The National Audit Office fails to ask the right questions of HMRC on the tax gap

Posted on

The National Audit Office has published another dire report, only a day after it completely missed the target with its review of NHS trusts. This time their subject is close to my heart: it is on HMRC and their work to tackle tax evasion.

The trouble with the comments they have to make starts with the opening lines of their press release, which says:

According to a report by the National Audit Office published today, HMRC estimates that losses to tax fraud amount to £16 billion each year. This is nearly half of HMRC's estimate of the tax gap (£34 billion): the difference between the amount of tax HMRC should collect each year and the amount it actually collects.

Today’s report is the first in a series of reports which will evaluate how effectively HMRC tackles different aspects of tax fraud, a longstanding problem not only for HMRC but for tax administrations around the world. Reducing the amount of tax that is lost due to tax fraud is a high priority for HMRC. To do this it will need to make better use of its data and develop its analysis.

What is clearly implied by the first sentence is that HMRC's estimate of the extent of tax evasion is being accepted by the NAO. And that's what is most wrong with this report, for two reasons.

First, I think that estimate of the tax gap is wrong, and despite the fact that the NAO know of my work (I have discussed it with them) there is no hint of a reference to it in their work.

That would be forgivable if they had clearly satisfied themselves that HMRC's estimate was fair before considering whether the work done to address the issue was appropriate. Candidly, I do not see how they could do anything vaguely approaching something called an audit without doing this: the first job in any audit process is to establish that the population being considered is fairly stated. But the NAO notes of HMRC's work:

 The tax gap is only an estimate, but it is the best data available on the amount of tax lost, including through tax fraud. The losses HMRC attributes to evasion, the hidden economy and criminal attacks have fluctuated from year to year but the overall trend in tax fraud is flat. HMRC estimates that losses from tax fraud accounted for 3.2% of all tax due in 2009-10 and 3.0% in 2013-14.

The NAO should have been on guard straight away: the Bank of England say that the UK shadow economy is 10.3% of GDP. This estimate is so far from the implied HMRC data that the NAO were, surely, duty bound to c0mment? But no, they just say the HMRC figure is the 'best available data'. That's the sort of audit error I would hope not even the Big 4 make.

The concern is, however, compounded when it is noted that:

HMRC has identified over 50 large and strategic risks to the collection of tax. 21 of these risks relate to tax fraud. An example of such a risk is small businesses failing to register for VAT when their turnover exceeds the threshold for registration. Of the 21 risks, 8 relate to organised crime and 9 involve medium-sized, small or micro-businesses. HMRC believes that these businesses are responsible for tax losses of £17 billion, almost half of its estimate of the total tax gap, but it does not consider its internal estimate of how much of this is the result of tax fraud robust enough for publication.

Hang on, the work is so bad they're too embarrassed to publish it but we're meant to accept it?

Then note this:

In 2010, HMRC was set a target to increase the yield from its compliance work by £7 billion by 2014-15. This target included the additional revenue HMRC could generate from tackling tax fraud and all other forms of non-compliance, such as simple error and legal but contrived arrangements to avoid tax by exploiting loopholes in tax law. It reported additional revenue of £26.6 billion in 2014-15, exceeding its target by £0.6 billion. HMRC does not record compliance yield in a way that identifies what proportion of the total yield is derived from all its activities to counter tax fraud.

In other words, as I have long argued, these numbers might as well be said to be made up. That's an opinion the NAO pretty much endorse, saying:

For example it has more complete information on the yield from its work to tackle organised crime than tax evasion. We estimate that between 30% and 40% of compliance yield may be generated by HMRC’s activities to tackle tax fraud, based on HMRC’s analysis of the compliance yield it can attribute to the largest tax risks. This is a crude estimate based on partial evidence, and we will look for firmer evidence of HMRC’s impact.

And this:

HMRC has only partial data on how much of the total yield is derived from its work to counter tax fraud. For example it has more complete information on its work to tackle organised crime than tax evasion. We estimate that between 30% and 40% of total compliance yield is generated by HMRC’s activities to tackle tax fraud, but this is an estimate based on partial evidence.

So let's be blunt, HMRC not only use a dire methodology, estimating evasion from tax returns received and not from macro-economic data, meaning that by definition they are bound to understate the scale of tax evasion, but even then the estimates they have made are so bad that they are at best 'partial evidence'. Despite this HMRC management present the tax gap data as fact, as do ministers. I think it fair to say that's as close to misrepresentation of the facts as it now gets.

This report is, when read properly, a damning indictment of the NAO's work and of HMRC's too.

The sooner we have an Office for Tax Responsibility reporting direct to parliament and a minster who is accountable for what HMRC does in the government the better. Because right now it is quite clear that it is way out of control and no one is doing anything about it.

Which leads me to suggest that this may be deliberate: after all, where would the austerity agenda be if I was right and there was ample more tax to be recovered?