Poverty is conventionally defined as the lack of some minimum level of income, so the “obvious” policy prescription is to shift more incomes to households deemed poor. Many observers spanning the full political-ideological spectrum have advocated scrapping the welfare system and instituting a broad program of cash transfers to eliminate or at least alleviate poverty.1
Proposals include British writer and politician Lady Juliet Rhys-Williams’s credit income tax in the 1940s; conservative economist Milton Friedman’s negative income tax in the 1960s; liberal economists James Tobin, Paul Samuelson and John Kenneth Galbraith’s 1968 call for a guaranteed income; and British economist James Meade’s citizen’s income. Conservative economist Friedrich Hayek stated, “I have always said that I am in favor of a minimum income for everyone in the country.”2 Other examples include the economic philosopher Philippe Van Parijs’s basic income in the 1990s and Canadian Senator Hugh Segal’s recent call for a guaranteed income. The U.S. Green Party has also promoted a guaranteed income. A Basic Income Earth Network holds biennial congresses and sponsors a scholarly journal Basic Income Studies; many countries have affiliates including the Basic Income Canada Network/Réseau Canadien pour le Revenu Garanti.3
Proponents of a guaranteed income or basic income generally share a distaste for the intrusive and paternalistic welfare bureaucracy, the pernicious distinctions between “worthy” and “unworthy” beneficiaries and the discretionary and often arbitrary aspects of the current medley of cash and in-kind transfer programs. They cite the various disincentive features of existing programs (while overlooking or downplaying the potential disincentives of their own schemes). They also critique the complex web of programs each with its own tangle of rules and regulations. For example, in an essay for the Literary Review of Canada, Senator Segal asserts that the problem demands a simple and direct solution: transferring the requisite incomes to those living below the poverty line.4 Complexity, he complains, is just society’s excuse for not addressing the problem.
I take a sharply different perspective from that of the proponents of an all-cash solution to poverty. My analysis leads to the conclusion that schemes to eliminate poverty purely by cash transfers are beset by many problems. A basic or guaranteed income that would raise all households above the poverty line carries severe hurdles of economic incentives, public finance and political feasibility that proponents typically neglect.5 A more multifaceted approach will be not only more effective but also much more likely to obtain the public support essential for progress. Complexity is unavoidable and indeed is the essence of the policy problem: better to face reality than deny it.
Here I outline some priorities – though not the design or scale – of an effective program to combat both the short-run and longer-run aspects of poverty. But before doing that, we need to understand the deficiencies of a pure cash approach to eliminating poverty. Enhanced cash transfers can be a part of the solution, but they should not be the sole policy instrument or the policy centrepiece.
Welfare programs don’t have to remain as they are: they can be made less parsimonious and less intrusive in the lives of beneficiaries. But contrary to some claims, welfare programs are unlikely to be scrapped altogether. The two most thorough studies of Canada’s income security for the working poor, both in 2006, concluded that welfare must remain at least a residual component of the system. The report of a large, diversely constituted Ontario task force titled “Modernizing Income Security for Working-Age Adults,” while proposing modest new refundable tax credits, recommended leaving welfare in place, albeit with loosened asset limits.6 The report of the Caledon Institute, a progressive Canadian think tank, proposed more generous temporary and long-term income supports for employable adults, but it too would retain welfare as a residual program.7
An important feature of welfare-type programs absent in general cash transfers operated through the tax system is the ability to distinguish between employable perople and those not able to or expected to work. This distinction allows for more generous cash payments to people not expected to work and work-facilitating services to those who are employable; it also enhances public support for such programs.
What are the barriers to proceeding with an all-cash approach to eliminating poverty? To assess this issue, I consider three basic forms that can be used to transfer money to the poor:
- “fill-the-gap” approach that provides each poor household benefits equal to its shortfall from the poverty line;
- “guaranteed income” that pays an amount equal to the poverty line for a household with zero income and phases out the benefit with higher incomes;
- “basic income” that pays every household an assured amount irrespective of its income.8
A major difference among these three structures is the implicit tax-back rate on reported earnings: 100 per cent for fill-the-gap, an intermediate figure such as 50 per cent for the guaranteed income and 0 per cent for the basic income. However, as I show here, all three formats face serious incentive, financial and political barriers, particularly if applied to an all-cash attack on poverty.
Fill-the-gap: A total disincentive to work
A fill-the-gap structure might appear to be the least costly method of eliminating poverty through an all-cash approach. If a household’s income were $20,000 and the poverty line for a household of that size and locale were $30,000, then just transfer $10,000 in cash. However, paying benefits in this manner means that the household has no incentive to earn anything – or a strong incentive to work off the books in the underground economy, so that it reports zero income for tax and benefit purposes. Regardless of how little this household earns, it will be topped up to the full $30,000 through transfers. The implicit 100 per cent tax-back rate on reported earnings reproduces the extreme disincentive effects of many provincial welfare programs for employable adults. Senator Segal cites a $30 billion cost for a Canadian income transfer program that uses this top-up approach and thus implicitly assumes a fill-the-gap structure.
As a result of the total disincentive to work or to report any earnings, a fill-the-gap structure would be much more costly than the naive estimate. Instead of needing to fill the so-called poverty gap – the difference between poverty lines and the actual incomes of the poor – the program would need to pay out the entire amount of the poverty line for individuals initially below it. In addition, some individuals initially above the poverty line would find attractions in either quitting work or concealing their earnings so as to receive benefits equal to the full amount. Together, these effects would sharply increase the budgetary cost of the program, possibly making it double or even triple the naive estimate. This scheme’s impact on taxpayer burdens plus the prospect of everyone being lifted from poverty without having to work, or a massive shift of workers to the underground economy, would dash any chance for public acceptance.
Guaranteed income: Costly, skewed and burdensome
A guaranteed income structure for cash transfers to the poor would at least preserve some work incentives by reducing benefits by less than 100 per cent of a beneficiary’s earnings. The most commonly proposed benefit-reduction or tax-back rate is 50 per cent, and I use this figure in my illustrations. Such a scheme has a guarantee amount that is paid to a household with zero income and a break-even income level at which the benefit is fully phased out; with a 50 per cent tax-back rate, the break-even income is twice the guarantee amount.9 Thus, the goal of guaranteeing every Canadian an income above the poverty line entails a break-even income level equal to twice the poverty line applicable to that person.
The standard measure of poverty lines is Statistics Canada’s Low Income Cut-Offs (LICOs), which vary by family and community size. For a four-person family residing in a metropolitan area of 500,000 or more, the before-tax LICO is about $45,000 in 2013.10 Accordingly, achieving this target would imply a break-even income of $90,000 for families of four in larger cities, or well above median family incomes. (For single individuals in larger cities, the corresponding break-even income would be $48,000.)
The implications of this illustration for a guaranteed income are striking. First, more than half of all Canadian households would be recipients of some net benefits, even if the benefit amounts would be small for many. Second, the enormous increase in the number of beneficiaries would sharply increase the budgetary cost of the program. Third, the guaranteed income’s tax-back rate of 50 per cent would apply not just to the poor but would encompass more than half of the population, with signficant disincentives to work, save, invest and comply with the benefit and tax systems.
Most Canadians would be beneficiaries and taxpayers simultaneously, subject to both the 50 per cent tax-back rate of the guaranteed income plus marginal rates of income tax and payroll taxes in the 25 to 40 per cent range. Their total “marginal effective tax rate” on incremental earnings would thus fall in the 75 to 90 per cent range. And these figures ignore the need for increasing income tax rates to finance the huge costs of the guaranteed income. A similar oversight about the overlap between tax-back and income tax rates beset an ill-fated proposal for a guaranteed income by the Macdonald Royal Commission in the 1980s.11
Additionally, a guaranteed income operating via the income tax, as most proponents envisage, would be extremely slow in responding to fluctuations in a household’s income – up to two years. The scheme would pay inadequate benefits to those with sudden need and excessive benefits to those no longer in need. Existing refundable tax credits that operate though the income tax – such as those for National Child Benefits and the GST/HST credits – have a similar failing, but this problem would be far more serious in a scheme intended as the primary income support program. Only a much more complex and burdensome system of frequent income reporting between annual income tax filings would overcome this deficiency. Suddenly the welfare system looks a bit better, at least in terms of responsiveness!
Basic income: Gargantuan costs, unacceptable tax hikes
Yet another way to structure a cash-transfer program is the basic income, which has gained advocates in recent years. This approach would provide regular payments to all individuals and families regardless of their income. Its advocates, such as Van Parijs, cite the virtue of the “real freedom” that people would enjoy if unconstrained in pursuing their own notion of the good life, whether that entailed paid work, family duties, volunteering, pursuing hobbies, or contemplating nature. One advantage of this benefit structure is that all individuals would receive the full payments regularly and not have to wait for an income test; the personal tax system would be applied to finance the system and thereby apply a kind of ex post income test.
Even if the basic income were set well below the poverty line, these universal payments would carry gargantuan gross budgetary costs. For example, assume benefits averaging $10,000 per capita – less for children and more for seniors to mirror the current maximum Old Age Security plus Guaranteed Income Supplement benefits. Basic income benefits at this level would be far below the poverty line but still well above welfare benefits for all but people with disabilities. With Canada’s population of 35 million, the gross budgetary cost of this basic income clocks in at a massive $350 billion. Even offsetting this figure by eliminating seniors’ cash benefits and provincial welfare, the implied additional cost to taxpayers would be enormous – greater than current total federal spending. Income taxes on individuals and businesses as well as other taxes would need to be sharply increased. The general public would not tolerate such tax hikes, particularly given that benefits would be paid to individuals even if they were able to work but chose not to do so. Evidence on this point arises with the public’s insistence that employable people be subjected to work requirements in obtaining even miserly welfare benefits.
Moreover, advocates of guaranteed income and related schemes for a negative income tax often overstate the ability to “cash out” other programs to help finance their proposed schemes. An early advocate of the negative income tax, Milton Friedman, was attracted to the idea by the prospect of firing lots of social workers and dismantling most existing social programs including welfare. He proposed to set the benefit level by taking the income tax threshold and converting it to a refundable credit using a maximum benefit such as 50 per cent of the threshold. For the contemporary Canadian federal income tax, this would be 50 per cent of the $10,500 threshold, or a maximum cash benefit of about $440 per month – which is less than the welfare benefit for single employable persons in every province except New Brunswick. Some advocates of basic or guaranteed income schemes cite the potential savings from scaling back other programs, such as Employment Insurance. However, this ignores the fact that EI premiums would also be reduced – since the program is self-financed – so those funds for the guaranteed income would have to be raised via increases in general tax sources. Moreover, the EI program serves purposes other than setting a floor under incomes, such as maintaining a portion of a worker’s accustomed earnings during temporary jobless spells and facilitating the search for new work.
The problem of high marginal tax rates
In progressing from the fill-the-gap method to the guaranteed income to the basic income, one can observe that the disincentive effects remain but are merely shifted around the population. With fill-the-gap transfers, the extreme disincentives of 100 per cent tax-back rates affect all of the poor and likely some workers above the poverty line. Moving to a guaranteed income reduces the disincentive to a 50 per cent tax-back rate for the poor but extends that disincentive to all workers with earnings up to twice the poverty line, while most of those workers additionally face effective income plus payroll tax rates of 25 to 40 per cent on incremental earnings. The basic income eliminates disincentive effects from all people in their role as recipients of the universal benefits, but the scheme shifts these disincentives to the entire population in their capacity as taxpayers who must finance the scheme’s gargantuan costs.
The disincentive effects of high marginal effective tax rates on incremental earnings – the loss of various cash and in-kind benefits as well as the payment of additional taxes – are pervasive. They affect the incentive to work, to take overtime hours, to accept more demanding or less pleasant work, to move to take a job that pays more, to undertake training to upgrade skills, to report earnings fully, to save to cover future needs, to start or expand a small business, and to maintain a family and support one’s children.
With existing programs, the marginal effective tax rate has been calculated to be high already for many families. It typically exceeds 60 per cent for households with children and incomes in the clawback range of $20,000–$40,000, and it can be considerably higher for those receiving income-tested benefits such as daycare or housing subsidies.12 Some advocates of increased cash transfers argue that the indirect and hidden nature of some income tests and clawbacks minimizes any disincentive effects,13 but this assertion is contrary to extensive empirical evidence on labour-supply and other responses. Hence, high marginal effective tax rates constitute an important constraint on desirable formats and structures for antipoverty policies.
Indeed, already high marginal effective tax rates with existing cash transfer and in-kind programs illustrate the barriers to moving significantly further toward reducing poverty via the cash approach. Raising the guaranteed benefit levels would be prohibitively costly without further increases in program tax-back rates, approaching or exceeding the 100 per cent rate of the fill-the-gap format. Moreover, simultaneously moderating the tax-back rate would extend break-even income levels so as to greatly expand both the beneficiary population and the program cost. Thus, my earlier critique of the all-cash approach to eliminating poverty also extends to even partial further moves in that direction. Only by limiting the target group for increased cash transfers to segments of the population without primary workforce attachments – such as the elderly and people with severe disabilities – can both the budgetary costs and disincentive effects be controlled.
A dollar of transfers isn’t a dollar earned
Yet another critical issue in the choice between cash redistribution and other policy approaches relates to the intergenerational transmission of poverty. Transferring cash to families through public programs may have different impacts on behaviour than raising incomes through work-related policies. A longitudinal study tracked the experience of children growing up in impoverished U.S. families. It found that work and income earned by parents (particularly the mother) was much more effective in reducing the incidence of various adverse events among their offspring as teens and young adults (such as dropping out of high school, out-of-wedlock teenage births and lack of work attachment) than equivalent amounts of public cash benefits.14 The educational level of parents was another important factor influencing favourable long-run outcomes for their children.
Thus, a dollar of transfers is not fully equivalent to a dollar of earned income; earned income carries not only greater autonomy and dignity but also more favourable long-run effects on poverty incidence. Antipoverty policies should take to heart this education- and work-oriented message, at least for the employable among the poor.
Another fundamental problem with all variants of an all-cash approach to eliminating poverty is that their enormous budgetary costs – even if acceptable to taxpayers – would inevitably crowd out high-priority in-kind social benefits. When well designed, these benefits are of crucial importance to the most disadvantaged and vulnerable among the poor, and in many cases the value of their in-kind benefits exceeds that of their cash benefits. Some common examples include health care, public and social housing, housing subsidies for private rentals, long-term care facilities, pharmaceutical subsidies, daycare subsidies, mental health services, drug and alcohol rehabilitation, child protection services, legal aid and many others. Funding for these in-kind social programs already faces great pressure in most provinces, and the availability and quality of services is often compromised.
This critique of proposals for major expansion of cash benefits was effectively posed by noted feminist economist Barbara Bergmann in her assessment of basic income proposals, and her critique applies equally to the other all-cash proposals:
State provision of “merit goods” and of narrowly targeted cash payments has higher priority than large universal cash grants … Advanced countries do not have the taxing capacity to do both at once. Other problems with cash payments schemes include the disincentive to work for pay, reducing taxpaying capacity, and retrograde effects on gender equality.15
Bergmann’s term “merit goods” covers the full range of in-kind benefits of both goods and public services. My analysis concurs with her essential conclusion that enhancements to more targeted cash benefits and in-kind benefits should take priority over the grander schemes for expanding cash benefits. I sketch these priorities in very broad terms, and clearly a detailed analysis of the structure, scale, finance, operation, and public acceptance of each item is warranted before proceeding to implementation. Much smaller sums than required by any of the monolithic cash proposals could be effectively applied to enhancing these programs, and this approach would also enjoy far greater public support.
Expanding targeted and in-kind benefits
Canada already provides relatively generous targeted cash benefits for two groups in the format of a guaranteed income. The maximum amount a senior with little income receives through the OAS/GIS is about $15,400 annually, while the total annual cash benefits for a child in a low-income family can exceed $5,000 through federal and provincial programs apart from welfare. In contrast, non-aged people with disabilities in most provinces receive welfare benefits lower than the maximum level for seniors, an unjustified discrepancy. A small first step to redress this imbalance would be to make the existing disability tax credit refundable – that is, payable to eligible people with disabilities with incomes so low as to be nontaxable; further steps might be a federal or federally financed provincial supplement for this group.
For employable people on welfare, particularly singles, benefits are miserly to the point of almost requiring beggary and thievery for bare sustenance. These welfare benefits need to be increased along with a relaxation of the 100 per cent tax-back on earnings that some provinces apply. Additionally, training, education and other work-supportive services need to be augmented to facilitate labour force re-entry and higher earnings. Enhancement of the current Working Income Tax Benefit is one way to expand cash support while simultaneously supporting work effort. Special public employment projects also have a potential role if well designed and targeted.16
In-kind benefits of both targeted and universal varieties are a crucial part of the current antipoverty policy package, and many warrant high priority for expansion. These benefits are often targeted at subgroups of the poor needing supports of particular kinds that cash transfers dispersed generally to the poor would not enable them to purchase in private markets. Additionally, the in-kind nature of many such benefits provides an implicit targeting mechanism that can reduce the overreliance on income tests that cumulate to exert strong disincentive effects.
Some examples include legal aid, child protection services, immigrant settlement services, prenatal counselling and support, drug and alcohol rehabilitation, supportive services for mental health and counselling of at-risk high-school students. Other in-kind benefits that necessarily apply income tests include public and social housing, subsidies for private rentals, residential care, daycare services, and extended medical, dental and pharmaceutical coverage. Many of these programs have been compromised by deficit-pressed provincial governments, and they warrant higher priority for funds than preliminary steps toward a full guaranteed income.
Even within unchanged budgets, some antipoverty programs can be redesigned to provide better and fairer support. For example, public housing in most provinces has long waiting lists on account of limited funding and the attractive implicit subsidies in their rent charges. A common standard for setting rents in public housing is 30 per cent of the tenant’s income. When many low- and moderate-income families are paying a larger share of their income for rentals in the private market and unable to access public housing, this situation seems inequitable. Raising the rents on public housing to 35 per cent of income would be fairer, finance the provision of more units and reduce the waiting time of applicants.
Other areas of social programming should also be reviewed with an eye toward improved incentives and economy; this will inevitably produce some losers as well as winners. While any realistic antipoverty program will require increased overall budgetary commitments, all methods of using available dollars more effectively still need to be pursued.
Ambitious strategies are also needed to go beyond ameliorating current poverty and reduce the incidence of poverty and low incomes in future generations. Children in poorer neighbourhoods need expanded programs of early child development and free school breakfasts. Major steps can be taken to improve counselling and other services to at-risk youth to encourage their high-school completion and subsequent education or training. Enhanced community centre activities and sporting facilities can further aid youth development.
Moreover, the provinces should emulate the systems in countries that have extensive occupational programs beginning in high school and extending to workplace apprenticeships. Such moves could play an important part in addressing Canada’s looming skill shortages, while raising many youth from lives of earnings insecurity and poverty. These issues are of particular salience for First Nations youth – the fastest-growing segment of the population and the group at greatest risk of poverty.
I conclude that a guaranteed income and other proposals for major expansion of general cash transfers fail all three tests of feasibility: incentives, finances and politics. Taking even modest initial steps toward a guaranteed income is likely to further starve critically needed targeted cash transfers and diverse in-kind benefits. These initial steps are also unlikely to proceed very far or to be fundamentally effective in combating poverty. Thus, the siren call of a simple cash fix for poverty is likely to divert both policy developments and political efforts away from more realistic and effective paths. A far better strategy is to refine and better support policy instruments known to improve the lives of the poor in meaningful ways.
An all-cash strategy such as a guaranteed income is touted as a simple solution to the complex problem of poverty. In this context, we should all heed H.L. Mencken’s famous aphorism: “For every complex problem, there is a solution that is simple, neat, and wrong.”
1 For a detailed chronology, see basicincome.org/bien/aboutbasicincome.html#history
2 Stephen Kresge and Leif Wenar, eds., Hayek on Hayek: An Autobiographical Dialogue by F.A. Hayek (Chicago: University of Chicago Press, 1994), p. 114. Note that Friedman, Tobin, Samuelson, Meade and Hayek were all recipients of the Nobel Prize in Economic Sciences.
4 Hugh Segal, “Scrapping Welfare,” Literary Review of Canada, December 2012 (Vol. 20, No. 10).
5 Terminology varies among the proponents and schemes. The basic income has also been called a universal basic income, basic minimum income, credit income tax, demogrant and citizen’s income. The guaranteed income has also been called a guaranteed annual income, negative income tax and refundable tax credits. Sometimes the variation in terminology does connote differences in program structure or operation.
6 Task Force on Modernizing Income Security for Working-Age Adults, Report: Time for a Fair Deal (Toronto, 2006).
7 Ken Battle, Michael Mendelson and Sherri Torjman, Towards a New Architecture for Canada’s Adult Benefits (Ottawa: Caledon Institute of Social Policy, 2006).
8 My analysis draws in part on that of Timothy Besley and Ravi Kanbur, “The Principles of Targeting,” in Michael Lipton and Jacques Van Der Gaag, eds., Including the Poor (Washington, DC: World Bank, 1993), pp. 67–82.
9 The relation between the break-even income (BE), the guarantee level at zero income (G), and the tax-back rate (as a pure number such as 0.50, not a percentage) is: BE = G/t.
10 These figures are taken from Statistics Canada, Low Income Lines, 2011–2012, Income Research Paper Series, Cat. No. 75F0002M-No.002, with updating for consumer inflation to 2013.
11 See Jonathan R. Kesselman, “The Royal Commission’s Proposals for Income Security Reform,” Canadian Public Policy, Vol. 12, Supplement (February 1986), pp. 101–112.
12 This figure was calculated by Alexandre Laurin and Finn Poschmann, Treading Water: The Impact of High METRs on Working Families in Canada, C.D. Howe Institute E-Brief No. 160 (Toronto: C.D. Howe Institute, 2013). For an analysis providing early warning about these issues, see Jonathan R. Kesselman, “Public Policies to Combat Child Poverty: Goals and Options,” in Keith Banting and Ken Battle, eds., A New Social Vision for Canada? Perspectives on the Federal Discussion Paper on Social Policy Reform (Kingston, ON: Queen’s University, School of Policy Studies, 1994), pp. 73–97.
13 For example, a leading early proponent of the National Child Benefit Supplement asserted, “Incentives-obsessed critics on the right allege that the imposition of high marginal tax rates on some working poor families jeopardizes ” (Ken Battle, “Child Benefit Reform: A Case Study in Tax-Transfer Integration,” Canadian Tax Journal, Vol. 47, No. 5 , p. 1222). For a dated but comprehensive review, see Robert Moffitt, “Incentive Effects of the U.S. Welfare System,” Journal of Economic Literature, Vol. 30, No. 1 (1992), pp. 1–61.
14 Robert H. Haveman and Barbara L. Wolfe, Succeeding Generations: On the Effects of Investments in Children (New York: Russell Sage Foundation, 1994).
15 Barbara R. Bergmann, “A Swedish-Style Welfare State or Basic Income: Which Should Have Priority?” Politics & Society, Vol. 32, No. 1 (March 2004), p. 107.
16 See the findings in Jonathan R. Kesselman, “Work Relief Programs in the Great Depression,” in John L. Palmer, ed., Creating Jobs: Public Employment Programs and Wage Subsidies (Washington, DC: Brookings Institution, 1978), pp. 153–229.