Under our second item of business, we will take evidence on the devolved taxes from two separate sets of witnesses. Our first panel will focus on the land and buildings transaction tax. I welcome to the meeting Kennedy Foster from the Council of Mortgage Lenders, Philip Hogg from Homes for Scotland, and Chris Stewart from the Scottish Property Federation, who was also here last week in another guise.
Members have received written submissions from each of our witnesses, so we will go straight to questions. As always, I will ask some opening questions before inviting colleagues to ask questions.
I will deal with the submissions in no particular order, starting with the Homes for Scotland submission. Your submission expresses concerns about the impact of the rates and bands. It says that
“the changes have given a small benefit to the majority of people purchasing a new home”
and
“have allowed for a smoother pattern of prices in the market, instead of having an artificial step at the £250,000 mark”.
However, it then says that
“the revised rates are having an adverse effect on the sale of properties in the middle to higher end of the ... Market”—
although, of course, that is based primarily on anecdotal evidence at this stage. Can you say a bit more about that?
The submission also says that some customers
“either requested that their transactions be brought forward to before April or delayed until after introduction of LBTT”.
Can you therefore also talk about the scale of the delays and the number of transactions that might be brought forward?
I recall sitting in front of the committee in January or February, when the rates had just been announced. The feedback that we had from our members at that time suggested that there was some concern that the rates on the higher end of the market could dissuade people from transacting, which would obviously have an impact on the tax take and could cause stagnation. I also recall mentioning that a fully functional market, with movement up and down the chain, is absolutely key; the concept of the ladder, with people going up and down it, is important. The feedback that we have had from our members in advance of today’s meeting has confirmed that. There has been no comment from our members that their concerns have been found not to have a basis.
The impact of the tax at the top-tier levels, where the higher tax bands come in, has been quite significant. Anecdotally, we are hearing a number of our members saying either that sales have been stalled or that they have not gone through. In our submission, we include feedback that shows how the delays in transactions have impacted on the situation, and statistics that show that there was activity to get sales of higher-value properties completed before the new tax came in, and a delay afterwards.
Also anecdotally, we hear that there has been lots of pressure on developers to help customers to absorb that tax by forms of incentive. Some developers can afford to do that, but others cannot. There is a question about the on-going viability of that assistance.
We have also detected, particularly in the north-east, a combination of “perfect storm” conditions. It is well known that the prices of oil and gas have tumbled. That has had an impact on employment, property prices and the market in general in the area. The uncertainty in the north-east, the new higher tax bands and the loss of the help-to-buy scheme have combined to create very difficult trading conditions in that part of the market.
A quite interesting comment in the editorial on the opening page of a Herald supplement from two or three weeks back caught my eye. It says:
“LBTT is becoming interesting. This week revealed news that business is booming for an upmarket Edinburgh bathroom company as owners in the million-pound plus price bracket opt to spruce up their homes rather than move.”
I am delighted for Mr Ronnie Scott’s bathroom company, but I think that that example shows in a nutshell the behavioural reaction to the tax. I am not saying that people in million-pound properties need assistance, but it is important to point out that purchases at that level are, on the whole, largely discretionary. In lots of instances, people probably do not have to make that move or that purchase.
I will give you an example of the impact of such a decision. If someone who was considering buying a £750,000 property decided to have their bathroom upgraded instead, 81 additional sales at £175,000 would be required to make up for the loss of that tax revenue—in other words, 81 additional sales would be required just to stand still. That brings into focus the high dependency on the high-tax levels, which we think not only is risky with regard to tax take but has a disproportionate impact on higher-value property levels.
I will go back to an earlier submission that we made, in which we argued for smoothing of the tax bands. We think that that would be a far more sensible approach to allocating the tax.
Thank you very much for that.
The Council of Mortgage Lenders has produced a short but excellent submission that in some ways contradicts a lot of what Philip Hogg has been saying, although not necessarily with regard to the high end of the market. It talks about
“a 39.2% increase in the number of loans and 27.5% in value on the first quarter of 2015”
and then says that
“Within these figures there were 8,000 first time buyer loans an increase of 50.9% and 5.3% on the first quarter of 2015.”
Just for clarification, are you talking about the first quarter of the current financial year or the first quarter of the calendar year?
I am talking about the quarter ending 30 June. In fact, we announced this morning our lending figures for the quarter to the end of September, which show further growth in the mortgage-lending market in Scotland. The number of loans that were granted for house purchase in the quarter to the end of September is 18,500, with a value of £2.4 billion. With regard to first-time buyers, the number is up to 8,500 and the amount is £920 million.
Those figures could be attributed to a number of things. Obviously, the mortgage market is currently quite competitive. We are still in a very low-interest-rate environment, and it looks as though interest-rate rises are being held off until 2016. Perhaps that is leading to current consumer confidence in taking out mortgages.
I am not really involved in the day-to-day sale and purchase of houses, but from time to time I speak to estate agents and others, so I echo what Philip Hogg said. People who are involved in the high end of the market will tell you that transactions were certainly brought forward to beat the introduction of LBTT. Compared with last year, there has since then been a much lower level of transactions in the higher end of the market. Whether that situation will return to something that is more like the norm remains to be seen.
Looking back at the history of stamp duty land tax, we can see that advance announcements of changes to the tax regime have always resulted in distortions to the market; the really interesting question is whether the higher end of the market will begin to recover. As Philip Hogg said, there is anecdotal evidence of folk not moving but instead upgrading their properties in whatever manner they consider to be appropriate.
Your submission also states:
“We believe it will be some time yet before the full impact of LBTT on the residential property market is understood but the first quarter following the introduction of LBTT appears to have been a positive one for the residential property market in terms of mortgage lending.”
If I have got this right, you are basically saying that in the lower end of the market—I suggest £325,000 or less—the LBTT changes have been positive, but you have concerns about the level above that.
Yes, absolutely.
Chris Stewart more or less agrees with that, as the last bullet point on the first page of his excellent submission suggests that the Scottish Government should
“increase the 5% rate threshold to £500,000 and ... abolish the 12% rate which we believe is making little positive contribution to LBTT revenues.”
You go on to say that the
“distortion has been replaced by a new distortion at the point where a 10% rate begins”.
Thank you again for having me along today.
We have some direct evidence from work that we have carried out recently. The point about the 12 per cent rate is that it is not generating tax take, so what is the point in having it? Our figures show that there has been a 22.5 per cent decline in sales of properties over £400,000 or above. That is—quarter on quarter—May to September 2014 compared to May to September 2015, so it is a like-for-like comparison. There has been a 54 per cent reduction in sales in the £1 million-plus part of the market.
As Philip Hogg has said, this is not about feeling sorry for people in that part of the market; it is about LBTT stopping the market from functioning properly. The key point is that people need to be able to move up and down, so it is a problem if a whole segment of the market is static. We rely on something like 74 per cent of the tax take from LBTT coming from properties above £325,000. A 22 to 25 per cent reduction in performance in that part of the market will have a fairly profound effect.
We advocate the introduction of a more progressive system that would introduce an interim band and raise the 5 per cent threshold. In that case we should overall, ultimately, get the best of both worlds because there would be an increased tax take and the market should perform better.
Thank you. I will stick with Chris Stewart. You obviously have an interest in the commercial end. Your submission states that
“the move to a progressive rate of taxation with a less radical change in rates and thresholds has acted to make the change in tax structure less marked and this appears to have contributed to a smoother conversion from SDLT to LBTT, supported by enhanced revenue for the Scottish Government in line with the strengthening wider economy.”
Can you say a wee bit more about how you feel the commercial sector has been affected?
In relation to the change from 4 to 4.5 per cent, the commercial market always prices in the tax level. The way that the system works in the commercial market is that, at the point of transaction, the sum is deducted from the value of the asset, so it affects the seller. In the commercial market, a 0.5 per cent change has not been material enough to lead to a wholesale change in activity. We have to guard against the backdrop of the fragility of that market. Take as an example Aberdeen, which has been a huge part of the Scottish commercial property market activity. The work that we have carried out shows that the value of transactions in Aberdeen for the previous quarter—quarter on quarter—fell from £201 million to £50 million. A downturn in that market can have a fairly profound effect on the tax take. Reliance on the commercial market when it is still fragile may not be the way—we need to get the residential market performing at the same time.
09:45
Thank you. I have another question for Chris Stewart, then I will ask Philip Hogg a similar question, about the role of Revenue Scotland. The timing is perhaps unfortunate, because we had Revenue Scotland in last week and could have put some of your questions to it at that point.
You have said that
“there is a feeling that the guidance and administrative support side of Revenue Scotland remains weak at this stage”
and that
“members feel restricted in being able to voice criticism because they need to have an on-going relationship with RS in order to serve their clients’ interests.”
That gives you concerns with respect to reassessment. I would like your comments on that.
Again, that is just around confidence in the market and the ability of commercial developers and investors to plan and to ensure that they understand the mechanisms behind how Revenue Scotland will work. We are keen to see further development of that and to ensure that there is a bit more transparency generally.
Homes for Scotland’s submission says that
“Concern was ... expressed about the approachability of Revenue Scotland”
and that
“it was noted that Revenue Scotland will not necessarily follow the HMRC guidance”.
We should keep this in perspective; we are talking about one or two comments that might be in isolation in a broad sample. I can relay such comments only as they are reported back to us.
I am looking at the feedback that we received and comments that there were early indications that Revenue Scotland would be more approachable regarding inquiries. However, we are hearing from a number of colleagues that that has not been the case, in practice. Our members were hopeful that it might be easier to discuss complex LBTT cases, but it is reported to us that it has not been possible to speak with anyone informally on the phone. It also appears to be difficult to obtain feedback by email. The remedy seems to be to seek a formal opinion from Revenue Scotland—a process that is, it has been reported to us, quite slow.
That may be because it is early days and there are teething problems, or whatever. Moreover, we should keep it in perspective that it is just one of our members who has reported that specific issue—albeit that that member undertakes quite a number of transactions with Revenue Scotland.
The Council of Mortgage Lenders has said that there has not been
“any ‘noise’ of issues in the collection and administration of LBTT”.
In paragraph 6 of your submission you say:
“we believe the full impact of LBTT needs to be better understood”
and you go on to talk about long-term impacts.
Given what you have said in your submission and what colleagues have said today, are you of the view that there should in the draft budget for 2016-17 be changes now to the rates and bands for residential transactions, or do we need to wait to see further evidence before changes are made?
History shows—particularly, as I said earlier, in relation to SDLT—that when tax changes are announced well in advance, that leads to a short-term distortion in the market. As I think we have all already identified, the impact is really at the higher end of the market. The question is whether that part of the market will recover to a more natural level over a period.
If there is one lesson from this, it is that in order not to distort the market we should not announce changes in tax regimes well in advance because people take advantage one way or the other of such change. We are only six months into the new tax regime; we need to see it operate for at least a full year.
I do not know so much about the commercial market, but the residential market is very cyclical. We are beginning to get into the quiet period in the housing market—Christmas and new year, January and February. The weather does not help people looking for houses. The market begins to pick up in the spring, dies off a bit during the school holidays and picks up again after that. It is a very cyclical market and we have not been through all the cycles under the new tax regime.
Thanks very much. That is very helpful.
We have now been joined by Jackie Baillie, so the committee is at full strength. I open out the questioning to committee members.
I want to follow on from one of the convener’s questions. I think that Chris Stewart talked about the 22 per cent differential in property sales at the higher end. I seem to remember something that was possibly in the same journal that Mr Hogg cited earlier about people having advance notice of the increase to 12 per cent. Is that 22 per cent differential a bit artificial? Would we really need to compare the figures for the past three or four years, perhaps, to get a feeling for whether the increase, which I believe there was, in the previous year in advance of the tax being imposed was true?
I suppose that the issue is that there will certainly be short-term advancing of transactions in the market when there is a tax change, but that will not roll through the entirety of the rest of the market. People who are thinking of selling in September do not just advance and sell in March to hit the tax point, but you will see—and we did see—in the figures a short-term distortion. In April and May, there were clearly fewer transactions, but the market then returned to the profile that we would consider to be normal.
The point is about a properly functioning market. All aspects of the market need to function for people to be able to downsize—to move into smaller properties—once they have had their families and for families to be able to move into higher-end properties. The threshold where that affects the market is not just at the £1 million mark; it is right down to the tax point at £325,000. That main band of tax take is being hit. To me, that seems very much out of kilter.
On sentiment, one of the things that I thought was quite telling was that in relation to £1 million sales, in 2014, 27 per cent of buyers were international, and in 2015 the figure is 18 per cent. So the 12 per cent rate is putting people off. Irrespective of whether it is meaningful for tax take, it is stopping people investing in Scotland. The whole tax structure affects people’s confidence about putting money into this country. How the market is perceived is critical.
We have referred to the downturn in the oil industry and the economy of the nation in general. Do you agree that it would not be right to place everything on the increase in tax?
It would not. We advocate a balanced and progressive system and we understand the need for the new LBTT system. There is no doubt about that. The old slab system did not work.
However, let us take the commercial market, for example. I believe that there were seven bids for the Atria Edinburgh building, which is just being sold to the Deka Immobilien German fund, and none was from a UK institution; they all involved international capital. If that asset had been in Manchester, there would have been 20 bids for it. Therefore, there is a regional distortion in our market.
It is about confidence and understanding policy and how tax will be structured. In the commercial market—it is not the same in the residential market—there is a dominance of international buyers, and they tend not to dive too much into the detail of how the market is operating; it is more about sentiment. In the residential market, they will pick up on the 12 per cent rate and ask whether it is a progression towards a higher tax structure and whether they want to be in that market. There was evidence last week on aligning policy with market sentiment, which is critical.
I will start on that point, Mr Stewart. You have given the example of the Atria Edinburgh building. You felt that there would have been more bids if that property had been in Manchester.
Overall, the tax take for commercial property seems to be slightly ahead of schedule, but I would like to drill down a bit. What impact has having a marginally higher top rate for commercial property of 4 per cent as opposed to 3.5 per cent had? Do you have any evidence on that, either anecdotal or otherwise?
As I said earlier, the transactions that are going ahead have not been affected by the 0.5 per cent increase. What happens is that the seller gets the price adjusted. In other words, there is a deduction from the price. Therefore, I think that the rate of 4.5 per cent for commercial property seems to be working.
However, please bear in mind that we should be looking at a profile of an increasing market—the market should be making progress no matter what the tax banding is. We have moved on from an all-time low and our market is starting to recover. I would suggest that simply meeting our expectations on tax take for LBTT relative to last year might not be good enough. Perhaps the commercial market, with increased transaction activity and a more confident market, should be delivering more tax take on the same banding, if that makes sense.
We would like there to be more transactional activity and more confidence in the market. That leads to more tax take. I would suggest that the fact that the tax take is only slightly ahead of or the same as it was previously does not show enough confidence. That is certainly borne out by the lack of depth in the commercial market for acquisitions, which is pretty significant.
The bigger point is that Edinburgh and Glasgow will do okay, but the international capital that goes into those cities does not roll into Kilmarnock, Kirkcaldy and Dunfermline. Commercial transactions in those areas need to be pushed on. It would be interesting to understand—I do not have the figures on this—the impact on secondary markets in Scotland, as opposed to on the prime ones in which buildings such as Atria are being sold.
Is it the SPF’s official position to call for parity with the rest of the UK at the top rate?
I think that our view is that that system is not creating the same issues as our current system is creating, so that would be sensible.
This question is for Mr Stewart and Mr Hogg, because Mr Foster has sort of answered it already.
On changes to the residential bands and rates, you have outlined that you would call for the threshold for the 5 per cent band to be increased so that the 10 per cent band kicks in at a higher level. It is no secret that I was sympathetic to that at the time of last year’s budget. If the Government were minded to give that serious consideration and made a change in the direction that you favour, when should such an announcement be made? Should it be made on 16 December when we get the Scottish draft budget, or should it happen, as Mr Foster suggested, right at the end of the budget process, shortly before it takes effect? From the point of view of looking after the overall health of the market, if such a change were to be made, when ought it to be announced?
10:00
One of the key points for the health of the market is people’s ability to project forward. For example, the house-building industry relies on being able to forward plan its activity in order to start development, to bring in labour, to start ordering, to contract with subcontractors and so on.
If the industry sees that it is not going to be able to build in certain segments of the market—because, say, there is no take-up at the £325,000 to £400,000 and above end—it will pull back on its programme. Philip Hogg can say a bit more about house building, but I think that certainty of environment and behavioural patterns, which have changed as a result of LBTT, are key and that if there were any inkling to communicate on that, the sooner that was done, the sooner the market would respond with more activity. After all, this is about everything from second-hand sales to the production of new homes, and the production of new homes is critical not just for the property market but for the wider economy. The earlier that such things are communicated, the better.
On your question about when any change should be announced, Kennedy Foster has already made the very good point that the further in advance the market is aware of things such as LBTT and stamp duty, the more buyer behaviour will be affected, both positively and negatively. It could be argued that it is probably best to give the least notice possible to minimise the number of what might be called artificial transactions.
In our submission, we say that it is too early to take a definitive view on this matter, but we remain concerned. It is still relatively early days, but there is nothing that has demonstrated to us that there should be no concern about the configuration of the tax structure. We welcome the removal of one of the bands, the introduction of the progressive system and the lower tax burden on purchasers at the lower end of the market. All of those things are excellent, and the Scottish Government ought to be applauded for bringing them in. However, our concerns remain: as we originally announced, we think that the tax is too steep at the higher end. I come back to my earlier example of the number of additional transactions needed to offset just one lost sale at the higher end, because that is where the risk lies.
We have to understand that transactions in the residential market broadly fall into two categories; the discretionary purchase, where people think, “We don’t need to move, but it would be nice to”; and moves that are essential because of family circumstances, change of job and so on. We need a tax system that encourages discretionary purchases and rewards people who want to move up the housing ladder, because that creates movement in the system and generates not only tax revenue but gross domestic product as a result of activity in the system. As we know, every home that is built creates more than four jobs, so we want a tax system that encourages people to want to move and to aspire to moving up the housing ladder.
Our fear is that the tax system does not do that and that, in fact, it does the opposite by making people stop, think twice and say, “Do you know what? We’ll stay where we are and get the bathroom upgraded instead.” That is great for the bathrooms industry, the smaller builders and all the rest of it, but I would suggest that it does not have the right impact and does not generate tax, GDP or any value for the Scottish economy.
As I have said, we remain concerned. We are not calling for an immediate change, but we think that the system should be kept under close review with a view to adjusting or tweaking it at some point in the future if the trends continue as the current evidence suggests that they will.
In the run-up to April, the volume of high-value sales increased and in the immediate period post April, the number reduced, which I guess was expected. It might be too early to tell, but six or seven months on, what is your sense about the likely volume of high-value sales going forward? Was that increase and decrease a temporary response to the change or does your evidence, anecdotal or otherwise, suggest that unless changes are made a reduced number of high-value sales will be an almost permanent feature of the housing market?
On the basis of the evidence that we have, I think that we will see the lower volume of activity at the middle and upper end of the market continue. Given the information that we have and the feedback that we get from our members, we think that there is evidence of a change in product mix on our members’ developments in recognition of the impact of LBTT—in addition to the impact of other market forces, which I should add into the mix. We think that there will be lower levels of activity in those parts of the market for the foreseeable future.
Are the other panellists hearing similar things?
Very much so. The trend is not a blip and it reflects not an adjustment but a behavioural change as a result of the cost of moving. The amount of activity that is going on to upgrade homes rather than move has been widely reported. The issue is the consequence of that on the broader market. It is not just about higher-end homes; it is about how the market functions as a whole, which is critical.
Mr Foster, in your paper you gave the current position on mortgage applications and mortgages being granted, which sounded quite buoyant, given the overall numbers—although you then explained that that does not necessarily correspond to tax take. Nonetheless, the mortgage market seems to be doing fairly well. Do you have a sense of how the figures compare with figures for the mortgage market across the UK? Are our figures broadly in line with UK figures or has LBTT made at least part of the mortgage market slightly healthier?
I do not have the UK figures in front of me, but I am sure that figures for first-time buyers are higher in Scotland than in other parts of the UK.
The mortgage market in the current year has followed roughly the same pattern: in the first three months of the year it was pretty depressed, it picked up in the second quarter, and our UK lending figures for October, which were announced last week, show the highest level of mortgage lending since September 2008. There is definitely an upturn in the mortgage market at the moment, which is caused by a range of factors, not just LBTT or SDLT.
I appreciate that you do not have the exact figures for the UK, but are you saying that your sense is that the mortgage market as a whole has performed similarly in Scotland and in the rest of the UK but that the figures for first-time buyers are higher in Scotland?
Yes.
This question is for any member of the panel. We are aware that forestalling happened—people brought transactions forward to avoid higher taxes—but to what extent did the opposite happen? Was there an incentive to hold off making transactions on some houses of lower financial value, because of the tax benefit, given that the threshold was going from £125,000 to £145,000? In some parts of the market, did we see people holding off, followed by a sudden increase in transactions after April?
We have no evidence that there was significant delaying in order to benefit from the new tax level for lower-end houses. I have some statistics with me, but they are not particularly revealing. As I said in the paper, 16.9 per cent of transactions in quarter 1 in 2015 were for properties below the £145,000 threshold, compared with 28 per cent in 2014. That does not give a massive picture. The feedback that we have had is that tax levels at entry-level prices were not in themselves game changers that persuaded people about moving one way or another.
I know that it is not my role to ask questions, but there is one point that I would like to ask Kennedy Foster about.
That is not really your role, but Kennedy Foster may wish to comment.
I would be interested to know whether he has any evidence or data on how many property purchases in the higher price bands are mortgage purchases, as opposed to cash purchases. That could be quite significant. We hear that there are a considerable number of transactions where mortgages are not required, so relying on just the mortgage figures may not paint the full picture. I am sorry to throw that at you, Kennedy.
I do not have the data on higher-end cash purchases. The figure that I have heard relates to the buy-to-let market—the private rented sector—where at least 20 per cent of transactions are cash purchases for which folk do not need to borrow. There is quite a high level of cash purchases.
A lot of investment in the property market in London, although not so much in Scotland, is through cash purchases from overseas investors—from China, the middle east and Russia.
I have a last follow-up question for Philip Hogg. You said that much of the evidence is not exact and is anecdotal, but your members have told you that they have sped up some transactions in order to complete them before the April deadline. Is it fair to say that your members did not tell you that some people were delaying until after April?
One or two did, but we are not talking big numbers. It is sometimes difficult to vary the numbers in build programmes. However, there was no evidence that there was a large swell or a large volume of transactions that were delayed until after the new tax system came in.
One of the pieces of anecdotal evidence is that in a market with low transaction activity, which is what we have at the higher end, you would expect an oversupply of stock—the market dynamic would be of oversupply—but the market is unique in that we currently have an undersupply of stock and buyers.
That relates to the behavioural activity of people who do not want to put their house on the market because of the cost of moving. That market dynamic is telling. The duration for which properties are sitting at the higher end of the market is markedly longer. It is substantially longer in areas of Glasgow and Edinburgh where we would expect movement of prime properties to happen very quickly. Properties are taking a year or a year and a half to sell, whereas previously, under SDLT, the market was performing. The lower levels of the market are working reasonably well, so that shows that there is a dysfunctional market.
Evidence that the committee has received and what we have heard today seem to suggest that there is market depression at all levels. For example, the National Association of Estate Agents said that there is a lack of stimulus at the lower end of the market, where there have been fewer transactions since the introduction of LBTT than there were in comparable periods in previous years. Today, we have heard evidence about higher-end market depression and the market for properties over £325,000. On the other hand, we received evidence from the Council of Mortgage Lenders about increases in mortgage lending and the number of mortgages. To the untrained eye, those two pieces of evidence do not seem to tally. What is the explanation behind that?
I do not know whether the evidence on mortgage numbers segments what are refinances—remortgages.
It does.
As Philip Hogg said, there is a question about the amount of cash activity at the top end of the market; unless those figures relate like for like, it is not easy to draw a direct comparison and say that mortgage lending is up and therefore the overall market is good—that is not the right way to look at it. We should be looking at a market that is moving out of its lowest position since 2008-09, rather than at a stalling market. The evidence does not necessarily align.
10:15
Mr Foster, I think that I heard you saying that your figures do disaggregate between what was mortgage lending and what was remortgaging.
Yes, our press releases split it into loans for house purchases and loans for first-time-buyer purchases, with a separate section for remortgages. The remortgage market is pretty flat at the moment.
So the increases that you have been speaking about are for first-time-buyer purchases and house purchases.
Yes.
Okay.
Does the data just show that there been an increase in mortgage lending, with no disaggregation according to value?
Correct.
I wonder whether you might care to comment on another thing. I recall what the press reports said during the period leading up to the introduction of LBTT—apparently, if someone was going to purchase a house valued over £750,000, they would have to pay 12 per cent on the entire purchase. Of course, that is not how the banding works. In fact, one newspaper contained a correction to a previous report but, on the following page, there was a reiteration of that very same reporting. Do you have any anecdotal or other evidence on the behavioural impact of some of that reporting, which suggested quite significant hikes in tax on property purchases, even though that did not reflect how LBTT was to be introduced?
I am not aware of the specific reports to which you refer, so I cannot really comment on all of that. As with any new tax system, there is a period of time for the tax to be understood and communicated correctly.
I would encourage members to go online or pick up property supplements—perhaps after this meeting. A quick scan of the property adverts, particularly for new build, will probably show—I am taking a gamble by saying this, but I will hopefully be proved correct—that there are a fair number of incentives along the lines of “LBTT paid”. That seems to be the incentive that most new home purchasers are looking for at the moment. That is itself symptomatic. It reflects what the stumbling block appears to be for many people. That is the market responding to what is thrown back as the barriers to purchase.
On your specific comment, we are not aware of a misunderstanding of the LBTT amounts, so I cannot comment on that particular point.
At the higher level of transaction, people are very well advised by agents, lawyers or whatever. Any press reporting of a level would be quickly qualified. I do not see that there is any evidence of people fundamentally misunderstanding the position.
Mr Foster, mortgage lenders will have been approached by people asking about the impact of LBTT. Did you pick up any evidence in that regard when it was first—
None whatsoever.
That is fine.
I note that, of your three organisations, only the Scottish Property Federation has gone out on a limb in making a suggestion for changes that could be made. Mr Stewart, have you done any modelling or hypothesising about the likely impact on revenues if the changes that you have suggested were to take effect—in terms of both market activity and the revenues that would be likely to be collected if the changes were made?
We certainly have done some modelling. I do not have the specific numbers for you immediately to hand, but we would be happy to provide them.
That would be quite helpful, so that the committee can see what the impact would be.
I have a question that follows on from that. I know that Mr Foster and Mr Hogg have hedged their bets in terms of suggesting changes, perhaps because they feel that more data or information is required. Mr Baker and I are very familiar with the north-east, being representatives of areas in that corner of Scotland, where other external factors have been involved in the price downturn, such as employment issues. How confident can we be that simply readjusting LBTT rates will, in and of itself, stimulate the market, irrespective of other factors that might affect it? If changes were made to LBTT on the assumption that that would kick-start purchases but that did not happen, that would have a further effect on revenues collected, particularly if the threshold was reduced for part of the market. I would be grateful for your thoughts on that.
If we look at the level of activity in the 12 per cent band, for example, we can see that there has been a huge drop in volume. There is so little activity in that market in terms of the number of units that it is clear that putting that band in place has affected sentiment. The fact is that 74 per cent of the tax take is from properties above £325,000, which means relying on the tax take from only 7 per cent of the market. There will be a positive benefit if there is a downward adjustment of any of the rates, because you probably cannot get any less out of the top end than you are currently getting. I am sure that a realignment would make a difference.
Does Mr Hogg or Mr Foster want to comment on the broader issue rather than the specific proposal?
Just to close the loop on this point, I referred earlier to the press, and The Herald’s “Scotland’s Homes” supplement, which you can see me holding, has an example of how builders are trying to overcome where the objections to purchasing are. I am sure that there are plenty more examples of that.
To return to Mr McDonald’s question, I recall that the finance secretary’s objectives were for the new system to be revenue neutral. From the information that I have seen—I stand to be corrected on this—the forecasts are that that might not be achieved. If the system is not delivering the tax take, that surely suggests that it needs to be reviewed and investigated. On that basis, we have to ask why it is not delivering the forecast revenue and look at where the sensitive areas are.
On the question of developing an alternative system, we proposed widening the 5 per cent band, and it is still our view that that idea needs investigation. However, our colleagues at the Scottish Property Federation have come up with alternative systems, and I am sure that there will be many other different ways. We have not come here with a pre-prepared alternative, but our concerns remain as originally reported.
A range of factors are involved in taking the decision to invest in buying a house. If someone is going to borrow, one of the issues to consider is whether it is affordable for them to take on a mortgage. Another consideration is the prediction for interest rate rises going forward. Consumer confidence, employment prospects and house prices all play a role in the decision to invest in buying a house.
I echo Chris Stewart’s comments in that I think that the Government probably needs to focus on the issue. I am commenting not from a political perspective but purely from a risk perspective. When so much of the tax take is coming from such a small number of sales, market decline must be a risk.
I think that it was Mr Stewart who spoke about the impacts of the changes in the rest of the United Kingdom. Do you have any evidence on the impact of the changes that were made to stamp duty land tax south of the border?
I do not have the statistics for the residential market, but we certainly know from the commercial perspective that the transactional levels that we would expect in our regional markets—Edinburgh and Glasgow—are substantially reduced. We measure that by the likes of the comment about how many bidders would go for a prime asset in Manchester relative to one in Edinburgh.
Commercial property is valued on the basis of its income and a percentage yield is applied to that. There should be no difference between Manchester and Edinburgh because they are both strong regional cities, but Edinburgh trades at a discount to Manchester which, historically, should not be the case. The difference is 20 buyers who are interested in a prime Manchester asset compared with seven who are interested in an Edinburgh asset. Therefore, there is something not quite right that is not attracting those investors into the Scottish market. That is what creates its fragility. Changes in tax, or the sentiment around tax, add to that fragility. One of our overriding concerns is how potentially unstable the market is.
I am sure that you agree that it is difficult to draw a wider conclusion based on one or two examples so, if you have other data in that respect—the data that I asked about previously—it would be nice to see some of it. That would allow for a slightly more reflective picture of what is happening.
We would be delighted. We can pull together a lot of information to support the conclusion, because there are a number of different strands to the matter, not just LBTT.
Is there anything to suggest that the fragility will remain?
Yes, there is. It is pretty well known that UK institutions and institutional investors—the big funds that we all know—are not making the same level of investment that they did historically. In many instances, they have filled their books with investment, so their 7 per cent allocation of their funds, for example, is not extending. That means that there is an increased reliance on international capital, which the SPF has analysed as being about 70 per cent of the Scottish market.
We have to understand the profile of those investors. That international capital is not a domestically based institutional fund or bank; it is probably headquartered in New York or the middle east. Money comes into Scotland and it focuses on the prime asset in which it is investing. The Bank of Scotland of old and the Royal Bank of Scotland were locally based. They were in the market and had an interest in funding it. That capital now looks at investing in Edinburgh alongside multiple European cities, such as Barcelona, Madrid and London, and tends to come here as a result of not being able to invest in London because the market is too hot.
The reliance on that type of market is very fragile. We are trying to encourage a more consistent market in which that capital is encouraged to stay. What has happened in our banking system because of slotting and the changes in capital requirements means that our banks cannot invest more to support local businesses and there is a ceiling to what they can deliver. Our concern is that all those factors mean an increase in dependence on a more fickle international investor. The fragility is evident.
10:30
You would therefore be in a position to model that behaviour and estimate the likely risk if that information is out there already.
The behavioural assumptions become more a matter of global macroeconomic conditions so, if an investor in New York gets a better return in Spain for the same allocation of capital, why would he put it into Scotland? All of a sudden, it is about your competing structure, how your economy looks and how attractive it is to that capital. If we had more local capital, we would not be having the same conversation. To model it, you need to look at the global markets.
That is helpful. Others have explored the lower volume of transactions at the high end of the market and Mr Hogg should definitely be on commission with that bathroom company. [Laughter.] I have counted three mentions so far; let us see whether he can manage a fourth.
I want to look at the level of receipts for residential transactions in particular and I want to understand the world that you guys understand, which is seasonality. Another submission tells us that, typically, by the time you get to the end of October, you could expect 62 per cent of the receipts to be in, with 38 per cent still to come. Is that your feel? Can you describe to the committee what is likely to happen between November and March?
I will not mention the bathroom company.
You just did.
I can assure you that I have no personal or professional affiliation with it.
I think that Kennedy Foster mentioned seasonality fairly early on. I can articulate it by referring to behaviour that many of us will understand. Immediately after the new year, we find that consumers typically tend to have a period of post-festive depression, so we start to see holidays being booked and the start of the process of searching for and buying a new home. I have modelled that using internet traffic. It starts to pick up around 12, 13 and 14 January when we see massive hikes in internet property searches. That is the start of the cycle and it is when people start looking, visiting showhouses and looking online.
For a new build, the purchase process, from the day the person first looks to the day they get the keys to the door, depends on how advanced the construction programme is, but we can talk about approximately three months and maybe a little bit longer. We will find a peak in transactions—sales completed—around Easter, so there is a rapid rise in the level of activity at Easter and into early summer. It dips off during the summer months because of the school holidays and for the other reasons that Kennedy Foster described.
We then see a second peak of activity in September and October, immediately after the school holidays and before we get into what we call the tinselitis period, which seems to get earlier and earlier and is when the tinsel appears in the shops and buying a new home is not so important. People have targets and say things like, “If I am going to buy in autumn, I want to get in before Christmas,” but for some people that is too much of a logistical squeeze. We therefore see a second peak of activity, at a lower level, in the autumn and winter period.
The curve looks like that. We have modelled that using our transactions in years gone by and it does not change. It only really changes if there are any stimulants in the marketplace, such as help to buy or changes in the tax regime. They might accentuate some of those waves in the pattern, but the overall pattern is fairly predictable and consistent, and it is like that. I am sorry—that does not really work for the official report.
No, no, the visual works for me. I love the word “tinselitis”. I have learned a new word today.
So, typically, by now you will have had your two peaks already and they are contained within your figures.
In an earlier life, I ran a mortgage processing operation for a major lender and one of our busiest periods was in the lead-up to Christmas. That is due to people who are out in September and October looking to buy houses; they are then completing and wanting to be in their houses before Christmas. You have to bear in mind that LBTT is paid at completion, so I would not necessarily have seen it in those months. You could still have strong receipts through to just before Christmas, but they will die off as we go into the new year if they follow the pattern of mortgage borrowing.
So you would expect January, February and March to be very quiet but there still to be activity right the way up to Christmas.
Yes.
That is helpful.
I have a small point to add about the Scottish Government help-to-buy scheme, which may have a distortion effect this year. All the properties that are eligible under that scheme have to complete by Christmas. That will create an artificial peak, which will probably be higher than normal, because those transactions have to be completed to ensure that the funding is delivered.
That is helpful too. The Scottish Property Federation commented specifically on this area, but I do not think that the other witnesses did in their submissions. The SPF submission compared the receipts: there is a shortfall of 5 per cent against 2013-14 and a shortfall of 25 per cent against 2014-15. Can you explain that? Also, what is your estimate—however difficult it might be to estimate—for the year-end figures?
The projected take across the bands in the residential market should be your benchmark for this time in the year, which is 60 per cent of £246 million, if it follows the normal pattern.
I stand to be corrected on the figures but, broadly, £118 million has been delivered in that same period, so at this stage there seems to be a £30 million shortfall, which is 23 per cent—I suggest that it will not be caught up—from May to October, on the predicted amount. I do not know whether that 23 per cent shortfall will be maintained or whether the gap will widen, but it is enough to highlight the difference.
You have made no projections that take account of incentives that might distort the market.
I think that we will see that pattern continuing.
Thank you very much.
I have a short supplementary question. You say that there is a shortfall of about £30 million on what was expected. The Scottish Fiscal Commission submission states that the Scottish Government estimates a possible fall in revenue of between £12 million and £37 million due to forestalling effects. How much of that £30 million shortfall could be attributed to forestalling? Is that something that you have information on?
You need to try to separate the commercial market from the residential market because they are performing very differently. I know that they can be combined to get the figures, but the residential sector figure is certainly not going to be of the quantum of potentially £50 million or £60 million by the end of the year. There may be an element that would have been collected in the year previously, but the behavioural pattern was certainly not that people who would have been moving in September advanced the move to April. We saw that correction, with less transactional activity in May. I cannot say how much exactly, but it would not be a huge amount.
Thank you.
Following on from that point, we have been given three possible scenarios, and we have still to see the Scottish Fiscal Commission.
The actual position is behind two and ahead of one of the forecasts. How certain are you that the actual position is behind forecast?
In the residential market, there has been substantially less take every month. In April, the difference from the previous year was £18.3 million, which is substantial. That would be as a result of the transition to the new tax.
In May 2015, revenue was £11.4 million, compared with £19.8 million in May 2014. In June 2015, revenue was £18.5 million, compared with £24.2 million in June 2014.
You are comparing with the previous year rather than the forecast. The forecast was £235 million, and the forestalling could be £12 million, £20 million or £37 million.
Yes—60 per cent of the market activity or collection for that time of year—
Yes, we have been given the figure 63 per cent, but that is roughly—
Even with 63 per cent, you would have expected more than £150 million take. As far as we can see, the actual take was £118 million—
That is depending on how much forestalling there was. Have you taken forestalling into account?
I suppose that as an answer to the previous question, there will be an element of forestalling, but not to the extent of over £30 million.
One of the figures that we have been given is a forecast of £235 million, with forestalling of £37 million, which takes us to £153 million, which is ahead of what is happening.
How would you explain £37 million of forestalling?
We have still to discuss that. We have not had the Fiscal Commission in yet. The Scottish Government’s forecast for forestalling was between £12 million and £37 million and the Office of Budget Responsibility’s was £20 million. We have not yet clarified how much the forestalling was.
About £37 million—
—added to £116 million is £153 million. I do not want to get too bogged down in figures. My point is that there is a variety of figures.
But £37 million of take on forestalling, if it were on commercial property and a higher tax basis, would meant that £822 million of transactional activity was—
I am talking only about the residential sector.
That would be even worse. A huge number would, in that case, have forestalled.
We have still to discuss forestalling with the other witnesses. My main point is that the actual amount is behind some of the forecasts, and ahead of at least one.
The panel feels that LBTT has had a sizeable impact on sales and prices. Has anyone—a university or research body—done an objective study, or is that feeling really just subjective?
No. It is definitely not a “subjective” feeling. Our businesses cover the whole market, and not just individual segments. Homes for Scotland clearly covers the entire market.
This is not sentiment: there are actual delays in property sales—
Have you asked sellers and purchasers, or prospective sellers and purchasers?
That is what the Savills report covers. It is evidence; it is not—
Who is that based on? How has it found that out?
It is a combination of Savills’s own transactional research and information from myhouseprice.com. It is evidence based: it has—
Has it actually asked purchasers or sellers?
Yes.
It has done that.
It is a property agent—it deals with transactions daily.
I know that it deals with transactions and could therefore say how much they have gone up or down. My questions are about why they have gone up or down, and how it has found that out.
We are all in the market and we deal with buyers and sellers in commercial and residential real estate. It is not just discussion—it is fact. There are—
I question how it is fact. We are six months into a new system. Unless some kind of academic organisation has studied it properly—
There was a year-to-year difference in the number of sales of properties over £400,000. From May to September this year, there were 1,283 sales; last year, in the same period, there were 1,657 sales.
10:45
The fact is that the number of sales is down. However, my question is this: in Aberdeen, where the oil price is way down and there are a whole lot of other factors involved, has anyone really studied how much of the sales drop is due to the oil price and how much of it is due to LBTT? Is it just that you would like to see the tax come down, so you are arguing that the drop is all based on LBTT?
No. We would love to be able to say that LBTT has worked, because we are looking for a positive and buoyant market. This is not a political point: it is about having a successful market in which jobs are being created and people are able to move houses at whatever level.
Are jobs created because people move or because new houses are built?
That is down to both: there is employment in both sectors.
I am not sure how quickly all this will settle down. That point has been touched on by other members. Are we in a settled position now or, following your arguments, are some people at the more expensive end of the market waiting to see whether the rates will come down in April? Are we still in a forestalling period? Can we tell whether we are?
My guess is only as good as that of anyone else around this table. I am not aware of any press speculation that there will be an adjustment in rates; that is not what our members understand will happen, nor is it the message that we would put out to the marketplace. I do not think that there will be many people waiting and hoping that the rates will be adjusted. However, that is a personal opinion as opposed to one that is based on research.
The first part of your question was on whether forestalling has washed through the system and we are now in a normal market. The real test of that will be Q1 next year—taking Kennedy Foster’s valid point that tax revenues follow a little while after purchase activity. I think that, perhaps by Q1 and certainly by Q2 next year, we will know what the pattern is and whether there is a continued depression in activity at the upper levels. If it looks as though activity is still reduced, that will be the new norm, with whatever implications that has for tax.
I think that we are now getting through the adjustment period. Picking up on the point that John Mason made to Chris Stewart, I think that key estate agents such as Savills, Knight Frank and others would be well placed to give you an accurate picture of consumer behaviour at the sharp end. They speak with prospective purchasers and sellers daily and know how they are behaving and reacting. I can only report back from our members, who are telling me that in the conversations that they have with prospective purchasers in sales centres, LBTT is a big issue, especially for people who are buying in the price brackets that we have been speaking about.
You are saying that it will be the first and second quarters of next year before we will see whether things are settling down. It would, therefore, be best for us to leave the rates as they are at the moment. If we change the rates, that will just confuse things more, will it not?
That is correct. Hence, in our paper, we suggest that you keep the situation under review for a little bit longer. However, if by the end of Q2 next year there is evidence of that pattern emerging, the concerns will be valid and the matter will need a serious review.
Mr Stewart, you said that certainty is a good thing. Do you agree that it would be best to leave things as they are in this year’s budget and to review the situation again next year?
Yes. It is important to get a full year’s data in order to mop all that up. However, we would welcome a review after that full year to see what the effect has been.
Your submission also mentioned the effect on labour skills: it was suggested that we are losing out on some skills because people cannot afford to buy houses.
I suppose that that is to do with the development of the house building and construction industry generally. In order to invest in new developments, new land banks and the skills that are required to increase that activity, we need consistency and an understanding of where the market is. If sections of the market are not performing, house builders will not build homes in those sections. That does not mean that they will build more homes at the lower end of the market—that is not the way it works—so the consequence is that they are not investing in the jobs. We know that from speaking to the contractors with which we work to grow their businesses. What we fundamentally want is a healthy and well-performing real estate market that is supplied by a healthy and well-performing construction market.
If wealth was shared out more evenly across the country, instead of building a £1 million house, somebody might build 10 £100,000 houses, which I assume would create more jobs. Would it not be better to build more cheaper houses than fewer very expensive houses?
I do not think that it is as simple as that, to be perfectly honest. The problem that we find in the construction industry and the house-building market is that most house building and construction businesses are suffering from the financial crisis of 2008 and have yet to recover. Banks are not funding small and medium-sized businesses to the extent that they were—the local debt providers to which we would usually turn are not in the market, so activity is just not happening outwith the major house builders in the market. The capacity in the market is not there, whether it is in lower-end homes, as you suggest, or elsewhere. It is a confidence issue.
On commercial properties, a comparison was made between Edinburgh and Manchester and it was felt that Edinburgh is suffering because of LBTT. However, London prices are, I presume, higher, and even if SDLT were lower there the net effect would be that it is more expensive to invest in London, but people do invest in London. It is not quite as simple as saying that people will go to the cheaper location, is it?
That is not what I was saying.
Right. Can you explain what you were saying? Why is Edinburgh disadvantaged against Manchester?
The point is that there should be no disadvantage; there should not be a disparity between transactional activity in those two markets. Edinburgh has traditionally been second to London as a place for commercial real-estate investors to put their capital, but that is evidently not happening. I am painting a picture in which, for some reason, LBTT is partly, but not entirely, why activity is not coming into the Scottish market as it might otherwise do.
Has anyone studied that to see what the reasons are? Has anyone spoken to the developers?
There is a lot of anecdotal information about that activity. I do not know whether anybody has researched it formally—I do not think so. It is about the number of transactions and the extent of interest in the commercial property market, which is not the same here as it is in Manchester. More research needs to be done on that.
Fair enough.
My final question is for Mr Hogg, who talked about guidance from HM Revenue & Customs and the fact that Revenue Scotland is not following HMRC guidance. Given that our law is slightly different, should Revenue Scotland be following HMRC guidance?
Did I say that about following HMRC guidance?
Was it not you who said that?
Actually, I see that it is noted on the bottom of page 3 of our submission. The point that we are making is that it was reported by one of our members that Revenue Scotland is not following HMRC guidance, even where the rules or provisions for LBTT and SDLT are identical. Our member raised the concern because they could not see any logical reason why that guidance would not be followed.
We were invited to comment on the performance of Revenue Scotland. I am merely reporting what was reported to us.
Fair enough. I accept that. For me, one of the fundamental differences is that the general anti-avoidance rule is slightly different under Revenue Scotland to what it is under HMRC. Even though a little bit of the detail of the legislation might be the same, the overall picture is different. I suppose that I would question whether Revenue Scotland should always issue the same guidance to HMRC, even in a similar situation.
That goes back to the point that I raised before about the opportunity for people to have an informal discussion with Revenue Scotland to get that understanding. Apparently the issue evidenced itself in complex transactions, in which people are looking for some guidance on how particular cases would be handled by Revenue Scotland. Again, the evidence that came back to us is that there was not as much access as had been available previously to have discussions about how those transactions should be presented or calculated. The call from our members has been for more opportunity to have those discussions.
Are most of those people trying to avoid tax?
I have absolutely no idea, but I suspect not. It is lawyers who have presented that data to us. Any inference that they are trying to avoid tax would be misplaced.
Okay, thank you.
That concludes questions from the committee. Thank you for answering comrades’ questions.
10:57 Meeting suspended.
We will now continue our scrutiny of the devolved taxes by taking evidence from the Scottish Fiscal Commission. I welcome Lady Susan Rice, Professor Andrew Hughes Hallett and Professor Campbell Leith. Welcome once again to the Finance Committee.
Before we go to questions, I invite Lady Rice to make a short opening statement.
Thank you, convener, for the chance to provide some comment today. You specifically asked us to share some views on LBTT forecasts relative to the outturn figures to date and to comment, to the extent that we can, on the impact of forestalling. You also asked for commentary on the Government’s forecast on landfill tax. We have given you a short paper on those issues.
I have a couple of reminders. In relation to landfill tax, as I have said in writing, to date, we have only one quarter’s worth of data. We believe that the data on the second quarter comes out at the end of this week or next week. Because there is a seasonality factor with landfill data, that is very little material on which to base judgments. We have done what we can, but I just want to caution the committee. I am sure that all the members understand that it is an imperfect science at this stage.
My other comment relates to LBTT. I thank the committee for accepting the most up-to-date version of our paper yesterday. We prepared the paper for the point when we needed to submit it to the committee using the six months of data through the end of September. However, as the data for the seventh month—October—recently came out, we thought that it was important to add that in. That is the main change in the most up-to-date version of the paper, although there are also a couple of clarifications. The change is important, for reasons that we can go into more fully.
With LBTT, there are factors of seasonality and forestalling. Every additional month of data we have, the clearer some of those factors will become. Again, looking at the issue midway through the year is an imperfect science. I believe that the committee understands that, but it is important to note it for the record.
We are very happy to take questions.
Thank you very much for that brief opening statement. As the witnesses will know, I usually ask the opening questions, and then I open up the discussion to colleagues round the table.
In your letter of 19 November, you said:
“There is a challenge in making a judgement about a year’s forecast based on part-year data, where assumptions have to be made about the effects of seasonality and/or forestalling on the observed outturn data.”
The committee is fully aware of that. Between the letter of 19 November and the most recent paper, there has been quite a significant change in the outturn data that has been analysed. For example, the minimum figure in the first paper of £215.5 million has dropped considerably to £203.9 million, and the maximum figure has fallen from £247.4 million to £243.5 million. You say that that is because you had an extra month’s data. Have those figures been shared with the Scottish Government?
The most recent report was sent to the clerk at the very end of the afternoon yesterday. We also asked our office to share it with the Deputy First Minister first thing this morning. It is really late-breaking data. It has not been shared beyond the committee and, this morning, the Deputy First Minister’s office.
Thank you—I appreciate that.
You say:
“typically 62.98% of revenues would be expected to have been raised over this part of the year.”
That is used to produce the figures that I have already quoted. There is quite a massive difference. We are talking about an error of potentially almost £40 million. In fact, it says “£31.1m and -£8.9m”, which is £40 million, although the annualised figure in the graph seems to be £39.6 million. Either way, there is quite a huge difference. You are talking about a potential marginal difference of about 16 per cent. Is that right?
I ask Campbell Leith to respond on that.
Sorry, but which figures are you referring to, convener?
Sorry. In your most up-to-date paper, you talk about
“an implied estimated forecast error of between £31.1m and -£8.9m”.
We are looking at an overall tax take of somewhere between £203.9 million and £243.5 million, or a difference of 16-odd per cent. That is a huge potential difference in the forecast, despite seasonality and the fact that there is only a few month’s data. What margins of forecast error would you expect in a normal year for a normal tax? It seems that it is almost a guesstimate.
For this particular tax, in forecasting, the Scottish Government essentially extrapolates current prices and transactions to long-run averages. It then applies a distributional model of transactions throughout the distribution of house prices. From that, it can build up estimates of the projected tax take. No formal statistical modelling goes into the projection of prices or transactions, so it is not possible to assign standard statistical bounds of error to that forecast. Therefore, we cannot formally say that the standard error is X per cent, given the techniques that are employed.
On top of that, for this particular year, because there is a change in the tax regime, we have forestalling effects, which the committee discussed earlier. The outturn data is significantly below what one would expect, given the normal seasonality. That may be because of a temporary forestalling effect, or it may be that the change in the tax regime has permanently subdued parts of the market, so the trend will continue indefinitely. The extent to which there is an underlying forecast error depends on how much one believes that the shortfall in the early part of the year is a temporary phenomenon or an on-going one.
It seems as though the median in the forecasting error will err towards overoptimism, even though the Scottish Government’s prediction was significantly lower than that of the OBR, which means that the OBR prediction was way out, which had implications for the block grant adjustment.
You seem to have taken figures for landfill tax from the UK. When we were considering landfill tax during the passage of the Landfill Tax (Scotland) Act 2014, one of the criticisms that was made of the OBR was that it gave figures for Scotland that were just extrapolated from UK figures. It seems that you have done exactly the same. You talk about the “UK as a whole” in relation to Scottish Government revenues, saying:
“We therefore scale the outturn data assuming such seasonality also applied in Scotland”,
which gives a significant forecast error of between £39 million and £53 million.
The outturn data is Scotland-only data. We do not have the historical revenue data to be able to compute seasonality for Scotland at this stage. It will take a few years of observing seasonality before we can do that for Scotland.
Considering the difficulties that you have had with forestalling, seasonality, the fact that it is a new tax and so on, you have had to do your best within fairly broad parameters. Are you saying that you expect to be able to hone the process significantly as the years progress, making the forecast much more accurate?
Yes.
On the same theme, as you might have heard, I asked the previous witnesses how quickly we will be able to get a clearer picture. They seemed to say, “Well, at least we have one year’s figures.” Will one year’s figures give us a clear picture, or do we need to wait longer than that?
We can perhaps start to see something already. The second table in our updated paper breaks down what you would expect to receive from residential LBTT by month, the liabilities that are generated and the difference—how much we did not receive but were expecting to receive—which is shown in the second-last column. You can see that, in April, the figure for the difference is £10.6 million, which is massive. That figure falls throughout the early months until August, where it practically disappears. Had that continued in August—had the forecast error been zero for the next two months of data— you might have concluded that the data for the early four months was a result of the forestalling effect. However, the fact that it then bounces back a bit in September means that there is a question whether the forestalling effect is truly over or whether it is an on-going, permanent effect. From the data, the best guess is that there is a combination of a little bit of forestalling and perhaps a little bit of a longer-term effect. We simply will not know until it all plays out.
That is consistent with what the previous witnesses said. They did not seem to feel that there is a forestalling effect currently—in other words, people are not hoping that the rates might fall next year. Who knows? There are also other factors that could have an effect, such as the oil price situation affecting house prices in Aberdeen. It could be that LBTT is not having much of an effect, or it could be that other factors are having less of an effect, and LBTT more.
It could relate to the broad economic conditions, forestalling or a permanent effect of the change in regime. Any of those factors or a combination of them would explain the data.
I would make the awkward point that you will never know the answer. You cannot separate out the information, because the people involved in each transaction were not asked whether they were forestalling or whether they were going to do it anyway. Even when all the data is in, you will not know the answer to your question; you can only infer the best that you can from the data in the second table in our most recent paper.
We would have to ask all the buyers and sellers, as well as the people who did not sell, what their motives were, which would be impossible.
11:15
You would have to do that to separate out the forestalling effect from other things, such as oil prices bouncing up and down, or whatever they do. Even after all the data has come in on the actuals, we will still not know whether any variation was due to forestalling or to some regular uncertainty, if you like.
I presume that at some stage we—or somebody—should do such a study, but should we wait perhaps five years before we do it?
From an academic perspective, five years would be fine, because by then we would have five years of regular performance. However, from your perspective, you want it done after five minutes, but you cannot do it.
My initial thought when I looked at the numbers was that the forestalling part was over by the end of August, but September cut the ground from underneath my feet a wee bit, although the figures for October look to be more back on track. My guess is that forestalling is largely over, but I am saying nothing about what will happen next February and March.
Even if John Mason’s scenario was magically to take place and you were to ask all the buyers and sellers what their motives were, they might not say, “Oh yes, I did it to save a few bob in tax.”
It is rather like opinion polls.
Exactly.
I went back and read the commission’s October 2014 report, in which you gave a warning about buyer behaviour and the behavioural response to the higher tax rates. I was fascinated by that, because you went on to say—as you have said already today—that the current forecasting methodology does not allow you to forecast that. What needs to change, and how quickly can we change it, to start to model some of that buyer behaviour? I am not thinking so much about LBTT but about wider applications. With more powers coming to the Parliament, I am keen that we are able to model behaviour in relation to taxation.
Essentially, when we do economic analysis, we follow a continuum of approaches. Typically, for short-term forecasting, the techniques that we use have varying degrees of sophistication, but we use data-driven statistical techniques. We take a bunch of historical data and try to extrapolate that in as intelligent a way as we possibly can to get a short-term forecast.
That is fine if everything is behaving as normal, but if the policy regime changes or there is a big financial crisis—if something unusual happens—those techniques become less reliable. In those situations, we move away from data-driven statistical techniques and have to do a more formal kind of modelling that relies more heavily on economic theory to provide structure to the way that we analyse the economy. We can then look at behavioural responses to policy changes.
For example, in the case of the housing market, the current forecasting technique will start with historical data and move towards some long-run average. Of course, there may be booms or busts in the housing market, so we might need to use a more sophisticated model to try to capture them, and the turning points in them, such as the point when a boom starts to crash. That requires what are called dynamic stochastic general equilibrium models, which are far more theoretically driven and can be estimated from the data. Such models are probably less useful for day-to-day forecasting but they are better for analysing policy changes and big events in the housing market. We want to do both: we want to do the more sophisticated, fundamental stuff to inform the evaluation of the day-to-day, short-term statistical forecasting. Doing the latter takes a long time.
Does anybody else have anything to add. Do you all agree?
I do not disagree with what Campbell Leith said, because he has given the correct answer.
More practically, with some other forecasts, we have tried to find out—either the Scottish Government does this or we do back-of-the-envelope calculations—whether the behavioural factors have a statistical impact on the forecasts and to establish whether there is a robust relationship there. A lot of the time, we find that there is not. That is partly because we may have the wrong definition of the variable that we are using. There might be no Scotland-specific data and, of course, if behaviour here is a wee bit different from behaviour UK-wide, differences will emerge.
Such points might be obvious, but it is a bit of a hard job to nail these things down. For a start, you want longer samples. The unfortunate thing about humans is that they change their behaviour; even if some of these factors are always present, behaviour itself is not always regular.
Moreover, with regard to LBTT and one or two other taxes, there is also speculative behaviour to take into account. For example, we are dependent on house prices, which, as we know, zip up and down from time to time. It is very hard to deal with that. I do not think that we will ever get the complete answer, but we might get further in due course.
I am still very encouraged that you identified buyer behaviour as a concern way back in October 2014.
I turn to the cumulative difference figures in the updated assessment. With regard to forestalling, which the convener has already touched on, is the OBR estimate of £20 million reasonable, given that it seems to land in the middle of the Scottish Government’s estimated minimum and maximum?
When the estimates were given, the Scottish Government felt that the figure for forestalling would likely be more towards the lower rather than the upper end of its estimates, and the OBR gave that £20 million figure. I note that, in the latest challenge meeting that we had with the OBR and Scottish Government forecasters, the OBR was contemplating raising that estimate. Looking at the figures in the second table in the updated assessment, which show the trend tailing off from the early months of the year, I would say that it seems that there was substantial forestalling in those early months.
I think that people would recognise that, but I suppose that the question is the extent of that forestalling. Are you at liberty to reveal the order of magnitude of the difference between the original assessment and what is being talked about now?
My understanding is that the OBR is contemplating a change to the estimate, but I do not know what it is contemplating changing it to. However, after only the first four months of data, the cumulative difference is already £26.4 million, which takes us above the £20 million estimate.
I now want to explore with you the fascinating issue of tinselitis. In light of what we know about seasonality, do you think that the estimates of tax revenues from November to year end are reasonable?
The first column of the second table gives the percentage of the annual tax revenues that you should expect to receive in a given month. You can see the tinselitis effect in December, the figure for which is 9.8 per cent. People try to complete their transactions, pay the tax and move in before Christmas; in January, no one moves, so the figure drops to 5.5 per cent. That effect can be seen; in fact, I am going through that process myself right now.
I was just about to say that.
That is fantastic. So you think that the estimates are reasonable.
As far as seasonality is concerned, yes. A standard and consistent approach to adjusting for seasonality was employed.
One can get excited about tinselitis; I do not know what the equivalent effect for the summer holidays is called, but it is, in fact, larger.
I saw that and thought it quite interesting.
There are two peaks.
Everyone assumes that more houses are sold in the summer, but the reality is that house selling happens after that.
The same happens at both times of the year.
Indeed. Thank you very much.
I want to check a difference between your first paper and the second one that came in last night. The convener has already touched on this, but if we take the minimum figure for forestalling, which is £12 million, you say at the bottom of the table on page 1 of your first paper that the estimated annualised forecast error would be £19.5 million. However, according to the document that you sent us last night, your view now is that, again with the minimum forestalling figure of £12 million, the estimated forecast error would be £31.1 million. That is a change of more than £11 million in the forecast error.
The only difference that I can see is that you have—very helpfully, I must say—added in the October figures. I am grateful to have figures that are as up to date as possible, but according to the figures for October, we expected £21.3 million in revenues and got £20 million, which means that we are £1.3 million down in a single month. Nevertheless, your forecast error has suddenly shot up. Can you explain that?
There was an error in the previous table. We lifted the £20 million forestalling figure from the spreadsheets and labelled it as the £12 million forestalling figure. It is simply a mistransposed figure, for which we apologise. The minimum forestalling figure in the original paper was actually £20 million, not £12 million—it was mislabelled.
So the estimated forecast error should not have been £19.5 million; it should have been higher than that.
Yes. It would be closer to the £31.1 million that is in the revised figures.
Because everything is scaled up from the annualisation figure, if that is 60 per cent, we are not quite doubling the error. You have identified an error. You put that into the annual total and divide it by 0.6, which is nearly two times. That is unavoidable because we do not have the end of the year’s data.
The simple explanation is that there was an error in the original paper—
I am just warning you—
So I should just use the updated paper and ignore the earlier one.
In the original paper, with the £12 million forestalling figure, there should have been a forecast error of £30.8 million and now it is £31.1 million.
We apologise for that, but we did catch it.
Thank you. The OBR figure is £20 million, which has been out there publicly. I refer to the paper from the Scottish Parliament information centre. When I reviewed the documents, I did not notice the £12 million and £37 million figures having been published before. Can you give us some background on where those figures come from? Are they your workings or the Scottish Government’s workings?
They are the Scottish Government’s workings. There is an academic paper by two authors called Best and Kleven, which looked at the impact of the stamp duty holiday that was introduced in 2008, which reduced stamp duty tax rates temporarily. They were reduced for a certain period, which was then extended, and then they went back up again. The natural experiment of tax rates coming down and then being reapplied can give you a measure of the forestalling effects to some extent.
We encouraged the Scottish Government to look into that paper and to look at plots of Scottish data around those key events, the magnitude of the effect and how quickly it dissipated. The range of estimates that come out—the £12 million to £37 million—capture the range of estimates as to the magnitude of the behavioural effect and how long it lasts. That is where the figures come from.
I am grateful for the explanation. Are you at liberty to publish the workings behind that or would I need to ask the Scottish Government about that? Just plonking the figures down is marginally helpful, but it does not give us the whole picture.
I think that it is a Scottish Government document, so it would have to publish it.
That is fair enough. Obviously, I can ask the minister more about this, but to your knowledge is the figure a net figure? We know that some transactions were pushed through quickly ahead of April in order to reduce the tax bill. There is certainly evidence from a number of sources that, to a lesser extent, some transactions were delayed because the threshold was £145,000 instead of what it is in the rest of the UK. Presumably there is a balancing act between the two, coupled with the fact that, as you said well over a year ago, there would be some behavioural impact in the medium to longer term anyway if the tax rate was increased.
There are a couple of factors at play. To your knowledge, does the £12 million to £37 million range take into account all those factors, or did it look simply at the transactions that were rushed through before April?
It looks at the impact on the 2015-16 tax take, so it is the net effect of the tax regime change on revenues raised in 2015-16. It relates to all those transactions that did not take place, possibly because they occurred before March. It does not add back in any revenue gained prior to the fiscal year, nor does it include any permanent behavioural effects; it is purely an estimate of forestalling, which is a temporary phenomenon.
On page 2 of your updated submission, you outline the percentage of annual tax revenue that is expected each month, from April through to March. To take an example, on what basis did you think that we would get 7.5 per cent of the tax in April? Is that based on averaging five years or is it based on just the previous year?
11:30
Essentially, it is based on looking at the seasonality that can be observed in both house prices and transactions in the housing market. It looks at historical data for prices and for transactions. It was run using a technique called X-12, which is the standard technique that statistical agencies use to seasonally adjust data so that they can decompose the historical path of data into trends, cycles and seasonalities. The figures identify the seasonality bit.
The combined effect of prices and transactions being lower in April means that a relatively low amount of tax is expected to be generated in April. The combined effect of lots of transactions and of prices being more buoyant in July means that we get a lot more tax in July.
I completely understand the seasonality point. I was just trying to make sure that you were not basing those figures on, for example, one year’s figures.
It uses a longer time series.
That is helpful.
My next question is on something that perplexes me a bit. The heading on page 2 of your submission is “Updated Assessment of Forestalling”. You have looked at the months for the financial year so far, but you have not looked at January, February and March 2015—or at least you have not put them in the table. By definition, forestalling is a transaction that takes place in advance of a deadline. Surely we cannot have any idea of what happened in relation to forestalling if we do not look very closely at what happened between the announcements of the tax rates and the tax rates actually starting to bite. Why have you not published anything for the months leading up to April?
We have not looked at that. The original paper by Best and Kleven suggests that the effects may not actually be symmetric. If you look at what happens before, you cannot necessarily infer that the forestalling that you observe afterwards is the mirror image of that. There seems to be an asymmetry. Some transactions may be rushed forward; temporarily, more transactions may not happen at all. The level of forestalling that occurs may not actually be tied down.
To be more precise on that point, we do not have the actuals under the old regime and we do not know how much of a regime shift there might have been. I do not mean a shift in tax rates; I mean a shift from one type of tax to another type of tax, or from the old stamp duty to LBTT. That makes it a bit difficult to do.
I looked at this very quickly this morning over coffee at breakfast time, so it is a back-of-the-envelope calculation, but I can tell you that, based on the numbers that I got from your previous witnesses’ evidence—it is conditional on that—at the rough average of the LBTT system about £56 million-worth shifted before April. Therefore, as Campbell Leith said, the effect is not symmetric but there is clear evidence on that basis. The figures on how much transactions and house prices rose can be put together and we can see what the net increase was over what might otherwise have been expected. That is a very rough idea.
I am grateful for that. I was not suggesting that it would necessarily be symmetric, but I do not see how you can do a sensible forestalling exercise by completely ignoring what happened before. If, for example, in the lead-up to the tax, transactions and tax collecting were exactly as expected, presumably the differences that we are seeing would be down to something else entirely. I do not understand how that can be ignored completely.
There is substantial evidence that transactions went up before the tax and in the bands that were most affected. We have not quantified the effect here. Another way of looking at it is that, if it is purely forestalling, you should see the effects trickle out towards the end of the year. That is another way of assessing the effect, and that is the approach that we essentially followed.
I assure you that we discussed that point, and you are absolutely right to raise it. We have not ignored it, but we have not had the data in the detail that would be needed—for example, comparisons have to be made by tax band as well—so it is slightly more complicated than just the sum of money. The sum of revenue that Andrew Hughes Hallett has stated is at least a guidepost.
Another complicating factor on which we do not really have information is whether house prices change in the circumstances. If people who have houses that they had hoped to put on the market in certain bands are worried about selling, that might change. The exercise is rather more complex, but the point is well taken and we will continue to look at the issue.
I am grateful for that. In my mind, there has been a bit of confusion about the forestalling exercise. Who is leading on that? Is the Scottish Government doing all the work and asking you to review it, or is it asking you to do the work and to show it your workings? There might not be an obvious answer but it is not clear to me whether the commission or the Scottish Government is leading. I wondered about your interpretation of the situation.
The Scottish Government did the work to produce the £12 million to £37 million estimate. We recommended that it look at the Best and Kleven paper in doing so, and we reviewed its work extensively. We then critiqued the work in the style of an academic seminar by having it present its work. We asked lots of questions and got very picky with everything, to toughen up that forecast.
May I just go back to the wording of your question? The Scottish Government has not asked us to do work that it would then incorporate, and I would not expect it to. That is not appropriate as far as we are concerned. What we do is make challenges along the way.
On the Best and Kleven paper, we wanted to see a somewhat sophisticated view so we recommended that the Government take that forward. It did not have to do that, but it did.
In your in-tray at the moment, is there a task that has been given to you to look more at forestalling, or is it a case of waiting until the Scottish Government comes back to you?
You have given us that task, I think.
The table that we have is incomplete. As more data emerges, we will fill it up, which will give us a lot of information. As resources come on stream for the Fiscal Commission and we extend the depth of the modelling work that we are looking at, the behavioural effect on the housing market is an obvious area for us to explore.
Thank you.
You said that the maximum forecast error is likely to be £31.1 million, but the cumulative difference in your table is already £31.6 million and every month so far the difference has been higher. It has varied from £10.6 million to £0.1 million. What makes you think that every other month will not be exactly the same? You seem to have concluded that the cumulative difference will be down to £31.1 million. We have not been given any workings to show how you produced those estimates. Why do you think that everything in the second half of the year will be much rosier and come down to that maximum of £31.1 million, which is not exactly good news?
I can explain that. The original forecast of £235 million was a forecast before forestalling. The forecast errors are therefore the forecast errors after the estimated effects of forestalling have been added in, so it is the forecast error relative to a world without forestalling.
Right.
We hope that that will help you and us to understand the strength or reasonableness of the forecast methodology. Forestalling makes that harder to understand during this first year.
We were asked to evaluate the headline forecast figure of £235 million, and that figure does not include the effect of forestalling. In assessing that, we adjust the outturn figures to add back in any estimates of forestalling and we project that forward for the year and compare it to the forecast to get the forecast error.
Okay.
It might be relevant to the conversation that we will have later this morning, but this is an illustration of how we work. We have to assess whether a forecast that came from the Scottish Government is reasonable, and that forecast does not include the effects of forestalling. Normally, you would just compare the forecast to the actual figures as they come out but the actuals have had the forecasting taken out of them, so if we are to assess the Scottish Government’s original forecast, we have to put it back in, and the forecasting of forestalling adds errors.
I differentiate myself slightly from my colleagues on that. After next April, when the numbers are in on the actual outturns, we will turn out to have been assessing something that nobody is interested in because it is a counterfactual—it is something that will never happen. Forestalling is actually there and that raises a whole sack of questions.
What we are doing is making a kind of forecast of what we think the end of the year would have looked like had there been no forestalling, as an aide-memoire—I hesitate to use the word “benchmark”. It is a calculation for checking that what the Scottish Government is doing is about right even if, in the event, next year we will be talking about a completely different set of numbers. There are inevitably a lot of assumptions in it.
I just want to bottom that out. I am looking at the cumulative difference if forestalling were to be factored in. It is £31.6 million just now, and if you were to take the figure of £12 million you would be looking instead at a difference of £19.6 million; that would be if you added in the forestalling effect to the cumulative difference at present.
In the second table, the cumulative difference projects how much tax we would expect to receive, given seasonality and assuming that there is no forestalling. We compare that with the actuals in the second table and that can implicitly give a possible measure of forestalling—it could be those permanent effects or some other effect. If we took that cumulative difference to be the estimate of forestalling and then went back to table 1 and redid the calculation, assuming that all those effects had finished at the end of October and that the rest of the year was just going to proceed with normal seasonality and no further forestalling, the forecast error would effectively be zero.
Yes. That is what I was trying to get to.
The crucial question is whether forestalling has finished.
Gavin Brown asked about the comparison of the months leading on from the announcement of the bands. How easy would it be to get hold of that data? One would assume that, if it were possible to disaggregate Scottish data from the HMRC figures, the OBR would have come up with a very different forecast. It would have been able to look at Scottish data as opposed to just extrapolating from the UK-wide figure, which is what it appears to have done for its forecast. How easy or otherwise would it be to get hold of Scottish data for those first three months of the year?
I do not know whether you were asking me or Professor Leith, but I would have thought that it would be difficult because HMRC does not release its data very easily. The earlier period will be HMRC data because Revenue Scotland was not operating at that point. In this particular case, it might be rather difficult.
I want to ask a purely hypothetical question. Earlier, we heard evidence from the Scottish Property Federation, which thinks that the tax rates should change and that the Deputy First Minister should announce that as part of his budget. Would we see a different effect—purchases and sales being delayed rather than brought forward—before such a change came into effect, and might that have a knock-on effect on the expected and actual revenues in the months of January, February and March?
If people start expecting a tax cut, the forestalling effect works in reverse—so, yes.
Thank you.
I am looking for clarity on the question that Gavin Brown asked. I want to be clear about what the figures represent. Did the commission receive data from the Scottish Government and apply its own methodologies to come up with the forecast, or is it analysing the reasonableness of the figures and the forecast, so that the figures are, in effect, a Scottish Government forecast of LBTT revenues?
The figure of £235 million is the Government forecast for LBTT before forestalling. The figures of £12 million to £37 million are the Scottish Government’s forecast estimate of the extent of forestalling. The rest of the analysis is a matter of taking the outturn data that we have to date and trying to adjust that sensibly to account for seasonality and that range of estimates of the forestalling effects. From that, we try to infer what we might expect outturn numbers to be for the rest of the year, as a basis of comparison with the forecast.
Is that work that the commission has carried out, rather than the Scottish Government?
That is right.
Yes.
It is a forecast by the commission. If we asked ministers to explain their estimates, forecasts and figures, we could expect a different set of figures and forecasts from the ones that you brought forward today from that analysis.
We are extrapolating the part-year outturn numbers to make them into sensible annual outturn numbers. It is a short-run extrapolation of outturn numbers. It is not a forecast in the same way as the Scottish Government is forecasting what LBTT revenues will be for the next year.
If we asked the Scottish Government to make an extrapolation on the same basis, would it use similar models and analyses? Is that something that you are aware of? Could it be something very different?
That is the way that we are doing it—
The Government would choose its own way to do it.
Thank you very much. At this point, we would normally have a break, but the same set of witnesses are with us for the next item. Would colleagues prefer just to plough on—and the witnesses? The witnesses would like a break, so we will have a short break.
11:46 Meeting suspended.