A dividend the auditor challenged, then had to walk back

The situation

The letter came first, before any conversation. It proposed to reassess a dividend Somchai's holding company had paid to itself from its operating company two years earlier, treating the full amount as an artificial reduction of capital gain rather than an ordinary flow of retained earnings between related corporations. The number attached to the proposal sat in the roughly $15,000 to $50,000 range, and it landed with no warning that a review was even underway.

Somchai worked full time as a forklift operator at a distribution centre outside Collingwood. Over a decade he and his wife Kittipong, a dental assistant, had used a modest holding company structure to buy three small rental properties, each held through an operating company that fed dividends up to the holdco once a year. The arrangement had been set up early, on the advice they had at the time, and it had run quietly for years without drawing attention. The one piece of the arrangement that had never been neglected was the bookkeeping: Yaa, a family friend who had done the couple's corporate accounting since the first property was bought, filed both companies' returns every year without fail, even in years when nothing else about the structure got any attention at all.

The trigger was a sale. One of the rental properties had been sold the year before, producing a capital gain inside the operating company. Before the sale closed, the operating company paid a dividend up to the holding company equal to its accumulated retained earnings, a routine step meant to move value to the parent before a sale rather than after, which changes how much of a later gain is taxed at the corporate level. On paper it looked like ordinary corporate housekeeping.

The auditor's letter read the transaction differently. It treated the pre-sale dividend as if it had been engineered specifically to strip value out of the operating company ahead of the sale, reducing what would otherwise have been a taxable capital gain, and proposed to deny the dividend's tax-free treatment on that basis. Somchai and Kittipong had no idea what 'safe income' meant, only that a number in the tens of thousands of dollars was suddenly being asked of them, attached to a company they thought had been run correctly the whole time.

What the law actually said

The rule the auditor was applying exists for a real reason. Ordinarily, dividends paid between two Canadian corporations that are connected to each other move tax-free, because the underlying income was already taxed once inside the paying company. But there is a limit built into the Income Tax Act meant to stop shareholders from using that tax-free flow to artificially strip value out of a company right before a sale, converting what should be a taxable capital gain into a tax-free dividend instead. The dividing line is what tax practitioners call 'safe income' — the portion of a company's retained earnings that genuinely reflects income already taxed and available to distribute, as opposed to unrealized value manufactured for the transaction.

The question was never whether the rule existed. It plainly did, and it plainly could apply to a dividend paid shortly before a sale. The question was whether this particular dividend, on this particular set of numbers, actually exceeded the operating company's safe income, or whether it fell comfortably within it. That is a factual and accounting question as much as a legal one, and it turns on a careful year-by-year reconstruction of the company's taxed earnings, not on the timing alone.

What made this file different from a straightforward dispute was an early move by the auditor. The proposal letter had calculated safe income using only the most recent fiscal year's retained earnings, rather than the cumulative total built up across the company's full history of ownership — a narrower approach than the concept actually supports. That single choice understated the company's genuine safe income by a wide margin, and it was the kind of error that, once identified precisely, gave us solid ground to stand on rather than a vague sense that the assessment felt unfair.

It mattered, too, that the operating company's books were clean. Somchai and Kittipong had kept consistent annual filings, and the retained earnings figures in those filings, added up correctly across the years the properties were held, comfortably supported a dividend well above what had actually been paid. The dispute was not really about whether the money was legitimate. It was about whether the auditor's math had captured it properly.

What we did

1. Requested the auditor's full working papers rather than responding to the proposal letter's summary figure alone, because a safe income calculation is built line by line across every year a company has existed, and disputing a number without seeing how it was built risks arguing against the wrong target entirely. The working papers confirmed the single-year approach we had suspected from the letter's language.
2. Rebuilt the safe income calculation from the company's own filings, going back to the year the operating company first held the property, to establish the cumulative taxed retained earnings actually available for a tax-free dividend, rather than accepting the auditor's narrower starting point as given. This produced a figure well above the dividend that had actually been paid, which meant the core transaction was defensible on its face before a single word was written back to the auditor.
3. Prepared a written technical response setting out the correct cumulative method, drawing on financial statements Yaa pulled from every year since the operating company first held the property, because a bare assertion that the assessment was wrong carries no weight with a reviewer who has already committed a position to writing. The response walked through each year's retained earnings figure individually rather than presenting only a final total, so the auditor was responding to a documented, checkable position instead of a conclusion asking to be taken on faith.
4. Flagged the procedural narrowing directly and by name, identifying the single-year calculation as the specific source of the discrepancy rather than allowing the file to drift into a broad, unfocused disagreement about the couple's intentions in structuring the dividend. Naming the exact error the auditor had made, rather than arguing generally that the number felt wrong, kept the conversation anchored to something the auditor could actually check and correct without abandoning the file's underlying premise entirely.
5. Held a resolution conference with the auditor's team once the revised, cumulative calculation was on the table, walking through the year-by-year retained earnings figures line by line so the reviewer could see precisely where the original single-year number had understated the company's safe income, and where, once the full history was accounted for, a genuine but much smaller gap actually remained.
6. Identified the one component of the file that was genuinely unresolved — a modest adjustment relating to a prior year's undeducted expense that Yaa's own records showed had inflated the retained earnings figure slightly beyond what the underlying books supported. Conceding that narrow point immediately, rather than contesting every dollar out of principle, preserved credibility with the auditor on the much larger and genuinely stronger part of the position.
7. Negotiated a settlement covering only the conceded component, formalized in a reassessment that reflected the corrected cumulative safe income figure for everything else in the structure. Closing the file this way avoided a full audit reopening of the earlier years' returns, which both sides had reason to want to avoid, and left Somchai and Kittipong with a single, resolved number rather than an open-ended review hanging over the rest of the holding company structure.

The outcome

The final reassessment landed at the low end of the original range in dispute, with most of the dividend confirmed as properly supported by the operating company's genuine safe income. The portion that was conceded related to a real, if minor, overstatement in one earlier year's figures, and Somchai and Kittipong accepted that adjustment rather than pushing to relitigate it, since the underlying evidence supported the auditor on that narrow point.

Neither side got everything. CRA's original position, built on the single-year calculation, did not survive contact with the full accounting history, and the file closed well short of the number first proposed. At the same time, the couple paid something, and the process took several months of documentation and negotiation they had not budgeted for, on top of an operating company structure they will now review more carefully before any future sale. Kittipong, in particular, said afterward that the hardest part had not been the number but the wait between the letter arriving and finally understanding, in plain terms, why it had been sent at all.

What the file left behind was clarity about how the structure needs to be run going forward. Somchai and Kittipong now keep a running cumulative safe income schedule for both operating companies, with Yaa updating it as part of the annual filing routine rather than reconstructing it under audit pressure, so any future dividend paid ahead of a sale can be supported immediately rather than defended after the fact.

What you can learn from this