Following the release of the latest tax gap analysis, Wendy Bradley explores whether the mandation of MTD for VAT from next April is really justified by the figures.
HMRC has published its latest estimate of the tax gap and the methodology it used to arrive at the figures. I have argued before that Making Tax Digital (MTD) is a customer service initiative reframed as a revenue-raising intervention and I wondered if the latest figures changed anything. Is the mandation of MTD from next April for traders over the VAT threshold really justified by the tax gap figures?
How is the tax gap calculated?
The tax gap is calculated, to put it crudely, in two different ways: one for direct and one for indirect taxes. A direct tax is "a tax imposed on a person or business" such as SA or CT, and an indirect tax is "a tax imposed on a transaction" such as VAT or Excise duty. For direct taxes, the tax gap is calculated "bottom up" and for indirect taxes, it is "top down".
There seems to me to be a false equivalence here. VAT and other indirect tax gaps are worked out by taking the total transactions – derived largely from externally visible Office of National Statistics data sources – and then calculating what the total tax on these transactions could have been, making allowances for exceptions (small businesses below the VAT threshold, schemes, allowances etc) and then subtracting the actual VAT received from the theoretical maximum.
In contrast, for direct taxes there is a large degree of estimation based on information known only to HMRC. The calculation starts from an extremely small number of random enquiries – see table H1 in the Methodology document where we learn that there were 2,763 SA random enquiries, 925 random employer compliance audits and 362 random enquiries into corporation taxpayers.
We are told that the results are statistically valid and we must, of course, trust HMRC to mark its own homework in this regard. However, the results are then "adjusted", firstly by assuming that the random selections which aren't worked would give the same results as the ones which are.
Figures ‘not the end of the story’
It is a long time since I worked an HMRC enquiry personally, but I seem to recall that when random enquiry cases were allocated to you that was not the end of the story.
You could convince your manager that a particular random case should not be worked as a full enquiry because, for example, the taxpayer was too old, ill or otherwise vulnerable for it to be reasonable to put them through the stress of an enquiry without prima facie evidence. If this is still the case, we might wonder if it is reasonable for this kind of taxpayer to be assumed to be as likely to understate their profits as the rest of the population?
There's more, though. Even after "adjusting" for cases selected but not worked, the numbers are then subject to a "multiplier", on the assumption that even though the cases have been fully worked by an HMRC investigator the investigator won't have found everything.
Apparently, they do this in the States, so the UK has adopted and adapted their methodology (page 64 of Measuring the Tax Gaps: "these multipliers are generated through supplementary studies on particular tax return entries, together with econometric analysis of non-detection rates across IRS examiners".)
Will MTD narrow the gap?
The tax gaps for direct taxes are subject to a larger degree of secrecy and of adjustment than the "top-down" figures for VAT, and it was largely by arguing the speculative nature of the estimated tax yield from making direct taxes subject to Making Tax Digital that enabled the industry to argue its way out of mandation for the smallest taxpayers, at least for now.
Is the requirement for traders over the VAT limit to keep their records electronically and update HMRC with their results four times a year likely to close the tax gap for indirect taxes?
The VAT gap has gone down over the medium term: in 2005-6 it was 12.5% and in 2016-17 it was 8.9%. But it has gone up in the short term: in 2015-6 it was 8.4%. Figure 2.4 on page 31 of Measuring the Tax Gaps shows us that perhaps the figure is best thought of, like the administrative burden figure, as a score.
Just as the administrative burden figure doesn't really tell you whether your business will save £x or £y as a result of a particular change to administration but will tell you whether the overall trend is up or down, so the top-down tax gap figures seem to be capable of multiple revisions year-on-year as fresher or better data become available, but the overall shape remains the same.
The real question is, will imposing the huge administrative burden of Making Tax Digital have a real impact on the Tax Gaps figures? I can see nothing in the new calculations that makes me revise my opinion, that MTD is a customer service initiative that should be funded and carried out. But that making it compulsory in the mistaken belief it will, at a stroke, reduce the tax gaps is a delusion.