Compliance by Design

Part 2

Anticipated questions

Part 3

Supporting brief

1. The question before the committee

I began my career estimating the size of Canada’s underground economy (Giles & Tedds, 2002). Estimates of that kind indicate how large the problem is but say little about who fails to report income or why.

Statistics Canada estimates that underground economic activity amounted to $72.4 billion in 2023, or 2.5 per cent of gross domestic product, and describes the figure as an upper bound (Statistics Canada, 2025). Residential construction accounted for 32.7 per cent of that activity, followed by finance, insurance and real estate (largely unreported rent) at 15.3 per cent, retail trade at 10.5 per cent and accommodation and food services at 8.0 per cent. The same four industries have been the largest contributors since 1992. Almost half of unreported income ($32.7 billion, or 45.2 per cent) went to employees as undeclared wages and tips (Statistics Canada, 2025).

These figures describe the size and composition of the underground economy. This brief addresses the committee’s question about behaviour and then sets that question in the wider context of how Canada administers taxes and benefits together.

2. A self-assessment system dating from 1917

Canada introduced a federal income tax under the Income War Tax Act (1917). Under the administrative model in place since the first returns were filed, each person completes an annual return, reports their own income and calculates the amount owed or refundable. In work for the Social Sciences and Humanities Research Council (SSHRC) and the Future Skills Centre, Marin Flaim, Gillian Petit and I described this self-assessment process as having “remained fundamentally unchanged since 1918” (Tedds et al., 2022, p. 1).

For most employees the annual return largely confirms information the CRA already holds, because the employer has reported their pay on a T4 slip and withheld tax at source. For the self-employed the return is often the only record available to the CRA. The CRA draws the same distinction in its own analysis. It notes that the personal income tax system relies heavily on arm’s-length third-party reporting, treats income verified in this way as “tax assured,” and locates most of the reporting gap in income that is not assured, such as income earned in the underground economy (CRA, 2026a, section 3.1.1).

3. Third-party reporting and compliance

Tax administrations that measure compliance at the level of individual taxpayers report a consistent relationship between third-party reporting and accurate reporting.

• United States. The Internal Revenue Service’s random audit program finds a net misreporting rate of about 1 per cent for wages, which are subject to both information reporting and withholding. The rate is about 6 per cent for income subject to substantial information reporting without withholding and about 55 per cent for income with little or no information reporting, such as non-farm sole proprietor income (Internal Revenue Service [IRS], 2022).
• United Kingdom. The gap for income tax and National Insurance collected through real-time payroll (PAYE) was 1.0 per cent of liabilities in 2024–25, compared with 29.2 per cent for self-employed taxpayers and small partnerships filing Self Assessment. Random enquiries found under-declared liability on 42 per cent of self-employed returns in 2022–23 (HMRC, 2026).
• Denmark. A randomized audit experiment found evasion close to zero for income subject to third-party reporting and substantial evasion for self-reported income (Kleven et al., 2011). Kleven et al. (2016) argue that reporting by firms acting as fiscal intermediaries largely explains how modern governments are able to collect high levels of tax.
• Chile. Pomeranz (2015) shows that the paper trail created by the value-added tax, in which one firm’s reported purchase is another firm’s reported sale, discourages misreporting by both parties. Slemrod (2019) reviews the wider literature.

Canadian figures follow the same pattern. The CRA estimates that the underground economy accounted for $9.8 billion of the personal income tax gap in tax year 2022, or 4.7 per cent of federal personal income tax revenue. Around 80 per cent of this amount relates to compensation paid to employees and around 20 per cent to self-employment income, while the self-employed made up only 13 to 15 per cent of the population over the period (CRA, 2026a). Earlier Canadian research using household expenditure data also found substantial non-compliance among the self-employed, varying by occupation and age (Schuetze, 2002).

Wage earners are not more honest than the contractor who renovates a kitchen; their income is simply visible to the tax authority. Tax rates, complexity and attitudes toward the tax system matter at the margin, but the evidence indicates that whether income is reported depends mainly on whether a third party also reports it.

4. What Canada’s estimates can and cannot show

The CRA’s second overall tax gap report, published in March 2026, estimates a gross federal tax gap of $59.5 billion in tax year 2022 (16.0 per cent of federal tax revenue) and a projected net gap of $34.7 billion (9.3 per cent) after compliance and collection activity (CRA, 2026a). The report improves transparency considerably. Its methods also show how limited Canada’s direct evidence on business compliance remains.

• The underground economy component is derived from Statistics Canada’s estimates rather than observed in tax returns (CRA, 2026a). Statistics Canada builds those estimates within the national accounts according to international standards and presents them as upper bounds (Luo et al., 2024; Statistics Canada, 2025).
• The estimate for small and medium-sized enterprises (SMEs) rests on a single random audit of tax year 2011. In that audit, 37.6 per cent of audited SMEs were assessed additional federal tax, with average assessments of between $1,690 and $2,160 (CRA, 2022). Estimates for later years are projected using the growth of reported tax and the growth of Statistics Canada’s underground economy series (CRA, 2026a). The decline in SME non-compliance reported for 2021 and 2022 therefore reflects slower growth in the underground economy series relative to reported income, and does not rest on new audit evidence.
• The estimate for large corporations addresses a separate issue. Of the $22.6 billion corporate reporting gap, $17.2 billion is attributed to large corporations and $5.4 billion to SMEs. The large-corporation figure is estimated from risk-based audits with statistical corrections (CRA, 2026a). It matters for revenue and for fairness, but it is measured differently and raises different questions from income that goes unreported in the underground economy. It warrants separate study, which the committee may wish to undertake.

In effect, Canada’s measures of business non-compliance depend on one macroeconomic series and an audit now fifteen years old. The United Kingdom has published annual tax gap estimates since 2009, based on random enquiry programs covering Self Assessment, small companies and small employers (HMRC, 2026), and the United States maintains its National Research Program (IRS, 2022). Canada has a reasonable estimate of the size of the problem and little current evidence on who does not report or why.

The CRA’s own evaluation of construction-sector reporting shows the same gap in measurement. Under the Contract Payment Reporting System, construction businesses report payments to subcontractors on T5018 slips. The system began on a voluntary basis in 1995 and became mandatory on January 1, 1999 (CRA, 2005). By 2005 it had generated more than $650 million in assessments, 95 per cent of which came from identifying non-filers. The same evaluation found that penalties for late or missing returns had been applied in only two cases, that the accuracy of reported payments was not being verified and that the program’s effect on voluntary compliance had not been measured (CRA, 2005).

5. Designing for compliance

If third-party information drives compliance, governments can raise compliance through the administrative measures they adopt. These measures include:

• reporting by payers, through information slips from employers, financial institutions, contractors and platforms;
• withholding of tax at source;
• certified point-of-sale systems that transmit transactions to the tax authority;
• pre-filled returns based on information the tax authority already holds;
• proof of good tax standing as a condition of being awarded or paid for a contract; and
• matching of returns against third-party data. The CRA’s T1 matching program returns an average of $46 for every dollar spent (CRA, 2026a).

Only matching resembles enforcement in the conventional sense. The other measures operate before a taxpayer files, and most of them reduce compliance costs for businesses that already comply.

Quebec

Since November 1, 2011, Quebec has required restaurants covered by its mandatory billing measures to produce every bill through a certified sales recording module and to send Revenu Québec a monthly sales summary (Ainsworth, 2009; Revenu Québec, 2013). Revenu Québec estimated revenue losses in the restaurant sector at $420 million in 2008–09. By February 2013, nearly 33,000 modules had been installed in more than 19,000 establishments, and Revenu Québec projected more than $2.3 billion in additional revenue by 2018–19 (Revenu Québec, 2013). Quebec subsidized the purchase of modules and compatible equipment, with higher amounts for restaurants that activated their modules early (Ryan, 2011).

Boyer and d’Astous (2023) evaluate the measure using anonymized tax data, comparing restaurants in Quebec with those in the rest of eastern Canada. They estimate that the modules increased reported sales by 5.8 to 9.8 per cent on average. Reported expenses, including wages, rose by a similar amount, so taxable income changed little. The authors interpret this as evidence that enforcement at the firm level extended to employees and suppliers, whose income also became visible, and they find that the one-time cost of the device is small relative to the recurring benefit to tax authorities (Boyer & d’Astous, 2023).

Quebec has since moved to a cloud-based system (WEB-SRM), under which certified systems transmit transaction data to Revenu Québec continuously. Mandatory billing now also applies to remunerated passenger transportation, and taxi operators have been required to use second-generation certified systems since January 1, 2026 (Revenu Québec, n.d.-a). Quebec also makes tax compliance a condition of contracting. Contractors on public construction contracts worth $25,000 or more, and their subcontractors, must hold an attestation from Revenu Québec (Regulation to amend the Regulation respecting construction contracts of public bodies, 2011). Since 2016 the requirement has applied to licensed subcontractors on private construction contracts, and contractors must verify the attestation (Revenu Québec, n.d.-b).

Because Revenu Québec also administers the GST in Quebec, the federal government already shares in these results. Quebec’s experience demonstrates that these measures can be implemented within the Canadian federation.

Federal reporting requirements already in place

Much of the legal framework for third-party reporting already exists at the federal level.

• Federal departments, agencies and Crown corporations report service contract payments of more than $500 a year on the T1204 slip, a requirement announced in the February 1998 budget (CRA, 2005, n.d.-b).
• Construction businesses report payments to subcontractors on the T5018 slip (CRA, 2005, n.d.-c).
• Since January 1, 2024, digital platform operators have been required to collect and report information on sellers in the sharing and gig economy under rules based on the model of the Organisation for Economic Co-operation and Development (OECD). The first returns were due January 31, 2025 (CRA, n.d.-d; Ernst & Young, 2023).
• Businesses and organizations that pay another business more than $500 a year for services are legally required to report those fees, usually on a T4A slip. Since 2011, however, the CRA has maintained a moratorium on penalties for failing to report them, which it has lifted only for the trucking industry following Budget 2025 (CRA, n.d.-a).

The fees-for-service requirement amounts to broad business-to-business reporting that has gone unenforced for fifteen years. Ending the moratorium, together with automated reporting through accounting and payroll software, would be among the least costly reforms available.

Recording once, reporting to many

Most of the cost of compliance for small businesses comes from preparing the same information repeatedly for different obligations, including payroll remittances, information slips, GST/HST returns, Records of Employment and statistical surveys. Reporting built into the accounting, payroll and point-of-sale software that businesses already use, or should use, removes most of that cost. Australia introduced Standard Business Reporting in 2010 on the principle that information should be “recorded once, reported to many” (Australian Government, 2012, p. iv). Government reporting rules are built into business software, which prepares reports from data already recorded in the course of running the business (Standard Business Reporting, n.d.). The government projected savings to business of nearly $800 million a year once the system was fully implemented (Australian Treasury, 2010). An independent evaluation found that this projection was overstated and that awareness among businesses was low in the early years, when participation was voluntary (Evans et al., 2012). The United Kingdom has instead required digital record-keeping through Making Tax Digital, first for VAT and now for income tax (KPMG, 2025). The lesson for Canada is to adopt a single reporting standard, build it into certified software and make its use the norm rather than an option.

Limits

None of these measures reaches cash paid by a household directly to a tradesperson. The CRA’s 2005 evaluation acknowledged that construction reporting does not capture transactions deliberately concealed by both parties (CRA, 2005). Residential renovation remains the most difficult part of the problem in every jurisdiction, and administrative design can be expected to narrow the gap rather than eliminate it.

6. International comparisons

Several comparable jurisdictions have built this infrastructure, and some use it for benefit administration as well as taxation.

Estonia is widely regarded as a world leader in digital government and ranked sixth of 33 countries in the OECD’s 2023 Digital Government Index (e-Estonia, n.d.). Its personal income tax return is pre-filled from government registers, and taxpayers review, amend and sign it digitally, typically in three to five minutes (SCOOP4C, n.d.). The system rests on the X-Road, a secure data-exchange layer in use since 2001 that links public and private registers (Consul General of Estonia in San Francisco, 2024). The X-Road has been exported to other countries and federated with Finland’s system (TechCentral, n.d.), which suggests that the architecture, rather than the size of the country, is what transfers. The same infrastructure that pre-fills the tax return holds the employment register discussed below. In Estonia, tax compliance and digital public services are parts of one system.

Estonia introduced its employment register to address envelope wages, which, as in Canada, were concentrated in construction and in accommodation and food services (Eurofound, 2015). Survey-based estimates of revenue lost to envelope wages fell from €154 million in 2013 to €92.5 million in 2014, the register’s first year (WageIndicator, 2015). The register also allows workers to confirm that they have been reported. Estonia’s entrepreneur account has had modest voluntary uptake, with just over 6,000 users by mid-2022 (ERR News, 2022), and its effects have yet to be established.

The American new-hire system shows how a single report from employers can serve several programs. The same report is used to locate parents who owe child support and to identify people who continue to claim unemployment insurance after returning to work (Department of Justice Canada, 2002; U.S. Department of Labor, 2000). The Department of Justice Canada examined such programs for possible Canadian use in 2002 (Department of Justice Canada, 2002). Canada’s closest equivalent today is Service Canada’s voluntary Report on Hirings (Employment and Social Development Canada [ESDC], n.d.-b).

7. Tax administration as state infrastructure for benefits

The CRA delivers a substantial share of federal income support in addition to collecting tax. It administers the Canada Child Benefit, the Canada Workers Benefit and the Canada Groceries and Essentials Benefit, which replaced the GST/HST credit on July 3, 2026 (CRA, 2026b). Eligibility for the Canada Disability Benefit requires approval for the Disability Tax Credit and a filed income tax return (ESDC, n.d.-a; Robson & Tedds, 2024).

My recent work with Gillian Petit, Jennifer Robson, David Green, Jonathan Kesselman and Marin Flaim has examined this role. In 2021, Petit, Green, Kesselman and I argued that the CRA had become a benefit delivery agent without an administrative design suited to the task. Reliance on self-assessment misses many eligible people, and an annual return cannot respond to changes in income during the year. We proposed pre-filled returns and real-time reporting (Petit et al., 2021). Our SSHRC evidence brief found that about 12 per cent of working-age Canadians do not file a return, concentrated among people with incomes below the poverty line, and that at least $1.7 billion in benefits delivered through the tax system go unclaimed as a result. We recommended a move from self-assessment to administrative assessment, expanded and automated third-party reporting and real-time reporting through payroll software integrated with the tax authority (Tedds et al., 2022). In 2026, Petit, Robson and I examined the governance gaps that arise when tax filing serves as the administrative infrastructure for benefits without being governed as such (Petit et al., 2026).

The underground economy is one consequence of this broader arrangement. About 80 per cent of the underground personal income tax gap relates to wages paid to employees (CRA, 2026a). When those wages are not reported, the tax goes uncollected and the workers concerned lose access to Employment Insurance, Canada Pension Plan credits and income-tested benefits. Benefits based on annual returns reflect income from six to eighteen months earlier (Helfand, 2025), and during the pandemic the government was unable to verify lost income in real time, particularly for people in precarious and gig work (Tedds et al., 2022).

The CRA is considering an ePayroll system under which employers would report pay for each employee in real time (Helfand, 2025). It should be designed and evaluated as infrastructure for both compliance and benefit delivery, and its evaluation should consider public value, equity and welfare in addition to collection cost and revenue yield, as the MVPF-TaxAdmin framework proposes (Arshad et al., 2026).

8. Relevance to the committee’s mandate

Unreported income affects competition. A business that underreports sales or pays wages in cash can underprice competitors that comply, and the CRA’s 2005 evaluation recorded construction firms’ concern that exemptions from reporting gave some companies this kind of advantage (CRA, 2005).

It also affects access to credit. In its October 2026 report, this committee found that Canada’s small and medium-sized enterprises are “underbanked and overcharged,” receiving 11.6 per cent of outstanding loans in 2021 compared with an OECD average of 44 per cent (Standing Senate Committee on Banking, Commerce and the Economy, 2026, p. 21). A business that conceals part of its income cannot document that income to a lender, so greater visibility may also improve access to financing.

9. Recommendations

The first four recommendations correspond to those in my opening statement. The fifth is a longer-term step. Each identifies who would act and whether legislative change is likely to be needed.

1. Build ePayroll and the digital platform rules as shared infrastructure. ePayroll should be designed to serve compliance and benefit delivery together and should be linked with platform reporting. Evaluations of both should be published, beginning with the platform rules now that the first returns have been filed. Who acts: the CRA and Employment and Social Development Canada. Using payroll data for benefit delivery may require amendments to the confidentiality provisions of section 241 of the Income Tax Act, depending on design.
2. Extend visibility where cash sales to final consumers dominate. The federal government should work with the provinces to apply Quebec’s mandatory billing model for GST/HST purposes in sectors with high levels of underground activity. Who acts: the Department of Finance and the CRA, with interested provinces. A federal billing requirement would likely require legislation.
3. Measure what works. The CRA should restore a regular random audit program for small and medium-sized businesses and the self-employed and publish the results in its tax gap reports. Accredited researchers already have access to de-identified tax-filer data through Statistics Canada’s Research Data Centres (Canadian Research Data Centre Network [CRDCN], n.d.), but not to the CRA’s audit and compliance data, which is what studies of reporting behaviour require. Who acts: the CRA, administratively, for the audit program; the CRA and Statistics Canada for research access, which may require legislative authority.
4. Enforce existing reporting requirements and make compliance easy. The 2011 moratorium on penalties for unreported fees for services should end. Fees-for-service, T5018 and T4A reporting, together with other business reporting, should be delivered through a single standard built into certified payroll, accounting and point-of-sale software, so that information is recorded once and reported to every agency that needs it. Reporting obligations should rest with the platforms, payment processors and payers that already hold the information. Who acts: the CRA. Ending the moratorium is an administrative decision; developing the standard requires work with software providers.
5. Longer term: review tax and benefit administration as a single system. The review should examine a move from self-assessment toward administrative assessment, mandatory new-hire reporting and the governance required when the tax system serves as infrastructure for benefits. Who acts: the Government of Canada, through the Department of Finance, the CRA and Employment and Social Development Canada. Mandatory new-hire reporting would require legislation.

10. Safeguards

Expanded third-party reporting raises legitimate concerns about privacy, data security and administrative burden. Most of the measures proposed here extend the kind of information the CRA already receives for every employee through the T4 system. The central questions concern governance, including what information is collected, who may use it and for what purposes, and how it is protected (Petit et al., 2026). Reporting obligations should rest with intermediaries that already hold the data, and filing should be automated wherever possible. Designed on these principles, greater visibility can reduce the reporting burden on compliant businesses.

References