I wrote this yesterday about HMRC's tax gap estimates for this year, which had then not been published:
I gather HMRC will be publishing their latest tax gap figures this week. The last are here.
I can make a fairly confident prediction. The figure will be around £33 billion, simply because ever since this data was first published in 2010 the tax gap has never been below £27 billion and has never gone above £35 billion. So I predict about the same again.
I also predict much trumpeting that as a proportion of total tax paid this will be the lowest tax gap ever. Because this also seems to happen now, as a matter of ritual celebration.
Was I right? Yes, is the simple answer. The HMRC estimate of the tax gap has been published today. The tax gaps is reported to be £35 billion and there is much trumpeting that at 5.6% of supposed tax yield this is the lowest claimed figure ever.
And for all the reasons I explained yesterday, this is utter nonsense. The real tax gap is in my opinion about £90bn (and I stress, all these numbers are spot estimates in ranges).
I am also amused to note that I am not alone in saying so. TaxwatchUK have also questioned the data for reasons quite similar to my own. And they have noted this:
The claim that the UK has the lowest tax gap in the world is highly dubious. HMRC themselves say they are the only tax authority in the world to publish an annual estimate of the tax gap that covers a comprehensive range of taxes. Any international comparisons are not being made on a like for like basis and are therefore pretty meaningless.
We really do deserve better than this. When the tax gap is very obviously very much bigger than HMRC admit they undermine the whole credibility of their management of the tax system and the hard work of their own staff. £90bn is not the same as £35bn and I know which one is much more likely to be right, and it's not HMRC's data. If they can underestimate the tax gap by £55bn what else do they get wrong?
Thanks for reading this post.
You can share this post on social media of your choice by clicking these icons:
You can subscribe to this blog's daily email here.
And if you would like to support this blog you can, here:
Buy me a coffee!
Not withstanding that HMRC and others heavily criticise your “tax gap” methodology, I have some points of my own I would like to raise with you.
Firstly, HMRC have done a fairly comprehensive study on the shadow economy:
https://www.gov.uk/government/publications/the-hidden-economy-in-great-britain
They say that roughly 50% of it would not be due for any tax whatsoever, and then a large further part of that shadow economy would be liable for only some lower rate of tax. You have assumed that the whole of the shadow economy should be paying the average tax rate of 33.1%.
Is this 33.1% tax rate not a terrible assumption to make? Have you done any research to justify this assumption?
Secondly, I notice that all of your tax gap estimate relies on the work of others. Have you done any original research into this estimate, or is it solely the work of others?
Thirdly, you state that your £90bn number is an estimate. What is the potential range of error on this estimate and how have you calculated it?
Lastly, total UK tax receipts were around £600bn. Are you seriously suggesting that the tax gap is 15%? Suggesting that as a proportion, it is larger than the size of the UK shadow economy which is around 10% of GDP?
Jacqueline
Thank you for your comment
First, the HMRC review of the shadow economy is not comprehensive. It looks at those in marginal employment not declaring tax. It estimates rather more are doing so than the estimates in the tax gap imply. But it does not seek to explain how almost 10% of turnover is lost, and yet it is because VAT is not paid on it. So the shadow economy has to be vastly higher than the sum they estimate. If you set out to look at moonlighting and small scale evasion you will find a small answer, and they did. It did not explain the VAT gap. So this review massively understates the issue. It was, to be candid, misleading.
And as I suggest, the sums evaded as implied by the VAT gap are additional to GDP. Why should that additional near 10% have almost no tax on it? Why should it not bear tax at the same rate as the rest of the economy? Why might it not be the top p[art of many people’s incomes, as in fact it will be in a great many cases, and therefore should be likely to be subject to higher rate taxes? I think my estimate of tax rate is low, not high for that reason. Tax evasion is now known to be a game for the rich, not the poor, and you assume otherwise.
Second, I have relied on peer-reviewed work published by the EU, IMF and an academic journal. Is that a problem?
Third, I have indicated my range estimates.
And yes, fourth of course the tax loss is bigger than the rate on evasion alone as there are other causes of loss. And the rate of loss I suggest is in fact commonplace in many equivalent economies, as EU based work shows.
I rema9on very confident in my work.
Richard,
I will try and deal with the points in order again:
First,
HMRC’s review of the shadow economy includes (I quote):
key characteristics of those within the hidden economy
key business practices of those working within the hidden economy
the scale of the hidden economy and its distribution across key sub-groups
the monetary value of income that is not declared
behavioural drivers of those choosing to participate in the hidden economy
those who might be buying from the hidden economy and what is purchased
So seems to be fairly comprehensive and rigorous. It also seems that you have done no investigation of your own into the shadow economy past using other people’s estimates of it’s size. Simply stating your assertion that things must be the way you say they are is not sufficient. It is just your opinion until you can back it up with evidence.
For example HMRC go to great lengths to explain where their VAT gap comes from. Of the £12.5bn, 17.5% of the total is VAT debt and 30% is from government itself. So claiming all of the VAT gap somehow relates to the shadow economy as you do is immediately false.
So back to the question at hand. Is a 33.1% average tax rate a reasonable assumption to make, and what evidence do you have for it to be so? You have made another statement in it’s defense “Why should it not bear tax at the same rate as the rest of the economy?” but this is meaningless as we don’t tax people on average rates – we tax them on tax due.
Given HMRC’s findings, that half of the shadow economy would fall under the tax threshold, your 33.1% rate would immediately drop to 16.5%, and in reality would likely be lower still as much of the rest would fall into lower tax bands.
In this case, HMRC’s research shows that whilst some tax would be due from the shadow economy, it would be at much lower marginal rates for the most part, and a large portion of it would not be taxable at all.
Second,
It is fine to use other peer reviewed research as part of your work, but I think you are missing my point. You are making rather bold claims about the veracity of your work – saying it is correct and HMRC’s work is not.
But it seems to me that HMRC have taken a huge amount of time and effort to look at the large amount of data in significant detail. Unless I am mistaken, you have not done so. You have only really looked at 3 data points – none of which directly relate to the tax gap itself.
The only “original” work you have done is take an estimate of GDP, estimate of the size of the shadow economy and and an assumption of average tax rates and multiply them together.
None of those things actually in themselves have a direct correlation to how much tax might be missing.
Third,
Again I think you are missing my point. You have indicated a range of estimates you have used to craft your £90bn number, but you have not indicated the range of error.
It is normal in research work to make it clear the level of statistical uncertainty in the method, which HMRC do. Specifically so as not to mislead people by making a bold headline claim only to find that the confidence range is so large as to make the claim meaningless.
So please could you let us know the statistical confidence range you are using – percentage terms or +/-. It is a standard part of any research.
Fourth,
The answer you gave here makes little sense.
I asked you if it was sensible to make claims that the total tax gap is 15%, when the size of the shadow economy in the UK is only 10% of GDP. It does not seem realistic that the shadow economy is somehow proportionally more productive than the regular UK economy – which is what your claims suggest.
Also, given that the Tax Gap is purely tax that should have been paid, but has not been I think you are making another obvious error.
You make a claim that avoidance is £11bn, and add that into your £90bn total. Avoidance is legal and is not part of the tax gap, by definition. So adding it in means immediately you are changing the terms of reference.
To Finish,
I fear you haven’t actually done any research into the tax gap. Certainly this estimate of yours relies only on three numbers – GDP, the size of the shadow economy and using the average tax rate as an assumption. I highly doubt the reality of the situation is as simple – which is why HMRC go to such great lengths to investigate the tax gap (which the IMF and EU both commend as the best in the world) and the size and nature of the shadow economy. You have done no such task.
Given that you haven’t got any data of your own to suggest your estimate is more correct than HMRC, and there are obvious flaws and huge assumptions made in your methods, isn’t it incredibly arrogant to suggest that you are right and they are wrong?
Dear Jacqueline
It’s rather hard to know where to begin here, there are so many errors in your claims
Re the supposed shadow economy, what you are ignoring is that tax evasion is much broader in scope than the activity investigated in the HMRC report you used for your claim – and you then seem to be mixing your claims. With respect, you can’t do that
Just as you cannot say 30% if the VAT gap comes from government when the report says 30% of the VAT liabiloity comes from business and government activity – which is nothing like the same thing
And you cannot claim that the missing GDP is lowly taxed when evidence is that this is not true – those on low incomes have as high a tax rate as those on high incomes
And nor can you say using a peer reviewed paper is wrong, and that the IMF and EU were wrong to publish them
And I have indicated some ranges for error – but it is true that if this was an academic paper I might have done more. But it is not one, and so your claim is spurious, and nor am I going to go back now and revisit the underlying data for you – the references are all available if you want them in the linked report
But why am I really not going to do this? That is because it is apparent that you really do not either know what you are talking about or your claim is simply politically motivated. How do I know? Because you say I cannot include tax avoidance in the tax gap and that this changes the terms of reference, and I am therefore wrong. But HMRC do include tax avoidance in the tax gap, and legal interpretation too (which is also about avoidance). In that case your claim is obviously untrue. And you have not actually read the HMRC report at all. And nor do you really want answers.
So, with respect, I will not be engaging further as your claims are actually entirely unsubstantiated and not worth debating.
But they do rather suggest those saying my work is wrong have a great deal more to do to show why
Richard
Richard,
Claiming my questions to you are politically motivated is pure hypocrisy, as this tax gap estimate of yours is purely a politically motivated attack on HMRC and government, just as the last one you issued was.
I am concerned that when you make these claims they get used without proper scrutiny, and it is very apparent that they need a lot – because you haven’t done any research. All you have done is multiply a few numbers together, and then dissemble as to why this is somehow more accurate than HMRC, who actually look at the underlying tax data.
As to your specific points:
The HMRC report on the shadow economy is very comprehensive, as I have pointed out. It covers all the potential tax loss areas from the shadow economy. It is then wholly incorrect of you to claim that they don’t.
Your work doesn’t manage to make any breakdown of any activity within the shadow economy. All you do is make a broad claim of how much tax is lost to it, with no understanding or analysis of the shadow economy itself. So how are you able to claim that HMRC miss things? You have no evidence of your own whatsoever. You simply don’t know.
You also claim that I am wrong to say that the missing GDP should be taxed at lower rates, because people on low incomes have the same tax rates as those on high incomes. Indeed those on low incomes do often get taxed at relatively high marginal rates, but what you have tried to do is a deliberate reframing of the evidence. HMRC’s investigation of the shadow economy found that almost 50% of the activity would fall under the zero tax threshold.
Which means it is not taxable, which makes your average tax rate of 33.1% far to high.
In fact, taking this into account, if you were interested in accurate work and research, rather than slapdash headline grabbing numbers, you would immediately modify the size of the shadow economy you used by subtracting the part of it that wasn’t taxable. You would then make an effort to calculate what effective or average tax rate the rest of the shadow economy should have because as HMRC state much of it will be at marginal tax rates rather than headline tax rates, which will lower your numbers further.
I suspect you haven’t done this partly because you are too lazy to do so, and partly because it reduces the value for your answer by a huge amount – which lessens the political impact of your claim.
I asked about errors as you are presenting for the same reasons. You are presenting this work as fact, yet are unwilling to discuss the size of the potential errors. Indeed, the only admission to error you make is the range of initial data you use (of which you chose the higher number of two, because “it is a better assumption”). Just because you have used peer reviewed papers to get to your starting point doesn’t mean you haven’t made huge mistakes.
You have presented this number without research or evidence. It is wholly right for people to be able to question it, especially as you have in the past and I am sure will in the future attempt to make political capital out of it.
Regarding avoidance,
HMRC include tax avoidance in the tax gap, but your definition and theirs is totally different. They only include “exploiting tax rules to gain a tax advantage that parliament never intended”. Loopholes, in short. Your version of avoidance is much broader, and not what would normally be termed avoidance. For example, HMRC losing legal disputes (which is worth £5.3bn of your £11bn avoidance number) is not avoidance – it has been tested in court and declared legal. So by definition cannot be called avoidance.
If you had bothered to read the report properly, you would see that avoidance does cover BEPS already, so you have double counted there, and “incentivised incorporation” as you call it is perfectly legal tax planning, and therefore not avoidance.
So of your £11bn avoidance number:
£1.7bn comes from HMRC
£1bn comes from BEPS, which HMRC include above so you have double counted
£5.3bn comes from HMRC losing legal cases. Legal tax planning is not avoidance.
£3bn comes from incorporation. Legal tax planning is not avoidance.
So you seem to have an 85% error on this part of your estimate alone.
I have substantiated all of my claims. These are specifically that you have done no original work or research to come to your claims, and that the assumptions you use are wrong.
When pushed on this, all you have done is make more claims, but provide no evidence. You could simply state that you have done no research on this and that this estimate is purely based on the work of others.
I also notice that as soon as I have started questioning your work, you have become defensive and started attacking me, claimed that my questions don’t need answering and in your words “not worth debating”. These are all the hallmarks of the response people make when they have been caught red-handed.
So, to finish:
Your tax gap estimate is designed to aim for as high an answer as possible to grab as many headlines and score as many political points as possible.
Unfortunately, it stands up to very little scrutiny. The large figure for evasion from the shadow economy relies on no data of your own, and makes huge and obviously (from HMRC’s data) incorrect assumptions, notably on the average tax rates. Your avoidance numbers also make huge errors.
As far as I can see, you have done no original research so are reliant on the work of others for the inputs into your calculation. The methodology of the calculation itself is highly suspect.
This clearly deserves scrutiny because you are presenting the result as wholly accurate, and more so than HMRC who have looked at the real, underlying data to present their case. I can’t imagine the level of arrogance it would need to make that claim.
Thank you for agreeing that your attack is politically motivated
My own is motivated by a desire to get facts right
Yours are wrong. I will just address those you raise on avoidance
First, HMRC say avoidance is in the tax gap, so you are wrong
Second, the figure for £5.3bn is HMRC’s on issues not litigated bt disputed
Third, their figure for avoidance relates to DOTAs style schemes. BEPS issues are not covered. So your claim I double count is wrong
And the figure on incorporation came from the Office for Budget Responsibility and was used as reason for action on this issue
These are highly relaible sources, all given
You note none
And you get your facts wrong
And I now note you argue HMRC are wrong although it is not at all apparent that you realise that is the case, yet you maintain they are right and I am wrong
Actually, you have not a clue what you are talking about and have read nothing at all to really inform yourself
Please don’;t waste my time again with such nonsense, because that is what you have written
I demand a public enquiry please.
Going back around 6 years, this quote is attributed to you
“Richard Murphy’s report for the Public and Commercial Services Union (PCS) estimates the tax gap for 2013/14 at £119.4bn”
This is around £90bn for 2018/19. In your view, what are the two biggest factors to have caused the tax gap to shrink by a nominal quarter in 5 years ( more in real terms ) – more resources, better targeting, better data compare to what was available 5 years ago, better management, a better public understanding of their responsibilities, a specific area like consumption taxes, or taxes collected by employers, I can’t think of any specific improvement to explain it myself.
The biggest factor by far is you read the report from 2014/14 is that I have taken the low figure for bad debt HMRC say they suffer rather than the annual write off they actually make which they say arises not because the sum was due but because they overestimate sums owing and then have to provision it
I am being generous on this occasion
But the question must arise, why they overestimate sums owing to provision it? Did they think it was owing or not. But this is the main difference.
I appreciate the reply, but what’s the value of the change, and surely the measurement estimated in 2013/14 is a better one? It seems to me that in 2018/19 you’re not measuring it right and are currently accepting HMRC numbers over your own previous method calculation which was presumably superior.
What’s the 2nd biggest change over the last 5 years to swing the TaxResearch LLP tax gap measurement to the much lower figure.
I have pointed out the main reason for the change
That figure explained more than £20 billion by itself
And I have suggested that the size of the shadow economy has fallen
The overarching narrative behind the figures reported annually by HMRC is the need to demonstrate the success of the continual shrinkage of the department, as shown by two mantras trotted out by its senior leaders – “more with less” and “new ways of working.” Now if that isn’t political I don’t know what is.
The truth is that HMRC is creaking, and the primary reason for that is the continuous drive to shed more and more staff.
The other thing that needs to be said about these figures is that the comparison also needs to be made between what HMRC says is the annual tax gap, and how much of it says it manages to tackle in terms of the results of its compliance activities (excluding, of course, the part of the latter HMRC ascribes to estimates of future yield resulting from those activities).
Tax avoidance and evasion cause market distortions – evaders and avoiders can gain competitive advantage from not paying the right amount of tax. Are the apologists content with that?
I entirely agree with you as to the motive for this sham
The way academia works is that one person presents a case, and everyone else tries to pick holes in it. When the hole-picking stops, both sides should be happy that what is left is closer to the truth.
Two thirds of Richard’s estimate is based on four numbers
* the UK’s GDP
* an exchange rate
* the estimated size of the UK’s shadow economy (as a percentage of the real economy, in GDP terms)
* the tax rate payable on activity in the UK’s shadow economy (also in GDP terms).
I trust we won’t quibble too much on a Euro-Sterling exchange rate, but:
(a) GDP is heavily criticised: it is notoriously hard to measure with any accuracy, somewhat arbitrary in what it counts and does not count, and so misses substantial activity in the economy, but let’s take it as the best number available. So, what errors might there be in that figure? Might it be 5% or 10% larger or smaller? If not, why not?
(b) the size of the shadow economy is contestable. Is it 5%? 10%? 15? 20%? Why have you assumed it is 10% and what errors might there be in that figure? Might it be 7%? Or 13? If not, why not?
(c) the tax rate payable on activity in the shadow economy might be zero (below VAT thresholds, within income tax and NICs exemptions), or might be 45% or more. By its nature, much of it will be at the lower end of the scale. A considerable amount of work would be required to accept that the figure for the shadow economy is exactly the same as the 33% figure for the whole economy. So, why have you assumed it is 33%, and what errors might there be in that assumption? Might it be only 10% or 20%? If not, why not?
I’d be really interested in some proper, peer-reviewed, published work trying to work this out, rather than some back-of-the-envelope calculations. Can someone point me towards some published papers?
Richard’s strongest point, it seems to me, is suggesting that HMRC’s estimate could be an underestimate given the bottom-up way they calculate it, and that there is some worth in trying to do a top-down estimate as a double check. Indeed, the more estimates, using different methodologies, the better. One does not need to be politically motivated to suggest that both estimates could be improved.
Andrew
Thanks for your comment: I have blocked ‘Jacqueline’ – and the response was very cl;er evidence that she is a seasoned troll who simply repeats claims, time and again. I engaged with an open but suspicious mind; my suspicions were justified.
You, however, raise fair points.
You are right that GDP is dubious: I noted the point. I made clear I had undertaken an investigation on this. It is likely that the vast majority of the shadow economy is additional to GDP based on that enquiry
I noted the range in shadow economy estimates – which are surprisingly small. They confirm the HMRC VAT gap. The estimate I use is, then very fair
And the tax rate may be low – that paid by those on low earnings is at least as high as that on this with high earnings whilst to pretend that people are either in / oiut of the shadow economy is just wrong: most of it is top slicing i.e. not declaring taxable earnings that form the top part of income and so the tax rate should be above average, not below. My assumption is reasonable in that case – and it is not just income tax evaded, of course. It is VAT, IT, NIC, excise duties, CT and more.
More work is needed – but that needs access to data – and the ONS and Eurostat now officially will not confirm what part of GDP is from the shadow economy and so good luck with that
And candidly, I suspect it would simply say I am right
Because I am very sure I am
I have to agree with Jacqueline here regarding the £67bn number for the shadow economy.
For what’s worth, I’m a statistical economist and have a bit of experience in this topic.
Shadow economies are usually estimated using MIMIC models. Certainly the sources you have used are MIMIC.
MIMIC has an error of +/-15%.
Nominal GDP estimates typically have an error rate of +/-2.5%. I’m not sure if you have used nominals or PPPs but that would probably explain the large differences you have in the numbers used.
From work I have seen on Germany, which has an average take rate of 37.5% and a shadow economy size of 10.4%, the effective tax rate on the shadow economy is little more than 12% due to most of it being below taxable and marginal tax bands. Statistically the majority of actors in the shadow economy are low income deciles.
From this you would get a +/-70% error using your method.
Combining those errors and you get to a 100% error margin.
So in terms of your £67bn, you should really be stating a range of £0 to £134bn, but with a heavy accent to the lower side.
So explain why all sources closely agree on the UK shadow economy
And do not on your claim for the size of the German shadow economy, by the way?
And I did not just use MIMIC estimates
And instead of using stats – why not using the accounting logic? If VAT is not paid – and it is agreed it is not – how can the same income then be subject to any other tax?
And with respect, you too make the gross political assumption that the shadow economy is made up of those on low incomes and ignore evasion. The poorest do not evade VAT at the rate the EU report – VAT registered businesses do that – so you are quite simply wrong and fail to define your terms correctly
I regret to say that your claim simply makes no sense – because you are misdefining the shadow economy to consider it as unreported low paid employment and earnings when in fact it is vastly more than that or the VAT gap could not be what it is
I’m not sure which other estimates you have used for the size of the shadow economy, but the EU source is definitely a MIMIC model. Most are, and typically they are the best.
Because of the complexity of the problem, in statistical terms they are still in the educated guess category. Anything at the 70% confidence level like these things has to be treated with a great deal of care.
I don’t think your logic is correct by using VAT and then extending it to all taxes.
VAT is a transactional tax with no tax bands. So low income people who would otherwise pay little or no tax still pay the full rate of VAT. Which is why it is normally seen as a regressive tax.
You can’t combine two different tax types (transactional and boundary types) and treat them the same way. They are not homogeneous and not correlated.
I’m not sure what political assumptions you are talking about, but data is available on the composition of shadow economies. The ILO has a large amount which you can access. As Jacqueline pointed out, the UK’s shadow economy is to a great extent operating below the tax thresholds. in Germany too, the great proportion of economic activity in the shadow economy operates at the smaller end of the spectrum – but you have used a basic mean tax rate.
By the same logic, you have made the assumption that the shadow economy is just an exact extension of the regular economy. Do you have the data to justify this?
Let’s look at it a different way. Starting at your 33% midpoint, what are the likelihood that the actual average tax rate for the shadow economy be higher or lower, and what would those extreme limits be? It seems unlikely that this average tax rate would be much higher, but there is plenty of evidence to suggest that it could be lower.
Going back to the main point. You have a !5% error from MIMIC, 2.5% from GDP and let’s be generous and say a large but unquantified error from your average tax rate. Errors compound, not sum, so you are looking at a minimum of an 18% error before you add in the error from your average tax rate.
This means you are working at the 65% confidence level at best (assuming zero error on your tax rate), which is basically a stab in the dark.
A 3% difference (and my guess is that your error is going to be significantly larger than that, but I don’t think you have the data) in your average tax rate equates to a 10% error function. This would put your total error at 30% and confidence level down at 40%, making your estimate meaningless.
HMRC’s data is mostly +/-5% (60% of it) so at the 90% confidence level. The rest is at the 80% level.
Errors get bigger, not smaller, so how on earth, given that you start with data which is already more likely to be wrong (bigger error from MIMIC alone than HMRC’s data), do you make the claim that your estimate is more accurate than theirs?
Basic statistics say you just can’t make that claim.
With respect, you clearly do not understand this issue.
I accept there are margins for error in MIMIC
But the EU data used for VAT gaps is not MIMIC, as far as I know. Two sources I use are.
And you simply are not getting the point on VAT. If for round number’s sake 10% of VAt is lost that is not the result of people doing dog walking in the shadow economy because they would never be AVt registered – and the VAT threshold is allowed for in the VAT gap calculation. That gap happens because of a) some bad debt (I agree) and b) very largely unrecorded income that should be subject to VAT. So of larger enterprises.
And they cannot then reintroduce that into their accounts to pay an9y other tax on it so it remains unrecorded
It does pay people in the shadow economy – some of whom are picked up by the data to which you refer
But it should at least be subject to CT or PAYE (since most will be by shadow companies, not shadow people)
And you are wholly ignoring this and are pretending that the issue only relates to micro entity losses
It does not – they are a small part of it
The real issue is a massive amount of unrecorded business
And that will be top parts of income with CT, NIC, VAT and IT payable – hence why my tax rate is if anything too low
You can argue all you like about stats
But first you have to understand the issue and you are not. This is macro, not micro
And you’re ignoring why the VAT gap does define this issue – and it’s based on macro data
Stop doing bottom up, in other words: it does not work
With respect, I think I have a fairly good understanding of the issue, working with this kind of thing full time.
Unfortunately I am starting to get the impression you don’t have a good understanding of this subject.
MIMIC models are models to estimate the size of shadow economies. Nothing to do with VAT. I would have though you would have known this, as you have used the size of the shadow economy as a starting point for your estimate.
You are not getting the point with VAT at all. It is a transactional tax. You cannot directly compare it to income, capital gains or business taxation.
You are probably right that lost VAT means that business is not being recorded. But that does not mean that that business is all taxable. Certainly not in direct proportion to the amount of VAT lost. Which instantly blows your method up.
This is simply because to pay other taxes, you have to earn or make enough money to rise above a taxable threshold. Businesses only pay CT on profits. Income tax has various bands, etc.
You are just making a basic and untested assumption. If this is not an assumption, you will I am sure be able to provide the data to back it up, no?
You seem very keen on untested assumptions. How much unrecorded business is there? More importantly, how much of it is taxable? At what rates? You do not have any DATA for any of this. The data we do have is very clear in showing that much of this unrecorded business would fall below the threshold for taxation.
Which by default makes the average tax rate you use too high.
I actually think it is laughable that you are trying to claim that the average tax rate you use is too low. If so, by how much, exactly? Higher than the rest of the economy? What, is everybody in the shadow economy a top rate tax payer now? Again though, you have made a claim with no evidence to back it up.
As for the statistics – it’s not an argument. Mathematics is mathematics.
By definition, the starting point for your estimate (the size of the shadow economy) is statistically less accurate than the data HMRC uses. It is therefore impossible for your answer to be more accurate than HMRC’s. Yet you claim it is so.
I think you don’t understand the statistics, or in this case basic scientific method. The reason people spend so much time collecting and analysing huge datasets is that the error rate is more manageable, depending on sample size. Doing what you have done, taking other people’s answers and multiplying them together simply multiplies the size of the errors.
Which leads inevitably to the macro/micro issue. One close to my heart. For may things, macro is good.But not for tax or the shadow economy, because taxation is so dependent on personal situations, transactions and the micro-level data. You can’t get an accurate picture by using the broad brush strokes of macro as you have done. It just comes out with nonsensical answers. I won’t even go down the path of substitution analysis of the problem – I think it is beyond you.
By your logic, there are literally only three variables to the size of the tax gap. The size of the economy (GDP), the size of the shadow economy, and the average tax rate.
So what you are saying, is that all things being equal the tax gap with grow by the same amount as GDP each year. Or if the government raised tax rates overnight, the tax gap would increase by the same proportion. It is crazy to believe that the problem is this simple. Three variables control all of it. Miraculous.
But please, by all means submit this work as a paper to a reputable journal to be peer-reviewed. My guess is that if you do such a thing, it many more holes will get torn in it than the ones I have made.
So, we have a classic problem
A microeconomist tries to answer a macroeconomic tax question and has not a clue what they are talking about
It so happens I do understand them
The EU estimate I used is not a MIMIC estimate: it was based on the tax gap
And you clearly show that you do not understand VAT. It is a tax on top lines. Turnover. Income. GDP, if you like. And 10% of the tax is not paid. That is because the vast majority of that income is not declared. That is the only way around that. And if the top line is not declared all the spend out of that top line on wages, profits and everything else will not be paid either
And that is why using average rates is fine, and probably low – because this will be the top slice of income – which is what evasion – which as HMRC says is an issue much bigger than the shadow economny per se as you define it – is often about
But you are ignoring this and choosing to think we’re chasing missing dog walkers
We’re not
We’re chasing tax on up to 10% of GDP wich HMRC admit is not taxed
And my argument is that unless you understand that all your stats are wrong
And I’ll assure you, as an accountant who does actually know about this stuff from that angle as well – that they are
And just to shatter your illusions – I am not making the claims you suggest. Of course I know my method approximates. But it approximates a lot better than your method does
And the distinction is well known – it is top down v bottom up
And bottom up is simply wrong when appraising a macro issue but you are claiming otherwise and I am using the best method available given what we know
Now answer the VAT poi8nt – showing an appreciation of accounting in the process please and not p[retending that each tax exists in a silo- which they do not – or do not waste my time with more false claims
Or the absurd claims that maths is maths
You are making some pretty wild claims
This not a macroeconomic tax problem. It is a micro problem. Would probably be worth you looking up what macro and microeconomics actually are.
You have tried to frame it as a macro question because you don’t have any alternative method. By doing so you have ignored all of the complexity of the tax system and it’s interactions with people and the economy. Do you really think that this is likely to give you a good answer?
You used an EU estimate for the size of GDP. I know first hand that these estimates for the size of shadow economies use MIMIC or DYMIMIC models. But maybe you could point me to the data you used. Regardless, MIMIC models are the most accurate (everything is relative), so if you have used data from another source, your accuracy is going to be worse still.
“And you clearly show that you do not understand VAT. It is a tax on top lines.”
No, this is simply incorrect. It is not a tax on turnover, income or GDP. It is a transactional tax on the end user. Do you understand input and output VAT? If you did, it would immediately become clear to you that you are making a mistake. If you think that VAT is directly proportional to GDP as a top line tax, you again might want to look at the data.
Your claim that not declaring VAT must mean all other taxes are undeclared is also flawed – and I suspect you have no evidence for it. If so, please provide it.
Which is yet another reason using average tax rates is absurd. Apart from the data being available (not that you have looked at it), you are conflating the shadow economy with all other forms of evasion, and are now saying that this should be on top of all other taxes paid.
All you have provided is your own opinion – a lot of which doesn’t even make sense, and the rest of it full of errors.
“We’re chasing tax on up to 10% of GDP wich HMRC admit is not taxed”
Sure, the shadow economy. HMRC have been very clear on both the nature of the shadow economy and where tax gaps originate. They have provided data for both. They clearly say half of it would fall under any tax threshold, and much of the rest at lower bands. They provide data for it. You provide your own poorly argued and confused opinion.
“And my argument is that unless you understand that all your stats are wrong”
My stats?
“And just to shatter your illusions — I am not making the claims you suggest. Of course I know my method approximates. But it approximates a lot better than your method does”
I haven’t provided a method, because I’m not trying to estimate the tax gap. I have asked you questions about your method, and compared it to HMRC’s method. Which the IMF seem to think is the best in the world.
Your method is to take some numbers, make a few spurious arguments and then multiply those numbers together.
HMRCs method is to take the real, observed data and sample large sizes from it where necessary. They then measure those gaps and provide the level of confidence in their estimates.
Which is likely to provide a better answer? Guesswork using estimated data inputs and a highly simplistic methodology or access to the real underlying data itself?
“Now answer the VAT poi8nt — showing an appreciation of accounting in the process please and not p[retending that each tax exists in a silo- which they do not — or do not waste my time with more false claims”
Pretty sure I already have. There are plenty of models out there for the interactions between taxes, but none of them come out with your claim “And if the top line is not declared all the spend out of that top line on wages, profits and everything else will not be paid either”. What you are saying (and using in your model) is that VAT is directly proportional to every other tax in terms of it’s effect through the economy as a whole.
The studies show (and there are a lot of them) show that tax interactions with the economy are complex, a lot of substitution goes on and most importantly the relationship between taxes is not linear and not constant. Which throws your simplistic argument out of the window.
“Or the absurd claims that maths is maths”
So math isn’t maths? Or only when it doesn’t suit you?
There is no way around this. If you start with a data point with a certain error, that error only gets bigger every time you multiply it with another point with another error.
This is why people go to so much trouble to measure the particular sample under test, rather than inferring from precursor data. Because those errors keep getting bigger at every stage.
Using the data you have, you have started with a 15% error (now probably bigger if you didn’t use data from a MIMIC model) so it is impossible that your error got smaller. Literally impossible. HMRC’s errors are between 5% and 10% depending. So by definition HMRC are more accurate than you.
“You are making some pretty wild claims”
Really? I am claiming that data is available already, and is more likely to be accurate than yours. I am also claiming there are serious flaws in your method.
You are claiming that you can accurately predict the tax gap by multiplying three numbers together. You can also do this without bothering to actually do any research, look at any actual tax data or bother with details such analysing the nature of the shadow economy or any effective tax rates. All you do is pick numbers which suit you then waffle around trying to justify it. All within the space of a page of A4. Quite remarkable. Your word does not make it correct, so maybe you should actually try and present some data to back up your assumptions.
“And the distinction is well known — it is top down v bottom up
And bottom up is simply wrong when appraising a macro issue but you are claiming otherwise and I am using the best method available given what we know”
But as any economist knows, tax is a microeconomics problem. Macro focuses on economic aggregates. People paying or avoiding tax is a micro level decision. There are macro models out there, but they haven’t been that successful, because you don’t get the granularity you need by using economic aggregates.
Calling what you have done a model is a stretch, but things like Stochastic Tax Frontier or RA-GAP (both macro tax models) have proved limited in their usefulness. Your “model” would be classed as a latent variable model. Guess what – you are not the first to look at such a thing. Maybe read Dell’Anno and Schneider (2006), who acknowledged the limitations of their model (which was a great deal more complex and detailed than yours) and found that due to the scale of errors and assumptions needed “Latent variable, and other macro model estimates, of the shadow economy are unsuitable for tax gap estimates.”
But why believe me. Or them. Let’s hear what ISWGNA have to say about it. It’s members being the EU commission, Eurostat, the IMF, OECD, UN and the World Bank.
“Macro model methods have not produced reliable estimates of changes in the shadow economy or of the size of the shadow economy. They are of no value for work on the tax gap. This conclusion on macro model methods is very much in line with a declaration by the world’s statistical institutions.”
When I cut through the hyperbole what you are saying is two things
Trust HMRC
And trust conventional economics
I have news for you. On this issue, I do not
And let’s cut the crap: I acknowledge all the weaknesses in GDP: I bo9thered to talk to Eurostat about them and discussed these very issues.
So let’s stop the hyperbole on that too
And shall we stop the patronising as well, while we’re at it?
Instead, let’s deal with the very obvious point that you do not get. And why you’re wrong, just like so many economists who do not understand taxes and accounting are wrong.
VAT is, of course, in theory a transactional tax. Much as you like to patronise me, I know that. But what you do not seem to understand is that it is charged by VAT registered traders. It has no other way it can be levied. Note, that this is then a tax on trade, and not transactions. And if VAT is not charged when it should be there is unrecorded trade. And when there is unrecorded trade it is the top line of businesses that is understated. And whatever you might like to think in your university – which I know to be a university so far removed from economic reality that there is only one decent economist there and you won’t even make him a professor – what this means is that once unrecorded for VAT then this income is unrecorded for all other taxes. But you don’t see that, because you have no idea how accounting and tax actually works. So, if unrecorded for VAt then what is paid out of that unrecorded income is also taxed. And that means it is not just the VAT that is lost – it is the income tax on the receipt of that unrecorded money that is lost, and the NIC that they owe. It will not be corporation tax in this instance as there is no way the funds can be kept in a company. And this loss of other taxes follows like night does day: it cannot.
So in other words – know the VAT gap and you have the base for all tax gaps – not the tire base, because as I note there are, of course other factors and non-trade losses. But you have it. And you deny that. And so do all those other economists who actually do not know how tax works because you’ve never been in the real world. And that’s your problem – and the problem of those drawing then conclusions you reach – and with respect I am happy to say they’re all wrong as well. And Unless you can answer how turnoiver evaded for VAT is then taxed for other purposes nothing you can say will alter this.
So try getting down off your ivory tower and let’s look at reality – and how this actually happens in the real world – not in your models. Because the real world is where the problem is and what you are saying about it makes no sense at all
“And shall we stop the patronising as well, while we’re at it?”
“Instead, let’s deal with the very obvious point that you do not get. And why you’re wrong, just like so many economists who do not understand taxes and accounting are wrong.”
Do you ever read what you have just written? Let’s stop patronising, but then you patronise me and every other economist who doesn’t agree with you.
I honestly can’t believe the arrogance on display here. You are essentially now claiming that everyone else in the world is wrong, but you with a methodology so simplistic have managed to create something no-one else in decades of work can?
“Trust HMRC”
Why not? They have presented their data and errors, and have spent a lot of effort to collect it. From the real world, not from multiplying a few already uncertain estimates together. Their work is praised by the IMF as the best in the world.
“And trust conventional economics”
Calling what you have done economics is a stretch, but what proves you right when hundreds of other economists using a variety of different models say those top-down macro models are no good for tax gaps? Not withstanding hat you have used the output of those same models to get the estimate for your starting point size of shadow economy. You can’t seriously claim they are wrong and then use their data.
“Instead, let’s deal with the very obvious point that you do not get.”
Let’s. I can already see where you have made your mistakes. You now acknowledge VAT is a transactional tax at least.
“And if VAT is not charged when it should be there is unrecorded trade.”
Partly right. There will be some unrecorded trade. But you have simplified the problem, which leads to your first error.
Because VAT is charged at every level of a production cycle and then can be partly reclaimed, to analyse VAT properly you have to look at input AND output VAT. Not just the last transaction in the chain.
The business not paying VAT on it’s output will still have paid VAT on it’s it’s input in effect, so the whole of the VAT amount of the final transaction in a production chain is not lost. Only a marginal amount. Not only that, but VAT is charged and reclaimed throughout the chain on purpose partly because it makes it much harder for people to cheat. It’s a tool tax authorities use heavily for VAT enforcement.
Your second problem comes when you say that unrecorded business means the top line of the business is understated. Unfortunately not paying VAT means nothing of the sort.
The top line of the business means nothing to the tax man. Business taxes are not levied on turnover. They are levied on profit, which is the bottom line. A businesses which has not paid VAT might not be making a profit, which means it might not be liable for any corporation taxes. You simply can’t tell from VAT alone.
The same goes for income taxes. That business is going to be paying it’s staff and owners, but not paying VAT gives you no evidence regarding any other taxes they may or may not be avoiding.
For your claim “once unrecorded for VAT then this income is unrecorded for all other taxes” to be true ALL of the economic activity of that supply chain would have to remain in the shadow economy. This is a horrible assumption which we know not to be true.
You have also claimed that your average tax rate should be higher because VAT is the top line tax. You allude to the claim here again “And that means it is not just the VAT that is lost — it is the income tax on the receipt of that unrecorded money that is lost, and the NIC that they owe.”
As I have explained, this is simply wrong. Let’s take a quick example using your logic, that VAT is a top line tax. This means all of the top line, or revenue must be counted for tax purposes.
A business sells something for £120. They evade paying the VAT (20%).
Of the remaining £100, £18 should go to CT, and for sake of argument £20 should got to income tax and £12 should go to NI.
According to you, because of the VAT evasion none of this gets paid. Top line, right! So the total tax loss is £70, or 58%!. This is how you get to your wild claims that the average tax rate for the shadow economy should be higher.
Of course this is simply incorrect, because you need the input cost as well. How much profit did the company make on the transaction? How much VAT was really avoided?
Let’s say the business bought the item they sold for £120 for £90.
Now let’s see what that does to your numbers, factoring in the input costs.
VAT is value add, so all previous companies in the chain have paid VAT up to that £90. So HMRC are only missing revenue on the £30 markup, or as we call it “value add” – so only £6.
Businesses pay people out of net revenue, not gross. Even if there was evasion on all of the rest you now only have £5.4 of CT, £3.6 of NI and £6 of IT for a total of £21, or %17.5.
It is important to note that the VAT loss is also not at the headline 20% rate – it is going to be much smaller, because VAT is value added, and applies all the way through a production chain. For you to lose the full 20% every step of the chain would need to be evading.
Quite a difference, no? Even this is a gross simplification as CT, NI and IT are totally unrelated to the payment of VAT. CT might not be liable at all and the VAT tells you nothing about the levels for NI and IT. Like the sole trader who evades VAT but doesn’t make enough profit to fall into IT thresholds, or conversely the trader who sells zero rated VAT products and evades no VAT but goes on to evade other taxes. The issue is highly devolved and complex. You can’t simplify the problem in the way you have and expect it to be correct – which somehow in the fact of everything that goes on in the real world, you do.
In short, your methodology, logic and assumptions are wrong.
“So in other words — know the VAT gap and you have the base for all tax gaps”
And here we come to another problem. As I just pointed out, knowing the VAT gap does not mean you have a base for all tax types. On top of that though, you firstly have to know the VAT gap.
To do that you have to measure it. How is that done?
Well, it is currently done via a top down approach. Top down methods are already well known to be problematic. It takes the relevant sectors from national accounts and then applies VAT rates to them, creating the net VAT total THEORETICAL liability (VTTL). It then simply subtracts actual VAT receipts from the VTTL to get the VAT gap.
So you have a few problems here. The method itself is known to only give you a theoretical answer because of this. You have errors in the national accounts consumption data itself, problems matching that consumption data to the right VAT base and problems estimating the effects of legal exemptions. You also have problems on the VAT receipt side, receipts in a given year won’t match the national accounts data – late payment etc.
So you have data which is at best unreliable. The EU themselves are aware of this:
https://ec.europa.eu/taxation_customs/sites/taxation/files/docs/body/vat-gap.pdf
They say that the VAT gap is best used to illustrate trends over time, as the headline number is highly uncertain. They also say:
“….which indicate that as much as half of the estimated ‘compliance gap’ may sometimes be attributable to factors other than outright tax evasion suggest that caution should be exercised in using even the best compliance gap estimates as evidence of the extent of outright VAT evasion…”
and
“Further development of microdata approaches to measuring tax non-compliance within particular sectors appears to be the most promising path to develop usable and meaningful measures of the components of the VAT Gap from the perspective of assessing and improving tax administration performance.”
So we know that top-down VAT gap estimates are very uncertain, likely to have large errors and the EU themselves say they should only be used as an example of a trend, and are likely to provide a ceiling or maximum VAT gap, not the actual number.
Yet you have gone and used the VAT gap estimates, with all their problems, as the basis for your whole method and argument. See the issue?
“And so do all those other economists who actually do not know how tax works because you’ve never been in the real world. ”
Really? The basic concepts of taxation are fairly simple. All of economists live in the real world as well – just as much as you do. Yet it is you who have made the error here as I have pointed out above. You’ve taken VAT gaps as gospel, when they are very unreliable, and you have made horrible assumptions and errors about VAT itself. Forgetting input and output, claiming VAT is top line and then saying all taxes are directly proportional to it. Anyone can see those errors.
“and with respect I am happy to say they’re all wrong as well.”
I’m sure you are. Only you are capable of such high level thinking as to be able to solve a problem hundreds of others struggle with, and all in a couple of pages of A4 no less. With no research of your own or original data.
“And Unless you can answer how turnoiver evaded for VAT is then taxed for other purposes nothing you can say will alter this”
See above.
“So try getting down off your ivory tower and let’s look at reality — and how this actually happens in the real world — not in your models. Because the real world is where the problem is and what you are saying about it makes no sense at all”
Implicitly this statement is you are claiming everyone else’s models are wrong, yet yours is right.
Really? You haven’t looked at reality at all. You have done no investigation or research. You have just made some wild assumptions about how reality works and then claimed you must be right. Assumptions which I have explained above are wrong.
Is your middle name Hubris?
I’ll provide a response to the hopelessly inaccurate claims in here when I have time
Suffice to say, an economist who has clearly never been near the real world, accounts or a tax return wrote it
Have you looked into the interaction of tax evasion and benefit fraud? Cash in hand not declared for income tax will similarly not be declared when claiming benefits. It would be interesting to see the impact of this.
Sorry, I have not
[…] said in his / her latest comment, the following, and as it is the basis from which I will show that they are wrong, I will repeat […]