How to reform the Barnett formula

The UK has traditionally had a system of expenditure equalisation among local authorities and that has influenced the block grant to the devolved territories.  Revenue support grants for local authorities were not based on equalizing their revenue per inhabitant, as in some other countries. Instead they looked at policy areas for which the local authority was responsible and had formulae to calculate the need for spending in that area given relevant characteristics of the local population, such as  demographics and relative poverty.

When block grants were introduced for territories following some degree of devolution, expenditures remained the focus. Revenues were not the focus. Instead expenditures in England on those policy areas that were devolved to the territory in question were taken as the basis for the block grant. However, the explicit consideration of need was dropped and instead the same absolute increase on expenditure per head in England was added to the previous year’s block grant for the territory – the so-called Barnett formula.

The expenditure base was retained but the grant was no longer related to relative need and was fully fungible to the recipient authority. With time any relation to relative need became more remote. The grant is now the largely arbitrary result of a series of historical accidents and ignores changes in relative income, poverty and demographics that have occurred in the devolved territories since the 1970s.

In determining how to proceed in future there are two broad alternatives. Firstly, the UK could decide it is a federation and the origin of tax revenues matters for expenditure. Each territory would have a claim on the tax collected there. There would then, as in most federations, be a series of transfers to equalise the resources per head available in each territory. That would be a revenue equalisation that would not attempt to assess relative needs or ensure therefore that public services could be provided to the same standard in each part of the federation. A territory with greater needs, such as a different dependency ratio or greater poverty, would have the same revenue per head as everywhere else with no allowance for those extra needs. If little else, this system has simplicity and tractability to recommend it.

Under such a system Wales and Northern Ireland would be much worse off than they currently are. However, Scotland might be in a broadly similar position as it currently enjoys income from oil revenues, depending on the market price, and assuming Scotland gets credit for the revenue from oil production in its geographical sector.

Such a revenue-equalisation system exists in many federations but it would imply necessarily different standards of social services in different parts of the country and the inter-country equalisation system would differ from the intra-country system of revenue support for local authorities if that maintained a system of expenditure equalisation.

The second option would be to attempt to maintain a system of expenditure equalisation whereby territorial or regional revenue support was on the same basis as revenue support for local authorities. That means explicit consideration of need would be required given the areas of devolved policy for which the territories are responsible. If we maintain expenditure equalisation, there does not seem to be any alternative to driving the grant off the expenditures of the largest country, England. Those expenditures could be adjusted for relative need and that would define an expenditure envelop for each territory.

Tax devolution requires a further step before a block grant can be calculated. From the expenditure envelop, defined as above, would be subtracted the tax resources of the territory in order to arrive at the block grant. The tax resources would be of two kinds: those tax revenues assigned to the territory and the proceeds of any tax base devolved to the territory, assuming tax rates are levied at the same ‘standard’ rate as in England. Taxes imposed in the territories that are not levied in England would be ignored as would be any increments or decrements to revenues from having different tax rates from the standard English ones. The principle here is the block grant is calculated assuming a similar tax effort or sacrifice in each territory. If a territory increases or reduces its tax effort either by changing rates or introducing new taxes it takes the proceeds and the block grant is unchanged.

The above arrangement would allow the devolved territory to increase or reduce its tax rates and revenues without prejudicing the block grant. To be sure, there are some complexities involved in valuing the devolved tax base as it moves over time.  That valuation is necessary to update the deductions from the expenditure envelop to get the block grant. A design criterion for that valuation is that it should as far as possible leave the consequences of tax policy decisions by the devolved government to fall on the territory but that the consequences of decisions by the UK government should be born centrally. This is a particular issue where a tax base is shared, like income tax under the Scotland Act. The devolved share of the tax base remains susceptible to changes in allowances and thresholds and these remain under UK control.

Treatment of deductions from the block grant can take one of four forms:

1. Own-base deduction: the deduction is indexed to the assessed growth of the tax base itself.
2. Indexed deduction: the initial deduction is indexed to an external variable such as the relevant UK tax base.
3. Proportionate deduction: the grant is reduced by a given percentage so the initial deduction grows at the same rate as the grant itself.
4. Fixed (real) deduction: the grant is reduced by an agreed sum, which may then be indexed to inflation.

Own-base deduction has the disadvantage that it negates the effects that the devolved government’s policy may have on a tax base. For example, suppose a devolved government cuts taxes so its revenues initially fall. If reduced taxes result in faster economic growth and growth of the tax base, no net benefit is achieved because the deduction from the block grant will grow with the tax base. Similarly, if tax rates are increased and the base shrinks, the deduction will fall, increasing the block grant and buffering the devolved government from the effect of its policy. This system therefore introduces a bias to higher taxes as well as violating the design criteria. It offsets domestic policies while leaving the tax base vulnerable to actions by the UK government in the case of a shared base.

Indexed deduction counteracts these disadvantages if a reasonable index is to hand.  Using the UK tax base as the index means domestic policies are not offset but if the UK government alters the tax base it will compensate the effect in the deduction from the block grant. In my opinion this works well for income tax. However, it does not work well where there are reasons to think the UK tax base will grow at a very different rate from that of the devolved territory.

Proportionate deduction may be reasonable but leaves a shared tax base exposed to changes of policy in the UK and any compensation has to be negotiated ad hoc.

Fixed deduction is the simplest. If governments can agree a present value of expected future tax receipts that can be subtracted from the block grant in a one-time adjustment. That would work for a slow-growing or static tax base. Where the tax base grows with the economy there may still be a real annuity equivalent, that is there is an initial deduction that is then indexed to inflation. The UK Treasury may well be reluctant to agree to this for taxes whose base is growing strongly or may demand a very large initial deduction.

The various approaches are discussed in detail in the report of the Independent Commission on Funding and Finance for Wales Fairness and Accountability of July 2010. Note especially chapter 5 and Annexes 7 and 8.

The UK Treasury initially accepted Proportionate deduction, as recommended by the Calman Commission but subsequently agreed to the Indexed deduction  approach for income tax. Having initially argued that Indexed deduction was “too complicated” it has more recently wished to apply it across the board. This is not in the devolved territory’s interest if its own tax base is inevitably slower growing than that of the UK.

In discussions over corporation tax in the Northern Irish case, the Treasury has also argued for increasing the deduction for losses to the UK tax base from lower taxes in Northern Ireland. That is inappropriate in my view. If one spill-over is considered, all calculable spill-overs should be taken into account. Alternative solutions to questions of spill-over are available such as allocating company tax liability better and limiting tax differentials.

It remains to discuss the determination of relative need if expenditure envelopes are to account for differences in need. In practice needs formulae used in local government and health are generally very complicated, recognising a large number of different factors. That is explicable because they are determined in a political process where the concerns of different groups can be shown to be met if an appropriate variable is included in the formula. However, from a practical viewpoint many of the variables are redundant in fixing the size of revenue support because they are highly correlated with each other. They tend to reflect a few underlying factors. Those are:

• Dependency, that is the proportion of old people and children in the population.
• Poverty, the proportion of poor people.
• Sickness, the proportion of the population with chronic illness.
• Sparsity, which is the biggest determinant of the relative cost of providing services, at least outside London, since it it is more expensive to provide services to a scattered population.

The Independent Commission on Funding and Finance for Wales (See chapter 3 and Annex 4) report found over 95 per cent of the variation in revenue support grants to local authorities in Great Britain and in grants to area health boards could be explained by an equation of some half a dozen variables representing these four factors. A needs adjustment that allowed for these factors in determining relative expenditure per head would therefore be an enormous improvement on making no needs adjustment. On the otherhand, a more refined or complicated effort to identify relative need would be more expensive and more open to political challenge over detail. In my opinion a coarse adjustment should be easy to do and to justify, relative to the option of not doing it. An attempt at refined adjustment could get bogged down in interminable normative disputes.

Of course, even when we have secured agreement on the few basic factors that are relevant to a needs adjustment, there is considerable scope for disagreement over how much weight to attribute to each factor. The Independent Commission’s report simply ran a cross-section regression and obtained parameters from current practice.  In other words, it claimed to have identified the “revealed preference” of current governments.

Of course, that is not the last word. When the results are scrutinised governments may decide that the results are not what they meant and there will be different preferences among the devolved governments of the UK and Westminster.  Nonetheless, the existing dispensation and its summary representation seems a reasonable starting point for discussion.

The evidence suggests that an approach of this kind would leave Wales and Northern Ireland in a situation not massively different from their current one.  It would, however, tend to reduce the resources going to Scotland. To see that it is sufficient to note that Wales scores worse than Scotland on almost all the need factors listed above yet grant-financed expenditure per head in Scotland is considerably higher than in Wales.

It might be necessary to add a clarification here. When we talk of an expenditure envelop, driven off English expenditures, we are talking of a way of determining a block grant. The idea being that it is necessary to take account of relative need if we are to assure the same level of public services in different parts of the country. However, it is unreasonable to go beyond that and to expect English taxpayers to pay for better public services in other parts of the union than they are prepared to finance for themselves.

Devolution of taxes means that the devolved territory is not confined within that envelope. It can have better (or worse) public services if it chooses to raise (or lower) taxes. In either case it will not be penalised by changes to the block grant. That seems to me to be the best achievable situation in the context of a continuing union.