HMRC’s estimate of tax evasion in the shadow economy is ludicrously low

Posted on

HMRC's new tax gap report is a gift that just keeps on giving. Take this:

Ghosts and moonlighters

6.39 ‘Ghosts’ are individuals who receive income from employment or self-employment but are not known to HMRC because they and/or their employers fail to declare their earnings. Ghosts are not accurately recorded by any government agency or survey and therefore any estimate as to their number or the consequential loss of duty is approximate.

6.40 ‘Moonlighters’ are individuals who pay tax on their main job through PAYE but who fail to declare earnings from a second job or additional income from self-employment.

6.41 Using a series of assumptions, it has been possible to produce illustrative estimates of the tax gap from moonlighters and ghosts. Due to the extent of the assumptions used to produce this estimate and the inherent uncertainties in the methodologies, this estimate has a large margin of error and should be treated with due caution.

Let's put that another way: the numbers they've used are complete and utter nonsense.

Apparently in an economy with a GDP of coming on for £1.5 trillion we lose just £1.3 billion to ghosts - people who declare nothing at all to HMRC - but rather more to moonlighters - who have a PAYE employment and don't declare their other earnings.

Peer reviewed research suggests this is wrong - I reference it in this paper. Simple logic also suggests this is wrong. The VAT yield after (supposedly) 10% had gone astray in 2010-11 was meant to be £81 billion. Some £9.6 billion of VAT was lost - suggesting maybe £55 billion of black money floating around. OK, some was fraud. And this estimate ignores the fact that most moonlighters and ghosts will be working in businesses that need not be VAT registered. But out of that £55 billion to suggest just £4 billion is lost in income tax from self employment and £3.2 billion to ghosts and moonlighters is ludicrous.

I note HMRC say:

this estimate has a large margin of error and should be treated with due caution.

I agree. But the due caution is that it's just wrong, and I think they know it.