--- title: "Topping up low wages in Canada: the Canada Workers Benefit, employer capture, and the workers who never file" kind: "Article" topic: "Economics" date: "2026-10-10" url: "https://canopticon.com/articles/topping-up-low-wages-in-canada" author: "Chris Telford" license: "CC BY 4.0" --- The Canadian argument about income support is usually staged as a choice between a universal basic income and nothing. It skips the fact that a version of the alternative has been running for years, paid quarterly, through the tax return. ## The benefit Canada already pays The Canada Workers Benefit tops up low earnings rather than replacing them. It phases in at 27% of earnings above $3,000, reaches a 2025 maximum of $1,633 (singles) / $2,813 (families), then phases out at 15% above $26,855 (singles) / $30,639 (families). It arrives without an application — delivered automatically through the tax system with advance quarterly payments — and it does not care how the money was earned: self-employment/gig income counts as working income on the same terms as wages. That is the design academics call participation income, already built and running. The design has refinements most people never see. It replaced the Working Income Tax Benefit in 2019, it protects couples where one partner earns less: Secondary-earner exemption, and it pays ahead: Advance payments (ACWB): up to 50% of entitlement in three installments. ## Participation income: the idea, and the argument against it The concept has an author. The economist Anthony Atkinson proposed in 1996 a benefit "paid conditional on participation… not limited to labour market participation," in which, he wrote: The condition "involves neither payment nor work; it is a wider definition of social contribution." Caring, training, job search and approved volunteering would all count. Its political appeal, in the research's words: It preserves the reciprocity norm that UBI abandons, which is its central political advantage. Its critics come from both sides. The leading advocate of an unconditional income argues that "the more broadly this condition is to be interpreted, the less of a difference there is with a basic income." Others say the checking would swamp it, calling it an administrative nightmare. A Finnish researcher's answer is to narrow it to low-income people currently reliant on social assistance — precisely the design most transferable to Canada. Even Thomas Piketty, sceptical of a full basic income as "social justice on the cheap," argued for something close to Canada's credit: support paid essentially as refundable tax credits on payslips of the employed ## Part of the money ends up with employers A subsidy paid to workers moves wages, and not always in the direction intended. The clearest estimate comes from the American equivalent, where $1 in EITC spending increases after-tax incomes by $0.73, while $1 spent on the NIT yields $1.39. The mechanism is that a top-up lets employers offer less — in one simulation low-skill single mothers keep only $0.70 of every dollar they receive [and] employers of low-skill labor capture $0.72, $0.30 from single mothers plus $0.43 from ineligible workers — note the last part, which is the uncomfortable one. Workers who do not qualify for the credit have their wages pushed down by it too. The incidence estimates deserve their own caveat. They rest on simulations with assumed elasticities, not clean natural experiments, and their author has said the evidence is "quite limited and dated." The direction is well supported; the exact share is not. ## A wage floor is what stops the leak The research is unusually settled on the fix. Lee and Saez 2012… finds that, under reasonable assumptions, the optimal policy combines an EITC and a minimum wage, and the two complement one another, with the minimum wage is an effective complement that limits any dilution. A floor is what makes the subsidy unleakable: wages cannot fall to absorb it. The pairing is not costless, and the research says who pays. Coupling the two enhances the positive employment and earnings effect for single women with children — but the combination can reduce employment and earnings for childless workers and less-skilled minority men. A policy that helps one group and hurts another is still a choice worth making, but it should be made with the second group named. ## Britain ran the experiment, and the bill grew with low wages The closest large-scale test of a wage top-up outside North America ran in Britain from 1999. It showed the cost risk the incidence research predicts: "Low wages force Britain to spend £900m more on tax credits than planned", which the research calls direct evidence that supplements can balloon in cost when wages stagnate. Annual reconciliation produced its own trouble, with large overpayments clawed back from poor families, and the system was folded into Universal Credit. ## The people the benefit never reaches Delivering money through a tax return makes it cheap to run and invisible to anyone who does not file. Among eligible wage earners the non-filing rate was 8.4% in 2019, 9.0% in 2020 and 9.2% in 2021, and among people who had once filed and stopped it reaches 59.6%. The consequence is structural, not accidental: Because the CWB is claimed through the tax return, non-filers are systematically excluded — a critical design flaw for a poverty-reduction tool. The poorest workers are the least likely to file, which means the benefit is weakest exactly where it is aimed. ## The welfare wall: clawbacks that tax low earners like the rich The other design fact that decides whether a top-up works is how fast it is taken away. Benefit withdrawals stack, and benefit clawbacks push METRs above 50% for families with two children on employment income roughly $45,000–$65,000 in most provinces. A marginal effective tax rate is what a family loses from the next dollar earned, after taxes and lost benefits. Another think tank found many working families "face METRs as high as 40% and 50%, an effective rate many would assume is reserved only for those among the highest earners." At the bottom of the scale the Canadian credit does the opposite: At very low incomes ($3,000–$13,000) the CWB produces a negative METR (rewarding work). The research's recommendation follows: flatten the phase-out so that families keep more of each raise, coordinating federal and provincial clawbacks. ## What the universal version would cost The alternative has been costed by Parliament's budget officer. A national guaranteed income would run $107 billion in 2025 under the nuclear family unit definition… roughly cut in half to $53 billion when using the broader economic family unit, against a measured poverty reduction of 34% (nuclear) vs. 40% (economic). Canada's own experiment in Dauphin, Manitoba, in 1974–79, is the usual evidence on whether people stop working; a later re-analysis found an 11.3 percentage-point reduction in labour-market participation, ~30% attributable to community effects, driven disproportionately by young and single-headed households. The costings have moved over time. The budget office's 2018 estimate, on Ontario pilot parameters, was $76 billion gross, ~$43 billion net after replacing $32.9 billion of existing federal supports. The 2025 update adds that the net figure can be far smaller: net cost as low as $3.6 billion after eliminating offsetting tax credits, depending on which existing programmes it replaces. The research warns how much rides on those choices: The 2025 GBI cost swings from $53B to $107B purely on family definition Provinces have run smaller versions. Quebec's program for people with severe employment limitations reached ~86,289 beneficiaries in 2024-25 (~25% of the social-assistance caseload). A Prince Edward Island plan for a broad guarantee stalled on federal funding, leaving a pilot that tops up ~700 people to 85% of the market-basket measure. In Ottawa, bills keep appearing and stopping short: None would actually implement a basic income — they only require the finance minister to develop a "national framework." The left's caution about a universal cheque is about what it replaces. One collection of essays warns basic income "can be an excuse for governments to eliminate critically important public services and simply replace them with a cheque," and a social-democratic institute argues that employment insurance "should not… be folded into a universal basic income." > **Objection: This is Speenhamland, and that ended badly** > > The parish wage top-ups of 1795 are the standard warning, and the standard account of them is shaky. Revisionist work holds that few of the indictments of Speenhamland hold up against the evidence, calling the 1834 commission's report empirically fragile. What survives the revision is not the moral verdict but the mechanism — employers capturing the subsidy where no floor exists — which is the same thing the American incidence research measures in modern data, and the same thing a minimum wage fixes. ## What the research pairs with a top-up: shorter hours in a downturn The research's framework adds a companion for recessions rather than a replacement for the top-up. Germany's short-time work scheme is its model; firms cut hours instead of jobs and the state replaces part of the lost pay, which in 2008–09 stabilized the labour market ("German labour market miracle"). France's legislated shorter week is the warning, which a major study found "failed to raise aggregate employment and increased job turnover," Canada already has the tool and barely uses it. Its EI Work-Sharing take-up was just 0.2% of the labour force at the May 2020 pandemic peak, versus a 21% average across comparable countries, largely because employers did not know it existed. The German scheme carries its own warning for normal times: Kurzarbeit can reduce labour reallocation and productivity by keeping workers in dying jobs (labour hoarding). ## What this explains It explains why the Canadian debate feels stuck: it is conducted as though nothing exists, when the instrument exists, reaches millions automatically, and has two known faults — a share of it ends up with employers where wages can fall, and it misses about 9% of the people entitled to it. Both faults have fixes that need no new programme. A binding wage floor plugs the leak. Automatic filing, or paying the credit to people the tax system already knows about, closes the gap. Neither is as interesting an argument as universal basic income, which may be why they get less of one.