Share this content

The link between HMRC’s tax gap and Python’s duck

Tax officer-turned-gapfinder general Wendy Bradley reads the runes in this year’s report from HMRC on the size of its tax gap.

16th Jul 2020
Share this content
Bad news ahead for HMRC's tax gap
iStock_Rubber ducks_porcorex

Crystal ball in hand, I can confidently predict that next year’s tax gap will be significantly larger than this year’s.

Actually I don't need the crystal ball while I have HMRC's Measuring Tax Gaps to hand. There are two little words why next year's figures (2020/21) will be worse than those for 2018/19.

Missing words

The number of people whose tax liabilities don’t match their tax payments (because of bewilderment, incompetence, chancing their arm, minor naughtiness, major rule breaking, significant legal disputes or outright villainy) seems to remain about the same over time. That being so, my confidence in the expanding tax gap is because the words “Covid” and “Brexit” do not appear anywhere in the 98-page Measuring Tax Gaps 2020 edition report or its 53-page methodological annex.

Percentage of what?


However, as the economy and the total expected amount of tax due have grown year on year, £31bn represents an ever smaller percentage of the total amount of tax due to be collected. It is currently down to 4.7% of tax liabilities but a similar amount was 5% of total liabilities in the previous year, down from 6.7% in 2006/07.

There is a time lag in the figures as HMRC enquiries and analysis refine the results year by year, so it's a fair bet that the 2019/20 figures will be similar to the 2018/19 figures revealed in the 2020 tax gap document. It is in the 2020/21 figures that we will see a change.

After Brexit and Covid-19, next year’s economy and the tax take for the current year is likely to be significantly smaller and thus the steady state amount of the gap would represent a much larger percentage of a smaller whole.

I’m not going to say “lies, damned lies and statistics” because honestly HMRC really does have a world-beating go at calculating its tax gap year on year, and explains the workings as transparently as one could reasonably expect. My problem is that the tax department doesn’t seem to do much of anything useful with the results.

For example, there is clearly a gap in the methodology as far as larger businesses are concerned, because large companies, partnerships, employers and businesses aren’t covered by the Random Enquiry Sample.

HMRC doesn’t therefore have base figures from which to extrapolate the tax gaps. Well yes, but how many years has it being doing this? Why aren’t larger companies at the same risk of random audit as small ones? Why isn’t there a separate unit set up to throw resources at a random audit of a number of each group to give a baseline check?

If small businesses can be audited at random but large businesses can’t, well, where’s the fairness in that?

Mixed methodology

The headline figures are still worked out “top down” from national statistics for VAT and “bottom up” from random enquiries and a bit of statistical jiggery pokery for income and corporation taxes, and HMRC is still using an American “multiplier” in the mix to arrive at the IT and CT figures.

Page 35 of the methodological annex explains that the American IRS undertakes more random auditing than HMRC, but does not explain why the US figures are applicable to the very different UK system. I have written about this before when the Office for Statistics Regulation raised doubts about HMRC’s modelling: I cannot see why HMRC persists in using this methodology.

As we have learned from Covid-19, modelling can only get you so far: model on different premises and you arrive at enormously different answers.


I am not a fan of HMRC’s “segmentation” of its “customers” – in my view a taxpaying citizen is a taxpayer with the same rights and responsibilities, entitled to the same level of consideration. However, HMRC separates out the wealthy (defined as individuals with more than £200,000 in income or £2m in assets) and find that the “wealthy” category shows exactly the same tax gap as the wider population. Well colour me surprised.

Overall, the trend in the tax gap looks positive, with a fair degree of fingers crossed for next year.

Why does HMRC do it?

The 2020 tax gap report says: “Our tax gap analysis provides insight into which strategies are most effective at reducing the tax gap.”

Is this the same technique medieval witchfinders used to work out which strategies were most effective at reducing the prevalence of witches? Or, as Monty Python might have advised, never weigh the same as a duck?

Replies (3)

Please login or register to join the discussion.

By Justin Bryant
17th Jul 2020 09:25

Strange how HMRC stay quiet about taxpayers overpaying tax (by not claiming all their expenses/reliefs etc. or who are too late to appeal an excessive assessment etc.) and that must add up to quite a bit.

Thanks (0)
By AndyC555
17th Jul 2020 11:51

"HMRC really does have a world-beating go at calculating its tax gap year on year, and explains the workings as transparently as one could reasonably expect."

Indeed. Of course HMRC battles with alternative approaches with 'calculating' the tax gap such as that by blogger Richard Murphy which come up with horrendously higher numbers of over 100 billion. Of course, that includes tax not collected because of the existence of tax breaks Murphy thinks shouldn't exist (it's bizarre to criticise HMRC for not collecting tax they aren't legally entitled to collect) but his 2010 figure also included the complete howler of including £28 billion of tax debt as part of the tax gap. That £28bn was simply a snapshot of tax owed on 31 March 2019 of which over £25bn was subsequently collected. Murphy included £25bn in the tax gap which was in fact tax actually collected. A bit like an accountant writing off all debt at a year end despite the fact that most of it is later collected.

There were numerous other errors, (detailed on page 12 and onward of the House of Commons Briefing paper 7948 "tax avoidance and evasion" if you are interested) but by then the damage had been done and I still hear the £120bn figure bandied about as the real tax gap. As the HoC paper commented such alarmist and inaccurate figures can be “dangerous if not countered by HMRC’s published estimates … partly because they give a misleading view of HMRC’s effectiveness and the amount of uncollected revenues. But also because they encourage the perception that deliberate non-compliance in the UK is the norm—a perception which could encourage further non-compliance"

Thanks (1)
17th Jul 2020 12:04

So two obvious question....

Is there any indication of how much the tax gap will have reduced from the introduction of MTD for VAT and if that improvement has any relationship to HMRC overstated selling of MTD.

If there is any indication of the amounts of the tax gap that can be attributed to non-represented tax payers, largely trading cash in hand, who have no concept of the errors that they are making in their records.

Thanks (1)