Professional Documents
Culture Documents
Is Canada Ready For Basic Income
Is Canada Ready For Basic Income
WAyne sImPson
Department of Economics, University of Manitoba,Winnipeg, Manitoba, and School of Public Policy,
University of Calgary, Calgary, Alberta
Rhys Kesselman formule une réponse qui suscite la réflexion et s’inscrit dans un échange important et
opportun sur la politique de réduction de la pauvreté au moyen d’une forme quelconque de revenu mini
mum garanti. Les auteurs traitent d’abord dans leur réplique des préoccupations générales qu’exprime
Kesselman au sujet du revenu minimum et qui s’insèrent, depuis un demi-siècle, dans un débat politique
continu portant notamment sur l’expérimentation sociale au service de l’analyse des obstacles à l’emploi et
d’autres problématiques. Les auteurs réagissent ensuite aux préoccupations particulières de Kesselman à
l’égard de leur proposition de revenu minimum garanti universel (RMGU) par le truchement de réformes
fiscales qui convertiraient les crédits d’impôt non remboursables existants en crédits remboursables. Ils
illustrent comment leur concept équilibre les faibles taux d’imposition correspondant aux taux appliqués
aux crédits actuels, l’expérimentation sociale préalable et les discussions politiques avec un nouveau
soutien en matière de revenu qui élimine la pauvreté mesurée pour tous les groupes familiaux à l’exception
de celui des célibataires non âgés. À cet égard, les auteurs fournissent de nouvelles estimations qui mon
trent l’effet redistributif de leur RMGU entre les déciles de revenu familial. Enfin, les auteurs évaluent
les propositions de rechange de Kesselman relatives à l’adoption d’une approche multidimensionnelle à
l’égard de la réduction de la pauvreté et constatent que ces recommandations établies quant à une stratégie
de lutte contre la pauvreté comportent des lacunes. Il est aussi possible, font-ils remarquer, de mettre en
œuvre leur RMGU progressivement en insistant sur le remboursement des crédits d’impôt dans l’avenir.
Mots clés : crédits d’impôt, offre de travail, revenu garanti, revenu minimum, sécurité du revenu
Rhys Kesselman’s thought-provoking response is part of an important and timely conversation about pov
erty reduction policy through some form of guaranteed basic income. Our rejoinder first addresses his
general concerns about a basic income, which reflect an enduring policy debate over the past half century
that has included social experimentation to examine work disincentives and other issues. We then respond
to his specific concerns about our proposal for a Universal Guaranteed Basic Income (UGBI) through tax
reforms that would convert existing nonrefundable tax credits to refundable credits. We illustrate how
our proposed design balances low tax rates that are in line with the rates applied to current credits, prior
social experimentation, and policy discussion with effective new income support that eradicates measured
poverty for all family groups except non-elderly singles. In this regard, we provide new estimates to show
the redistributive effect of our UGBI across family income deciles. Finally, we appraise his alternative pro
posals for a multifaceted approach to poverty reduction and find these established prescriptions for anti-
poverty strategy wanting. We also point out that our UGBI can be implemented gradually by focusing on
tax credit refundability in the future.
Keywords: income security, tax credits, basic income, guaranteed income, labour supply
We welcome Rhys Kesselman’s (2018) thought- Because a wide range of plans are feasible, any deci
provoking response to our proposal for a universal sion to implement a BIG ultimately requires the choice of
guaranteed basic income (UGBI) that can be realized a specific plan. This choice trades off “adequacy, incen
relatively simply and directly by the conversion of ex tives, and cost” (Kesselman 2018, 430), but that choice is
isting non-refundable tax credits to a refundable format not unique to our plan for tax reform or to any other BIG
(Stevens and Simpson 2017). This is an important and proposal. Adequacy invites a high guarantee level that,
timely conversation, judging from the serious policy for a given budget, must imply high tax rates associated
interest in poverty reduction through some form of guar with greater work disincentives and economic inefficiency.
anteed basic income now emanating from federal and Concerns about these impacts were at the heart of the de
provincial government circles. We divide our rejoinder sign of the income maintenance experiments and remain
into three parts. First, we address Kesselman’s general prominent considerations, although the experiments were
concerns about a basic income, as some form of stan also interested in assessing the administrative benefits of
dalone plan corresponding to the proposal by Boadway, a negative income tax or BIG. Indeed, the “objective of
Cuff, and Koebel (forthcoming), our proposal, or some simplifying the administrative costs and complexity of
other generic negative income tax or basic income plan. income security” (Kesselman 2018, 430) while achieving
We then respond to Kesselman’s specific concerns about effective poverty reduction remains an attractive feature
our proposal for an UGBI through tax reform and provide of BIG policy proposals, including our own.
some new estimates of the impact of our proposal. Finally, Kesselman’s (2018) criticism of Boadway et al. (forth
we appraise Kesselman’s proposals for a multifaceted coming) and of our proposal starts with a recitation of
approach to poverty reduction that reflects many other the “traditional concerns” about a BIG that include “large
prescriptions for anti-poverty strategy. tax hikes” resulting in “very high marginal effective tax
rates (METRs) for most beneficiaries and many middle
Is a Basic Income Guarantee a Good Idea? earners, with pervasive disincentive effects and economic
The concept of a basic or guaranteed income is derived inefficiencies” (424). Kesselman refers to rising METRs
from the original concept of a negative income tax articu as “an unavoidable weakness of self-financing a BIG by
lated by Friedman (1963) during a period of heightened eliminating the basic taxable threshold” (427). He also
interest in poverty and welfare reform in the United States provides illustrative examples (428, Table 1) of the higher
that soon spilled over into Canada. Despite changes in efficiency costs that result from higher METRs. However,
nomenclature, the basic design of a basic income guarantee these impacts are a well-understood feature of any BIG,
(BIG) remains quite simple at its core and involves only as we discuss in this article and consider carefully at vari
two parameters, a family guarantee level G and a nega ous points in our UGBI proposal. Although Kesselman’s
tive income tax or benefit reduction rate r.1 For a given analysis suggests that these efficiency cost increases pose a
budget, an unlimited set of feasible [G, r] plans exists that significant disincentive, he does not provide any evidence
range from low G and r plans that provide benefits across as to their size or relative importance.
a wide family income spectrum to high G and r plans that These concerns about efficiency costs have been
focus on higher levels of benefits for a smaller, poorer raised since the early discussions of a negative income
population of families (Kesselman 2018, 425). The intense tax as one of its distinctive features and are not a conse
interest in poverty reduction through a negative income quence of our self-financing design. They underpin the
tax led to the implementation of four social experiments extensive research program on the disincentive effects
in the United States and one in Canada, the Manitoba of income maintenance programs that includes the five
Basic Annual Income Experiment, or Mincome (Simpson, social experiments in the United States and Canada. The
Mason, and Godwin 2017). These experiments collectively experiments provided consistent evidence that the labour
tested a variety of plans with guarantees ranging from supply response to the experimental negative income tax
50 percent to 140 percent of the poverty line and tax rates plans was modest (Burtless 1986; Hum and Simpson 1993;
ranging from 30 percent to 75 percent (Hum and Simpson Robins 1985). Indeed, an important consequence of this
1993, S275). vast research program was the emergence of a degree of
table 1: Net Redistribution of Federal and Provincial UGBI Benefits by Adjusted Household Income Decile ($M): Canada 2015
Stevens and Simpson
Beneficiaries, $ Contributors, $
% Redis
Adjusted No. of Total Change Total Change Contribu- tributed Benefits Redis-
Householda Familiesb Total Change in Taxes & in Taxes & tions to by Family tributed Net of
Income Who Benefit in Taxes & Benefits GSTC—GAI GSTC—GAI Redistri- Total % of Income Contributions to
Decile (Millions) GSTC Total GAI Redistributed Non-Recipients Recipients Total GAI bution, $ Familiesb Winners Group Redistribution, $
1 2.341 4,400.7 22,001.1 17,600.4 64.8 137.0 107.3 −94.4 2.388 98.0 79.6 17,506.0
2 1.848 6,232.8 14,661.4 8,428.6 16.4 1,547.0 1,744.4 181.1 2.154 85.8 51.4 8,247.5
3 1.070 3,619.0 8,095.7 4,476.6 116.7 4,986.5 4,173.8 −929.4 2.008 53.3 36.5 3,547.2
4 0.714 1,964.7 5,388.8 3,424.1 450.0 6,554.1 3,363.2 −3,640.8 2.018 35.4 39.1 −216.7
5 0.555 1,395.5 4,287.9 2,892.4 2,561.6 4,972.2 1,570.1 −5,963.7 1.989 27.9 49.4 −3,071.3
6 0.493 1,158.9 3,661.2 2,502.4 5,301.7 2,317.4 823.1 −6,795.9 1.959 25.2 55.8 −4,293.5
7 0.450 1,043.2 3,325.8 2,282.6 6,248.7 1,329.6 505.9 −7,072.4 1.946 23.1 59.6 −4,789.8
8 0.446 989.7 3,272.4 2,282.7 6,577.8 879.9 372.1 −7,085.6 1.943 22.9 62.6 −4,802.9
9 0.364 826.3 2,569.8 1,743.5 6,893.3 700.6 379.9 −7,214.0 1.910 19.1 59.1 −5,470.5
doi:10.3138/cpp.2018-043
Is Canada Ready for Real Poverty Reduction through a Universal Guaranteed Basic Income? 441
consensus on some of the biases associated with the esti provide more powerful identifying variation than previ
mation of labour supply response from nonexperimental ous studies. They find robust results that the elasticity
data (Hum and Simpson 1991, chapter 3; Jakubson 1988; of taxable income is quite modest and within a range
Mroz 1987), which favoured lower elasticity estimates. captured by the labour supply elasticity estimates we use
This consensus is reflected in the research program of the in our microsimulation analysis. They also find that their
US Congressional Budget Office and in McClelland and elasticity estimates are even smaller for smaller tax chan
Mok’s (2012) most recent review, which we have incorpor ges of the sort proposed by our UGBI. Consequently, we
ated into the microsimulation analysis of the UGBI plans do not see compelling evidence for efficiency costs beyond
in our article. This behavioural response includes both those we have captured using labour supply response.
effects at the intensive margin (hours) and the extensive Why does policy discussion of a BIG appear more
margin (participation) to address precisely the concerns serious recently at the provincial and federal levels, in
raised by Kesselman (2018, 428, 431). cluding Quebec’s proposed “revenu minimum garanti”
Although labour supply has been the focal point of con (Gouvernement du Québec 2018), Ontario’s recently
cerns about the disincentive effects of a negative income cancelled basic income pilot project (Ontario Ministry
tax, other aspects have been examined. Several studies of of Finance 2016), and the federal Liberal party platform
the impact of the US and Canadian experiments on marital proposal for a “minimum guaranteed income” (Liberal
and family stability have found no adverse effects (Cain Party of Canada 2016)? The answer, we think, is twofold.
1986; Cain and Wissoker 1990; Choudhry and Hum 1995); First, poverty persists in advanced economies de
Hum and Choudhry 1992. Hanushek (1986) did not find spite decades of discussion, research and policy action,
that the US experiments affected family consumption and Canada’s record on poverty is particularly weak.
patterns, although they did result in an increase in school The Conference Board of Canada (2018) recently ranked
attendance that appeared to offset the decline in labour Canada 13th among 16 peer Organisation for Economic
force participation for youth almost completely. In other Co-operation and Development (OECD) countries in over
words, many of the anticipated disincentive effects of a all poverty and 15th in child poverty, ahead only of the
BIG or negative income tax have been rigorously investi United States, despite a parliamentary resolution in 1989
gated and found to be either entirely absent or modest and to eliminate child poverty by 2000. A BIG is viewed as the
certainly not prohibitive in terms of the well-understood most appropriate general and direct measure to deal with
trade-off between equity and efficiency.2 this persistent stain on market-oriented economies, a matter
Although most of Kesselman’s (2018) concerns are to which we return in discussing the current multifaceted
traditional and have been studied, he does raise a new approach to poverty reduction.
issue. He cites recent research that suggests that “the A second reason for more serious consideration of
labour-type responses to the METR increases imposed a BIG is the growing evidence that the social impact of
by a BIG are much broader than total work hours” and, poverty is extensive and intergenerational. The Confer
in particular, cites research by Blomquist and Selin (2010) ence Board of Canada (2018) refers to a consensus that
that finds that “the wage-rate impacts of changes [to] poverty “can lead to higher crime rates, illness, substance
METRs are at least as large as the work-hours impacts” abuse, and poor educational outcomes, which, in turn, af
(Kesselman 2018, 427). In recent years, the literature on fect the economy through lost productivity [and] can also
the excess burden of taxation has extended beyond the lead to discrimination, inequity, and social exclusion.” An
consumption–leisure trade-off captured by labour sup impetus for this view of the detrimental social impact of
ply response to consider other margins of behavioural poverty is recent research on the long-term social out
response—such as effort and career choice that affect comes of the unique Mincome saturation site in Dauphin
wages but also tax avoidance and evasion—encapsulated that ran from 1974 to 1978. Forget (2011) uses a quasi-
in the elasticity of taxable income. After an extensive re experimental approach based on administrative health
view of this literature, Saez, Slemrod, and Giertz (2012) records to estimate that the single Mincome program with
conclude that, although the elasticity of taxable income a tax rate of 50 percent made available to all Dauphin
is likely higher than the elasticity of labour supply alone, residents was associated with an 8.5 percent reduction in
its application to the determination of efficiency loss has hospitalization rates and a reduction in physician claims
been overstated. Moreover, they argue that the empirical for mental health disorders, potentially important health
evidence, including Blomquist and Selin’s analysis, suf outcomes and savings to the health system.
fers from weak identification strategies associated with Kesselman’s (2018) general concerns about a BIG are
tax reforms for higher incomes that are very sensitive to part of a larger debate about its relative merits that will
specification and are therefore unreliable at this stage. continue and evolve because poverty is an important
More recently, Kleven and Schultz (2014) have used social problem. It is a problem that has defied other, less
rich administrative data for the entire Danish population direct solutions and shows no signs of disappearing. This
that exploits a series of across-the-board tax reforms that is the starting point of our article: What measures can
government introduce that would be truly effective in re wide swath of middle-income earners would bear large tax
ducing poverty in Canada with minimal loss in economic hikes on account of the withdrawn tax provisions” (424).
efficiency? In addition, what measures can government Another interpretation of these results in our article,
introduce that would improve the fairness of the tax however, is that they indicate the extent to which tax
system by addressing the asymmetric treatment faced payers have benefited from tax credits at the expense of
by taxpayers lacking sufficient taxable income to receive non-taxpayers, because the savings to government from
existing tax credits? Our proposal is for a UGBI that could making tax credits non-refundable permit greater credits
be achieved quite simply and directly by making an exist and lower taxes for taxpayers. This is not to argue with
ing slate of non-refundable tax credits refundable. Kesselman’s (2018) concern that our UGBI and the BIG
scheme proposed by Boadway et al. (forthcoming) “im
Is our uGBI Plan a Good Idea? poses the largest relative net burdens at middle incomes
Kesselman (2018) cites survey evidence that “59 percent and protects the highest income earners from bearing a
of respondents felt that a BIG would be too costly for commensurate share of the costs” (433), a consequence
government and would discourage people from work that would require additional tax measures or financing
ing” (429), which agrees with the point of departure of to overcome. Kesselman’s reference to “broadening the
our article (Stevens and Simpson 2017, 121). We started tax base measure of incomes, which could improve both
from the premise that the slow progress toward a BIG re equity and efficiency while also generating more revenues
flected the seemingly prohibitive challenges of justifying, without raising marginal tax rates” (431) falls into this
devising, and financing a grand guaranteed income plan category.
but that such a grand design is not necessary because a We also share Kesselman’s (2018) concern about the
BIG can be achieved through tax reform. These reforms, implications of our UGBI for METRs that affect economic
the development of refundable credits to replace exist efficiency and try to develop an UGBI that balances ef
ing non-refundable tax credits, have precedents in the ficiency and equity considerations, settling on a federal
introduction of the Goods and Services Tax Credit and plan with a low benefit reduction rate of 15 percent that
the recent conversion of dependent child tax credits to a coincides with the rate currently applied to the gross
refundable credit, the Canada Child Benefit. They are also value of federal non-refundable tax credits and results
consistent with the literature on optimal taxation, which in less adverse labour supply effects and efficiency loss
argues for negative tax rates at the bottom of the income than alternatives (Stevens and Simpson 2017, 128–9). We
spectrum. Perhaps the best expression of this principle is then use the same rationale to set the provincial benefit
found in the chapter by Brewer, Saez, and Shephard (2010) reduction rates to coincide with the tax rates currently
for the Mirrlees review of taxation in the United Kingdom: applied to the gross value of the credits in each province.
This design recognizes the importance of keeping federal
One of the key concepts of optimal tax theory is that of a
and provincial rates low, even when combined, so that the
net tax function, whereby people with high incomes pay
resulting METRs are consistent with the BIG literature.
some of that income in positive taxes to the government,
and people with a low income receive money from the Kesselman (2018, 426) uses the example of an un
government (by paying negative taxes); no conceptual attached individual facing the federal and median
distinction is made between net recipients from and net provincial tax rates whose METR under our UGBI would
contributors to the state’s finances. (94) be 48 percent from the first dollar of taxable income. Al
though this METR may be higher than these individuals
Their analysis (112) argues for a U-shaped pattern of opti currently face,3 it is consistent with any BIG design and
mal METRs with rates of about 70 percent at low income the 48 percent rate is consistent with all the experimental
levels declining to about 36 percent for moderate incomes. designs to date. Indeed, 50 percent is the archetypical
Our analysis picks up on this principle of suboptimal benchmark METR in the negative income tax experiments
taxation and asks what progress toward the alleviation and the Ontario Basic Income Pilot (Ontario Ministry of
of poverty could be made by converting existing non Community and Social Services 2018). Similarly, Kessel
refundable tax credits to refundable credits to correct man’s calculated post-BIG METRs for specific family types
what Boadway (2013, 3) labels the asymmetric treatment and provinces for our UGBI range from 40 percent to 59
of taxpayers and non-taxpayers. Although our UGBI need percent, consistent with the design of the negative income
not have been self-financing or budget neutral, our analy tax experiments and much of current policy considera
sis shows the significance of that asymmetric treatment in tion, which justifies the design choices we made in our
limiting the benefits of existing tax credits to those without illustrative plan. The justification of our UGBI design is
sufficient taxable income. Kesselman (2018) argues that also strengthened by Kesselman’s comparison of our tax
“the flip side of any self-financing or budget-neutral re rates with the standalone BIG plan articulated by Boadway
distribution scheme is that other households must pay for et al. (forthcoming) whose “total METRs could go to 70
the gains enjoyed by those who benefit” (428) and that “a percent or beyond” (Kesselman 2018, 426).
A major point of our research, then, has been to develop shield upper earners from commensurate tax increases”
an UGBI design that keeps tax rates low and in line with (429), although upper earners would clearly also make
both the rates applied to current credits and negative in a significant contribution. Although a BIG based on the
come tax experimentation and policy discussion but that conversion of tax credits to a refundable design need not
also provides effective income support. Our design, which be self-financing, the absence of new financing in our UGBI
includes a federal guarantee of $13,314 for a family of four proposal necessitates a reallocation of tax credits toward
and provincial guarantees that vary quite widely from lower income families at the expense of all higher income
$4,196 in British Columbia to $14,632 in Quebec under families (in the top seven deciles).
current tax arrangements, makes a truly important con
tribution to poverty reduction and income support across If not a BIG, then What?
the lower end of the family income spectrum.4 Poverty, Kesselman’s (2018) final two sections briefly discuss alterna
defined by Statistics Canada’s Low Income Cutoff after tives to a BIG. These fall into three categories: a wage or
tax, is eradicated in single- and two-parent families and earnings subsidy, an active labour market policy to enhance
among elderly individuals; it is virtually eradicated for employability and earnings, and public services and merit
non-elderly couples; and poverty and its depth declines goods directed to the poor (Kesselman 2018, 430–32). Each of
substantially for non-elderly singles (Stevens and Simpson these proposals is presumably combined with existing social
2017, 136). Families with income up to twice the poverty line assistance and unemployment insurance benefits, which
generally realize improvements in disposable income, and we have retained in our UGBI as well. Before we look at
income inequality measured by the Gini coefficient falls by these policies individually, we would note that none of these
7 percent across all Canadian families. This result reflects are new policies and all of them involve significant financing
the power of refundable tax credits or a BIG to address and (dis)incentives in their own right. However, we would
income redistribution and the enduring poverty problem. also argue, none of these policies can address poverty as dir
Also, our UGBI is truly redistributive. Although Kes ectly as an effective unconditional benefit program designed
selman (2018) cites the total tax credit budget of $84 billion along the lines of a negative income tax or BIG.
(actually $83.1 billion) inherent in our federal–provincial The earnings subsidy is traditionally considered by
UGBI plan, only $47.5 billion of the tax credit benefits economists as an alternative to the BIG. The basic argu
are redistributed across families as our new results pre ment is that an earnings subsidy conditioned on low
sented in Table 1 indicate, or about $0.57 of every dollar family income will lower the METR and enhance work
currently spent on these credits in their non-refundable incentives and earnings, although METRs will necessarily
form. When we break down household income by decile rise as the earnings subsidy is phased out. This argument
in Table 1 after adjusting for household size,5 $26 billion, underlies the development of the substantial Earned In
or more than half of the redistributed benefits, go to the come Tax Credit (EITC) in the United States and a more
bottom two deciles and $36.8 billion, or more than three- modest Working Income Tax Benefit (WITB) in Canada.
quarters of the redistribution, goes to those below the Although there is evidence that the EITC increases labour
adjusted household income median. In turn, $35.8 billion market participation, as expected, there is less evidence
or 77 percent of the contributions to redistribution (those that it has enhanced the hours worked by those who are
who realize smaller benefits as a result of the UGBI) are working, at least in the case of single mothers, and that is
supplied by families above the median adjusted household likely why the impact on poverty has been more difficult
income. As we reported in our article and considered in to find (Hoynes and Patel 2015; Meyer 2002). It is unlikely
our assessment of the best plan, a substantial proportion that simple expansion of the WITB will provide much dif
of families (42.7 percent) benefit from the UGBI even after ferent results and have a dramatic impact on poor families;
considering labour supply effects on earnings, with the that is, work participation and earnings in poor families
largest proportions benefiting in the first income decile would likely rise, but not enough to pull most of these
(98 percent) and the second income decile (85.8 percent). families out of poverty. Families without work continue
The last column of Table 1 offers an alternative, more to receive no benefit from an earnings subsidy, making
detailed look at the net benefits received across the ad its conditionality a limitation in any comprehensive anti
justed household income spectrum by decile as a result of poverty strategy. Nonetheless, expansion of the WITB is a
the UGBI. Benefits exceed contributions only for the first potentially valuable tool to be considered in reducing the
three income deciles, although benefits and contributions work disincentives inherent in our UGBI, other BIG plans,
are similar for the fourth decile. After that, contributions and existing programs to assist poor families.
significantly exceed benefits by amounts that rise by in “Active labour market reforms in social assistance,
come decile but are only slightly larger for deciles 9 and Employment Insurance benefits, and education and train
10 than for deciles 5–8. This provides some substance to ing at all levels are essential” (Kesselman 2018, 431), but
Kesselman’s (2018) concern that our UGBI “would shift the there is no discussion of what reforms might be effective
net burden disproportionately onto middle earners and that have not already been tried. The disincentive effects
of the “welfare wall” associated with the ineffective de services—housing, retraining, immunization, child care,
sign of existing social assistance benefits have been well legal aid, counselling, medical, and rehabilitation servi
documented and remain an issue (Milligan and Stabile ces—and others will make significant inroads into poverty
2007), although some progress has been made across the when poverty continues to be defined in monetary terms.
provinces to reduce METRs along the lines of a negative Although advocacy of these services is often made in
income tax system. We have not considered changes to so terms of poverty alleviation, we have seen virtually no
cial assistance programs despite their obvious links to the evidence that there is any direct relationship between their
provincial UGBI plans that we analyze, in part because the provision and improvement in the incidence and depth
programs provide emergency transfers and other in-kind of poverty as it is conventionally measured. However, a
assistance that are beyond the scope of our current pro BIG directly addresses the alleviation of poverty and, as
posal. We have similarly ignored Employment Insurance, we have shown in our article, can eradicate most of the
an important income maintenance program, but one that poverty that we currently measure. We do not dismiss
also involves significant work disincentive effects (Shan the importance of alternative strategies to tackle other
non and Kidd 2000) and cannot directly assist families who problems experienced by low-income populations, but a
lack regular sources of employment income. In the absence BIG is required to address its root cause, the lack of income
of any careful discussion of how these programs might be support. Moreover, many if not most of those forms of
expanded to make inroads on the alleviation of poverty, intervention are already in place, whereas cash poverty
our only comment is that these longstanding programs continues to pose serious difficulties for many families.
each introduce their own efficiency considerations whose What if our UGBI is too ambitious at this stage? As we
scale, in comparison with the efficiency considerations of said at the outset, we take the opposition to a BIG and its
a BIG enumerated by Kesselman, is difficult to determine. implied cost to be an important barrier. Poll results re
Kesselman (2018) quotes Bergmann, who advocates jecting new taxes to pay for a BIG (Kesselman 2018, 13) do
against “large universal cash grants” in favour of a not surprise us and the self-financing aspect of our UGBI
continuation of “welfare state provisions” that would provides a budgetary foundation to address this concern.
include “provision of ‘merit goods’ and of narrowly The key findings of our UGBI should be encouraging
targeted cash payments” (Kesselman 2018, 431) as well because we show that it can eliminate most of the cash
as the active labour market programs discussed earlier. poverty that Canadian families now face with losses in
He cites corresponding concerns that a BIG would crowd existing tax credit benefits that are spread across middle-
out existing social programs as a cost-saving measure and upper-income families. Although Kesselman (2018)
(Kesselman 2018, 426), as though any new and potentially characterizes the tax increases on middle-income earners
more effective policies to address longstanding poverty as large, they are neither inconsistent with the tax range
and other social ills should be greeted with immediate associated with BIG proposals and experimentation nor
suspicion rather than thoughtful scrutiny. One import do they fall heavily on any particular family type or
ant attraction of a BIG, self-financed or not, has always income level (Stevens and Simpson 2017, 135, Table 15,
been on the grounds that cash transfers improve welfare and Table 1 in this article). That said, it is not difficult
compared with equivalent in-kind transfers, a point that to understand why federal and provincial governments
Kesselman makes quite eloquently: might want to proceed more gradually. Our key mes
sage is that well-designed refundable tax credits offer
a beneficiary would always prefer (or at best be indiffer a powerful method to redistribute income and alleviate
ent between) a given amount of money versus a bundle poverty and that federal and provincial governments
of goods and services that could be purchased for the can steadily address cash poverty by following two
same amount. The reasoning is that a bundle dictated
simple rules:
by anyone else will be less valued than the bundle the
individual would choose—unless the two happened to 1. Make all future new tax credits refundable on the
coincide precisely. Thus, cash transfers are deemed to be
grounds of fairness (to non-taxpayers) and progressivity.
more efficient than in-kind transfers such as the medley of
housing, health care, child care, and other services that a
2. Convert existing non-refundable tax credits to refund
government might choose to provide or to subsidize. (432) able credits whenever circumstances permit.