When The Books Stop Answering

The first article in this series described CRA obtaining building permits and matching them against its own registrations. This one describes what happens after that match, when an auditor arrives and the records do not explain the difference.

A Note On Currency

Everything here is stated as verified in August 2026 and requires confirmation before reliance.

We have not read any of the judgments described here, and our accounts of Hickman Motors, Ramey, Fuhr and St Phillippe come entirely from professional commentary. We have not verified any statutory provision against the Act.

We accessed the structure of CRA's audit manual chapter on these techniques but not its operative content, so we describe what the chapter covers rather than what it says[1].

One source we cite is an International Monetary Fund technical note on audit practice generally rather than a statement of Canadian law, and we flag it where it appears[2].

This is not tax or legal advice. An assessment of this kind is a dispute, frequently involving several years and penalties, and anyone facing one should obtain representation rather than respond from a general description.

What It Is

The method, and when it appears.

Commentary describes indirect verification of income as an umbrella term CRA uses for audit techniques that attempt to estimate a taxpayer's earnings and spending by looking at the circumstances, where the taxpayer lacks records of them[3].

Another explains the trigger: in cases where CRA believes the taxpayer's normal books and records do not reflect an accurate picture of income earned, the Act provides that CRA may assess based on indirect methods using information obtained from third parties or the taxpayer[4].

A third describes the context more broadly. Where a taxpayer fails to file a return, CRA frequently makes an arbitrary assessment based on an estimate of income. Where an auditor suspects unreported income, an indirect determination may be used to estimate income from the taxpayer's lifestyle and assets[5].

Commentary describes the methods as an extremely powerful tool, particularly when dealing with small businesses or individuals who do not have an extensive paper trail[3].

Two features define the situation and this is our own analysis.

The method is reactive. It is deployed because the direct evidence has failed, which means the businesses exposed to it are those whose records are weakest.

And it works from the outside in. Rather than building income from transactions, it infers income from what the money did, which means the taxpayer's household, assets and spending become the evidence.

Supporting Or Assessing

A distinction that determines what an examination actually means, and it is not widely understood.

Commentary separates two categories. Supporting tests are used to quickly determine the risk of a taxpayer underreporting income, or to back up audits using more thorough methods. It states that CRA does not use them to alter people's income and raise their taxes, though it may apply more rigorous versions as assessing techniques[3].

Assessing techniques are described as more strenuous, requiring more time and more exhaustive, and as techniques that may be used to reassess someone's income and raise their taxes[3].

CRA's own audit manual reflects the same division. Its chapter includes sections headed the assessing net worth technique, the assessing unidentified bank deposits technique and the assessing projections technique[1].

The practical importance, and this is our own reading, is that a taxpayer should establish which is happening.

A supporting test producing a discrepancy is a risk signal that may go no further, or may prompt a request for explanation. It is not itself an assessment.

An assessing technique is a reconstruction of income intended to support a reassessment, and it carries the procedural consequences described later in this article.

Being asked to explain a ratio is a different event from being told your income has been reconstructed, and treating the first as the second wastes resources while treating the second as the first loses time that matters.

The Four Supporting Tests

What CRA runs early, as commentary describes them.

It identifies four primary supportive tests: bank deposit analyses, described as mandatory in many circumstances; rough net worth analyses; source and application of funds; and ratio analysis. It notes these are not exhaustive and other tests may be used[3].

It also observes, more colourfully, that supporting techniques may extend to speaking with neighbours or to how lavish a lifestyle appears on social media[3].

We report that as commentary characterising practice rather than as a documented CRA procedure.

Two of the four deserve a note, and these observations are ours.

That bank deposit analysis is described as mandatory in many circumstances means a business under examination should assume its deposits have been totalled, whether or not the auditor has raised it. That is a cheap test performed from records CRA can obtain directly.

And ratio analysis is the one businesses can anticipate. It compares a taxpayer's margins, cost ratios and turnover against sector norms, which is why the restaurant article in this series noted food cost ratios and the cash-to-card split as standard enquiries.

A business whose ratios sit outside the range for its sector should be able to explain why, and there are frequently good reasons: a different product mix, an unusual lease, a start-up period, wastage, theft, or a location that does not resemble the comparator group.

Bank Deposit Analysis

The simplest technique, and the one most likely to be encountered.

Commentary sets out the theory: it is based on the logic that the taxpayer's income is equal to the amount of money deposited into their bank account. The auditor requests bank statements from the relevant financial institutions and adds all deposits to calculate income. If there is a discrepancy between that estimate and reported income, the auditor will assume there was unreported income[5].

It then describes the qualification: auditors will usually remove any transactions that are clearly non-taxable, such as transfers between accounts or loans, but rarely go to any great length[5].

And the consequence: the taxpayer then bears the burden of establishing which deposits were in fact non-taxable, and even a taxpayer with good records will struggle to remember and substantiate every deposit[5].

That last sentence is worth pausing on, and this is our own emphasis.

It is not saying that a taxpayer with poor records will struggle. It is saying that a taxpayer with good records will, because the task is not producing records of business transactions. It is accounting for the origin of every credit to every account across several years.

Deposits from a spouse, repayment of a personal loan to a friend, proceeds from selling a car, an insurance settlement, a gift at a wedding, a transfer from a line of credit, the redeposit of cash previously withdrawn. Each is non-taxable. Almost none is documented.

The Flaw In The Logic

Why the technique overstates, described by commentary and extended by us.

Commentary states that these types of assessments routinely overinflate tax liability and are difficult to challenge[5].

It also gives a practical instruction that is unusually specific: if a client receives an assessment based on a bank deposit analysis, it is worth investing time in a detailed review of the auditor's calculations and logic, because the auditor often makes errors, which can help demonstrate the unreliability of the figures underlying the assessment[5].

Another identifies a specific recurring error, noting that proper audit techniques require the auditor to reconcile and eliminate transfers between accounts, and that this often does not happen[6].

The structural reason, and this is our own analysis, is that the technique is asymmetric.

Every deposit counts toward income unless someone removes it. The auditor removes what is obvious. The taxpayer must remove the rest.

So an ordinary household with several accounts, a line of credit, a mortgage, family transfers and occasional asset sales generates a large volume of credits that are not income, and the default treatment of each is to be treated as income.

A business owner who moves money between a business account and a personal one, in both directions, can generate double counting from a single dollar circulating.

That is not an abuse of the technique. It is the technique operating as designed on facts it handles poorly.

What Unremoved Transfers Cost

The magnitude, computed by us on a hypothetical file at an assumed marginal rate of 43 percent.

Take total deposits of $400,000 across the period against reported income of $180,000. The assumed unreported income is $220,000, and the tax on it is roughly $94,600.

Now assume that some proportion of that gap consists of transfers, loans, gifts and asset proceeds that the auditor did not remove.

If ten percent, the true gap is $198,000 and the assessment overstates tax by approximately $9,460.

If twenty percent, the true gap is $176,000 and the overstatement is roughly $18,920.

If thirty percent, the true gap is $154,000 and the overstatement is roughly $28,380.

If forty percent, the overstatement is approximately $37,840.

Two points follow and both are ours.

The relationship is linear, which means every dollar of deposit successfully explained is worth its full tax value. There is no threshold to reach and no all-or-nothing outcome, so partial success is proportionate success.

And the work is granular. Explaining thirty percent of a gap is not one argument; it is a large number of small ones, each requiring its own evidence. That is why this exercise is laborious and why it is nonetheless usually worth doing.

The Net Worth Technique

The most powerful of the methods, and the premise behind it.

Commentary describes the net worth audit as one of the most common, powerful and, from a taxpayer's point of view, troublesome methods CRA employs, proceeding from the premise that a taxpayer has unreported income the auditor may not be capable of identifying through direct means[4].

It explains the logic: the method relies on the assumption that when a taxpayer accumulates wealth in a taxation year, there are two options, to either spend or invest the income, and the auditor will examine in depth the taxpayer's expenditures during the selected period[4].

So the computation compares net worth at the start and end of a period, adds personal expenditure, subtracts amounts from non-taxable sources, and treats the remainder as income.

CRA's own manual chapter includes sections on the shareholder's investment, the reason for a discrepancy in net worth, the analysis of a discrepancy in income, completing the assessing net worth technique, the working papers and the audit report form[1].

Note what the technique requires the auditor to establish, and this is our own observation.

Opening net worth, closing net worth, personal expenditure for every year, and all non-taxable receipts. Four quantities, each estimated, each across several years, and each capable of being wrong.

Getting any one materially wrong distorts the result, and an error in opening net worth alone shifts income between years even where the total is right.

Where The Errors Compound

The criticism that matters most, and it is a structural one.

Commentary states that while the rationale underlying a net worth assessment is sound in principle, the technique too easily permits a tax auditor to aggregate the practical flaws of each other indirect method of income verification[6].

It explains the mechanism. When applying a net worth assessment, an auditor will often use a bank statement analysis to derive personal expenditures and growth in net worth. In doing so, the auditor may err not only by mischaracterising business-related withdrawals as personal expenditures, described as a common mistake in computing personal expenditures, but also by failing to correct for transfers between accounts, described as a common mistake in bank deposit analysis[6].

It adds that auditors often overlook liabilities incurred by the taxpayer that would offset an increase in assets[6].

That third error deserves separate emphasis, and this is our own analysis of why it is so consequential.

A taxpayer who buys a $60,000 vehicle appears to have accumulated $60,000 of wealth. If $50,000 of it was financed, the true increase in net worth is $10,000.

Omitting the loan overstates the year's income by the amount of the borrowing, and borrowing is invisible in the asset record. It appears only in a liability the auditor must think to look for.

So the single most valuable thing a taxpayer can produce in a net worth dispute is evidence of debt: mortgages, vehicle finance, lines of credit, family loans, credit card balances and business borrowings, at each year end.

Those documents exist, are obtainable from lenders, and directly reduce the computed income.

Unsatisfactory And Imprecise

What the Tax Court has said about the method it nonetheless upholds.

Commentary records that the Tax Court of Canada has long recognised that the net worth method of estimating income is an unsatisfactory and imprecise way of determining a taxpayer's income for the year, citing Ramey v The Queen[6].

We report that characterisation as commentary quotes it, not having read the judgment.

Two things follow and they pull in opposite directions, which is our own reading.

A judicial acknowledgement that a method is unsatisfactory and imprecise is useful material for a taxpayer. It establishes that the resulting figure is an estimate rather than a measurement, and that the Court knows it.

But it has not caused the method to be rejected. Courts continue to uphold net worth assessments, because the alternative where records are absent is no assessment at all.

The correct inference, and it governs the strategy sections below, is that criticising the method is not a defence. Its imprecision is judicially acknowledged and it survives anyway.

What the acknowledgement supports is a narrower and more useful argument: that where a particular computation contains identifiable errors, its imprecision is not merely theoretical, and the specific figure should not stand.

That argument requires the errors to be found and shown, which is a matter of work rather than advocacy.

Projections

The third assessing technique, which CRA's manual treats separately.

Its chapter includes an assessing projections technique with sections covering audit evidence, indirect versus direct verification, projections to determine income, circumstances that warrant proceeding to the technique, reasonable projections, assumptions, comparison with the net worth technique, sample selection and types of projections[1].

Notably, it also includes a section headed circumstances that do not warrant proceeding to the assessing projections technique[1].

We did not obtain the content of any of those sections. Their headings are nonetheless informative, and this is our own analysis.

The technique works by taking a sample and extrapolating it across a period. A sample of days, transactions, or items is examined and the result projected to the whole.

That structure means the projection is only as good as the sample, and the manual's inclusion of a dedicated section on sample selection indicates the Agency's own procedures address this.

Two practical implications for a taxpayer facing a projection.

The representativeness of the sample is a legitimate line of enquiry. A sample drawn from a peak period, an atypical week, or a period with unusual conditions produces a projection that overstates.

And the existence of a manual section on circumstances that do not warrant the technique means CRA's own guidance contemplates cases where it should not be used. A taxpayer's representative should establish whether the file falls within them.

Demolish The Assumptions

The procedural position, which is the reason these assessments are hard to displace.

Commentary sets out the principle established by the Supreme Court in Hickman Motors Ltd. v. Canada, [1997] 2 SCR 336: the Minister's factual assumptions are presumed valid, and the initial onus rests with the taxpayer to demolish those assumptions by way of a prima facie case. A prima facie case in this context means evidence sufficiently credible and weighty that, if unchallenged, would be accepted by the court as tending to disprove the Minister's assumptions[7].

Another puts the same point more bluntly, stating that once CRA has used an indirect verification audit to lay out the assumptions underlying an assessment, it turns to the taxpayer to demolish those assumptions, and observing that not many people understand that in tax law you are guilty until you prove yourself innocent[8].

We would soften that framing slightly, and this is our own view. The presumption is not a finding of wrongdoing; it is an allocation of the evidentiary onus in a civil dispute, and it reflects that the taxpayer holds the facts.

The practical effect is nonetheless what the commentary describes. The figure in the assessment stands unless displaced, and displacing it is the taxpayer's job.

It is worth noting what a prima facie case requires. It is not proof beyond doubt. It is evidence credible and weighty enough that, unanswered, it would tend to disprove the assumption.

A Web Of Assumptions

Why the burden bites harder here than in an ordinary dispute.

Commentary observes that CRA's ability to rely on its assumptions is particularly powerful when an estimative method is used, since indirect income verification techniques are underpinned by a web of assumptions that the taxpayer must individually address[7].

That is the central procedural difficulty and this is our own analysis of it.

In an ordinary dispute about a deduction there is one assumption to address: that the expense was not incurred for the stated purpose. The taxpayer produces the invoice.

A net worth assessment rests on assumptions about opening net worth, closing net worth, the existence and value of every asset, the absence or amount of every liability, personal expenditure in each year, and the non-existence of non-taxable receipts.

Each is a separate assumption and each must be addressed on its own evidence.

Commentary elsewhere notes the related difficulty that it is always difficult to prove a negative, and that once CRA resorts to these techniques the subject's records tend to be fairly limited[8].

So the taxpayer is asked to displace many assumptions, individually, using records whose inadequacy is the reason the method was used at all.

That is a genuinely difficult position, which is why the next section matters so much.

The Burden Splits Three Ways

The observation we consider most useful in this article, and it is our own synthesis of the sources.

A taxpayer receiving an indirect verification assessment typically receives one document covering several years with penalties. It reads as a single demand and is naturally treated as a single fight.

It is not. The burden is allocated differently across three components.

On open years, the taxpayer must demolish the Minister's assumptions[7].

On statute-barred years, commentary states the burden shifts to CRA, which must demonstrate that the taxpayer made a misrepresentation attributable to neglect, carelessness or wilful default before it can validly reassess beyond the normal period[6].

On gross negligence penalties, commentary states CRA bears the burden of proof[6].

Consider what that means arithmetically. On our own illustration, take a five-year assessment producing a $220,000 total adjustment, of which two years are statute-barred, at an assumed 43 percent rate and with gross negligence penalties at an assumed 50 percent of the tax.

Roughly $56,760 of tax sits on the three open years, which the taxpayer must address. Roughly $37,840 sits on the statute-barred years, which CRA must justify opening at all. And roughly $47,300 of penalty sits across the whole, which CRA must prove.

So on those assumptions, 40 percent of the years and the entire penalty rest on CRA rather than on the taxpayer.

A taxpayer who treats the assessment as one problem, and spends all their resources reconstructing deposits, is contesting the portion where the odds are worst while leaving the portion where the odds are best unaddressed.

The Reversal On Old Years

The most important of those three, developed.

Commentary states that limitation periods act as a safeguard for taxpayers, that once a tax year is statute-barred, which typically occurs three years from the date of the original assessment, the burden of proof shifts to CRA, and that this reversed burden puts CRA's indirect methods, such as the net worth method, under a microscope[6].

It reports that in Fuhr v The King, 2024 TCC 43, the Tax Court invalidated CRA's statute-barred reassessments because of the auditor's speculative assumptions and flawed net worth calculations[6].

Another notes that CRA's auditors routinely attempt to reassess statute-barred taxation years without first discharging their burden of proving misrepresentation on a balance of probabilities[7].

Three observations, ours.

The reversal changes the character of the argument entirely. On open years the taxpayer must build a case; on statute-barred years the taxpayer's task is to show that CRA has not built one.

The weaknesses of the method become the taxpayer's ally rather than an obstacle. Speculative assumptions and flawed calculations, which on an open year are hurdles to be cleared, are on a statute-barred year evidence that the Agency has not met its onus.

And the observation that auditors routinely attempt this without discharging the burden means the point should be raised early and explicitly, not assumed to have been considered.

A Gap Is Not Proof

The formulation that captures the statute-barred principle, and it deserves its own section.

Commentary states that when estimative methods intersect with statute-barred years, CRA must do far more than present a discrepancy between reported income and the results of an indirect verification method, and that the mere fact that a net worth assessment produces a gap does not, without more, constitute proof of misrepresentation[7].

That sentence is the whole argument on an old year, and this is our own reading of why it is correct in principle.

A net worth computation produces a gap whenever its inputs are imperfect, which the Tax Court has acknowledged they inherently are. A gap is therefore consistent with unreported income and equally consistent with the method's known imprecision.

Treating the output of an admittedly imprecise estimate as proof that the taxpayer misrepresented would allow the Agency to satisfy its burden simply by running the calculation.

Commentary adds a realistic caveat, noting that the analysis is discretionary and turns in large measure on credibility and the particular record before the Court, and that settling with CRA will in many instances be a viable alternative to testing these questions in court. It stresses the importance of retaining experienced tax litigation counsel when facing reassessment of statute-barred years[7].

We would repeat that caveat rather than soften it. The principle is favourable; the outcome depends on the record and on credibility.

Penalties Are CRA's To Prove

The third burden, and a point commentary makes crisply.

It states that CRA also bears the burden of proof when assessing gross negligence penalties, so that even if the court upholds CRA's use of an alternative assessment technique, the court will vacate the penalties where that burden is not met[6].

That is a genuinely important structural point and this is our own analysis.

It means the tax and the penalty can be decided differently on the same facts. A taxpayer can lose on the quantum, because they could not displace the assumptions, and still succeed on the penalty, because CRA could not prove gross negligence.

Given that gross negligence penalties are substantial, that outcome can represent a large proportion of the total exposure.

Two practical consequences.

A taxpayer should not concede the penalty as a consequence of conceding the tax. They are separate determinations with different onuses, and treating the penalty as following automatically forfeits the stronger position.

And the evidence relevant to each differs. The tax turns on the numbers; the penalty turns on the taxpayer's state of mind and conduct, including their education, experience, reliance on advisors and the nature of the errors.

What Decidedly Does Not Work

A warning stated plainly by commentary, and it corrects a natural instinct.

It says that one thing that will decidedly not work is claiming that because you were assessed by an indirect method, the assessment should not count, and that the Tax Court's records are littered with taxpayers who have tried to just poke holes in CRA's methods, without offering alternatives of their own. It gives as an example St Phillippe, concerning a cab driver who kept limited records of his driving[8].

We report the case as commentary describes it and did not obtain its outcome beyond that framing.

The lesson is clear and this is our own reading of it.

Establishing that a method is imprecise does not establish what the correct figure is. A court presented with an imperfect estimate and no alternative has only the imperfect estimate.

So criticism must be paired with construction. A taxpayer's position should be: this computation contains these specific errors, and correcting them produces this figure.

That is a materially harder task than criticism, and it is what distinguishes a successful challenge from an unsuccessful one.

It also explains why the arithmetic section earlier in this article emphasised granularity. An alternative computation is built from many individually evidenced corrections, not from an argument about the method.

What Actually Works

The constructive response, drawing the sources together. This section is our own synthesis.

Commentary states that where CRA uses these methods, the best possible response is to produce exhaustive records, while acknowledging that this is a challenge because the records tend to be limited[8]. It also emphasises the value of consistent and rational explanations for non-taxable deposits to strengthen credibility if the matter proceeds to appeal[5].

An IMF technical note on audit practice generally, which we flag is not a statement of Canadian law, records among the principles that an auditor must investigate all reasonable evidence presented by the taxpayer refuting the computation of income, while also noting that determinations are presumed correct, that the auditor may not be arbitrary, and that the auditor is not required to negate every possible non-taxable source[2].

Five things follow, ours.

Obtain the auditor's working papers and computation, since errors are described as common and cannot be found without seeing the arithmetic.

Reconstruct liabilities at each year end, from lenders rather than memory, since omitted debt directly overstates income.

Account for inter-account transfers systematically, since failing to eliminate them is a named common error.

Evidence non-taxable receipts individually, with consistent explanations, since credibility across the whole depends on the coherence of each.

Produce an alternative computation, not merely a critique.

The Recharacterisation Option

A second line of argument that is easy to overlook.

Commentary observes that it may be possible to recharacterise the income CRA discovers, giving the example of a person trading stocks successfully outside work and not reporting the gains, who might be subject to a bank deposit analysis with CRA deciding they were operating a business. It notes that establishing the gains were capital gains rather than profits from a business could nearly halve the tax owing[8].

The general principle, and this is our own extension, is that an indirect method establishes that money arrived. It does not establish what the money was.

So a taxpayer has two distinct arguments available: that the amount is smaller than computed, and that the amount is of a different character than assumed.

Several characterisations may be available on the same facts and produce materially different tax. Capital rather than income. A different year, where amounts have been allocated across a period. A shareholder loan repayment rather than a benefit. Proceeds of a personal-use asset. Amounts belonging to another person or entity.

Each requires evidence and none is available on assertion. The point is that a taxpayer whose entire effort goes into disputing quantum may be leaving the more valuable argument unmade.

This publication made a related observation in its article on real estate flipping, where characterisation rather than amount determined the outcome. The same logic applies to money an indirect method has located.

The Cheapest Defence

What a business can do before any of this happens. This section is our own analysis.

Every difficulty in this article follows from the same root: the taxpayer cannot account for the origin of money that moved.

That problem is created over years and cannot be solved after an assessment. It can be prevented at almost no cost.

Keep business and personal banking rigidly separate. Commingling is what generates the transfers, the double counting and the mischaracterised withdrawals that these techniques handle badly.

Document every non-business deposit at the time. A one-line note with a date and a source, attached to the statement, is worth more than any reconstruction attempted years later.

Paper family loans. Money from a relative to buy a house or a vehicle is the single most common non-taxable receipt and the hardest to evidence afterwards. A signed note costs nothing.

Retain financing documents for major assets. They are the direct answer to the omitted-liability error.

Keep a personal balance sheet annually. Assets, liabilities and their basis, prepared once a year. It is the taxpayer's own version of the computation the auditor will build, and having a contemporaneous one is decisive.

That last item is unusual advice for a small business owner and we would press it. A net worth assessment is a reconstruction of the owner's personal financial position. An owner who has been recording that position all along is not reconstructing anything.

What The Auditor Actually Examines

The enquiry in practice. This section is our own analysis, drawn from the methods described.

All bank accounts, personal and business, obtained from financial institutions directly, with every deposit totalled.

Assets acquired during the period, including property, vehicles, investments and improvements, from registries and third-party records.

Personal expenditure, derived from statements and from published cost-of-living data where records are absent.

Liabilities, to the extent identified, which is where omissions favouring the Agency arise.

Ratios against sector norms, including margins, cost percentages and turnover.

Lifestyle indicators, which commentary describes as extending to informal enquiry and publicly visible information.

Third-party records, from suppliers, customers and platforms, which this series has repeatedly identified as the strongest evidence available.

The third item is worth isolating. Where personal expenditure records do not exist, an estimate must be used, and an estimate of household spending applied to a family that lives frugally will overstate income for every year of the period.

What Records Survive

Separate bank accounts for business and personal, maintained without exception.

Annotated bank statements, with non-business deposits identified at the time.

Signed loan documents for family and private borrowing, dated.

Financing agreements and payout statements for every major asset, retained across the whole period of ownership.

An annual personal statement of assets and liabilities, prepared contemporaneously.

Evidence of one-off non-taxable receipts: inheritances, settlements, gifts, asset sales, insurance proceeds.

Records explaining unusual ratios, including wastage, theft, product mix and pricing decisions.

The auditor's own working papers and computation, obtained as soon as an assessment is proposed.

What To Do

Establish whether the technique is supporting or assessing. One is a risk signal; the other is a reconstruction intended to raise tax, and they call for different responses.

Split the file by burden immediately. Open years, statute-barred years and penalties carry different onuses, and on our illustration a substantial share rests on CRA.

Raise the statute-barred point explicitly and early. Commentary states auditors routinely attempt to reassess old years without first discharging their burden.

Do not concede the penalty with the tax. They are separate determinations, and a court may uphold the assessment while vacating the penalty.

Get the auditor's computation and check the arithmetic. Commentary states errors are common and that finding them demonstrates the unreliability of the figures.

Reconstruct liabilities from lenders, not memory. Omitted debt overstates income directly, and the documents are obtainable.

Eliminate inter-account transfers systematically. Failing to correct for them is a named common error running against the taxpayer.

Offer an alternative computation, not a critique. The Tax Court's records are described as littered with taxpayers who poked holes without offering alternatives.

Consider characterisation as well as quantum. An indirect method establishes that money arrived, not what it was.

If it has not happened yet, separate your banking and keep an annual personal balance sheet. The whole exercise is a reconstruction of a position you could simply have recorded.

The Limits Of This Analysis

Several caveats matter. This is not tax or legal advice; an assessment of this kind is a dispute and anyone facing one should obtain representation. Everything is stated as verified in August 2026 and requires confirmation. We have read none of the judgments described, and our accounts of Hickman Motors, Ramey, Fuhr and St Phillippe come entirely from professional commentary; we did not obtain the outcome of St Phillippe beyond a description of its framing. We have not verified any statutory provision against the Act, including the normal reassessment period, the misrepresentation exception or the gross negligence penalty provision. We accessed the structure of CRA's audit manual chapter but not its operative content, so our description of the projections technique and the net worth technique reflects section headings rather than guidance. One source is an International Monetary Fund technical note on audit practice generally and is not a statement of Canadian law. The characterisation of supporting tests as not being used to raise tax comes from a single commentary source. Descriptions of practice including informal enquiry and lifestyle observation are commentary characterisations rather than documented CRA procedure. All arithmetic is our own, applies assumed rates to hypothetical figures, adopts an assumed gross negligence penalty proportion for illustration, and takes no account of interest or provincial variation. The three-burden framing, the emphasis on omitted liabilities, the observation that criticism must be paired with an alternative computation, the prevention analysis and the audit examination structure are our own. This article does not address the process for objecting or appealing, the voluntary disclosures programme, criminal investigation, GST/HST assessments made on a similar basis, or provincial equivalents.

Related Reading

CRA Audits By Industry Half The Interest Turns On One Letter From October 2025 the programme runs on two streams. The difference between them is always exactly fifty percent of the accrued interest, and it turns... CRA Audits By Industry What The Money Replaces The tax treatment of a settlement is decided by what it was intended to replace. That question is usually answered by wording drafted at two in the... Compliance & Reporting 365 Days Is A Floor, Not A Safe Harbour Where characterisation rather than amount decided the outcome, as it can here.

Frequently Asked Questions

What is indirect verification of income?

An umbrella term CRA uses for techniques estimating income where books and records do not reflect an accurate picture. Commentary describes it as estimating income from a taxpayer's lifestyle and assets, and as particularly powerful against small businesses and individuals without an extensive paper trail.

Why do these assessments come out so high?

The technique is asymmetric. Every deposit counts as income unless someone removes it; auditors remove what is obvious but commentary says they rarely go to great length, leaving the taxpayer to substantiate the rest. Commentary states such assessments routinely overinflate liability, and that even a taxpayer with good records will struggle to substantiate every deposit.

Is arguing that the method is unreliable a defence?

Not on its own. Commentary states plainly that claiming an assessment should not count because an indirect method was used will decidedly not work, and that the Tax Court's records are littered with taxpayers who poked holes without offering alternatives of their own. Criticism must be paired with an alternative computation.

Does the taxpayer always carry the burden?

No, and this is the most useful point. On open years the taxpayer must demolish the Minister's assumptions. On statute-barred years CRA must prove misrepresentation attributable to neglect, carelessness or wilful default. And CRA bears the burden on gross negligence penalties. One assessment, three different onuses.

What single item most reduces a net worth assessment?

Evidence of debt. Commentary notes auditors often overlook liabilities that would offset an increase in assets. A financed vehicle looks like accumulated wealth unless the loan is shown, and borrowing appears nowhere in an asset record. Lender documents are obtainable and directly reduce computed income.

What should a business do before this ever happens?

Separate business and personal banking without exception, annotate non-business deposits at the time, paper family loans, keep financing documents, and prepare an annual personal statement of assets and liabilities. A net worth assessment reconstructs a position you could simply have been recording all along.

References

1. Canada Revenue Agency. Income Tax Audit Manual, Domestic Compliance Programs Branch, Chapter 13, for the structure of its treatment of indirect verification techniques, including sections on the shareholder's investment, the reason for a discrepancy in net worth, the analysis of a discrepancy in income, completing the assessing net worth technique, working papers and the audit report form; the assessing unidentified bank deposits technique; and the assessing projections technique, with sections on audit evidence, indirect versus direct verification, projections to determine income, circumstances that warrant proceeding to the technique, reasonable projections, assumptions, comparison with the net worth technique, circumstances that do not warrant proceeding, sample selection and types of projections. Note: a CRA primary publication; we accessed the chapter structure rather than its operative content. canada.ca — Income Tax Audit Manual ch. 13
2. International Monetary Fund. Biber, E. Taxpayer Audit: Use of Indirect Methods, Technical Notes and Manuals, on principles including that an auditor must construct income that is reasonable under the circumstances, may not be arbitrary in the use of the authority, may use an indirect method to test the accuracy of books and records, must investigate all reasonable evidence presented by the taxpayer refuting the computation of income, that determinations are presumed correct with the taxpayer bearing the burden of proving otherwise, that the auditor is not required to negate every possible non-taxable source, and that third party records may be used in reconstruction; and on the burden being on the taxpayer to prove that receipts came from a non-taxable source. Note: an international technical note on audit practice generally, not a statement of Canadian law. imf.org
3. Rosen & Associates Tax Law. Income Verification: CRA Audits of Indirect Income, on indirect verification of income being an umbrella term CRA uses for techniques estimating earnings and spending from the circumstances where a taxpayer lacks records; on the methods being an extremely powerful tool particularly against small businesses and individuals without an extensive paper trail; on the distinction between supporting tests, used to determine risk or back up other audits and not used to alter income and raise taxes, and assessing techniques, which are more strenuous and exhaustive and may be used to reassess income and raise taxes; on the four primary supportive tests being bank deposit analyses, described as mandatory in many circumstances, rough net worth analyses, source and application of funds, and ratio analysis; and on supporting techniques extending to informal enquiry and publicly visible lifestyle information. Note: a Canadian tax law firm publication; the characterisation of supporting tests and of informal practice comes from this source alone. rosentaxlaw.com
4. Rotfleisch & Samulovitch PC. CRA Net Worth Audit, on CRA being able to assess based on indirect methods using information obtained from third parties or the taxpayer where it believes normal books and records do not reflect an accurate picture of income earned; on the net worth audit being one of the most common, powerful and troublesome methods employed, proceeding from the premise that a taxpayer has unreported income the auditor may not be capable of identifying through direct means; and on the method relying on the assumption that when a taxpayer accumulates wealth there are two options, to spend or invest the income, with the auditor examining expenditures in depth. Note: a Canadian tax law firm publication. taxpage.com
5. Counter Tax Litigators. (2024). Indirect Determinations of Income in the Tax Audit Process, on CRA frequently making arbitrary assessments based on an estimate of income where a return is not filed, and on indirect determinations being used to estimate income from lifestyle and assets where unreported income is suspected; on bank deposit analysis theory being that income equals amounts deposited, with the auditor requesting statements and adding all deposits; on auditors usually removing clearly non-taxable transactions such as transfers between accounts or loans but rarely going to any great length; on the taxpayer bearing the burden of establishing which deposits were non-taxable, with even a taxpayer with good records struggling to substantiate every deposit; on such assessments routinely overinflating tax liability and being difficult to challenge; on the value of a detailed review of the auditor's calculations and logic because the auditor often makes errors; and on the importance of consistent and rational explanations for non-taxable deposits to strengthen credibility on appeal. Note: a Canadian tax litigation practice publication. countertax.ca
6. Rotfleisch & Samulovitch PC, as published by Canadian Accountant and syndicated. How to Beat CRA Alternative Tax Assessments, on limitation periods acting as a safeguard, with the burden shifting to CRA once a year is statute-barred, typically three years from the original assessment, and CRA being required to demonstrate misrepresentation attributable to neglect, carelessness or wilful default; on that reversed burden putting indirect methods under a microscope; on Fuhr v The King, 2024 TCC 43, in which the Tax Court invalidated statute-barred reassessments because of speculative assumptions and flawed net worth calculations; on CRA bearing the burden of proof when assessing gross negligence penalties, so that a court may uphold the technique yet vacate the penalties; on the net worth technique aggregating the practical flaws of other indirect methods, including mischaracterising business withdrawals as personal expenditures and failing to correct for transfers between accounts; on auditors often overlooking liabilities that would offset an increase in assets; on proper technique requiring reconciliation and elimination of inter-account transfers, which often does not happen; and on the Tax Court having long recognised, in Ramey v The Queen, that the net worth method is an unsatisfactory and imprecise way of determining income. Note: a Canadian tax law firm publication; we have read none of the judgments. canadian-accountant.com
7. Miller Thomson LLP. Estimative Methods, Statute-Barred Years, and the Burden of Proof, on the principle established in Hickman Motors Ltd. v. Canada, [1997] 2 SCR 336, that the Minister's factual assumptions are presumed valid with the initial onus on the taxpayer to demolish them by a prima facie case, meaning evidence sufficiently credible and weighty that if unchallenged would be accepted as tending to disprove the assumptions; on CRA's ability to rely on assumptions being particularly powerful where an estimative method is used, since such techniques are underpinned by a web of assumptions the taxpayer must individually address; on CRA's auditors routinely attempting to reassess statute-barred years without first discharging their burden of proving misrepresentation on a balance of probabilities; on CRA needing to do far more than present a discrepancy, with the mere fact that a net worth assessment produces a gap not constituting proof of misrepresentation without more; on a case in which CRA applied both a net worth analysis and a bank deposit analysis across three statute-barred years; and on the analysis being discretionary and turning largely on credibility, such that settling may be a viable alternative, with experienced tax litigation counsel important. Note: a Canadian law firm publication; we have read none of the judgments. millerthomson.com
8. Rosen & Associates Tax Law. CRA Audit: Indirect Verification of Income, on it turning to the taxpayer to demolish CRA's assumptions once an indirect verification audit has laid them out; on the best possible response being to produce exhaustive records, while noting that once CRA resorts to these techniques the subject's records tend to be limited and it is always difficult to prove a negative; on explanations being able to remove items from a net worth audit or a bank deposit analysis; on the possibility of recharacterising income CRA discovers, with the example of trading gains treated as business profits which if established as capital gains could nearly halve the tax owing; on it decidedly not working to claim that an assessment should not count because an indirect method was used; and on the Tax Court's records being littered with taxpayers who tried to poke holes in CRA's methods without offering alternatives of their own, referring to St Phillippe concerning a cab driver who kept limited records. Note: a Canadian tax law firm publication; we did not obtain the outcome of the case referred to. rosentaxlaw.com

This article is provided for general informational purposes and is not tax or legal advice. The authors have read none of the judgments described and have not verified any statutory provision against the Act. CRA's audit manual was accessed for chapter structure rather than operative content. One source is an international technical note on audit practice, not Canadian law. All arithmetic is the authors' own, applies assumed rates and an assumed penalty proportion to hypothetical figures, and excludes interest.