 Gw etiqu. I welcome to the 20th meeting in 2017 of the Finance and Constitution Committee. I can just remind members as normal to switch off your phones or at least put them in a mode that will interfere with proceedings. As this is a first public meeting of the committee that Alexander Burnett has attended, our first item is a declaration of interests from Alexander and I welcome him very warmly to the committee and invite him to declare any relevant interests. Thank you, convener. I take this opportunity to refer members to my register of interests. Specifically, I declare an interest as an owner and manager of property, including agricultural, residential and commercial lethings, recreational and sporting usage and forestry, as a shareholder in a renewable energy company and as the holder of remunerated positions in companies related to these matters. Thank you, Alexander. The second item on agenda is to decide whether to take item 5 and any future discussion in the work programme on private. Are members agreed? нимra. The third item of our agenda is to take evidence this morning in relation to the Scottish Fiscal Commission forecast evaluation report 2017 and we are joined for this item by Lady Susan Rice, the chair of the Scottish Fiscal Commission, Professor Alistair Smith who is one of the commissioners. David Wilson who is also one of the commissioners on John Ireland, the chief executive. I warm welcome to the meeting this morning and I invite Lady Rice to make a short opening statement. Good morning, convener, and committee, and thank you. In fact, I probably should thank you three times over, first of all, for inviting us to give evidence today, also for engaging with us at various times earlier in the year, especially when the OECD was over and had their conference, which was held here in the Parliament, and in preparing for this session it suddenly dawned on me that we hadn't officially and formally been in front of this committee since the end of last year, so I wanted to thank you for the reprieve. But could I say that it doesn't really make a difference because you are always on our minds, and I gather this is your first session after the recess and we're your first public conversation, but this is also our first appearance as a statutory body of the Fiscal Commission, and I'm joined by the first time by my colleagues whom you've just named. As you know, on April 1, another first this year, we assumed responsibility for independently forecasting Scottish GDP, devolved tax receipts and devolved demand-led social security expenditure. The transition from a non-statutory body scrutinising the Scottish Government's forecasts to becoming an independent non-ministerial department has been a fair piece of work, and we'll leave it at that. But I think you'll want to know that we've agreed a formal protocol with the Scottish Government, which is set out in a framework document, and we've made similar or in the process of finalising similar arrangements with the OBR, HMRC, Revenue Scotland and other bodies. Since April, in addition to induction from my colleagues here, and I joined those sessions as well because I think it never hurts to be reminded. We've also recruited 15 or so analytical staff, and they come from both the Scottish and UK civil service, from academia, from the private sector. So, quite varying backgrounds, their experience includes fiscal forecasting, macroeconomic modelling, housing market analysis and public sector finances, and they are a good team. We're very pleased and we've been getting down to work proper. We published our draft corporate plan yesterday, which you may have seen or possibly not. We hope you will. Last week, of course, we produced our first publication, the forecast evaluation report, which we're here to discuss with you today. This is for 16-17. A paper setting out how we propose to approach our forecast at the time of the draft budget later this year. We also have kicked off a programme of external engagement because it's important that the people who care about what we do know us. That includes yourselves, and we were pleased that a number of you were able to attend the session we held for parliamentarians in June. Over the summer, we also had some experts, forecasting experts from around the UK, engage with our teams, going over the forecast, turning them inside out, commenting on our work that we're doing right now, but also for the future. We'll be meeting informally with a number of other economists next month to take them through our approaches and our ideas. As with the journalists we met earlier in the year, we want to keep our stakeholders as well-informed as we can. In part because it's the right thing to do, in part because that reflects our adherence to the OECD principles for independent fiscal institutions, IFIs, which is what we are. That brings us full circle to this evidence session in which we'll be happy to answer questions about our forecast evaluation report. I'm sure you want to get into the details, so let me just leave you with a couple of what I call keepers, just a couple of high-level thoughts. The first is that forecasting is challenging, but it's also an inexact science. At any point in time there can be a range of valid and reasonable forecasts that could be made. There's no one right forecast that you pull out of the pot. Forecasting therefore typically involves judgment and judgments change over time. As a committee I know you've been grappling with the changing relationship between the UK and the EU, wearing your other hats. This also is an example where the commission will have to make broad judgments about the flight path, if you will, in order to produce our forecast that will come later in the autumn. The second keeper, the second thought, is that forecast benefit increasingly as data accumulate. You can see from the report how the approaches to our devolved taxes developed between 2015 and 2016. We expect that we get more data over time under the new taxes, especially under a newly structured tax such as LBTT, that the models will be further enhanced and developed. One example, additional dwelling supplement ADS, was very challenging to forecast in the first instance because there were no data to start with, so there was the best approximation of what might be the baseline numbers. We now have first full year of data of how many properties fell under that category, and over time we'll have more so that will improve. For us, the insights from the evaluation report have been really helpful as we develop our models for the forecasting that we'll be doing later this year. We hope that you found it helpful as well and we're happy now to try to answer your questions. Thank you for listening. I remember that the whole issue of forecasting became clear in my mind when Robert Chote said it was a bit like spot the ball, but somebody is always moving the ball on the spot the ball picture for those of you who are old enough to remember spot the ball. On that note, your executive summary at paragraph 4 on LBTT states that residential forecast is sensitive to changes in house prices and volume of purchases brings with it overall forecast errors, but I'm not sure that fully explains the forecast changes between December 15 and December 16 in table 1. It would be helpful if you could explain to the committee the main reason why the Scottish Government forecast, if you can, for residential LBTT for 2016-17 was £282 million in December 15, and then £181 million in December 16 is indicated, as I said in that table 1 in your executive summary, just so that I can fully understand it would be helpful. If I can have a go at that, the main difference between the two Scottish Government forecasts is the basis on which they forecast house prices. In the earlier forecast period, as I think it says somewhere in the report, the Government forecasts were looking at a rather long period of experience with house prices, including house prices, before the global financial crisis of 2008 and expecting that the economy and the housing market was going to revert to historical patterns in due course. By 2015, I think we all understood that the effects of the global financial crisis on the housing market and elsewhere were pretty long lasting and that it was no longer reasonable to suppose that the housing market was going to go back to its pre-2008 period. So there's quite a substantial change in view of what data about the housing market was relevant for the forecast, and that's the big change that took place between the two Government forecasts. In the later forecast, no attention was given to prices before 2008 because that was regarded rightly in our view as ancient history by that point. That sets the scene. Ash, I think that you used some questions around LBTT as well, so we'll just stick on the LBT issue at the moment for us. Yes, I think that some of my questions have been partially covered by that answer, but I'm just wondering, can I get a bit of clarity on the log normal distribution model? In your report, it suggests that the model did fit the data, but then it goes on to say that the forecast error in the top two tax brackets, that's the highest two, was partially due to the forecast error in the top two tax brackets, that's the highest two. Was partially due to the fit of the log normal distribution. Could you explain that for me a little? Yes, I'll try. There's nothing magic about the log normal distribution, it's just a mathematical way of describing certain kinds of distributions of which house prices is one, or indeed house transactions is one. And it happens to work pretty well in the Scottish housing market as another housing market, except at the very top of the distribution. So it isn't that there's something extraordinary happening up there, it happens to be that this little piece of mathematical kit doesn't fit so well up at the top and it's necessary to make an adjustment. And the adjustment that was needed was slightly larger in the second period than in the first period. But it's just a statistical way of describing the numbers, which proves to be very useful. And making the adjustment makes it more accurate, it's not perfectly accurate, but it's a very good walking tool that helps the forecasters make their forecasts. So, as a tool, this would be the tool that you would use to make these forecasts. You've got no plans to move to, is there anything better out there or is this kind of the best practice? Our current plans are to, that the adjusted log normal will still play a central role in the forecast, but we're looking to develop the tools further. For example, and maybe this is something we've come to by looking in greater depth at behavioural effects. But the log normal approach we envisage will still play a central role. Okay, thank you. Alexander, I think you had a question about LBTT in particular in the northeast. Do you want to deal with that now? Thank you. As you said, one of the components of the errors is the number of property transactions, and I think the error accounted for 6 million last year, which was a 23 million swing on the errors of the previous year. I was wondering how, what breakdown of figures of a number of transactions there were for geographical regions, and whether those were in line with forecasts across Scotland. And specifically, if I'm understanding the figures report right, in 234 are you saying, talking about the northeast and the Aberdeen housing market, are you saying it was in line with forecasts? And if so, I wonder where that comes from. So let me give you the first part of the answer and turn to one of my colleagues to follow up. Last year, in our previous guys' commission, in having challenged meetings with the Scottish Government forecasters, we saw variation in the numbers that were coming. This is all work in progress, probably summer of 2016. And we speculated maybe there's an impact because of the economic constraints that were happening in the northeast and suggested that that be looked at specifically. We don't do regional analysis or sort of 32 local authorities or anything of that sort, but we felt that that was a significant area to explore, and that's why that was teased out. But actually what we found was that house price growth was, you know, there were not noticeable impacts from the what happened in the northeast, at least at this point in time. Any further comments? Paragraph 234 and 235 in the report said what we did. And it is the case that house prices in Aberdeen and Aberdeenshire have had lower growth than in the rest of Scotland. And in figure 2.8, it shows the effects in 2015 and 2016 forecasts of Aberdeen and Aberdeenshire having been different of the way it does it. Had Aberdeen and Aberdeenshire followed the Scottish average, there would have been this much more LBTT revenue. And as is set out in Paragraph 235, the differences, there would have been 2.5% and 7% more revenue in 2015 and 2016 had house prices in Aberdeen and Aberdeenshire risen at the Scottish average. So there is an effect. Okay, thank you if I can focus on the transactions element. So you're saying that there's no link, but when you're doing the modelling, you're not collecting data from the 32 councils. How do you predict if you're not taking planning data and consents and all the rest of it, how are you predicting how many transactions you think should take place? That's done from data covering the whole of Scotland rather than broken down by reading. Potential transactions come through the planning process? No, we have historical data on the number of transactions and we use that to predict the path of transactions in the future. So it's a Scottish wide transaction prediction equation which is used there. Rather than very fine detailed information about planning consents and things like that because that would just be an awful lot of work and in comparison with this more sort of standard forecasting approach of using historical transaction data. So that's the way you're saying there's no link between the model and what's physically being built on the ground? No, there is a link. There's a very definitive link because what's being built on the ground feeds into that transactions data. But remember this is transactions data for the whole market, not just new build. So those links are there but relatively weak. Thank you. Okay, I think you're on the same theme or at least on LBTT, am I right? Yes, you're on tying up the forecast. I mean, thanks, for coming along. I'm just looking at your charts where you're bridging between the forecast and the tax raised in particular 2.3 for December 2015 forecast for 2016-17. ..on lo charwn y r emanau iawn i eu teuluau i elioedd ar glimorthiaeth llwyr, a fe allan o drwsllyf o'r rhesymu a'i'r ddest Charliell a'i drafodaeth y maradw i ein darllen o Unae gyda'r cyfnodr hwn mae'n gyntaf i'r prifyffael ar y prif alterio y cyfnodd yn dda. Rwy oedd oedd oeddeithas wneud, bei'r tyfu ar y rhai osgym珍. Rwy oedd wedi gwneud o'r gweithio cyffredinol, o'r ffordd ymgweithinol, o'r ffordd cyffredinol o'r dweud cyffredinol, oes unig sefydliad i'r llyfr oherwydd gyda'r rôl ymlaen i amddangos ar hyn yn golygu lawer o'r fudio hyn, i gyda'r rôl fawr谢谢a mor hynna. The behaviour effects include longer run effects as well as for stalling. We will continue to look at both. In some ways, for stalling is easier to look at because for stalling felly mae'r dyniad ar y ddau sy'n digwydd ar y cwrs o'r ffieinidol ar y ddau a'r ddau a'r gweithio ar y llunio'r ddau. Mae'r ddau'r gweithio o'r ffieinidol yn ddigwydd ar gyfer gyda'r ysgriffig, lle mae'n ddau'r ddau'r ddau yn ei ddweud o'r ddau, that that in this particular forecast, that 13 million represents a behavioural adjustment that the government made to its forecasts so it's not an estimate of the behavioural effects by the Commission, it's the adjustment that the Government made when he was producing its forecasts? Okay, okay. Thank you. Right, LFT questions. I mean, I'll come back to, well, Lily before I move on to murder on the non-domestics Lleidwyrd i—LBTT? Yr adeiladau hyn oherwydd LBTT yn cwestiynau unrhyw ymlaen nhw'n gweithio gyd, ac yn cael ei wneud o'r adeiladau ymlaen nhw, ac mae'n gweithio'r adeiladau gan cael ei wneud o'r adeiladau ymlaen nhw, ac yn gyfnodd y prif, y prif yn cael ei wneud o'r adeiladau ymlaen nhw'n gweithio'r adeiladau. Felly, mae'n gwneud o'r adeiladau ac mae'n gwneud o'r adeiladau y fawr meddwl y trafnodd yn ymddangos i gyd yn ddod. Rwy'n fawr hynny yw'r dda, mae'n fawr hwn yn ddod. Mae'n fawr oedd ymddangos i'r newid ychydigol yng Nghymru yn ddod. Mae'r ddysgu'r ystafell sydd ymddangos i'r problemau ond y rhan o'r rhan o'r LBTT yw'r ddesgol hon i'r gael i gael eu cyfrifolol sy'n ddysgu ei ddweud yn ffath o'r pethau. Mae erbyn yn oed, mae'r drwy o'n gwirio'r syniadau strategiol yw'r dyfodol yma i'r gweld yma i'r gweld. Ond ydych yn gofod i medion wedi'i modd i'r ddweud? lle i chi nhw'n… Roedd yna'r amser i ddim yn hawdd diolch maen nhw mewn at ystafell hynny fel ymdweithio'r bair fforddau cyflog yma. Ond yng Nghymru yw'r awtgyn o'r ffordd a'r hyn o'r gwmhau fel cymdeithasol yn eich peth a'r cymdeithasol yn eich peth a rywbeth y gallai'n leirio gyda'u digwyddau. Ond g Pol engellwch i'w gwneud mightwyr o ffordd wedi'i wneud o'r hyn o gwneud Acred is forming a good sense of how much uncertainty there is. That is an important part of forecasting. It's as important to tell the user of forecasts how much confidence you have in your central forecast as it is to tell them what is your central forecast. In this area, necessarily however good you are at focusing on whatever fancy tricks you use, Mae yw'r ysgrifennu oedd ymddangos yn ddiweddol, dyma o'r byw ymddangos yn ddigonol. Ddod. Felly, mae'n ddigonol o'r LBT. Felly, os ydych yn y cyfnod, gyda ni'n ddynnu rhywbeth ffordd. Fynyrddol. Ac mae'n ddigonol o'r ddynnu rhywbeth. Felly, mae'n ffordd i'r cyfnod i'r ysgrifennu in total rateable value for 2016-17. The forecast growth was 1 per cent, whereas the out turn was 0.33 per cent. I appreciate in cash terms that relatively doesn't make a huge sum, it's about £10 million difference. Within percentage terms, it's quite a major gap, it's a third of the forecast. Can you shed any light on why the out turn is so much lower than the forecast? To cover that, this gets to the question of the buoyancy of the change in the increase in non-domestic rates from year to year. We have to make an assumption about that, and the 1 per cent figure that we're evaluating against was the assumption that the government made, rather than the fiscal commission made. Last year, that seemed to us a reasonable assessment based on what we know about the trends in buoyancy. However, I would quickly say that we feel and the Government statisticians feel that they are trying to provide a statistical estimate of buoyancy based on some form of economic determinants is an extremely difficult thing to do. In fact, increasingly there's a view that in the past the rate of change of non-domestic rate income can somehow reflect economic circumstances or economic determinants. I think looking at the data then, while in one sense the economy must have some impact on this, but it seems to be so far removed from the year on year changes that you end up in a situation where the estimate of buoyancy is actually a residual rather than something that's based on economic determinants. By definition, perhaps similar to the previous issue on LBTT, it's very difficult to develop statistical measures to forecast something like this when it's actually just something that falls out of a whole series of factors that are built into the system around appeals and changes in the overall system. I think that the main message was it was a reasonable estimate to make at the time on something that is an extremely difficult thing to actually forecast. As you rightly say, it has made some difference to the outturn, but we're not reading too much into it as this reflects a need for a substantial change to how we make these forecasts in the future. It sounds, Mr Wilson, from what you're saying, that this is a very difficult area to forecast. So, are you just guessing really where the figures will end up? The term just guessing is perhaps pejorative, in that judgment needs to be made based on the best possible evidence that you can bring to bear. I think what we are very keen to get across to you that these are difficult judgments to make, but we recognise that the commission going forward, part of the reason for setting up the commission, was that the overall budget process requires somebody to make a decision on the basis of the best estimate we've got, however difficult it is, and that's the role that we need to play. Just one follow-up question, if that's okay just for clarity, because there is mention in your report that one reason why the forecast level of growth could have been lower than the estimate was that there were several significant removals from the valuation role. What was this relate to? Were these properties being demolished? Does that explain that? Again, just to draw the comparison with LBTT instead of large tracts actions, without going into the detail of its individual cases, there are large properties, large facilities that are coming, that are leaving, are no longer in use, and there are new buildings that are being replaced. I'm not sure if I can quite see it, but the St James Centre is an example of a large property that is changing its status and its payments, and there's a number of these developments both leaving and entering into the register. One issue of one area that we could look at is a more detailed listing of major properties that are moving in and out of the system, and tracking that may help to give an indication of year-on-year changes. That's one area that we've been looking at, but inevitably the timing of the precise changes given reviews, appeals, and construction timing has been that. It's a fairly demanding task to undertake. Can I just add a tiny point, just to put it in very simple terms? There is what is happening on the ground to buildings, as David has just explained, but over this period there was also a court decision that moved properties into different categories, so there's an administrative impact here as well, which also makes it rather complicated. Thank you. Nobody has some non-domestic rates. In that case, we'll just go to a more general forecasting issue and the length between this and the fiscal framework. I think it was Marie. We were talking at the beginning about the, it isn't particularly the differential between the forecast and the outturn. That's significant. It is maybe in the difference between the outturn and the block grant adjustment. Given the complexity that we have as a devolved nation, there is an added layer of complication. I just wondered if you would be able to give us a little bit more explanation on how your forecasting fits back in with the fiscal framework. Sorry. I'm just trying to pinpoint what you're looking for. How does our forecasting affect the block grant negotiation or outcome? It feeds back into the, relates back to the fiscal framework, so there is the amount of money, I'm not getting this right at all, the amount of money that you forecast and then there's the amount of money that we collect and there's the amount of money that they forecast in the UK government and there's the amount of money that they collect and then there's an adjustment that happens between the two governments and I just wondered if you could give us a little bit more, sorry, to be so simplistic, a little bit more explanation around that. I think you've got it. So the block grant adjustment depends upon the difference between the Scottish revenue raised and the rest of the UK revenue raised. We forecast the Scottish component of that. The OBR, by default, forecasts the rest of the UK part of that and those two things will give you the block grant adjustment which is the government's additional money. So I think you've got exactly. In that case, move on to readiness of the organisation and more general terms, James. Thanks a lot convener and good morning panel. Lady Rice has spoken in your opening statement about readiness for producing your first official forecast later in the year. And in terms of being able to do that, you know, as I see it, there are three strands, you know, there's a resource you need to do that and you outlined the 15 staff that you've taken on. There are the models that you use and other methodologies that underpin those models. Can you maybe just give a bit of an overview of how robust you feel all that is as you move towards the production of your first official forecast? I'll give you perhaps just an overview comment and then turn to John, who's been so key in pulling so much of this together over recent months as the chief executive. So, as I said, I think we have a really good team. We still have a couple more people due to join us. So, you know, it is forming still, but they're working well together. They really do bounce off each other well and we're pleased with the people we have. We do think that we have the resources we need for our remit and our responsibilities this year. If we find in a few years' time that the remit grows, we will obviously need to have more resource. We don't have anybody's spare at all at the moment. So, we do think we have the people and we have helped us develop our own views as commissioners because ultimately we are responsible for the forecasts that are produced in December by bringing in the experts that I mentioned in my opening comments. Their job was just to bring another independent view. They were doing this at the behest of ourselves by going into the forecasts and then they reported back to us as commissioners and to members of our senior team. Overall, with lots of thoughts and suggestions and interesting points, but overall, in all cases, they felt that we were on track. So, we take some comfort from that in addition to our own personal judgments that would be on track. In terms of the models used, in essence, and we have the report called our current approach to forecasting which summarizes some of this, we're building on models for the taxes that already existed in a devolved form. We're building on those models. We did judge these to be reasonable with some challenges in the past few years and we think they're a good place to start. That doesn't mean that they will be the same forever because what we're doing is enhancing and changing over time. For some of the new devolved taxes, so a lot of work is being done on air departure tax, that we have to build ourselves, obviously. We're looking at social security expenditure for those categories that we're responsible for in the first instance and we're building that new, obviously. John, do you want to say anything more about our resources overall? Susan's probably covered the resources that we have access to and also the evolution of our models from those that we inherited from the government and those that we built in-house. Perhaps I could just add a little bit about quality assurance. We have three levels of quality assurance of the modelling work that we've been doing over the course of the past year. First of all, there's an internal challenge. The commissioners have been working very closely with the staff to go through the models, so we have that internal challenge process. We've been working very closely with the government analysts, so we've been taking back the models that we took from them and talking through our changes and explaining them and getting their feedback on those changes. We've been working with external people such as David Eisner at the Fraser Under Institute on things like our income tax model. The other thing that we've done, as Susan has said in some detail, we employed three academics from the south to come up and give us over the course of three or four days to give us their insights into the modelling that we've been doing. We've gone through that process of quality assurance and we're reasonably confident that, at the moment, we're in a good starting place for our forecasting. Just in terms of that quality assurance check, is that in terms of, if you like, the actual process of the model? Have you done any testing based on numbers? Yes. The experts and our quality assurance have been looked at both at the ability of the models to forecast and also their methodological approach. Is this a state of the art approach or are there alternatives? We've assessed that, but for sure we've been looking at how well the models perform. The forecast evaluation report we've just been discussing is an important part of that. We have internally been doing very similar exercises. A new area of forecasting for us is forecasting the macroeconomy. We've constructed models and we've also been using the NISA SGM model, which was built for the Scottish Government as a framework. Over the summer we've been producing dummy macro forecasts. We're now, I think, in about our third round of internal forecasting using that sort of framework. So yes, we've been producing numbers and running our eyes over them. So it's an iterative process, in the sense that you've taken soundings from experts on the methodology, you've run actual tests or, in some cases, live data that you'll have available through it and you continue to check that through. Thank you, convener. I just wanted to ask a follow-up on one aspect of methodology, which takes us a little bit beyond the evaluation report. So it might be that you're not prepared for this and you want to come back to us in writing, and if that's the case, that's absolutely fine. But, ladies and gentlemen, you've mentioned a couple of times this morning, both in your opening remarks and in your answer to Mr Kelly just there, that the Fiscal Commission is taking on new responsibilities, has already taken on new responsibilities with regard to making official forecasts for spending in devolved social security. I wanted to ask what kind of methodological challenges you've had in terms of developing models for that and how you're overcoming them. Do you want to pick up on social security or... 80? Let me take air passenger, Judy. Social security will be responsible in the first instance for about 2.5 billion, 2.8 billion, about 40 billion spent. So it's a few programs within social security. And part of what you have to do to begin with is to understand how those programs will work, how we anticipate or we hear that they may change as they're devolved and administered from Scotland. So I'll put that aside just for the moment though because air departure tax or passenger, Judy, as it might have been called, is an interesting space, understanding what numbers are included and what is the base number of passengers departing, what are the categories. It's just literally starting at the beginning. So some of the information on passenger numbers relate to surveys that are done, you may have been stopped in an airport at some point. So it's where are you going? And if you say London, you go into the counter, if you say you're on route to somewhere else, you don't. It's understanding just what the baseline numbers are just as with the additional dwelling supplement, the first year we had to have proxies for understanding that. So I'm not sure if that's what your question is, but we have to start there. I guess at this stage you don't know what the baseline numbers are in the social security field, I guess that's what you're telling us there. And John, I can see you beavering away trying to find an answer there, but I think it's better if you go and reflect on that and just let us know about what your early preparation is to begin dealing with these. I can point to where that information is now if that's helpful. So the paper that Susan referred to that we published, the current approach to forecasting and the method paper, that has a section 8 which outlines where we are on social security modelling. The government has been working on this for some time, we've been working pretty closely with the government. The government has access to DWP data which gives us a fair amount of historical data. There are, I think, particularly difficult methodological issues when it comes to things like take-up, and I think that's where our judgment and our work is going to be really focused, that you can get reasonable data on health characteristics of the population, not perfect but reasonable, but there are, if the aim in the sense of the Scottish approach to social security is to operate a kinder system, to reflect some of the language in the programme for government, and how you actually model that in terms of hard take-up data is quite a challenge, I think, and I think that's where our judgment is going to be most needed. Okay, we have signed post-arrest if we have the materials available at this stage. Very brief point, which is, I think, hopefully one of the things that has become clear is that we take a different approach to forecasting the different areas of our responsibilities, depending on the needs of that particular approach. I mean, for example, the developing work on forecasting onshore GDP is very much a combination of statistical models and overall judgment about the development of the economy, which both requires an expertise in modelling and data. Also quite a significant amount of engagement with external commentators with people who are business representatives and others who are also thinking very actively about these issues, and that will all feed into the work that we're doing. So it's not all just about the modelling, it's also about external engagement on that one through a variety of different processes. Just to pick up the work on social security, much of that is about engagement with the teams who are implementing the new approaches that the Scottish Government is going to take, and developing a very specific set of expertise and understanding about the possible expenditure. So it's partly the development of statistical models to do that, but a big part of our work at the moment is active engagement with the teams who are developing and implementing those systems so that we can work with them to make sure that there's an understanding of expenditure. So I just wanted to clarify, it's not all just about having a single model that will give us all the answers. There's a significant amount of engagement and working collaboratively with various organisations in order to produce the estimates. That's helpful, thanks. I've got one final question before I go there. Thank you. Good morning. I wanted to ask a question not about the doing of the work, which is extremely complicated for most people, including ourselves, to understand, but about the communicating of it. We're all very familiar, for example, with the political impact of something like the GERS publication, which whatever you think about why it is, what it is, it tends to further the polarization about the stories that we tell each other about the Scottish economy, rather than really shedding light on things. I'm hopeful that the publication of your forecasting doesn't become a similarly polarising political event in the year, and also that the publication of the block grant adjustment doesn't become just an opportunity for us all as politicians to say, well, that proves exactly what we've been saying all along. That's not a very helpful dynamic generally. We're all guilty of that, but there's an opportunity for you, as an independent body, to become more of a source of publicly accessible authoritative information. Most people, to be fair, are not going to have hours to spend reading and wrapping their heads around the detailed reports that you publish. So I'm wondering about how you intend to go about communicating in more accessible ways the key findings that are coming out of your reports or, indeed, whether you're going to engage with moments like the confirmation of what the block grant adjustment will be and whether you see part of your role as informing the public about how those things have come about and what they really mean. Your social media account, for example, is active, but it mostly includes just links to your main publications rather than anything that's perhaps more digestible. We have a social media account now, so we're moving on in that score. Let me just say that the issue you raise is a really important one and one that we're conscious of, we care about, and we speak about amongst ourselves a lot. I'll give you a couple of answers, but look to my colleagues as well. Last year in the report we issued alongside the draft budget, so those were not our forecasts, but we commented on them. We did put an executive summary in the front, and the purpose of that was when we read the content of the report, we thought, gosh, there will be a lot of people who won't find this really accessible. Let's try to articulate what we're talking about with those few small number of devolved taxes in a way that might be more accessible. We think that's important. We intend to use our website as well to clarify and to speak in easy language, if you will. We also have those who are highly technical who want the technical side, so we need to do both. In terms of commenting, we will be able to comment about the work that we're responsible to do because we understand that and we can comment on that. We will not be commenting on other aspects of the fiscal infrastructure if we are not expert in that space, so we won't be making commentators in that sense about other matters that are outside our purview. We wouldn't be expert at that, and I don't think that would be helpful. Anything about communications, John? You think a lot about that. On top of the structure of the Executive Summary and the main reports trying to give the Executive Summary very non-technical, I think there's enormous value in charts. I think today's questions have illustrated particularly those LGBT decomposition charts that some carefully chosen charts can do an awful lot, so we are spending quite a lot of time thinking about the charts that we produce and how we use those in social media as well. No, I think that's helpful. Last year, I think some of these are from December last year, you produced a series of infographics. Charts would be one stage beyond that simplistic presentation of a single statistic, but I think there'd be great value in ensuring that the commission is seen as a source, not just of detailed information for government and for academics and for people who want to explore that detail, but also a source of authoritative clear information for people who might only be reading the headlines. We endorse that view. David, did you want to add a comment? I'd very much want to say a couple of things. We're not very much welcome what you're saying in terms of the role of the commission. To build on what Susan has said, I think we're very conscious that there are some very major issues which will underpin the forecast that we'll make of which there's significant political interest. For example, on page 12 of the report, there's a graph about the current position on productivity trends, which it's often said in terms of both growth of incomes, wage income in people's pockets, and also government fiscal revenues. It's not all about productivity, but most of it actually is. Issues about the fundamentals of what's happening in the economy around productivity, about behavioural responses to any tax changes, about how the flight path, as Susan put it earlier, of how things will develop, we recognise and very much want to take on the role of communicating some of the major issues that they underpin, but perhaps I do want to add a slight caveat in terms of the wider issues about the block grant adjustment. Our role we see very much as principally about assessing the income side of the balance sheet in terms of where the incomes that the Scottish Government will receive from the developed taxes, rather than the wider set of questions about the fiscal framework, which is very much for the Governments respectively to sort out in themselves. We play a key part of that overall framework, but it's perhaps for others to be developing the more detailed understanding of the precise minutiae of how the block grant adjustment will be taken forward rather than us. That's helpful. Thank you very much. Thank you very much. That actually goes into some of the area that I was going to cover myself, so I think that all that remains to say is thank you very much for coming along this morning. It was very helpful. It's thrown us not on a light on your forecasting figures, but you're beginning to always understand some of the journey you're on as well, so we're very grateful for your attendance this morning. I now suspend the meeting to allow changelor of witnesses. Hello. The next item on our agenda is to take evidence from the Scottish current bill team on Social Security Scotland Bill's financial memorandum. I welcome to the meeting Chris Boyle and the legislation delivery team leader, Chris Stevens, who is the senior finance business parter. David, now forgive me if I don't get this right, David. Seniori o signori? Seniorini. I didn't get it right either term, so thank you for helping out. James Wallace, who's the head of finance and social security division, I think all members have received copies of the written submissions received by the committee along with a briefing from Spice and a paper from the clerks, so I think we'll just go straight to the questions. I thank you for coming along this morning to give us evidence as part of the financial memorandum on the social security bill. The written submissions, which I've already talked about, highlighted some of the demand-led nature of the majority of devolved security benefits, and that will bring with it uncertainty. And a new budget risk, which the Scottish Government will have to manage, but specifically there are also new risks associated with the block grant adjustment mechanism, as each benefit is devolved. I just wondered how the Scottish Government intends to manage those budgetary risks and ensure that it can meet the costs of providing those social security benefits when devolved to Scotland, particularly in regard to the block grant adjustment process. I know that Ash Denham has got a question on the wider range of risks, so if you could concentrate on the beginning on the block grant adjustment risks, you'd be most grateful. We want to take that on. James, on you go. So there is a risk created as a result of the block grant adjustment. The fiscal framework agreement between the UK and the Scottish Government sets out the arrangements by which the block grant adjustment will operate. So the block grant is essentially based on expenditure in Scotland in the year prior to devolution, so if we were to develop the Volvo benefit in 1819, it would be based on DWP's forecast from 1718. That adjustment is made as an initial adjustment to the block grant for the Scottish Government as a whole, and then it would be reconciled in line with the fiscal framework. The technical annex for the fiscal framework sets out that that reconciliation can be done in year. So if there were variations down south from forecast when expenditure was tracking against a different trajectory to the forecast, then that adjustment could be made through the autumn budget down south in the rest of the UK and adjusted through supplementary estimates in year. It does create a risk, however, in terms of cash management for the Scottish Government in year. Should the forecasts track away from expenditure, we may have a requirement in year to pay more in cash than is being transferred through the budget, and then those two would converge together later in the year. The method by which we attempt to manage that risk—I will perhaps comment on that and then ask Evan to comment further on it—is around our forecasts. There are a number of analysts working in the Scottish Government preparing detailed forecasts of likely expenditure on benefits, which will be used to inform our view on whether the block grant adjustment is appropriate. The Scottish Fiscal Commission will prepare forecasts and the OBR will prepare forecasts on what the block grant adjustment is based on. It is a new risk. It is something that we are very live to, and we are putting in place arrangements to manage. I do not know if Kevin would care to add. Do not worry about it. Sorry, right. We are building procedures and processes in different ways. There are three main areas that we are looking at. In the social security directorate, we have a finance hub that will be staffed by appropriately and professionally qualified people to manage the finances there. Central Finance will be working closely with the Social Security Finance Hub. We have good working relationships there. We are developing good working relationships with Communities Analysis Division so that we can make sure that we understand the internal Scottish Government forecasts and that we embed new processes into the business as usual financial management of the Scottish Government. We have good working relationships with HM Treasury as well, and, indeed, clearly there will be links to the work that the Scottish Fiscal Commission will do. It is clearly a complex area, but we are looking at setting up robust, transparent processes where all those areas that I have outlined can all work together productively. Thanks, James. Do you have a bit more? We will have a risk to manage, and then there will be a reconciliation process at some stage. Can you just talk us through about that? If there is either an up or down for the Scottish Government, it will need to work its way through the system and it will come to date when all that will be reconciled. I do not know who wants to take that on. I will again. Essentially, the reconciliation process is described in the Fiscal Framework and its technical annex. It was agreed or recognised between the UK Government and the Scottish Government when the fiscal framework was negotiated that there was a risk of variations. I should say that it is variations for a particular reason. It is where expenditure is deviating from the forecast. It is not, as a result of policy changes, demographic changes, or divergence between the way expenditure grows in Scotland and expenditure grows in the rest of the UK. That will need to be accounted for differently. It is particularly forecast error as a result of the way the initial block grant adjustment is calculated, because it is calculated for the year preceding devolution. We know straight away that there will be an indexation adjustment to account for the movement in a year forward, but we are also aware that the nature of the expenditure as demand-led means that we are not working with an expenditure limit. We are working with expenditure based on demand. It will be what it is based on the factors that affect demand. We will require to make sure that we are adjusting the block grant accordingly. Linked to that, the Fiscal Framework describes a number of new cash management and resource borrowing powers, in particular for cash management and forecast error that allow the Scottish Government to draw on those resource borrowing powers along with the Scotland reserve, to ensure that we can manage the cash flow of the Scottish Government prior to the in-year adjustment, if we choose an in-year adjustment, but prior to that in-year adjustment taking place through the supplementary estimate process down south. On the last item, I had questions specifically about the demand-led risk slash. Yes, so you have just mentioned in your answer there, James, about the factors that will affect the demand. Clearly, there are going to be risks to the public finances that are created by managing a demand-led expenditure on this type of scale. Could you maybe outline for us what you see those risks as being? The main risks, I would say, are policy and demographic. I think that the financial memorandum goes into it in detail. The way in which population changes in Scotland and ageing population could encourage a greater growth in some of the devolved benefits over the rest of the UK. I think that the Scottish Minister has a desire to change policy in some areas. Some of those are outlined in the financial memorandum. That will cause divergence over time from the UK Government's position. Those differences will not be accounted for through the block grant adjustment. It will be for the Scottish Government to make up any shortfall through the Scottish Government's own resources through the spending review process and through the annual budget process. I should add for balance that there is also the possibility of a variance downwards that it is not necessarily the case that expenditure will rise above comparable UK spending. It could also fall below UK spending depending again on policy and demographic differences, which would release money into the Scottish Government budget effectively. You have mentioned there things like ageing population. Obviously, we can, to a degree, predict that, what that might be in the future, and policy change, you would know before you were going to change it what that likely impact of that would have. Is there anything in there that you would see as being less predictable that maybe we would not be able to predict in advance? There are difficulties in prediction. There is always an inherent uncertainty in forecasting. That is one thing that our analysts colleagues are working through to understand where those variances are going to occur. How can we improve our modelling to ensure that we are in the best position to forecast expenditure? There will be occasions where it is difficult to forecast, particularly some of the areas around uptake. Uptake on benefits, there are incomplete datasets. I would describe them as and where we have incomplete datasets, that obviously affects the validity of the forecasting that we are able to do. However, our analysts colleagues are well aware of the limitations, and are working on new models to enable them to gather the data or to get around the data limitations. I may see if Kevin is got anything to add there on forecasting. As the social security system in Scotland diverges from the UK level, assessing the impact of changes to the Scottish system in terms of take-up will be a new area and an inherent source of uncertainty. That is why the analysts are developing their capability to produce models to forecast the effect of those changes. Acknowledging that the current evidence base of take-up rates in the existing system is limited, as James says. A bit more on that line, on demographic profile. I am understanding if I am wrong, but the way that the BGA, certainly on the taxide debates, works, and I am assuming that it is the same here, is as an adjustment made to take account of overall population. The issue that you are talking about in demographics is the profile of the population within that overall population number, and specifically around the ageing population, or conversely the lack of working age population. Is there a risk there in terms of where we are with population growth, specifically on immigration, that could impact that, and also potentially an opportunity there, in terms of if Scotland had a differential immigration policy opposed Brexit that would allow us to grow that working age population, that would give us an advantage potentially significant in terms of the way that the BGA is calculated, just looking at the way that the numbers are calculated, obviously without commenting on the policy as such, but just the way that those numbers flow through. Is that correct? There hasn't been specific modelling done on the scenario you describe as far as I'm aware, but my understanding would be that you're correct, that if the population were to grow, then it could increase, in a wider economic sense, it could increase tax receipts, which would push more receipts into the Scottish Consolidated Fund, which would make more funding available to pay for benefits, but the population growth should be accounted for in that way through the index per capital model. Kevin, do you want to comment on that? I think that's a fair comment, James, and I'd also highlight that the benefits which are being devolved as set out in the financial memorandum are related to disability and age-related benefits. They are not primarily benefits that are related to economic activity, so our focus is very much on understanding the behaviours of those benefits that are being devolved in the Scottish system. To just clarify, in that scenario, where the working age population is a percentage of the top population grew, meaning that the elderly population, as a percentage of the top population, was reducing, not only would you get increased tax tech, but you would also get a benefit from the BGA, specifically on the social security aspect. We're moving on to issues around set-up costs, and I think that Adam Tomkins has a question around that. Thank you, convener. Good morning. I just want to see if I've understood the numbers correctly. It's highly likely that I haven't, and you can correct me. As I understand it, the projected set-up costs for the new Scottish Social Security Agency are estimated at £308 million over four years. In the physical framework, the UK Government agreed to make a one-off transfer of £200 million, being quite a lot smaller as a number than £308 million, and also that the estimated running costs of the agency are around £150 million annually, which again exceeds by some considerable margin the value of the £66 million annually to be transferred from the UK Government. Are those numbers broadly correct? Yes, they are. Why are the numbers that the Scottish Government is giving so much greater than the amount of money that will be transferred from the UK? I think that others before me have commented widely on the physical framework. My understanding of the physical framework is that it represents the Scottish Government's fair financial settlement. It wasn't the total that the Scottish Government asked for, and the transfers were only ever intended to represent a share of implementation costs and administration costs. They were never expected to be the total amounts, so I don't see a relationship between the two in terms of one should cover the other. That was not the way that the physical framework was designed. As I said, we do believe it as a fair financial settlement. The £200 million on implementation and the £66 million on administration, which are for all of the Scotland Act 2016, powers not just social security, can be topped up by the Scottish Government's own resource to enable us to implement the system and run the system for social security in Scotland. That's very helpful. The margin between those two sets of figures, 308 versus 200 and 150 versus 66, is to be met from within the Scottish Government's own budget. It will require to be yes. Can I just ask you a slightly unrelated follow-up question about what is not included within the financial memorandum unless I've missed it and if I have missed it, then forgive me, but I couldn't find anything in the financial memorandum about costs associated with the proposed charter on social security rights. Is that correct? That is correct, to an extent. The agency running costs, which are detailed in the financial memorandum, are based on the outline business case for social security in Scotland, which was published by the Scottish Government. The outline business case analyzed the costs from DWP to build what we call the current activity-based model. DWP supply does with detailed activity-based information on cost, which we were able to build a model to work out in a robust evidence-based way what an agency in Scotland is likely to cost us. Our view is that within those costs would be the cost of reviewing the charter, but we do not believe that those costs would be significantly above a material threshold that would cause a requirement to update the figures beyond what we have already prepared. That puzzles me, and it puzzles me for this reason. One of the principal features of the social security bill is that it seeks to put devolved Scottish social security on a human rights footing. One of the most important of all human rights is the right to have your rights enforced in a court of law. It puzzles me that there has been no thought given to the extent to which there will be inevitably increased litigation costs that will have to be borne by the Scottish budget because of the way in which social security is being put on a human rights footing in this bill. I do not really understand why that is not above what you just described as a material threshold. Our current view is that the cost will not be significant. However, perhaps I can expand on how we will take the costs forward from where they are at the moment in the financial memorandum. As I have described, we have used the current activity-based model to define the costs as recorded in the financial memorandum. The outline business case goes on to discuss the future activity-based model, which our analysts colleagues are developing. The best information that we have available is based on historic information. We are conscious that we will do things differently from the way DWP do things, that there may be variations in costs as a result. At the moment, our colleagues are working on a future activity-based model that maps out the 2B systems and processes of a new agency for Scotland, which will enable us to wrap cost information around those future processes and systems. That may flush out some variances, the likes of which you describe. However, as I say, at the current time, our view is that the cost will not be significant. We have introduced issues about information provided by the DWP, and I know that Patrick had a specific question around that from the SPICE paper that we provided about estimates of DWP costs. Patrick, you want to go on that area or bring you back in later on in your wider area. Thank you very much. Good morning. The activity-based model that you describe is the basis for the Scottish Government's estimates of DWP costs of administering devolved benefits in Scotland. Just so that I am clear that I understand, that is about calculating a cost that DWP will be provided by the Scottish Government for its continuing administration. This is after the period of DWP administration. The DWP provided us with detailed cost information, which we have used to build a model to model a potential social security agency for Scotland. We have used this information as a historic basis of how much administering benefits costs the DWP and applied that to the Scottish situation. Does this not at all relate to what DWP will continue to administer in the short term? Well, no, sorry. Yes, absolutely. Put that way, yes. It is based on their cost information, so the cost information that DWP is supplied is in relation to their operations. If their operations continue, it will, in a way, relate—an element of it anyway—to the benefits that they continue to administer. So have they agreed to that estimate? To our estimate. Have they agreed with the estimate that you have made? I do not believe so. Is not that going to be important to get agreement on what the cost of administration is going to be? For the agency in Scotland. We have not used this information to estimate funding flows from the Scottish Government to the DWP. It is their cost information, so what it costs them, I would assume it will continue to cost them, but that does not relate to funding flows between the Scottish Government and the UK Government. What is the reason why the estimate of the Scottish agency's running costs would be lower as a percentage of the overall benefits being administered? You have said that it is 6.3% as the current DWP costs. 6.3% of the benefits they administer is the administrative cost. You are suggesting that it is going to be 5% for a Scottish agency. Why can you say with confidence that that is going to be that lower administrative cost? I think that the analysis is intended to show that we are within a margin of error. We are forecasting here, so there are margins of errors involved. It is to show that the methodology by which we have estimated the cost of the agency in Scotland has been done on a robust basis and as comparable we DWP. I do not believe that it is intended to show that we are going to be lower than DWP. Related to that, I think, as the future activity-based model, these will not be the final costs of the agency in Scotland. They are our current estimates based on our current design assumptions. As the programme to implement social security in Scotland matures, as decisions are taken on how we will administer benefits on the exact setup of the agency, processes may change and costs will require to be reviewed. Can you say anything about the current nature of the administration that DWP has been doing that in future will be done in Scotland? Does any of this relate to benefits that are currently administered by way of paper records in warehouses in London that would need to be moved up here and separated out by geography when, at the moment, they are only stored according to name rather than having a separate batch that can be easily moved that relate to Scotland? Am I being at Branclus and there, if that's okay? The industrial injuries disablement benefit is currently entirely paper-based. I don't know where the warehouse is, but that one, I know with confidence, is currently a paper-based administration. Is the intention simply to move the paper or to digitise it in the process or administer it in a completely different way? We don't, I think, have a decision on that point. Before we would do anything we would need to identify the records relating to Scottish recipients. The sorting job would be the first and arguably the most important part of that process. Presumably DWP would have to do that because it's wherever the warehouse is, it's their warehouse. I'm not sure that they would physically have to provide the staff to do that. I couldn't be definite on that point. It does just leave me still a little bit unclear as to how the Scottish Government can be confident of a lower administrative cost compared with the scale of the benefit being administered if even principal decisions like how it's going to be done physically haven't been made yet. We did a detailed analytical piece of work that took DWP's database around the current activity-based model that they provided that we built from their data. Our analysts looked at that in a lot of detail in terms of case load, new cases coming into the system, cases leaving the system. We built up a very granular understanding as to how the flows might work for the different benefits. We made assumptions around corporate costs like running costs, audit costs, internal audit costs. We built up a picture of what we think it would cost DWP to administer the current system in Scotland because DWP do not provide separate statistics on Scottish running costs. We did that exercise and overlaid assumptions on how it would run in Scotland to provide a broad indication of the figure. That gives us a good benchmark as to what we think it might cost in Scotland. The important point moving forward will be the future activity-based model. We have service design experts on the social security programme that are designing new fit-for-purpose processes to support a new organisation. We have analysts in communities analysis division who are building detailed models to model the future. The links between the service design and the analytical piece will enable us to understand how much it will cost to administer the system in the future. However, we know that the assumptions that we have made are reasonable because they broadly tie in with what we estimate costs to run the system at the moment. James, you wanted to talk about IT issues and the costs around that specifically. In the financial memorandum, you have £190 million of IT setup costs. Clearly in previous projects, both in Scotland and the UK, IT costs have been an issue that can sometimes spiral out of control, so it's important that that figure is robust. Can you give some indication as to what does £190 million entail? I can. I think that it's probably important to say first off that the financial memorandum makes clear that there are a number of decisions on detailed design still to be made that costs will vary materially as a result of decisions on what we buy, how we buy it and when we buy it. The Scottish Government has taken the view that we wish to provide the committee with as much information as possible on implementation to enable scrutiny of costs, but I think it's important to mention the caveats within the financial memorandum. The figure of £190 million within the financial memorandum is based on a specific set of assumptions that we will design and build our own system for IT for a social security agency in Scotland. That may not be the method by which we do it. We may reuse where appropriate. We may buy customisable off-the-shelf packages which could push costs down. In terms of coming to this figure, our colleagues within the social security programme have been able to scope out exactly what they would intend to build or possibly build for a system in Scotland and wrap costs around those. We've then taken the Treasury Green Book and applied appropriate optimism bias within the cost figures that they are presenting. As the process moves forward and evolves and some of the decisions are taken around detailed design of the system, more information will become available on IT. We would expect that to be supported and we will ensure that it is supported by a business case that follows the Green Book and the five-case model in the best-standard methodology. We will apply optimism bias where necessary and the appropriate amounts per the Green Book to make sure that we are accounting for the possibility that costs may rise beyond the business case at the beginning. Within the programme, there will be business cases at project level. We will be happy to share those business cases with committee and publish them in due course so that we can provide the evidence base for spending decisions. Sorry, if it is at all helpful, the Minister for Social Security wrote to the convener of the Social Security Committee last week attaching a couple of things, one of which was a relatively brief, fairly short update on the IT implementation for social security. We could obviously arrange that members receive that as well. Just pick up on a couple of points there. Are you saying that the assumption is that you will build a system in the house? You won't outsource it. The cost within the financial memorandum uses the assumption that the system will be built in house, but no decision within the programme has been taken on what we will actually do in practice. It will depend, and it may be that different parts are done in different ways. Is the case that you don't at this moment in time have a system specification or a business case? We do not at the moment, no. We are not at the stage where we are spending money on the detailed design of IT systems. There will be for some of the benefits, though, that are about to be procured. The Wave 1 benefits were obviously pushing towards creating a system to enable us to administer those benefits. There will be a business case for that. I do not understand how you are able to arrive at a figure if you have not got a system specification or a business case, because it will provide the component parts of any system, and obviously the detailed costs, or even the high-level costs, would drive from that. You are correct. The detailed costs that we actually spend on IT will be driven by detailed system specification and the business case. However, we wanted to give the committee an indication of likely costs. As I said, Wave made a number of assumptions around building the system in-house that have costied those out. Using a broad design for a potential system have wrapped costs around the potential components of those systems and have applied optimism, bias and contingency as appropriate in order to come to this £190 million cost. The line of your questioning is why the financial memorandum is so heavily caveated in regard to the implementation costs. There will continue to be uncertainty as the programme moves along and becomes more defined and defines precisely what we are going to build or buy. It might be helpful, convener, if the panel are able to write to us, just providing more detail as to how £190 million figure is arrived at, because I think there is an awful lot of subjection going on there. Hi, thank you very much, convener. Clearly, it is a substantial set-aside figure for the IT procurement, but it does look because it is over half of the total implementation cost of the entire programme. I am just keen to find out. Can you just remind me when the system must go live? Where are we now? Were it pre-specification and what you are saying? When does it need to be live? Different elements of the system will need to be live at different times. I think that the Cabinet Secretary for Communities, Social Security and Equality has announced the Wave 1 benefits that is cares allowance supplement by summer 2018, funeral expense assistance and best start grant by summer 2019. They will require to be systems to administer those benefits within those timescales. Further benefits or the timescales are yet to be announced, but there will be elements of IT that I would imagine are required to support those elements as they are announced. When are we going to start the process of speccing the requirements so that the system is ready on time for the first benefit that needs to be live at the earliest? I might invite Kevin to describe our programme's agile methodology, which I think might assist with that. The low income benefit project is undertaking a discovery exercise to scope out what the requirements might be for that. It is important to say that the fundamental approach is that a phased approach is being taken to the development of the different parts of the system in order to de-risk the programme more generally. It is not a big bang approach, as it were. Looking at the phasing of the Wave 1 benefits, the individual components that are required will be developed in due course. They are gathering the requirements in a discovery phase. Does that help? Sort of, but it does not really answer when is the process going to begin to spec what we need so that it is ready for summer 18? The work is on going at the moment in the discovery phase. Specking the requirements, as Mr Kelly was referring to, is that under way at the moment? One of the fundamentals of any software development is specifying the requirements, understanding what they are, and delivering that sticking to it and delivering it. Just again, if it would be helpful, the first Scottish benefits to be paid is the carers allowance supplement, which will be paid in the form of two roughly six monthly payments annually in order to, between the two of them, account for a year's worth of the carers allowance supplement. That benefit has been designed the way it has been designed. My understanding is so that it can actually be delivered through Scottish Government's existing C's payment system. The system design, service design requirements for the first benefit are very, quite deliberately intended to require less in terms of new build, in terms of new systems, and so on. As Kevin has said, the service and system design for way one, which is 2019, is currently ongoing at the moment. They're going into discovery phase. They'll go from discovery through the alpha build, the beta build, and so on and so forth. If the question pertains to the first payments, the carers allowance supplement payment, and the system requirements work for that, that benefit has been designed in order to make it as manageable as possible within the envelope that we can currently deliver with our own systems, without needing new build for that part of it. In terms of the overall numbers of transactions, the financial memorandum says here quite clearly that the number of transactions that the Scottish Government will carry out in a week with new systems is more than what we currently do in a year. That's a huge volume of transactions that we're anticipating with the system. Are you confident that the systems will be able to cope with the volume that's going to come through this system to be processed? Most of that won't happen until 2019. On you go. Perhaps beyond 2019, as I think Chris said, the initial way one benefits are on the smaller side in terms of the greater whole of the number of transactions that are actually being devolved. I think there is work on going now looking at a payment platform to ensure that we have a robust system to make payments to the people of Scotland. I think that it is viewed by the programme as being an absolute critical factor in our work. We must ensure that payments go when we say they're going to go. There cannot be failure there. The work has started now. We are looking at a number of options for systems, and that work will develop over time to ensure that we have a system in place when we need it. The likes of the Carers Allowance Supplement, we are using existing systems. We would plan to pay that through the Scottish Executive Accounting System payment platform, because we know it is a robust and reliable system to ensure that we are at no risk of not meeting our payment dates. Okay. Thank you. We're going to another area, which I'm going to bring in Neil Bibion in terms of COSLA cost. I think the point that James Kelly raised and followed up by Willie Coffey, I think, is right. I think it would be best if the Government were to write to us laying out over the period of time that this whole programme will roll out what you know about the procurement process and the various stages that will be required between now and the final benefit becoming the responsibility of the Scottish Government. I think that that would be helpful. I recognise that that's technically not what this bill is about. This is an enabling bill, and that goes beyond what the support for this particular bill is concerned, but I think the legitimate questions have still been asked. Neil Bibion. In terms of COSLA, COSLA have told us on the figure of £21 million for local delivery that they believe it's unclear what assumptions have been made and what might be in or out for local delivery, given the levels of uncertainty that still exist. More detailed scoping of local delivery services and costings will be needed to understand what is actually required to deliver the principles outlined in the bill. Can I ask what has been done or what will be done to address the concerns of local councils going forward? Chris, I think that's probably a... We're currently engaged in a programme of engagement activity with all of Scotland's local authorities. I think we are up to the point where local delivery colleagues will have visited 16 out of the 32, and the intention is to continue that engagement until we have met with and discussed these with all 32 of our local authorities. The points that we have made so far are firstly that the agency's local presence will be agency staff, there will not be local government staff, we're not talking about rebadging existing local government staff resource, we are talking about adding a new agency resource into local government areas. We've been talking about local government areas partly because it is a reasonably easy way of describing the locality part of it. I think we have been reasonably clear that we're not necessarily talking about using local government premises or sites, facilities in all instances. The wording that the minister has used is that we intend to base our local presence in places where people currently visit. It may be local government premises, it may also be NHS Scotland, health or social care premises. The decisions on these matters will be taken on an area by area basis dependent on essentially what we believe will work for that specific local area. This to an extent is where it becomes useful to talk about local authorities because it's a good way of breaking the country up and it's a good way of describing what you mean when you're talking about a specific local area. But we're not necessarily talking about using local government facilities in all instances, in every instance. Where we do, there will be recharging, Scottish government will meet any costs on the local authority that are incurred as a result of us occupying or using their premises. But until we have a local area by area breakdown and a model for each area is precisely what we will be putting where, it's perhaps not possible to interrogate that impact on each LA in detail. Patrick, I think you installed one area that you wanted to talk about and that was about potential value of savings elsewhere on budgets and how that's going to be accounted for. Thank you. Again, this memorandum doesn't only address the immediate impacts of the bill itself, but it does touch on the longer-term operation of the newly devolved powers. Yet it seems to me that for the most part it's still cast in terms purely of additional costs that come from paying benefits. I'm wondering how and to what extent the Scottish Government intends to estimate the additional value to the public purse of paying benefits through reducing the demand on other public services, whether they're the healthcare system or anything else, just to give not just financial costs but to place a value on other benefits as well. One example would be the proposal that we put forward and that the Government seems to be open to on a young carers allowance, specific to young carers. If that's delivered well and achieves its objectives, it should enable young people who are also carers to get greater value from their education to remove some of the barriers that they currently face to that publicly funded education service achieving what the money is being spent for. How and to what extent are those other benefits going to be assessed in terms of the impact on public finances or the effectiveness of spending public money through more than just the extra financial costs of paying benefits? I'll perhaps start and bring Kevin in if that's okay. I think it's a good point you raise. When our colleagues prepared the outline business case for the agency in Scotland, it was an issue that we began to look at. The OBC goes into detail on the marginal utility gained as a result of social security expenditure that, for every pound that we spend in the lower income deciles, there is a greater marginal utility gained by those individuals. It details that in attempts to quantify that benefit in a weighted way to show the benefit of social security expenditure in Scotland. Related to that, in more detail and more specifically other points that you raise around, how do we quantify the impacts on the other areas of the public sector? There is an element in the fiscal framework where we try to do that or where we describe how we will do that. The fiscal framework covers the interaction between the UK public sector and the Scottish public sector. It accounts for spillovers where we make a policy decision that has an impact down in the rest of the UK. How marginal savings that might be gained might be transferred between the Governments. I will perhaps bring Kevin in there to give you. The outline business case used a multi-criteria analysis framework to weigh up the relative merits of the different options for the design of the agency, so factors such as dignity and respect, implementability and risk and other factors that were used in that scoring process. I would imagine that moving forward when on-going decisions are made on the programme, that kind of framework is used. When we then transition into measuring what we call measurable improvements that the programme is delivering, we will have a measurable improvement strategy that will ensure that the money that is spent on the programme is aligned with the measurable improvements that are being delivered, so we can assess the value for money of the programme. It is important to bring in the wider societal impacts so that we are not just setting up a payments platform to pay money to people, but setting up an agency that will treat people with dignity and respect. Finding wider measures to demonstrate the measurable improvements and the benefits that the agency is delivering will be important. That is an on-going piece of work or area of thinking. Yes, the measurable improvements manager on the programme is—there is a role on the programme whose job it is to bring this kind of thing together. It feeds into the business cases, it feeds into the socio-economic case of the programme business case. By including financial and non-financial measures, we can establish the wider value for money of a particular project or initiative. I have captured everybody who wanted to contribute in this and ask questions. I thank the Government Bill team for coming along today. We will now reflect on the evidence that they have provided to us, and we will be writing in due course to the lead committee to let them know some of the outcomes of this discussion. I am very grateful for you coming along. I now suspend this meeting. I close it because we are now moving to the public part and the private part.