulted about how that policy should be shaped. I think that that is what you are asking.
Yes. Thank you.
The committee has looked at this subject before—we are grappling with better understanding it. In relation to behavioural change and financial consequences, is one of the basic principles that, if the Scottish Government were to drive up uptake of reserved benefits in Scotland, the consequential financial impact on the UK budget could lead to discussions between the Scottish and UK Governments, and the Scottish Government might have to financially recompense the UK Government? In previous evidence sessions, we have tried to get our heads around that. Is that part of the behavioural change?
That would be a matter for the two Governments. It would go through the behavioural spillover process, which they have not yet agreed. Potentially, that would be the route for that.
I apologise for repeating myself, but committee members want to be clear. If, next week, the Scottish Government was to successfully encourage more people to claim pension credit, with more people in Scotland claiming that benefit, could there be a financial cost to the Scottish Government even though pension credit is a reserved benefit?
As Claire Murdoch said, that depends on the two Governments agreeing where the financial consequence should fall. If the Scottish Government followed a policy that led to increased take-up of UK-financed benefits in Scotland, it is still not agreed between the two Governments which Government should pay for that.
That is the bit that most of us find absolutely staggering. However, that leads me into the realms of politics rather than scrutiny of the Scottish Government budget.
This is a fascinating area. There is a massive learning curve for the committee and the Parliament in relation to the new responsibilities. I want to pursue two issues. The first is about the information that you need to forecast beyond this year to two, three, four or five years ahead. Dame Susan said in her opening statement that the criteria are key to your forecasting. There has been a delay in the Scottish Government’s introduction of the regulations for the replacement for PIP in Scotland, so we do not know what the criteria will be or what the regulations will look like. Is it almost impossible for you to forecast if you do not know the criteria?
We forecast using the information that we have, with an extensive proviso about the risks to accuracy if we have limited information. For some but not all benefits, we normally need data such as the number of people who currently receive the reserved benefit. We need that data in a timely way that fits with our timetable for our forecasts. We need to know the launch date for new benefits or new features of a benefit from the Scottish Government, and again we need that in a timely way. We need to know how the Government plans to communicate and drum up interest in the benefit. It is helpful to understand the process design and how people can apply. I believe that the Scottish Government intends to make the application process easier for people over time. All those factors will have an impact on how many people receive the benefit.
We make an estimate of what we call the take-up rate, and an estimate of that estimate then goes into our forecast, so those are not hard numbers. However, over time, once a new benefit is in place and we get outturn numbers and see what has actually happened, we adjust our forecasts and our view based on that reality. Over several years, our forecast errors should reduce.
I do not know whether that answers the question.
I am talking about forecasting for the next two or three years. We are unclear about the Scottish Government’s policy on the criteria for the replacement for PIP. We do not know whether the criteria will be different from or similar to the current ones. I presume that forecasting for 2023-24 is very hard without the information on the criteria. To take the extremes, in theory, the new benefit could be awarded to everybody if the criteria were set that way or it could be awarded to nobody if the criteria were very limited. How do you forecast for 2023-24 without knowing that information?
I will ask Claire Mellor to come in, as she is our expert on that.
It is really challenging. Forecasts are an iterative process. If a policy is being introduced in the next financial year, we work with the Scottish Government and we have a protocol with it that enables us to request as much information as we can on the policy. You are right that information will sometimes not be available because some things are so far ahead that the plans are not available for us to look at. In that case, we have to make assumptions. In our publications, we try to highlight the assumptions that we have made and the uncertainties around those. As Susan Rice said, in the past we have made assumptions about start dates when we have not had that information. However, we have highlighted that the spend could be very different if the start dates changed.
That is helpful.
I turn to the second area that I want to ask about. The take-up for the best start grant was twice what you expected it to be. Most of us thought that it would be a fairly straightforward benefit to deliver. I do not want to put words in your mouth, so please correct me if I am oversimplifying what you say in your submission, but it seems that you did not take into account the publicity that the Scottish Government gave to the best start grant, which resulted in more people applying for it more quickly than you expected. Was that because of a lack of information from the Government or was it because an old model, which was perhaps better suited to the UK, had been applied to Scotland?
In future, how confident can we be about the forecast for the new award that will replace attendance allowance, which will come in next year? The take-up for that might be a lot higher than expected. Have you remodelled your forecasting in the light of your experience with the best start grant?
The number of applications for the best start grant on day 1 was pretty much equal to what we would normally see over most of the year, so it was much higher than we anticipated. That was partly because of the six-month window. Women who had had babies six months before the benefit’s start date received extensive communication that it was coming. It looked financially attractive, so many of them waited and applied for it. That is why I have said that communication is such an important element in the process.
Quite honestly, those who administer such benefits are still learning about them, because they are brand new. A lot of learning could be taken away from the best start grant experience and applied to future benefits. That was a big factor in what happened with the grant.
Very quickly—
Your point will have to be very brief, I am afraid.
I want to follow up on what Dame Susan has said, because we will soon have to set budgets for the new benefits that will come in next year. How confident are you in your forecasts?
We are happy with our current level of communication with the Scottish Government. We have open and frank discussions about its plans, insofar as it has finalised plans that it can tell us about. There is no question of its somehow deliberately holding back information that we need—that is not the right way to think about it.
However, we are aware that we have a common problem—for the Government in implementing the policy and for us in forecasting its financial effects—which is that, for the reasons that we have tried to outline, it is a challenging job for us all to do, and we are still learning. The best start grant was a good experience, because we all learned how important it is for us to share as much information as possible and understand as much detail as we can about the Scottish Government’s plans. As is implied in Jeremy Balfour’s question, even quite small details can have big financial consequences. I stress that that is a common challenge for us to face, rather than an oppositional issue.
I want to develop the points about lessons learned and whether different things could be done to mitigate risks. In the light of what you have seen in the past year, what might you do to support improvements in the budget-setting process?
We have talked about the challenges of managing demand-led spend, particularly against a backdrop of tight budgets. Are there other approaches, or is the approach that you have used the only way? For example, could a range of spending, rather than a fixed point, be used, or would that be insufficiently specific? As you can probably tell from my line of questioning, I do not have a background in finance. Is the current approach the only way that it can be done? Could there be a range of spend rather than a fixed target that is either hit or not, or would that complicate matters too much?
10:15
In general, forecasters might show a range. The Bank of England often does that in its economic forecasting, when it has years of data and can look back over a long time period, factor in the data and understand what the range might be. If we did that with something brand new, we would have no evidence or basis from which to create a range.
Might it be doable in future, when data over time exists?
I am not sure. It might be possible for some of the devolved taxes, but the benefits vary. There are categories of people who might be eligible for a given benefit and we will not know until a given year how many people will be eligible.
The other way to look at it is to say that, even if a range is put out there, the Government still needs to have a budget with a fixed number. If the Government takes the middle of the range, is that any different from our forecasts? I am not sure that it would achieve what I think that you are trying to achieve.
I presume that, from your experience to date, there are things that you would want to pick up with Government, such as the timing and flow of information. Are those the most important lessons that have been learned?
Yes—those are the lessons from the early benefits. We have developed a good working relationship with the Scottish Government and Social Security Scotland. I believe that they have come to understand our needs and the timetable for them, which is important. It will be important to keep those conversations going. We have a formal protocol with the Scottish Government, which outlines how and when we will have those conversations and exchanges of information. It is important that that protocol is public and that it is revisited most years and changed and updated. We have a formal process for that communication and we learn.
We also produce an annual statement of data needs. We have just produced this year’s statement, which includes some specifics that would help us with social security forecasting. Claire Murdoch can explain.
It is a largely positive statement of data needs. When we wrote last year’s statement, we were having difficulties with the Department for Work and Pensions and we did not have a memorandum of understanding. We now have that MOU with the DWP and we are getting the information and data. As Social Security Scotland takes on administration and payment of benefits, it will collect lots of good data that we can use to refine and improve our forecasts. This year’s statement of data needs is about the data and information that Social Security Scotland will publish and make available to us. As Susan Rice and Alasdair Smith said, we have good relationships with the agency. Many of the issues are because we are at the beginning and the agency is administering its first benefits, and we know that the situation will keep improving over time.
You talked about a 2.5 per cent differential between the outturn figures and the forecast being relatively acceptable and normal in the world of forecasting. Do you build any tolerance into your forecasts? Is an acceptable tolerance level built into them?
Not as such, in that we do not sit down and say that we will be happy if we are within 5 per cent or something; that is not how it is done. We need to make judgments when we do not have firm evidence-based information, and that will cause variation. We do not set a target for what that variation might be. Have I understood your question?
Yes. I was wondering whether you make assumptions about that when you calculate your figures. For example, if you think that something will cost £2.5 million at outturn but it could potentially vary by 2.5 per cent either way, do you forecast it as £2.5 million plus 2.5 per cent?
No. We come up with a number, because the Scottish Government needs that to create its budget. However, we are explicit about the areas where judgment has been used and the evidence is slight. We point to the potential variability.
So you risk rate the forecasts.
Yes.
We always try to produce central forecasts, which forecast what we think is most likely to occur, but the figures could end up being higher or lower. If you were budgeting, you might think about it differently and say, “I want to make sure that I have enough money, so I will add to that central figure.” However, we are trying to achieve a central forecast.
In all forecasts, there are known unknowns and unknown unknowns. If we forecast something that is established, we might have a reasonably good sense of the amount of risk or variation that there is likely to be in the forecast. That relates to known unknowns. If we forecast what will happen with a new benefit, for example, there will be unknown unknowns, because there will be no established basis on which to judge what a reasonable margin of error would be.
Does your forecasting inform policy, or do you forecast only after policy decisions are made? Those are obviously quite different approaches.
Under our operating rules, we cost stated Government policies—recently, we published a report on how we do that—but we do not make speculative forecasts.
You make the forecasts only after all the decisions are made.
That is right. However, the Scottish Government has resources to look at those questions along the way.
In the lead-up to a Scottish Government budget, a process takes place that involves dialogue between the Government and the commission about the development of Government policies. The development of the Scottish Government’s policies is influenced by the dialogue in the run-up to the budget. We do policy costing on the basis of the Government’s policy plan, and that might lead the Government to adjust its policies.
Right—so there is involvement.
Yes, but, along the route, we are not asked, “Try this one out. Try that one out. Try another one out.”
I understand that.
There is on-going dialogue between the Scottish Government and the commission but, once the Government picks its budget date, it starts to share with the commission some of the details of the policies that might be in the budget, including the cost assumptions, and then the commission does its independent forecasting. I assume that all that work is confidential. The work is independent but private, and the Government will amend or tweak and accept or otherwise your forecasts when it publishes its draft budget.
We do not really need to understand the process; the important question is whether the commission thinks that it gets adequate information, data and communication from the Scottish Government when the Government sets the budget date, before the commission enters into that process. Does the commission get all the data, information and support that it requires to make as informed a decision as possible through what you call the dart arts of forecasting?
Overall, we have a good suite of conversations with a lot of different teams. The process has improved over time. Some of that is down to trust and to people from the Government and the commission understanding that they can share where they are. In those meetings, we can be challenged and asked, “Why are you thinking about this in this way?”, and we can share our thoughts about our judgment. We have good dialogue.
Data is really important to what we do. Of course, data does not come only from the Scottish Government; it comes from the various agencies that collect it. Data will come from Social Security Scotland in due course, and it comes from the DWP. The data on taxes might come from Revenue Scotland or Her Majesty’s Revenue and Customs. We need to develop relationships with all the agencies and data providers so that we get information from them in a timely way and in the best and most useful format for our forecasts. We say to them, “This is what works best for us—please help.” There is an iterative process.
There is a learning curve for the commission and the Scottish Government. Are there no concerns about the process?
There are no concerns. We are not lying awake at night, fretting about the issue. The process has improved over time. Our recently published statement of data needs shows that we have received a lot of the data that we wanted from various agencies, but we are now asking for more. We are perhaps asking for data that is more refined or on a different timeline. We will always look for more. If I answer your question by saying, “Everything is fine,” that does not mean that things will just continue as they are now. We are constantly improving, we hope.
So everything is fine, but you are constantly looking to enhance and improve data to develop the process. However, the main thing is that you are not raising any flags with us about how the process currently works.
I will ask Claire Murdoch to comment, because she is directly involved.
We have a good relationship with the Scottish Government. With the benefits that are administered by the DWP under executive competence, changes that the UK Government makes will affect how much is spent, so we need to have regular conversations with the OBR and the DWP. We are having those conversations, but I flag up that we have slightly less control over that part of the process than we do over the part involving the Scottish Government and Social Security Scotland.
Because of time constraints, perhaps you could write to us after the meeting about any aspects of that part of the process that raise concerns for you, as I am conscious of colleagues who want to get in. However, that is helpful.
It is pretty obvious from the discussion how complex the system is. Susan Rice started by taking a bird’s-eye view. I am just thinking back to Adam Smith’s exposition of how simple taxes should be so that people can understand them. The system seems an awfully long way from that simplicity. We have talked about hypothetical assumptions, financial reconciliations, the fiscal framework and various other constraints. My concern is that the more constraints there are, the less government—in its broader sense—it is possible to get.
I know that you cannot comment on much of the other stuff that I have mentioned, but perhaps the biggest concern is what the public will make of the issue. If the public were trying to work out whether the Scottish Government or the UK Government was acting responsibly or was doing what they expected in relation to social security spending, how could they possibly find their way through the process? What level of public understanding is there of the process? Is it important that the public understands it?
Those are a couple of good questions. As I said, we publish everything that we do so, if any member of the public is interested, they can approach us or look at our website. We periodically hold public events to explain things to people who are interested.
My colleagues may have a different view, but my view is that the public are less interested in the technical activity that sits behind all this than they are in the Government statement that it wants as many people as possible who are eligible for a benefit to receive that benefit and that it is taking steps to try to make that happen. I think that people are interested from their point of view as recipients, and I am not sure how much interest there would be in the mechanics, although Alasdair Smith may have a different view.
As we have discussed this morning, aspects of the system are complicated. I have been on the Fiscal Commission for over two years and I have a little bit of professional expertise in dealing with economics, but I confess that I still stumble over aspects of the block grant adjustment and reconciliation process, as my colleagues will attest.
It would be ambitious to hope that ordinary members of the public might find it easy to understand how the details of the system of devolved finance work. However, as Susan Rice said, that is not the important thing; the important thing is that we are moving to a new arrangement where very large sums of money will be spent by the Scottish Government out of the Scottish Government budget on social security. Those levels of expenditure will be determined in Scotland and the fiscal consequences will fall on Scottish taxpayers. That is not a complicated thing to understand.
We hope that that new arrangement will lead to political dialogue in Scotland between the Government and voters about what levels of social security expenditure Scotland can afford and wants to have. That is the central and crucial issue and that is what we should aim to discuss rather than have a detailed discussion about the more arcane details of reconciliations and so on.
10:30
My point is that the complexity of the system itself might, to some extent, account for the lack of public interest. I do not know who could ever have invented a system like the one with which we are lumbered.
My second question is, again, hypothetical. You have mentioned a number of times the different approach that has been taken to benefits in Scotland, and the larger-than-expected uptake of some of those benefits, which is attributed in part to greater communication and a different ethos about benefits in general.
Dame Susan Rice mentioned at the start of the session that there are three elements that can drive the cost of social security. One of those elements is take-up. Let us say that the different ethos in the benefits landscape in Scotland, brought about through devolved benefits, resulted in an increased consciousness of benefits in general, which had the effect of increasing the take-up of reserved benefits. Am I right in thinking that that would be a problem for the Scottish Government? I think that, under the fiscal framework, how that was dealt with would be a matter for discussion between the two Governments. There is a chance that the Scottish Government would be penalised and end up having to be responsible for the increased take-up.
Before you answer, I would be keen, if the answer is not clear—we have wrestled with the issue a number of times—to have a written explanation, because I cannot quite follow how this works. Basically, if there was an increased take-up of reserved benefits after the cut-off time, would there be a penalty on the Scottish Government? Does that mean that there would be a disincentive for the Scottish Government to preside over an increased take-up of reserved benefits in Scotland?
If you are talking about spillover, I think that we should give you the information in writing. Perhaps Alasdair Smith can say something about that.
The question, as Keith Brown has put it, is absolutely clear. If Scottish Government policy affects the take-up of UK-financed benefits, and there is an increased bill, the question is whether that bill falls on the UK Government, because it is a UK benefit, or on the Scottish Government because, by hypothesis, it would be a change in Scottish Government policy that had led to the increase in take-up of reserved benefits. It is a simple, straightforward clear question to which we do not have the answer, because it is a question for the two Governments to decide between them.
I add only that if the outcome of that discussion were that the Scottish Government had to pay the bill, you could use the word “disincentive” to describe that, or you could simply say that that is what is going to happen with Scottish Government-financed benefits. If the Scottish Government’s policy of increasing take-up and reducing stigma is successful, the Scottish social security bill will increase. You might say that it is up to politicians to decide what language that they want to use, but you could say that that would be a policy success: it is what you wanted to happen.
It is obviously in the Scottish Government’s interest to have the UK Government pay for the increased cost of reserved benefits. Equally, if it is Scottish Government policy to increase the take-up of social security benefits, we have to recognise that somebody has to pay.
For clarity, with regard to my request for written information, I am asking specifically about an indirect action from the Scottish Government. It seems to me that the Scottish Government would have to be very careful that any action that it took had an impact only on devolved benefits. I am happy to get a written explanation.
We will undertake to send you something that will be helpful. That is really a developing area.
Our briefing ahead of today’s meeting was clear but open-ended. It would appear to be the case that, yes, such a situation would be to the detriment of the Scottish Government’s financial position, but that cannot be forced by the UK Government; it would have to be done by agreement. That leaves hanging the $64,000 question: how can the Scottish Government be forced to sign up to that agreement? We, as a committee, need more information on that.
If the Scottish Government made a policy change, I imagine that the UK Government would say that it had to pick up the tab because it had changed the policy in Scotland. However, if a campaign on uptake of reserved benefits was run by the Scottish Government or the third sector, who would pick up the tab? That is not a policy change; it is just encouraging people to take up their benefits.
I am sorry to be repetitive, but we do not have a view on that. It is not our job to have a view on it. It is an important and interesting issue, but it is clearly the responsibility of the two Governments. Although we have been happy to give you whatever benefit you get from our understanding of the issue, it is not for us to pronounce on how that question should be answered.
There is another point about take-up that would be good to discuss. Take-up is really difficult to quantify, because we often do not know the size of the eligible population, so we try to estimate that and then estimate take-up. Quantifying take-up is challenging in itself, but it is even more challenging to disaggregate that to understand the reasons for take-up.
I want to understand the numbers a bit better. Another aspect of the underpinning finances is that—if I understand it correctly—after a benefit has been devolved, an element of that is put into the Scottish block grant, which provides on-going core funding towards it, and there are subsequent block grant adjustments.
For the executive competence of the carers allowance—I do not know whether I am picking the correct benefit as an example—there is a service level agreement between the DWP and the Scottish Government, because, although the DWP pays the carers allowance, the Scottish Government picks up the tab for it. It is a devolved benefit in that respect.
The UK Government’s impact on the squeezing of benefits that are devolved in Scotland is that if they squeeze the same benefits in England, there is a consequential block grant adjustment and Scotland’s block grant goes down.
So, even after benefits are devolved, if the UK Government makes a financial decision about benefits in England that are at the devolved level, there is a consequential decline in the Scottish block grant. Have I understood that correctly?
Pretty much, yes.
The block grant adjustment is what we call the funding that comes to Scotland. It reflects the amount that was spent on carers allowance, to use your example, in the year before it was devolved. It is grown in line with spending in England and Wales, so it is a hypothetical that shows what would have been spent in Scotland had the benefit not been devolved. If, as you say, the UK Government had a policy that cut spending on that benefit in England and Wales, you would see that effect through the block grant adjustment, in that there would be less funding for the Scottish Government.
That brings us back to Alasdair Allan’s question about divergence. As part of the statements that the Scottish Fiscal Commission produces, could you provide the committee with a separate social security statement on block grant adjustments that are based on a divergence with benefits that have already been devolved to Scotland but for which there have been policy changes in the rest of the UK?
Since May, as part of our publications on the budget, we have started to include a fiscal overview chapter in which we try to add clarity to the extremely misty area of block grant adjustments in the fiscal framework. As part of that, we publish the block grant adjustment figures, including in comparison to the spending in Scotland.
We have no responsibility for the block grant adjustments, which are based on OBR forecasts and agreed with the Treasury. However, we can show how Scottish Government policy and our forecasts of Scottish Government spending vary from the block grant adjustments.
Any hypothetical discussions about what block grant adjustments might have looked like under different policies are not for us, unfortunately. We can only look at what the current Scottish Government policy is, what that means for Scottish Government spending and compare it to the block grant adjustments.
I think that I understand that. I apologise, but the committee is grappling with the best way to fit the new framework into the budget scrutiny process.
On the points that Claire Mellor was making, obviously, the better the data, the easier it is to do forecasts with some degree of accuracy. We must have some idea of how many people are eligible for the best-start grant, but I notice in the “Statement of Data Needs—September 2019”, the commission says:
“There are no data sources that reliably show the number of people eligible for social security support. For example, we do not know how many carers in Scotland are caring for someone, meeting the Carer’s Allowance (CA) eligibility criteria, but not taking up support.”
This is, clearly, an important area for the committee, as we are investigating benefit take-up. I know that this is not your specific role, but that question of data is key. Is there anything that we could be doing in Scotland that has not been happening with the DWP and the OBR to make the process more robust?
You are right to say that, in the current process, the information that is available is survey driven. That means that, when we try to work out an eligible population, we use survey information, and the survey questions might not perfectly align with whether someone is eligible to have a certain benefit. That makes it difficult to assess take-up. We are awaiting the Scottish Government’s take-up strategy, which will be announced later in the year. It involves a different approach for Scotland, and will involve plans for take-up.
I suppose that that is one thing that we will have that could make a difference. Is there anything else that we could be doing?
Do you mean is there anything that this committee could do?
I mean that we are relying on surveys, and there are issues around who the survey reaches and how targeted the questions are. Carers and people who have been applying for the best-start grant must be continually engaging with the health service and other services. I just wonder whether we are missing something here.
Jeremy Balfour mentioned PIP earlier. Because of the criteria for that, there is a very involved process for someone who is going through the assessment. Trying to get information about how many people might meet the exact eligibility criteria is challenging. Even if you had the survey questions, it would be difficult to reflect that eligible group in the population. It is especially challenging to work out how to do that in relation to something such as disability benefits.
I think that we have almost finished our questions. We are trying to work out how we use this evidence session as part of our budget scrutiny.
I understand the pitfalls around forecasting, and we are starting to get a glimmer of how challenging that can be. Are the uncertainties that the Scottish Fiscal Commission has to consider understandable uncertainties? For example, if the Scottish Government decided to introduce the child income supplement earlier than initially intended, that would introduce more forecasting uncertainty. I would describe that as an understandable and acceptable uncertainty, because that is what happens when Governments take policy decisions.
However, there could also be uncertainties out there that are not acceptable, if that makes sense. I am trying to work out whether you are content that the uncertainties that you have to grapple with when you are forecasting are reasonable, understandable and acceptable—this is starting to sound like something out of “Yes, Minister”—or whether there are some uncertainties out there in relation to which we need to do more in order to give you the necessary tools to enable you to do the forecasting?
That is a very thoughtful question. A year ago, we might have given you a different answer. We think that we have reasonable streams of information coming to us at reasonable times, and we have reasonable relationships with the relevant bodies that we need to deal with and with the Scottish Government. If we felt that we had a problem that we had not been able to address ourselves, we might talk to you in a future evidence session and say that that was a problem, and it might be that you could be helpful.
That might be a good place to end the discussion. You are independent of Government and of this committee, and we want to find out how we can best fit into the overall process. If we can provide you with any assistance, we are happy to do that. Hopefully, we will build up this relationship. I have a feeling that the committee is just finding its feet with regard to how we operate in that regard.
Would you like to make any comments before we close the evidence session?
10:45
Claire Murdoch has said what I was going to say. As I said in my opening comments, we would be happy to respond to questions. If you write a letter to me at any time in between our appearances before this committee, I will write a letter back to you. It is important that we all mature our understanding of what we are doing.
I thank you for the comment that you have just made because, as you were speaking, it occurred to me that the one word that I have not used today is, “independence”, with a lower case “i”, because we operate—
You can use a capital “I” if you wish—that is not a problem.
No, no—we are not political.
We operate independently and, ultimately, we are responsible to the Parliament. That is why I said that I appreciated your questioning as you try to understand your role and also try to find out whether there is anything that we need. We would certainly come to you if there was.
Thank you. I thank everyone for their time this morning. That ends agenda item 2. We will take agenda item 3, which still concerns our pre-budget scrutiny, in private.
10:46 Meeting continued in private until 11:24.