HMRC suggest I am abusing their VAT gap data. The trouble is the IMF do not agree with them

Posted on

I have already blogged on one response made by HMRC to my new reports on the tax gap as reported in the Guardian this morning. The first response was to my comment on tax abuse by companies. The other response was to my use of VAT gap data to suggest the total value of unrecorded sales in the UK. Of this work HMRC are reported to say:

This is a seriously flawed study based on, among other things, a significant overestimate of the UK’s VAT gap. HMRC’s estimates, produced under the Code of Practice for Official Statistics, show the overall tax gap is £35 billion and falling as a percentage of tax owed.

I am amused first of all by the claim that HMRC must have their data right as it confirms to an official code when, as must be obvious, that code is basically written and published by them. I do not take such comments seriously in that case. It's as good as me saying my work is right because I wrote it, and no one with any sense would ever do that.

More important is the suggestion that my work is seriously flawed. My response to that is that if it is then HMRC must be largely responsible because I have used their data, with the odd twist as noted here.

Firstly, I adjusted HMRC's data because they said that the VAT gap was 11.4% in their latest tax gap estimate. I have only used a figure of 9.7%, which is somewhat lower. That's because I cut out their figures for criminal attacks on the VAT system and bad debt  because I felt it inappropriate to extrapolate such data to lost sales; clearly that would be inappropriate. On the other hand I did ignore VAT avoidance, because HMRC suggest it is virtually irrelevant. And nor could I ignore EU estimates of the VAT gap and the trends that they showed. Their data is referred to on page 51 of my detailed report and the differences are summarised here:

Screen Shot 2014-05-20 at 10.12.04

The EU actually, on average and before bad debt and criminal attacks in both cases, thinks the UK has a lower tax gap than HMRC does, but the trend is very different. The marked recent downward trend HMRC has found is not matched in HMRC data. Given that overall the two strongly agree I took the HMRC data for VAT lost to criminal attacks and bad debt off the EU estimate then averaged the two over five years to come up with my estimate as a base for projection of 9.7%. HMRC may say that's wrong: I would suggest it's fair use of data. I think the likely result from doing so statistically appropriate.

Next, I took the EU's estimate that only 53% of the theoretically possible VAT yield is collected in the UK because of exemptions, allowances and reliefs. Taking the factors into account, and then assuming that the VAT gap occurs at the standard VAT rate of 20% (possible because of the compensating adjustment in the EU data for allowances, exemptions and reliefs) I could gross up the 2011/12 VAT gap net of bad debt and criminal attacks to suggest the total missing sales in the UK - which happened to come to £100 billion. Full explanation is in the report.

The question to then ask is whether it is wrong to use this top down approach and to even then apply it the resulting loss of income tax, national insurance and corporation tax, which is what HMRC clearly imply. Unfortunately for them in doing so they ignore advice provided to them by the IMF when that body reported on HMRC's tax gap methodology in October 2013. As the IMF noted (page 45 and on), HMRC really should in their opinion try top down modelling of the direct tax gaps to prove the logic of their claims made based on the analysis of tax returns they actually receive (which is the basis of much of what HMRC do now and which is bound as a result, ad by virtual definition, to omit a loot of tax evasion by those not submitting returns). How do they suggest this should be done? As they say (para 69):

One possible approach would be to start with the top-down VAT gap analysis 

That's exactly what I did HMRC. It's a shame you won't also follow the advice you've been given because then you would find, as I have, that your tax gap estimates are ridiculously low. I'd suggest as a result that it is HMRC who does not understand the data or its implications. You could of course ask why that is. Is it, as I suggest in my report, that they want to turn a blind eye to the obvious, even when those who advise them tell them to think again? It does look very much like it. And, in the meantime, I'm happy to say HMRC are very simply wrong, without saying that my figures are emphatically right because I am well aware that they are indicative of outcomes that are bound to fall in a range. That range would not, however, plausibly include HMRC's estimates.