HMRC published its tax gap data for 2018/19 yesterday. I have long criticised this annual debacle which has always published a number between £29bn and £38bn or so over the last decade. It's so predictable that I now wonder why they do it, as this data shows:

And the methodology remains as flawed as ever. In essence, tax evaders pretty much fall outside the scope of much of the tax gap because HMRC refuse to believe they can exist, even when the evidence is so obviously there to see - not least in the 3 million companies that exist in the UK and do not declare any tax liabilities and who are not chased by HMRC to do so and who instead turn a willing blind eye. In addition, the definition of tax avoidance is so narrow that much of what everyone agrees to be tax avoidance - like the widely condemned abuse by multinational companies - does not get into the tax gap definition that HMRC use, which is absurd.

Rather than reiterate all my arguments try this from last year, except please note it now looks like they have also changed the way of doing VAT and I suspect that I now have reservations about that too.

But there is a dimension to this that makes the report this year even more absurd than usual. To keep the number in the required and seemingly pre-determined range (four years in a row with an almost identical number is utterly implausible) that suits HMRC's management HMRC has to make the most staggering claims about their increases in productivity to counter the impact of inflation and growth. In other words, the tax gap as a proportion of tax take has to keep falling to hit the required number range, as is very apparent from the above chart, and also this one on the VAT gap:

We are apparently almost twice as honest now as we were only a decade ago.

Or rather, HMRC assume we are. This chart shows the revisions to the data as a result of their own changes in methodology:

Last year the VAT gap was deteriorating as the grey line shows. The trend was markedly upwards.

Now, apparently, HMRC realise we were much worse in the past, but suddenly are getting so much better after all. That's what the orange line shows. And apparently, instead of getting worse, were all reformed and complaint people.

There are, then, some simple questions to ask of this data. The most obvious by far is how likely is it that attitudes to tax abuse changed so radically in such a short period of time without there being a great many obvious reasons for them to do so?

On international issues I think there is a chance that this happened. I have no doubt that multinational companies have changed their behaviour since Margaret Hodge savaged Google, Amazon and Starbucks in 2012. Knowing that country-by-country reporting was coming will also have hastened change, although abuse remains, as I note elsewhere this morning.

And for individuals using tax havens I am sure that automatic information exchange of data will have had a significant impact on compliance.

That's two gains for tax campaigners then and not much for HMRC. There is also a third. We know that the crackdown on VAT abuse through Amazon and eBay had begun by 2018/19. But I stress, it had only begun. And then only due to the relentless pressure from Richard Allen on this issue. Another win for campaigners and not for HMRC then.

But only the Amazon / eBay case impacted VAT: offshore did not. And yet HMRC claim that compliance rates on VAT have improved staggeringly over a very short period of time, when during that period they were closing as many tax offices as possible, pushing their staff's morale through the floor, and were losing their most experience people who did not want to move to the new regional hubs that HMRC has insisted on creating. Inspection of trader's records had almost entirely ceased as well. All my feedback from HMRC is that compliance got worse in this period.

The only move HMRC can claim might have helped is Making Tax Digital and very small amounts of extra data exchange from credit card suppliers and trading platforms. I don't dismiss the impact of the latter, but I very much doubt that these were widely known or understood amongst the tax evading community. If they were then that community would have made even greater use of anonymous ‘throwaway' companies that are so easy to create and get rid of in the UK - and which represent the biggest threat to the tax take in this country, and which HMRC persistently ignore.

So, we're down to Making Tax Digital, which HMRC has to say is a success. No one else does. I get no feedback through accounting web sites, for example, that anyone thinks that this has helped productivity, error rates, compliance of anything else. The feedback there is that it has been embraced as a new way of working by those already compliant, but it has not changed compliance rates. Indeed, in many ways it will have helped force the non-compliant out of the system altogether.

But I very strongly suspect that this tax gap data reflects assumptions that Making Tax Digital works and has produced massive savings.

I stress when saying so that I accept that the tax gap is improving. I think it is. But by the amount HMRC claim? I have to say that this change of behaviour is implausible.From 2007 to 2013 the compliance rate hardly changed. But now it's tumbling. I simply cannot believe that. It strikes me that this makes the tax gap data as dodgy as the tax return of the proverbial ‘man down the pub' who has a new BMW each year and says he never pays a penny in tax.

HMRC has a duty to be credible. It has a duty to publish information that can be believed. And I am a long way from believing what they are producing. Put bluntly, the UK is not the most tax compliant country on earth, which it would have to be if these numbers were true. And it does HMRC no good to claim that we are.