Couldn’t help but laugh at this:
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 [sic – EU?] data. Given that overall the two strongly agree….”
… I’m sorry, what?
The two sets shows “very different” trends? But they “strongly agree”?
Let’s hope nothing significant is riding on that unsubstantiated observation… The rest of the paragraph reads:
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.
Right… Just the entire basis of his calculations….
And I don’t think it is fair use of data. Openly observing that two sets of data are “very different” and then deciding that is irrelevant in order to use figures from one set of data to manipulate the other? I might want to justify that choice a bit more…
To be fair, I think he’s talking about agreement in overall magnitude as opposed to difference in trend. But I think that when one data set suggests that the other is overstated by as much as a third in some years and understated by almost a third in other years , I would really think about providing some actual evidence before declaring that they “strongly agree”.