In a series of blog posts pre-dating that FT editorial, I argued that the non-dom rule is an unsightly bribe to those with some foreign connection to come to or remain in the UK; that the structure of the remittance basis (which taxes income and capital gains brought into the UK but not those left outside) discourages the very thing (inward investment) that we should want to encourage; that establishing someone’s non-domicile status is incredibly difficult – a fact which may be responsible for the fact that in the last decade HMRC have only taken one domicile case to the Tribunal; and I also asked why, if the introduction of the remittance basis charge didn’t reduce the number of registered non-doms, it should be thought that abolition of the remittance basis should cause the sky to fall in.

I also advanced a policy suggestion:

But what I want to look at here is the prospective yield from certain changes to the non-dom rule proposed today by the Labour Party.

***

What proponents of the remittance basis of taxation for non-doms say is this: yes, we know it’s unfair that those with some happenstance foreign connection should pay less UK income tax and capital gains tax than those who are UK resident through and through. But unless we lower their tax bills, they’ll leave and take such tax as they pay with them.

You need to start with that statement because, embedded within it, is a rather startling argument: that it doesn’t matter if a tax result is unfair, so long as it increases yield. This is the logic of a fiscal dutch auction: that we should offer foreigners a lower and lower income tax bill until they agree to move (or stay) here. Whilst this approach will have enthusiastic supporters – not least as Charles Walker MP (Con) has pointed out, high end estate agents – it’s not a proposition many of us would find palatable. Not even Deloitte who noted, rather drily, and of the increase to the remittance basis charge in the 2014 Autumn Statement:

It’s important not to forget this. Proposals to reform non-dom status must, of course, have regard to yield – but not only to yield. Fairness too.

Everyone will place their own subjective value on fairness in the tax system – I’ll leave that to the politicians and the voters. But what I want to do is examine the available data on yield.

***

Calculating the yield to Treasury from changes to the tax regime is typically a two stage process. Stage one is to calculate the theoretical yield: if things remained the same, what would the extra measure raise (or cost)? Stage two is then to factor in the ‘behavioural effects’: i.e. how people will adjust their behaviour in light of the changed regime.

To take an example, if you reduced the duty on fuel by 30% you wouldn’t expect this to cost you 30% of your present rake. Fuel consumption would go up and so, although you’d now only get 70%, it would be 70% of a larger pot. The same is true of tax rises: it’s behavioural effects that enable Labour to say that cutting the top rate of income tax from 50% to 45% was a £3bn tax cut for the highest earners; and the Tories to say ‘but it only cost Treasury £100m’. People change their behaviour when circumstances change.

On the cut to the additional rate of tax, stage one was unproblematic. We knew how many people were liable to the “additional rate” of income tax and what they earned. What we were less clear on was stage two – how they would change their behaviour in light of increasing it to 50% – a question which formed the subject of this lengthy Treasury report.

But when you calculate the yield from changes to non-dom status you have a further difficulty. It’s not merely that you don’t know what the behavioural effects will be. It’s also that, as I observed here, we don’t collect data on the income and capital gains enjoyed by non-doms and not remitted to the UK. So you have a stage one difficulty as well. These difficulties led even the IFS to observe back in 2007 (about the effects of rival political proposals for the remittance basis charge) that no-one knows what the yield might be.

There are, however, a few propositions from which one might, cautiously, proceed.

(1) As to the size of the theoretical yield, the only data we have on the amount of unremitted gains shows average foreign earnings not remitted of £50,000. This average is drawn from a data set of 17,000 who voluntarily chose to disclose this information. However, this average is likely considerably to understate the theoretical yield. As it seems to me, the higher your unremitted gains the lower the prospect that you will voluntarily disclose them to HMRC – because disclosing high unremitted income or gains would invite close scrutiny of your entitlement to the remittance basis. Moreover, this figure looks to be just unremitted income – ignoring unremitted capital gains.

• If you assume 50,000 people claiming the remittance basis each with £50,000 of unremitted income – and conservatively no gains – the theoretical yield would be 50,000 x £50,000 x 47% (including 2% Class 4 NICS) = £1.175bn per annum.
• If, for the sake of argument, you assume 17,000 with an average of £50,000, 17,000 with an average of £100,000 and 16,000 with an average of £250,000 and again ignored capital gains the theoretical yield would be £3.1bn.
• If you take the figures at the second bullet point above and also assumed average capital gains of 50% of income taxed at 28% the theoretical yield would be £4bn per annum.

(2) As to behavioural effects, we’ve introduced a number of measures of late that have rendered the UK less tax-attractive for non-doms: not merely the remittance basis charge (and increases to it) but also the annual tax on enveloped dwellings, a series of measures widening the scope of what constitutes a remittance and the forfeiting of the personal allowance from income tax and annual exempt amount from capital gains tax. I’ve looked at some historical data here but as I read it, it shows a gently rising trend of non-doms in the UK. Whilst these measures might (on one view) have effected a short term interruption to the trend, the trend remains in place. Contrary to the warnings of many, and as Charles Walker rightly predicted, the changes have not caused the sky to fall in.

Looking forward, there is good reason to believe the trend might even strengthen. For example, as the Telegraph has reported, there has been a 69% increase in the number of applications for investor visas in the UK from Russians.

This – it seems to me – provides powerful support for the proposition that one should not over-estimate behavioural effects in calculating the likely yield.

(3) Further support for not overstating behavioural effects comes from looking at the yield calculations for earlier restrictions to the remittance basis charge.

The introduction of the remittance basis charge in 2008 generated a yield in the hundreds of millions of pounds (figures of £800m and £500m were given – see page 164 – but with only limited detail). The Budget 2011 increase in the size of the charge generated a median yield of £70m (see page 36 which went on to observe, inaccurately as it transpired: “There will be no other substantive changes to these rules for the remainder of this Parliament”). And the increase in the remittance basis charge in the Autumn Statement 2014 generated a steady state yield of £90m (made up of a £120m theoretical yield reduced by 25% for behavioural effects – see page 41).

If you have three data points, each demonstrating a positive yield from a restriction to the non-dom basis, it becomes quite difficult to contend that the behavioural consequences of a fourth such restriction will be such as to swamp the theoretical yield.

(4) If you proceed from that relatively modest evidential base and assume a 25% reduction due to behavioural effects (as Treasury did in 2014), then the annual yield numbers at (2) above become £880m per annum; £2.325bn per annum; and £3bn per annum. Other reductions for different behavioural effects can be relatively easily calculated.

***

Ultimately – and again leaving aside questions of basic tax fairness – a good change to the remittance basis is one which constrains the likelihood of huge behavioural effects. To analyse the likelihood of big behavioural effects from any particular measure what, it seems to me, you need to do is look at two groups of people: prospective immigrants to the UK and prospective emigrees from the UK.

As to prospective immigrants, what’s attractive about Labour’s proposal is that it gives people a decent period in the UK to enjoy the remittance basis. Long enough to put down roots by buying property, forming social networks, putting their children into local schools, and so on, such that they might be disinclined to leave when the incentive lapses.

And as to prospective emigrees, the data analysed above suggests that, whilst some will leave, one shouldn’t expect too many to do so. This cautionary tale – of the hugely deleterious effects on Guy Hands of his ceasing to be UK resident – might explain why.

***

If I proceed from the above and stick a finger in the air – an exercise that you’ll have to take it from me is not so dissimilar to that which Treasury does when it forecasts the effects of tax measures – where do I get to in terms of yield? I’m not an economist – and the data is poor. But my instinct is that the stage one theoretical yield figures will tend towards the top end – towards the £4bn end – of the spectrum. But I also think 25% is rather low as a behavioural effect: 50% or even more might well be more realistic, depending on the detail of Labour’s measures. But that would still leave a yield well north of £1bn.