A basic income is an income paid by a political community to all its members on an individual basis, without means test or work requirement. This definition does not fit all actual uses of the English expression “basic income”, or of its most common translations in other European languages, such as “Bürgergeld”, “allocation universelle”, “renda basica”, “reddito di cittadinanza”, “basisinkomen”, or “borgerlon”. Some of these actual uses are broader : they also cover, for example, benefits whose level is affected by one’s household situation or which are administered in the form of tax credits. Other uses are narrower: they also require, for example, that the level of the basic income should coincide with what is required to satisfy basic needs or that it should replace all other transfers. The aim of the above definition is not to police usage but to clarify arguments. Each of its components are explained in more detail below.
One can conceive of a benefit that would have all other features of a basic income but be provided in kind, for example in the form of a standardised bundle of food, or the use of a plot of land. Or it could be provided in the form of a special currency with restricted uses, for example food stamps or housing grants, or more broadly consumption in the current period only without any possibility of saving it, as in Jacques Duboin’s (1945) “distributive economy”. A basic income, instead, is provided in cash, without any restriction as to the nature or timing of the consumption or investment it helps fund. In most variants, it supplements, rather than substitutes, existing in-kind transfers such as free education or basic health insurance.
Is Basic Income paid on a regular basis, rather than as a one-off endowment?
A basic income consists in purchasing power provided at regular intervals, such as a week, a month, a term or a year, depending on the proposal. One can also conceive of a benefit that would have all other features of a basic income but be provided on a one-off basis, for example at the beginning of adult life. This has occasionally been proposed, for example long ago by Thomas Paine (1796) and far more recently by Bruce Ackerman and Anne Alstott (1999). There is a significant difference between a regular basic income and such a basic endowment. Yet, it should not be overstated. Firstly, the basic endowment can be invested to generate an actuarially equivalent annual or monthly income up to the recipient’s death, which would amount to a regular basic income. If left to the insurance market, the level of this annuity would be negatively affected by the length of a person’s life expectancy. Women, for example, would receive a lower annuity than men. However, the advocates of a basic endowment (including Paine and Ackerman and Alstott) usually supplement it with a uniform basic pension from a certain age, which erases most of this difference. Secondly, while other uses can be made of a basic endowment than turning it into an annuity, the resulting difference with a basic income would be essentially annulled if the latter’s recipients could freely borrow against their future basic income stream. Even if one wisely protects basic income against seizure by creditors, the security it provides will make it easier for its beneficiaries to take loans at every stage and will thereby reduce the gap between the ranges of options opened, respectively, by a one-off basic endowment and a regular basic income.
Is Basic Income restricted to the Nation-state?
By definition, a basic income is paid by a government of some sort out of publicly controlled resources. In most proposals, the basic income is supposed to be paid, and therefore funded, at the level of a Nation-state, as sometimes indicated by the very choice of such labels as “state bonus”, “national dividend” or “citizen’s wage”. However, it can in principle also be paid and funded at the level of a politically organised part of a Nation-state, such as a province or a commune. Indeed, the only political unit which has ever introduced a genuine basic income, as defined, is the state of Alaska in the United States (see Palmer 1997). A basic income can also conceivably be paid by a supra-national political unit. Several proposals have been made at the level of the European Union (see Genet and Van Parijs 1992) and some also, more speculatively, at the level of the United Nations (see e.g. Kooistra 1994, Frankman 1998, Barrez 1999).
The basic income may, but need not, be funded in a specific, ear-marked way. If it is not, it is simply funded along with all other government expenditures out of a common pool of revenues from a variety of sources. Among those who advocated ear-marked funding, most are thinking of a specific tax. Some want it funded out of a land tax or a tax on natural resources (from Thomas Paine (1796) to Raymond Crotty (1987), Marc Davidson (1995) or James Robertson (1999) for example). Others prefer a specific levy on a very broadly defined income base (for example, Pelzer 1998, 1999) or a massively expanded value-added tax (for example, Duchatelet 1992, 1998). And some of those who are thinking of a worldwide basic income stress the potential of new tax instruments such as “Tobin taxes” on speculative capital movements (see Bresson 1999) or “bit taxes” on transfers of information (see Soete & Kamp 1996).
Redistributive taxation need not be the only source of funding for basic income. Alaska’s dividend scheme (O’Brien & Olson 1990, Palmer 1997) is funded out of part of the return on a diversified investment fund which the state built up using the royalties on Alaska’s vast oil fields. In the same vein, James Meade’s (1989, 1993, 1994, 1995) blueprint of a fair and efficient economy comprises a social dividend funded out of the return on publicly owned productive assets. Finally, there has been a whole sequence of proposals to fund a basic income out of money creation, from Major Douglas’s Social Credit movement (see Van Trier 1997) and Jacques and Marie-Louise Duboin’s (1945, 1985) Mouvement français pour l’abondance to the recent writings of Joseph Huber (1998, 1999, 2000 with J. Robertson).
There can be more or less inclusive conceptions of the membership of a political community. Some, especially among those who prefer the label “citizen’s income”, conceive of membership as restricted to nationals, or citizens in a legal sense. The right to a basic income is then of a piece with the whole package of rights and duties associated with full citizenship, as in the conception of the French philosopher Jean-Marc Ferry (1995). Others, especially among those who view basic income as a general policy against exclusion need to conceive of membership in a broader sense that tends to include all legal permanent residents. The operational criterion may be, for non-citizens, a minimum length of past residence, or it may simply be provided by the conditions which currently define residence for tax purposes.
There can also be a more or less inclusive conception of membership along the age dimension. Some restrict basic income, by definition, to adult members of the population, but then tend to propose it side by side with a universal, i.e. non-means-tested, child benefit system, with a level of benefit that may or may not be differentiated as a (positive or negative) function of the rank of the child or as a (positive) function of the child’s age. Others conceive of basic income as an entitlement from the first to the last breath and therefore view it as a full substitute for the child benefit system. The level of the benefit then needs to be independent of the child’s family situation, in particular of his or her rank. Some also want it to be the same as for adults, and hence independent of age, as is actually the case in the modest Alaskan dividend scheme and as would be the case under some more generous proposals (for example Miller 1983). But the majority of those who propose an integration of child benefits into the basic income scheme differentiate the latter’s level according to age, with the maximum level not being granted until majority, or later.
Analogous to the case of children, some restrict basic income to members of the population which have not reached retirement age and then see it as a natural complement to an individual, non-means-tested, non-contributory basic pension pitched at a higher level, of a sort that already exists in some European countries, like Sweden or the Netherlands. In most proposals, however, the basic income is granted beyond retirement age, either at the same level as for younger adults or at a somewhat higher level. In all cases, this basic income for the elderly can be supplemented by income from public or private contributory pension schemes, as well as from private savings and from employment.
Even on the most inclusive definition of the relevant notion of membership, any population is still likely to contain some people who will not be paid a basic income. Detaining criminals in prison is far more expensive to the community than paying them a modest basic income, even if full account is taken of any productive work they may be made to perform. Unless the detention turns out to have been ill-founded, it is therefore obvious that prison inmates should lose the benefit of their basic income for the duration of their imprisonment. But they can get it back as soon as they are released. The same may apply to the long-term inmates of other institutions, such as institutions for the mentally ill or elderly, to the extent that the full cost of their stay is directly picked up by the community rather than paid for by the inmates themselves.
Does Basic Income imply uniform levels of benefit?
A basic income is paid to each individual member of the community, rather than to each household taken as a whole, or to its head, as is the case under most existing guaranteed minimum schemes. Even if a benefit is paid to each individual, its level could still be affected by the composition of the household. To take account of the fact that the per capita cost of living decreases with the size of the household, existing guaranteed minimum income schemes grant a smaller per capita income to the members of a couple than to a person living alone. A fair and effective operation of such schemes therefore supposes that the administration should have the power to check the living arrangements of their beneficiaries. A basic income, instead, is paid on a strictly individual basis. Not only in the sense that each individual member of the community is a recipient, but also in the sense that how much (s)he receives is independent of what type of household she belongs to. The operation of a basic income scheme therefore dispenses with any control over living arrangements, and it preserves the full advantages of reducing the cost of one’s living by sharing one’s accommodation with others. Precisely because of its strictly individualistic nature, a basic income tends to remove isolation traps and foster communal life.
Relative to existing guaranteed minimum income schemes, the most striking feature of a basic income is no doubt that it is paid, indeed paid at the same level, to rich and poor alike, irrespective of their income level. Under the simplest variant of the existing schemes, a minimum level of income is specified for each type of household (single adult, childless couple, single parent of one child, etc.), the household’s total income from other sources is assessed, and the difference between this income and the stipulated minimum is paid to each household as a cash benefit. In this sense, existing schemes operate ex post, on the basis of a prior assessment, be it provisional, of the beneficiaries’ income. A basic income scheme, instead, operates ex ante, irrespective of any income test. The benefit is given in full to those whose income exceeds the stipulated minimum no less than to those whose income falls short of it. Nor are any other means taken into account when determining the level of benefit a person is entitled to: neither a person’s informal income, nor the help she could claim from relatives, nor the value of her belongings. Taxable “means” may need to be taxed at a higher average rate in order to fund the basic income. But the tax-and-benefit system no longer rests on a dichotomy between two notions of “means”: a broad one for the poor, by reference to which benefits are cut, and a narrow one for the better off, by reference to which income tax is levied.
From the fact that rich and poor receive the same basic income, it does not follow, however, that the introduction of a basic income would make both rich and poor richer than before. A basic income needs to be funded. If a basic income were simply added to existing tax-and-benefit systems, it is clear that the comparatively rich would need to pay both for their own basic income and for much of the basic income of the comparatively poor. This would clearly hold if the funding were through a progressive income tax, but would also hold under a flat tax or even a regressive consumption tax. For the ex nihilo introduction of a basic income to work to the financial advantage of the poor, the key condition is simply that, relative to their numbers (not necessarily to their incomes), the relatively rich should contribute more to its funding than the relatively poor. In most proposals, however, the introduction of a basic income is combined with a partial abolition of existing benefits and tax reductions. If the proposed reform simply consisted in spreading more thinly among all citizens the non-contributory benefits currently concentrated on the poor, the latter would clearly lose out. But no one is making such an absurd proposal. In most proposals that rely on direct taxation, the basic income replaces only the bottom part of the non-contributory benefits, but also the exemptions or reduced tax rates on every taxpayer’s lower income brackets. The immediate impact on the income distribution can then be kept within fairly narrow bounds for a modest basic income. But the higher its level, the higher the average rate of income tax and therefore the greater the redistribution from the comparatively rich to the comparatively poor.
Thus, giving to all, rich and poor, is nor meant to make things better for the rich. But, for a given level of minimum income, is there any reason to believe that it is better for the poor than a means-tested guaranteed income? Yes, for at least three interconnected reasons. Firstly, the rate of take up of benefits is likely to be higher under a universal scheme than if a means test is in place. Fewer among the poor will fail to be informed about their entitlements and to avail themselves of the benefits they have a right to. Secondly, there is nothing humiliating about benefits given to all as a matter of citizenship. This cannot be said, even with the least demeaning and intrusive procedures, about benefits reserved for the needy, the destitute, those identified as unable to fend for themselves. From the standpoint of the poor, this may count as an advantage in itself, because of the lesser stigma associated with a universal basic income. It also matters indirectly because of the effect of the stigma on the rate of take up. Thirdly, the regular, reliable payment of the benefit is not interrupted when accepting a job under a basic income scheme, whereas it would be under a standard means-tested scheme. Compared to means-tested schemes guaranteeing the same level of minimum income, this opens up real prospects for poor people who have good reasons not to take risks. This amounts to removing one aspect of the unemployment trap commonly associated with conventional benefit systems, an aspect to which social workers are usually far more sensitive than economists.
The other aspect of the unemployment trap generated by means-tested guaranteed minimum schemes is the one most commonly stressed by economists. It consists in the lack of a significant positive income differential between no work and low-paid work. At the bottom end of the earnings distribution, if each Euro of earnings is offset, or practically offset, or more than offset, by a loss of one Euro in benefits, one does not need to be particularly lazy to turn down a job that would yield such earnings, or to actively look for such jobs. Given the additional costs, travelling time or child care problems involved, one may not be able to afford to work under such circumstances. Moreover, it would generally not make much sense for employers to design and offer such jobs, as people who would be grateful for being sacked are unlikely to constitute a conscientious and reliable work force. A minimum wage legislation may anyway prevent full-time jobs from being offered a wage lower than the income guarantee, in which case the latter consideration only applies to part-time jobs. The replacement of a means-tested guaranteed income by a universal basic income is often presented as a way of tackling this second aspect of the unemployment trap too. If one gave everyone a universal basic income but taxed at 100% the portion of everyone’s earnings that does not exceed the minimum guarantee (see for example Salverda 1984), the unemployment trap would be the same, in this respect, as under a means-tested guaranteed minimum income. [Fig.1 and Fig.3] But if one makes the mild assumption that the explicit tax rate applying to the lowest income brackets must remain noticeably lower than 100%, then the following statement holds. Since you can keep the full amount of your basic income, whether working or not, whether rich or poor, you are bound to be better off when working than out of work. [Fig. 2]
Is a Basic Income equivalent to a negative income tax?
If you can keep the full amount of your basic income, whether working or not, whether rich or poor, you are bound to be better off when working than out of work. [Fig. 2] But this aspect of the unemployment trap can be removed just as effectively, it would seem, by a means-tested scheme that would phase out the benefit less steeply as earnings rise. This is achieved through the so-called negative income tax, a uniform and refundable tax credit. The notion of a negative income tax first appears in the writings of the French economist Augustin Cournot (1838). It was briefly proposed by Milton Friedman (1962) as a way of trimming down the welfare state, and explored at more depth by James Tobin (1965, 1966, 1967, 1968) and his associates as a way of fighting poverty while preserving work incentives. On the background of an explicit tax schedule which taxes no income at 100% and which can be, but need not by definition be, linear, a negative income tax amounts to reducing the income tax liability of every household (of a given composition) by the same fixed magnitude, while paying as a cash benefit the difference between this magnitude and the tax liability whenever this difference is positive. [Fig. 3] Suppose the fixed magnitude of the tax credit is pitched at the same level as under some basic income scheme under consideration. Someone with no income, and hence no income tax liability will then receive an amount equal to the basic income. As the income rises, the benefit will shrink, as in the case of conventional means-tested schemes, but a slower rate, indeed at a rate that will keep post-and-transfer income at exactly the same level as under the corresponding basic income scheme. [Fig. 3 and Fig. 4] The NIT variant simply consists in netting out taxes and benefits. Under a basic income scheme, the revenues needed to fund the NIT’s universal tax credit are actually raised and paid back to all. Under NIT, transfers are all one-way only: positive transfers (or negative taxes) for households under the so-called break even point, negative transfers (or positive taxes) for households above. [Fig. 3]
How much of a real difference there is between the a basic income and a negative income tax depends on further specification of administrative procedures. It shrinks, for example, if taxes are levied at source on a pay-as-you-earn basis (rather than only after tax returns have been processed), or if tax liabilities are assessed on a weekly or monthly, rather than an annual basis, or if everyone is entitled, under a NIT scheme, to an advance payment of the presumptive tax credit (subject to subsequent correction), or if everyone is entitled, under a BI scheme, to get the BI as a tax discount rather than in cash. But even in the closest variant, there remains a difference between a system that operates, by default, “ex ante”, and one that operates, by default, “ex post”. Any remaining difference would count as an advantage for the basic income variant with respect with the first, uncertainty-linked dimension of the unemployment trap. Yet, with a rudimentary benefit payment technology (coins carried by the postman!) or with a tax collection administration plagued with corruption or inefficiency, the case for the NIT variant, which does away with the back-and-forth of tax money, may be overwhelming. In an era of technological transfers and with a reasonably well run tax administration, on the other hand, the bulk of the administrative cost associated with an effective guaranteed minimum income scheme is the cost of information and control: the expenditure needed to inform all potential beneficiaries about what their entitlements are and to check whether those applying meet the eligibility conditions. In these respects, a universal system is bound to perform better than a means-tested one. As automaticity and reliability increase on both the payment and the collection side, it is therefore, in this administrative sense, increasingly likely to be the cheaper of the two, for a given degree of effectiveness at reaching all the poor. In is for this sort of reason that James Tobin (1997), for example, preferred a universal “demogrant” to its negative-income-tax variant.
Should Basic Income be paid irrespective of present work performance?
The right to a guaranteed minimum income is by definition not restricted to those who have worked enough in the past, or paid in enough social security contributions to be entitled to some insurance benefits. From Juan Luis Vives (1526) onwards, however, its earliest variants were often linked to the obligation to perform some toil, whether in the old-fashioned and ill-famed workhouses or in a more varied gamut of contemporary private and public workfare settings. Being unconditional, a basic income sharply contrasts with these forms of guaranteed income intimately linked to guaranteed employment. It also diverges from in-work benefits restricted to households at least one member of which is in paid employment, such as the American Earned Income tax Credit or the UK’s more recent Working Families Tax Credit. By virtue of removing the unemployment trap – i.e. by providing its net beneficiaries with an incentive to work – a basic income (or a negative income tax) can be understood and used as an in-work benefit or a top-up on earnings. But it not restricted to this role. Its unconditionality marks it off from any type of employment subsidy, however broadly conceived.
Should Basic Income be paid irrespective of willingness to work?
It also marks it off from conventional guaranteed minimum income schemes, which tend to restrict entitlement to those willing to work in some sense. The exact content of this restriction varies a great deal from country to country, indeed sometimes from one local authority to another within the same country. It may involve that one must accept a suitable job if offered, with significant administrative discretion as to what “suitable” may means in terms of location or skill requirements; or that one must give proof of an active interest in finding a job; or that one must accept and respect an “insertion contract”, whether connected to paid employment, to training or to some other useful activity. By contrast, a basic income is paid as a matter of right – and not under false pretences – to homemakers, students, break-takers and permanent tramps. Some intermediate proposals, such as Anthony Atkinson’s (1993a, 1993b, 1996, 1998) “participation income”, impose a broad condition of social contribution, which can be fulfilled by full- or part-time waged employment or self-employment, by education, training or active job search, by home care for infant children or frail elderly people, or by regular voluntary work in a recognised association. The more broadly this condition is to be interpreted, the less of a difference there is with a basic income.
Are means-unconditionality and work-unconditionality linked?
If we want no means test, it is important to drop the work test. Bringing together the last two unconditionalities discussed – the absence of the means test and the absence of the work test – makes it possible to briefly formulate the core of what makes basic income particularly relevant under present circumstances. At first sight, there is total independence between these two unconditionalities, between the absence of an income test and the absence of a work test. But the strength of the basic income proposal crucially hinges on their being combined. The abolition of the means test, as we have seen, is intimately linked to the removal of the unemployment trap (in its two main dimensions), and hence to the creation of a potential for offering and accepting low-paid jobs which currently do not exist. But some of these jobs can be lousy, degrading dead-end jobs, which should not be promoted. Others are pleasant, enriching stepping-stone jobs, which are worth taking even at low pay because of their intrinsic value or the training they provide. Who can tell the difference? Not legislators or bureaucrats, but the individual workers who can be relied upon to know far more than what is known “at the top” about the countless facets of the job they do or consider taking. They have the knowledge that would enable them to be discriminating, but not always the power to do so, especially if they have poorly valued skills or limited mobility. A work-unconditional basic income endows the weakest with bargaining power in a way a work-conditional guaranteed income does not. Put differently, work-unconditionality is a key instrument to prevent means-unconditionality from leading to the expansion of lousy jobs. If there is no means test, no work test is needed. At the same time the work incentives associated by means-unconditionality makes work-conditionality less tempting as a way of alleviating the fear that benefits without a counterpart would nurture an idle underclass. In the absence of a means test, the tax and benefit structure can be expected to be such that beneficiaries can significantly increase their disposable incomes by working, even at a low rate and on a part-time basis, and without being trapped in such jobs once their skills improve or once they can improve their working time. Moving (back) into the work sphere will therefore be facilitated and encouraged, and, for those who fear a dualisation of society into workers and non-workers, there will therefore be far less of a need to insist on coupling the right to the benefit to some obligation to (be available for) work. To put it (somewhat too) succinctly: Just as work-unconditionality prevents means-unconditionality from unacceptably supporting exploitation (which it would do by subsidising unworthy low-paid jobs accepted under the threat of losing the benefit), similarly means-unconditionality prevents work-unconditionality from unacceptably fostering exclusion (which it would do by inviting one to no longer regard as problematic a system that durably disconnects the less productive from any labour participation by effectively killing off low-productive jobs). The two key unconditionalities of basic income are logically independent, but they are intrinsically linked as components of a strong proposal.
This solidarity between means-unconditionality and work-unconditionality underlies the central case for basic income as a specific way of handling the joint challenge of poverty and unemployment. Compared to guaranteed income schemes of the conventional sort, the crucial argument in favour of the desirability of basic income rests on the widely shared view that social justice is not only a matter of right to an income, but also of access to (paid and unpaid) activity. The most effective way of taking care of both the income and the activity dimension consists in maintaining the income transfer (in gross terms) whatever the person’s activity, thereby “activating” benefits, i.e. extending them, beyond forced inactivity, to low-paid activity. It may correctly be objected that there are other schemes – such as earned income tax credit or employment subsidies – that could serve better, or more cheaply, the objective of securing the viability of low-productive jobs and thereby providing a paid job to the worst off. However, if the concern is not to keep poor people busy at all cost, but rather to provide them with access to meaningful paid activity, the very unconditional nature of a basic income is a crucial advantage: it makes it possible to spread bargaining power so as to enable (as much as is sustainable) the less advantaged to discriminate between attractive or promising and lousy jobs. It is therefore on the basis of a comprehensive conception of social justice, which gives work the importance it deserves, and not in spite of it, that the right to a basic income should be as unconditional as is sustainably generalisable to all. (See Van Parijs ed. 1995 for a variety of ethical justifications of basic income, Van Parijs 1995 for a systematic statement of the argument just sketched, and Elkin ed. 1997, Krebs ed. 2000 and Williams ed. 2001 for sets of critical assessments of this argument.)
Phrased in this very general way, the question makes no sense. Let us bear in mind that it is not part of the definition of a basic income that it should be sufficient to satisfy the beneficiaries’ basic needs: consistently with its definition, the level of the basic income could be more and it could be less. Nor is it part of the definition of a basic income that it should replace all other cash benefits: a universal benefit need not be a single benefit. A meaningful answer can only start being given to the question of affordability if one specifies the level at which the basic income is to be pitched and stipulates which benefits, if any, it is to replace. Under some specifications – for example “abolish all existing benefits and redistribute the corresponding revenues in the form of an equal low benefit for all” -, the answer is trivially yes. Under other specifications – for example “keep all existing benefits and supplement them with an equal benefit for all citizens at a level sufficient for a single person to live comfortably” -, the answer is obviously “no”. Each of these absurd extreme proposals is sometimes equated, by definition, with basic income. But neither has, to my knowledge, been proposed by anyone. Every serious proposal lies somewhere in between, and whether some basic income proposal is affordable must therefore be assessed case by case.
Is Basic Income more expensive because work-unconditional?
Are there general reasons why a basic income would not be affordable at a level at which a conventional guaranteed income would? One obvious reason might simply be that a basic income is given to all, whether or not they are willing to work, whereas a conventional guaranteed minimum income is subordinated to a willingness-to-work test. As a result, it is claimed, more poor people will be receiving a basic income than a conventional guaranteed income, or, if the number beneficiaries is not much greater, they will be doing less work than would be the case under a work-conditional benefit system. In net terms, therefore, a basic income scheme is certain to cost more.
Is there a dilemma in “Job seeker’s allowance” versus “state-sponsored workfare”?
Closer scrutiny reveals that this expectation rests on feeble grounds indeed. For suppose first that the work test is conceived as an obligation to accept work if offered by some (private or public) employer concerned to get value for money. If the worker has no desire to take or keep the job, her expected and actual productivity is unlikely to be such that the employer will want to hire and keep her. But if the worker is formally available for work, the fact that she is not hired or that she is sacked (owing to too low a productivity, not to anything identifiable as misconduct) cannot disqualify her from a work-tested guaranteed income any more than from an unconditional basic income. The only real difference between the former and the latter is then simply that the former involves a waste of both the employers’ and the workers’ time. Alternatively, suppose that the work test is conceived as an obligation to accept a fall-back job provided by the state for this very purpose. Rounding up the unemployable and unmotivated is not exactly a recipe for high productivity, and even leaving aside the long-term damage on the morale of the conscripted and on the image of the public sector, the net cost of fitting this recalcitrant human material into the workfare mould might just about manage to remain lower than plain prison, with the cost of supervision and blunder correction overshadowing the work-shy workers’ contribution to the national product. The economic case for the work test is just about as strong as the economic case for prisons.
As is fully recognised by no-nonsense advocates of workfare (e.g. Kaus 1990), if a willingness-to-work condition is to be imposed, it must be justified on moral or political grounds, not on the basis of a flimsy cost argument inspired by the shaky presumption that a benefit coupled with work is necessarily cheaper than the same benefit taken alone. From the fact that workfare is likely to be costlier than welfare, it does not follow that the “unemployable” should be left to rot in their isolation and idleness. There can and must be a way of helping them out of it, namely by creating a suitable structure of incentives and opportunities of a sort a universal basic income aims to help create, whether or not a willingness-to-work test is coupled with it. Setting up such a structure is costly, as we shall shortly see, but adding a work test will not make it any cheaper – quite the contrary. And the absence of such a test, therefore, cannot be what jeopardises basic income’s affordability.
Is Basic Income more expensive because income-unconditional?
Instead of resting on the fact that a basic income is paid to all, whether or not they show any willingness to work, the claim that a basic income is unaffordable invokes even more often the fact that it is paid to rich and poor alike. The discussion of the means test should have made plain that this allegation is wrong, misled as it is by too superficial a notion of cost. As the comparison of Fig.1 and Fig. 2 shows, it is in principle possible to achieve with a basic income exactly the same relationship between gross and net income as with a conventional guaranteed minimum income. If this relationship is the same, it means that the cost to those taxpayers who net contributors to the scheme is the same in both cases. If one is politically affordable, therefore, the other should be too. If the relationship is the same, it also means that the marginal tax on earnings at any level of earnings is the same in both cases. If one of the two schemes is economically affordable, therefore, the other should be too.
Of course, the budgetary cost is hugely different in the two cases, and if one could sensibly reason about transfers in the same way as about other public expenditures, there would indeed be a strong presumption that a basic income may be “unaffordable” when a conventional guaranteed minimum income is within our means. But transfers are not net expenditures. They are reallocations of purchasing power. This does not mean that they are costless. They do have a distributive cost to the net contributors, and they do have an economic cost through the disincentives they create. But both costs, we have seen, can be the same under either scheme. In addition, there are administrative costs. But, as also pointed out earlier, assuming a computerised and efficient tax-collection and transfer-payment technology, these are likely to be lower under a universal, ex ante scheme, than under a means-tested, ex post one, at least for a given level of effectiveness at reaching the poor. Paradoxically, therefore, giving to all is not more expensive but cheaper than giving only to the poor.
Is Basic Income more expensive because it creates work incentives at the bottom ?
Marginal rates at the bottom and in the middle: the big trade off. To be fair, however, the fact that the basic income is not means-tested naturally combines with the mild requirement that the explicit rate of tax should fall short of 100%. Which means that the sort of basic income proposal we should be looking at is not represented by Figure 2, but rather by Figure 4, or at least by Figure 6. [Figures 1, 4, 6] Relative to the conventional guaranteed minimum scheme represented by Figure 1, it can then no longer be said that there is not genuinely higher cost. True, it does not uniquely stick to the universal nature of the benefit, since the corresponding means-tested negative-income-tax variants share exactly the same feature. In particular, a linear tax combined with a uniform refundable tax credit at the current level of the minimum guaranteed income would be very expensive in this sense. [Fig. 3] But that the problem should be entirely shared with negative income schemes does not make it less of a problem, which needs to be faced squarely. The basic fact is that the more material incentives one wishes to provide (for a given minimum income) to people earning at the bottom of the earnings scale, the more one needs to decrease the material incentives higher up. There is a sharp trade off here. [Cf. Example]
Is it better for the poor that the poor be taxed more?
This is not as terrible as it sounds. The modestly paid workers whose marginal tax rate would need to go up are also among the main beneficiaries of the introduction of a basic income, as the increased taxation of their wage falls short of the level of the basic income which they henceforth receive. The concern, therefore, need not be distributive. Even if one ends up, as in some proposals, with a linear income tax, i.e. if the lowest earnings are taxed at the same rate as the highest ones currently are, the reform would still redistribute downwards from the higher earners (whose tax increase on all income layers would exceed their basic income). However, there is some ground for a legitimate concern about the impact such a reform would have on incentives. As stressed by some opponents of basic income and negative income tax (e.g. the marginal rates would be lowered in a range in which there is a possibly growing, but still comparatively small proportion of the economy’s marginal earnings, while being raised in a range in which far more workers would be affected. The incentive to work and train, to be conscientious and innovative would be increased in the very lowest range of incomes (say, between 0 and 500 Euro per month), but it would be decreased upward of this threshold, where the bulk of society’s work force, and particularly of its most productive work force, is concentrated. We would therefore be well advised not to rush too quickly to a system in which the effective marginal tax rate on the lowest incomes would not be higher than those higher up (see Piketty 1997).
Should we prefer a low earners’ overcharge or a partial basic income?
We would be well advised not to rush too quickly to a system in which the effective marginal tax rate on the lowest incomes would not be higher than those higher up. There are two ways of accommodating this advice in a basic income proposal. One consists in correcting a linear, or even a progressive system with an “overcharge” for the net beneficiaries of the basic income, as suggested for example by James Meade (1989). [Fig. 6] Another is a “partial basic income”, as proposed for example by the Dutch Scientific Council for Government Policy (WRR 1985) and explored at length since, both in the Netherlands (Dekkers & Noteboom 1988, de Beer 1993, van der Veen & Pels eds. 1995, Groot 1999) and in other European countries (Atkinson 1989, Parker ed. 1991, Lahtinen 1992, Brittan 1995, Gilain & Van Parijs 1995, Clark & Healy 1997). A partial basic income would fall short of the level of income currently guaranteed to a single person, but it may approach or even exceed half the level currently guaranteed to a couple, and it would go hand in hand with the maintenance of a residual means-tested guaranteed income scheme. It would therefore imply the preservation of a 100% effective tax rate on a shrunk lower range. [Fig. 7] Under either variant, the earlier paradox becomes sharper: it is not only better for the poor that the rich should receive the same as the poor. It is also better for them that they should be taxed more than the rich.
Is Basic Income more expensive because strictly individual?
It cannot be denied that the lifting of the means test raises a genuine cost problem, not as such by virtue of the fact that the basic income is given to the rich as well as to the poor but because (part of) its point is to provide the poor with stronger material incentives. It is not the only genuine cost problem intrinsic to basic income proposals. Another directly stems from the fact that, unlike most existing guaranteed minimum income schemes, basic income is meant to be strictly individual. These schemes typically provide a lower level of income support to each of the two members of a couple than to a single person, especially when account is taken of the housing subsidy, sometimes administered as a separate benefit. Why? Obviously because it is cheaper per capita to share a house, durable goods (cooker, washing machine, car, bed?) and some services (child care) with one or more other people than to shoulder the cost individually. The cheapest way of covering a given definition of fundamental needs therefore involves tracking the household composition and modulating the per capita level of the income guarantee accordingly. Of course, the corollary of this household-conditionality is that economies of scale are discouraged, fake domiciles rewarded and hence checks on people’s living arrangements required. One of the blatant advantages of basic income is precisely that it would do away with all that. People who put up with each other and thereby make society save on accommodation and consumer durables would be entitled to the benefits of the economies of scale they generate. There would therefore also be no bonus for those pretending to live apart when they do not, and no need to check who lives where and with whom.
Should a Basic Income be inadequate rather than household-based?
At what level would the individual and unconditional basic income be pitched. If it is at the level of the guaranteed income currently enjoyed by each member of a couple, the amount is bound to fall far short of what is needed by someone who has no option but to live alone. If it is at the level currently awarded to a single person, the cost implications, in some countries at any rate, are phenomenal. This is again not just a matter of budgetary cost. There is an irreducible distributive cost in the sense of a dramatic shift of purchasing power from one-adult to bi- or multi-adult households. And there is also an irreducible economic cost, owing mainly to a substantial increase in the marginal rates required in order to fund the outlays for this enhanced basic income. There is therefore, in the short term at any rate, a dilemma between giving a fully individualised but inadequate basic income and giving a sufficient but household-modulated one (see Brittan & Webb 1991, Brittan 1995). Note, however, that this dilemma is not to be confused with a dilemma between making some households unacceptably poor (with too low an individual basic income) and subjecting all households for an indefinite period to a control of their living arrangements (with an adequate, but household-dependent basic income). Even under short-term cost constraints, the latter dilemma does not hold, for it is possible to conceive of a strictly individual but inadequate “partial” basic income for all, combined with a much shrunk residual means-tested household-tested social assistance for the reduced number of those who, despite the floor provided by the household’s basic income(s), do not earn enough to reach the income threshold as from which means-tested assistance is switched off. [Fig. 7] Providing it is not conceived as an immediate full substitute for existing social assistance, such a partial basic income thus provides an attractive way of handling both of the real cost problems – those stemming from incentives for low earners and individualisation – which a full basic income would raise (see e.g. Gilain & Van Parijs 1995 for a microsimulation of the distributive impact of such a partial basic income in the case of Belgium).
For reasons explained at length elsewhere (Van Parijs 1995), a coherent and plausible conception of social justice requires us to aim, with some important qualifications, for an unconditional basic income at the highest level that is economically and ecologically sustainable, and on the highest scale that is politically imaginable. But while a defensible long-term vision is important, precise proposals for modest, immediately beneficial and politically feasible steps are no less essential. The sort of general but household-tested, means-tested and willingness-to-work-tested guaranteed minimum scheme that is now in place with many variants in most EU countries (including, most recently, Portugal) is a fundamental step in the right direction. But whatever the well-meaning “insertion” or “integration” conditions, it cannot avoid generating traps whose depth increases with the generosity of the scheme and whose threat increases as so-called “globalisation” sharpens inequalities in market earning power. In countries in which guaranteed minimum schemes have been operating for a while, these traps and the dependency culture said to be associated with it risk triggering off a political backlash and the dismantling of what has been achieved. But they have also been prompting progressive moves in the form of basic income and related proposals. Like the fight for universal suffrage, the fight for basic income is not an all-or-nothing affair. This is no game for purists and fetishists, but for tinkerers and opportunists. Without going all the way to even a partial basic income, the following three types of proposals are plausible candidates – more or less plausible, depending on each country’s institutions, and in particular its tax and social security context – as the most promising next step: (1) an individual tax credit, (2) a household-based regressive negative income tax, (3) a modest participation income.
The Netherlands already have universal (i.e. non-means tested) systems of child benefits, of student grants and of non-contributory basic pensions, in addition to one of the world’s most generous and comprehensive means-tested guaranteed income schemes. In January 2000, the Dutch Parliament approved the essentials of the government’s plan for a comprehensive tax reform incorporating the replacement of the exemption on the lower income layer by a strictly individual tax credit at a level of about Euro 140 per month for all families with at least one worker (see Boerlage 1999). Gradually increased and made individually refundable (so that a worker’s non-working partner, for example, would be entitled to a cash payment equivalent to the credit rather than have the working partner doubly credited), this “negative income tax” for working families would provide the last missing element for the provision of a universal income floor. It could then be painlessly integrated into a low, but strictly individual, universal and unconditional basic income. Of course, even at a significantly increased level, this would remain a partial basic income, which would need to keep being supplemented, et any rate for single-adult households, with residual means-tested assistance.
What role for a household-based regressive negative income tax?
Despite the forbidding label, this would definitely be a major change in the right direction. Under the more enticing name of “Bürgergeld”, it has been been advocated for many years in Germany by Joachim Mitschke (1985, 1995), professor of public finance at the University of Frankfurt. Ulrich Mückenberger, Claus Offe and Ilona Ostner (1989) argued for a less specific version of the same proposal, and Fritz Scharpf (1994, 2000), director of Cologne’s Max Planck Institute, endorsed it as his preferred option. More recently, under the clumsier label “allocation compensatrice de revenu”, a variant of it has been defended in France by Roger Godino (1999), former Dean of the management school INSEAD, and has been cautiously supported by sociologist Robert Castel (1999) and economists François Bourguignon (1999) and Laurent Caussat (2000). The idea is simply to take as given the household modulation of the current guaranteed minimum income and, instead of withdrawing the benefit at a 100% rate as earnings increase, to withdraw them at a somewhat lower rate, say 70 or even 50%, so as to create material incentives to work for any household, however low its earning power. In Godino’s proposal for France, for example, the rate is calculated so that the benefit would be entirely phased out for single people as their earnings reached the level of the guaranteed minimum wage, as opposed to the much lower level of the guaranteed minimum income, as is currently the case. In the case of a larger household, the starting level is higher. If the same reduced rate of benefit withdrawal applies, the benefit is completely phased out only at a level of earnings that exceeds the minimum wage. One major political advantage of this formula is that it can be presented as taking the current guaranteed minimum income as its point of departure and strengthening it by getting rid of the absurd penalisation of any effort to get out of the trap by taking on some low-paid activity. One major administrative disadvantage is that it implies not just that a much expanded number of households will be on benefit (admittedly at a far lower average rate), but, more awkwardly, that how high a benefit the households are entitled to receive depends on their living arrangements, which the administration must therefore be allowed to control.
It is possible to build upon existing parental, study or care leave schemes and integrate them, jointly with tax credits for the employed, into a universal basic income subjected to a very broad condition of social contribution, as proposed for example by Anthony Atkinson (1993a, 1993b, 1996, 1998) under the label “participation income”. “In order to secure political support”, Atkinson (1993a) argues, “it may be necessary for the proponents of basic income to compromise. To compromise not on the principle that there is no means test, nor on the principle of independence [i.e., the idea that no one should be directly dependent on any particular person or group], but on the unconditional payment”. A participation income would be a non-means-tested allowance paid to every person who actively participates in economic activity, whether paid or unpaid. Persons who care for young or elderly persons, undertake approved voluntary work or training, or are disabled due to sickness or handicap, would also be eligible for it. After a while, one may well realise that paying controllers to try to catch the few really work-shy would cost more, and create more resentment all over than just giving this modest floor income to all, no questions asked. But in the meanwhile the participation income will have politically bootstrapped a universal basic income into position. Compared to the income-tax-reform approach and the social-assistance-reform approach, this third approach would be particularly appropriate if some specific funding were set aside for basic income: a tax on energy consumption, or a dividend on some public asset, or simply some broadly based levy on the national product. But it could also be combined with either of the first two approaches.