A brief review of ‘The Cost of Basic Income in the United Kingdom: A Microsimulation Analysis’ by Karl Widerquist and Georg Arndt

There is a translation of this review in French

The article by Widerquist and Arndt can be read here; and a pdf can be downloaded here.

Widerquist and Arndt use microsimulation analysis to estimate that the net cost of a poverty-level Basic Income for the United Kingdom (£7,706 per annum for adults and £3,853 for children) is about £67 billion per year or 3.4% of GDP. The paper makes a useful contribution to the current debate about financially feasible Basic Income schemes for the UK.

The authors correctly recognise that their scheme is ‘not optimised for political feasibility’. This is true. British benefits policy is quite path dependent—that is, it tends to continue in its existing direction—mainly because of the extreme complexity of the tax and benefits systems. Change is normally incremental, and this will be particularly true during the next few decades because of the difficulties that the Government has experienced attempting to combine a handful of means-tested benefits into the new means-tested ‘Universal Credit’. Only a Basic Income that made a very small number of changes to the existing systems would be likely to be considered for implementation during the next few decades, whatever the political ideology of the Government. Widerquist’s and Arndt’s scheme makes multiple major changes all across the systems, and so would be unlikely to be considered for implementation.

There are particular aspects of the scheme that would make the scheme even less politically feasible. The authors are commendably honest about the political infeasibility of their suggested 50 per cent basic rate of Income Tax. Equally politically impossible would be the abolition of National Insurance Contributions and National Insurance benefits such as the Retirement Pension, even if that would be replaced by a new Basic Income for elderly people. National Insurance is a concept deeply embedded in the British psyche, and any government that tampered with it would suffer the consequences. It is not insignificant that when the new Single Tier State Pension, which is very close to a Citizen’s Pension, was implemented, a National Insurance Contribution record conditionality was retained. Under the circumstances, the only option is to retain the National Insurance system: although making it fairer, for instance by charging contributions at the same rate across the entire earnings range, rather than at a reduced rate for higher earners, might be politically feasible.  

The authors propose a ‘hold harmless’ mechanism to ensure that low income households that would otherwise have suffered losses on the implementation of the proposed scheme would not in fact do so. The losses occur because the authors have decided to abolish all existing means-tested benefits except for Housing Benefit and various disability benefits. The problem is that they have not specified how the ‘hold harmless’ proposal would be administered. The only feasible way of achieving such a mechanism would be to reintroduce a means-tested benefit that would mirror the benefits that had been abolished. It might be objected that simply recording each household’s disposable income at the point at which the scheme was implemented would be sufficient: but that would only tell the administrators how much to pay on the day after implementation. If would not protect households from subsequent differences between the old system and the new once household circumstances started to change. Only a means-tested benefit could do that. The UK Government has found it difficult enough to combine a few existing means-tested benefits into a single new one. To abolish most existing benefits and then to implement a wholly new one with a particular aim, all at the same time as implementing Basic Incomes and changing the tax system, would be administratively and therefore politically impossible.

The paper contains a useful discussion of the difference between the gross cost and net cost of a Basic Income scheme, and quite rightly points out that rather too many commentators fail to understand the importance of the difference. The authors calculate that the net cost of their scheme would be £67bn per annum. This might be only 3.4 per cent of GDP, but it would still have to be found from somewhere. Suggestions are made at the end of the paper, one of which is that the whole £67bn should be paid by the top 30 per cent of earners. No government would want to alienate that group of voters to that extent, rendering the proposal politically feasibility. We are therefore left with sharing the burden across the earnings range, although not necessarily via an additional Income Tax. What the authors don’t seem to realise is that wherever the money was found it would impact household living standards. For instance, let us suppose that additional consumption taxes or a new carbon tax were to be employed to fill the gap. In either case prices would rise, household disposable incomes would be affected, and the statistics given in the paper for poverty levels and household gains and losses would no longer be correct. Only a Basic Income scheme that fully specifies the funding method, and that calculates the reductions in the Income Tax Personal Allowance, increases in Income Tax rate changes, and other changes, that would achieve a net cost of zero, can be sure that statistics on poverty levels, household gains and losses, etc., generated by microsimulation would be those that would be seen when the scheme was implemented. A scheme with a net cost above zero leaves us entirely in doubt as to the effects of the scheme. This means that in the case of the scheme researched for this paper none of the outcome statistics can be believed.

There are two ways of approaching the question of the financial feasibility of Basic Income schemes: 1. to research the effects of a particular scheme and then decide whether it would be feasible, or 2. to set feasibility criteria and then seek a scheme that would fit those criteria. The second of those approaches is the one that follows scientific method and is therefore the method that ought to recommend itself to social science researchers. The paper under review employs the first of the two methods. As far as the author of this review knows, only his own research follows the second method—a fact that never surprises him because the second method can take days of testing of alternative schemes before a scheme that fits the criteria can be found, and there is always the possibility that no scheme will be found that is anywhere near to being feasible. Widerquist and Arndt reference a 2019 paper that employs the second of the two methods. The most recent such research can be found here.

Widerquist and Arndt are to be commended for contributing to the UK’s strong research tradition on the financial feasibility of Basic Income schemes. If further research were to take account of the hesitations discussed in this review, and were to employ the second of the two methods outlined in the previous paragraph, then their future research would make an even bigger contribution.