Gerald Holtham examines the complexities of devolving income tax to Wales.
The Welsh government is facing a difficult, even grueling, negotiation with Westminster and especially with the Treasury. Discussions are afoot over a Welsh “fiscal framework” and particularly over how the block grant to Wales is to be altered to take account of taxes being devolved to Wales. The talks will be made all the trickier because the UK and Scottish governments have just agreed a fiscal framework for Scotland and a messy old fudge it is. No doubt this precedent will be aimed at the heads of Welsh politicians and officials. They would do well to duck and agree something more sensible. Over the next decade or two billions of pounds will be at stake.
Up to now the Welsh government has been financed almost entirely by a block grant, determined by the infamous Barnett formula. The U.K. Government has agreed to supplement the formula with a “floor” to stop Welsh spending falling below a certain percentage of the English average. Currently the floor is set at a perfectly reasonable 115 per cent but that commitment lasts only for the current parliament after which we do not know what the floor will be, nor how it will be determined.
When taxes are devolved, in any event, the block grant has to be reduced because Wales then keeps money that would have been going to the Treasury. In the first year of devolution there is nothing to argue about. You make an estimate of the revenue from the devolved tax and take that off the block grant. If the outcome is different from the estimate you subsequently adjust the block grant for the discrepancy. The issue comes in following years. You cannot repeat that procedure because the actual revenue will depend on the policies of the Welsh government. If it reduces tax rates, for example, revenues will fall. If the deduction from the block grant falls too, HMG ends up financing a Welsh tax cut! Similarly it must not take away revenue resulting from a tax increase or tax devolution is defeated.
Ideally what we want to know is: what revenues would HMG have collected from those devolved taxes in a parallel universe where no devolution had occurred. That is the amount we are supposed to deduct from the Welsh block grant in this universe. Evidently those revenues, and so the deduction, will depend on the policies of the U.K. Government but not on the policies of the Welsh government, which has no power in the parallel universe. Unfortunately if there are parallel universes the laws of physics prevent us from inspecting them. We just have to estimate the revenue foregone by the Treasury with the indicators we have.
Consider income tax. Historically while Welsh taxes revenues per head have been lower than those in the rest of the UK they have grow at much the same rate. One approach therefore is to take the initial deduction and then make it grow at the same rate as income taxes in the rest of the U.K. (Known as the Index deduction method).This suggestion hit a snag in Scotland. Population there grows at a slower rate than in the rest of the U.K. so even if incomes and tax payments per head grew at the same rate as in England the total tax would grow more slowly. The Scots therefore argued that their deduction should be indexed to (that is grow at the same rate as) tax revenues per head in the rest of the U.K., not at the rate of total revenues.(Known as the Per capita Index deduction method).
HMG and the Treasury would not agree. The arguments were arcane and in my view specious and there is no space to rehearse them here. The Treasury came up with a different formula whose effects are a bit better than simple indexation to aggregate tax receipts but do not protect properly against relatively slow population growth. (Known as the Tax-base adjusted levels deduction). The compromise was the Treasury imposed its formula but agreed to top up the Scottish grant for five years to give the same effect as per capita indexation. Then there will be a review and if there is no agreement the Treasury formula will continue. The Scots get their way for five years but then they are at the mercy of the Treasury and its formula.
This matters hugely for Wales. The Office of National Statistics projects that the Welsh population will grow at an average rate of 0.21 per cent a year in the 20 years from 2018-19. It projects that population in the UK will grow at an average 0.56 per cent a year over the same period, that is more than twice as fast as in Wales. Note that the ONS is making those projections now on the basis of existing demographic trends. There is no assumption in there about different economic policies in Wales and England affecting, for example, rates of migration. In the economic jargon those projections are of exogenous population trends.
They mean that even if income tax receipts per head grow at the same rate in England and Wales, income tax receipts in England will grow faster, purely because of a relatively rising population. If income tax receipts per head went up at 4 per cent in both countries, after 20 years Welsh total income tax receipts would be up by 29 per cent but receipts in the rest of UK would be up by 45 per cent.
If the adjustment to the Welsh block uses the Treasury formula instead of being indexed to per capita income tax receipts, , Wales will lose £2.6 billion pounds over those 20 years compared with the status quo. And that is even if the Welsh economy matches that of the rest of the UK and its per capita income tax receipts move in line.
In devolving income tax, the UK must in any case impose a new risk on the Welsh government: the risk that its income tax base rises more slowly than that in the rest of the UK because the economy underperforms. If that happens, because tax receipts are partially replacing grant money, Wales will lose relative to the current situation. If Tata closes the Port Talbot works after income tax devolution, for example, or reasons beyond Welsh government control, Wales could lose up to £150 million a year in income tax revenue. However devolution also confers an opportunity. To the extent that Wales can stimulate its economy its tax receipts will benefit, whereas now an acceleration in Welsh growth increases tax revenues only for the UK Treasury. Evidently the ideal situation is one where the Welsh government takes on the responsibility and the risks over which it has some control but is not subjected to risks where it has little or no control and which are better born at the UK level.
Some people oppose income tax devolution on any terms because they argue that the Welsh government does not have the power to affect relative economic growth rates. But that is to accept Wales’ role as a pensioner with no ability even in principle to improve its public finances by improving economic performance. Polls have shown the Welsh people are prepared to shoulder more responsibility. Yet if Wales is to bet on its own ability to improve its condition, the bet should surely be on fair terms. If it is to be penalised for demographic trends over which it has no control, the bet is not fair and only a mug punter would take it on.
While population is the main stumbling block, the recent report by the Wales Governance Centre at Cardiff University points to another. Welsh income tax revenues have grown more slowly than those of the rest of the U.K. since 2010-11. That is at least partly owing to the UK government steeply raising the personal allowance. Because more Welsh taxpayers are at the lower end of the income distribution, those increases in the allowance took a higher proportion of Welsh taxpayers out of tax than the proportion in England. If that was repeated in future after tax devolution, per capita indexation would reflect the reduction in English tax payers but that would not fully compensate the Welsh government for the loss of Welsh tax revenues – which were occurring as the result of a UK policy decision.
More to the point, think of the parallel universe. The rise in the allowance would have cost the Treasury what it will now cost the Welsh government, not the lower figure coming from indexation to England. So even per capita indexation would fail to identify the true opportunity cost to HMG. It would understate the drop in revenue and so would overstate the opportunity cost of ceding that revenue to Wales.
Of course in future, increases in personal allowance are not likely to rise as steeply as in the recent past. The recent past is not typical of the longer run history and is unlikely to be typical of the future. If we want to avoid that risk, however, we have to make things a little more complicated, which the Treasury may resist.
It is easy enough to set up a simple model of the Welsh income-tax-paying population and to simulate the first-round effect of a given policy change. HMRC can do it; the Institute of Fiscal Studies can do it and the Welsh government had better acquire or buy the capacity to do it. Those calculations are what enable the papers to report at Budget time that such and such a measure will raise or cost x billion. If the model is agreed the effect of a UK policy change can be simulated and the result for revenues used to adjust the block grant. To avoid marginal fiddling this procedure would be restricted to changes above a certain size. That would be in accord with the principle of “no detriment” from devolution, to which HMG is committed.
The Welsh Government should therefore insist on full per capita indexation to allow for relative population movements, and should insist that an agreed model be selected to simulate the effect of UK policy decisions on revenues from Wales. In this negotiation Wales is, for once, in a stronger position than Scotland. The SNP government really wanted tax devolution and could not be seen to refuse it. The Welsh Labour government has mixed feelings about income tax devolution anyway and many Ministers would be quite happy to knock it back if the deal is no good. It would be hard for the UK to impose tax devolution on a reluctant Wales when it has waived the necessity of a referendum so if income tax devolution is to happen HMG has to agree to something reasonable for block-grant adjustments.
By the way, Treasury proposals for the block-grant adjustment for Stamp Duty devolution are far worse than what they propose for income tax. They move from the realm of the stingy to the country of the outrageous. The only suitable reply is John McEnroe’s: “you cannot be serious”. But that’s the story for another day.
*Correction: This article first stated that the Barnett Floor was set at 116 per cent. This was corrected to 115 per cent on the 26th February*