A Written-Off Invoice Turns Into an HST Timing Fight

The situation

What kept Menachem awake was not the number itself. It was what would happen to his operating line if the number stayed frozen. He ran his electrical contracting business out of a small unit in North York, and like most contractors his size, he financed the gap between finishing a job and getting paid with a modest line of credit tied to his HST filings and his revenue history. If the Canada Revenue Agency held his return in review for months, the bank would see an unresolved audit before it saw a healthy business, and the renewal he needed before his next large job started would be in question.

The return itself involved a bad debt adjustment of roughly ninety thousand dollars. Two years earlier, Menachem had completed a rewiring contract for a property management firm run by a client named Nikos, work that took his small crew nearly four months to finish. Nikos's firm ran into serious cash trouble partway through a subsequent expansion, and the invoice for the rewiring job went unpaid past ninety days, then past a year. Menachem eventually wrote it off as uncollectible after repeated attempts to collect, and claimed the HST bad debt adjustment that the rules allow when a business has remitted tax on a sale it never actually got paid for.

His spouse, Shira, worked as a surveyor but handled the books for the business on evenings and weekends, and had made the claim on the return she believed matched when the debt had genuinely become uncollectible, based on the last real communication from Nikos's firm and the point at which further collection effort stopped making sense. That judgment call is exactly the kind of thing bad debt claims turn on, and it is exactly the kind of thing an auditor can see differently with the benefit of hindsight and a different set of assumptions about when a debt should have been recognized as lost.

The review letter that followed did not simply ask for documentation. It suggested the claim belonged in an earlier period, one where Menachem had already filed, which meant the current claim risked outright denial rather than a straightforward shift. By the time Menachem called our office, two phone calls with the auditor had already gone badly, and he was convinced, rightly or not, that the file had become personal rather than technical.

What the review found

Once we had the file, the first job was separating what the auditor was actually entitled to ask from what the conversation had drifted into. The technical question was narrow: in which reporting period had the debt become bad, in the sense the rules use that word, meaning the point at which it was reasonable to conclude the amount would never be collected. Menachem and Shira had picked the date Nikos's firm stopped responding to calls. The auditor's position was that the debt should have been recognized earlier, closer to when the invoice first went significantly overdue, which would have placed the adjustment in a period already filed and closed.

Reviewing the underlying file changed the picture. Nikos's firm had not simply gone silent after the invoice became overdue. There had been a partial payment roughly seven months in, along with two written promises to pay the balance once a receivable of their own came through. Those exchanges mattered, because a partial payment and an active, if unreliable, promise to pay are evidence that the debt had not yet crossed into uncollectible at the earlier date the auditor was pointing to. The claim that mattered was not when the invoice became overdue, it was when a reasonable business owner would have concluded collection was no longer realistic.

The difficulty was that most of this evidence lived in text messages and a handful of emails Shira had saved inconsistently, not in a tidy collections log. Some of the record supported the later date Menachem had used. Some of it, read uncharitably, could support the auditor's earlier date. A file that is genuinely mixed, rather than clearly in the taxpayer's favour, is not a file where a client should walk into a third phone call with an auditor he has already alienated and expect the ambiguity to resolve his way.

The other thing the review turned up was smaller but relevant to timing: a portion of the write-off, in the low five figures, related to a separate and older invoice for minor repair work that genuinely had gone bad earlier, in the period the auditor was proposing. That amount had been folded into the same claim by mistake. It was a real error, not a matter of interpretation, and it needed to be conceded rather than argued.

What we did

1. Took over direct contact with the auditor. The two prior calls had left both sides talking past each other, with Menachem reading routine questions as accusations and the auditor, understandably, becoming more formal and less flexible in response as the calls went on. Having our office become the sole point of contact removed the personal charge from the exchange entirely and let the review proceed as a document question rather than a running dispute between two people who had already stopped listening to each other.
2. Rebuilt a proper collections timeline. We assembled every message, invoice, and partial payment record we could find into a single chronological file, rather than leaving the auditor to piece together scattered texts on their own or, worse, to draw their own inferences from whatever fragments Shira had happened to save. A clear, dated timeline does more to support a bad debt date than any argument about what the rules should mean in the abstract, because it lets the auditor verify the story rather than simply weigh whose account sounds more credible.
3. Separated the two invoices. We identified the smaller, older invoice that had been mistakenly bundled into the same adjustment and pulled it out of the claim before the auditor had a chance to raise it themselves as evidence of sloppier bookkeeping across the whole file. Conceding a genuine, isolated error early does more for credibility on the larger, defensible amount than contesting everything on principle and hoping the small mistake goes unnoticed.
4. Presented the partial payment and promises to pay as the operative evidence. We framed the seven-month partial payment and the two written promises to pay as the reason the debt could not reasonably have been treated as lost at the earlier date the auditor was proposing, which is the actual legal question the bad debt adjustment rules turn on, not simply how many days the invoice had sat overdue on the books.
5. Proposed a split date as a landing point. Rather than insisting on Menachem's original date or simply accepting the auditor's earlier one, we proposed a middle point tied to the last genuine promise to pay, the date it lapsed without any further response from Nikos's firm. That date was defensible on the actual record and gave the auditor a basis to close the file on reasoned grounds, without either side treating the outcome as a simple capitulation.
6. Corrected and refiled the adjustment. Once a landing date was agreed in principle, we prepared a corrected calculation removing the mistaken invoice entirely and shifting the remaining balance to the agreed period, so the auditor was reviewing one clean, internally consistent number rather than a moving target that kept requiring further explanation. Submitting a single, reconciled figure also meant there was nothing left for a second reviewer to second-guess if the file changed hands before it closed.
7. Kept Menachem out of the remaining exchanges. We handled the final rounds of correspondence directly ourselves, including the closing confirmation, so the file closed on the strength of the record rather than on the tone of any further conversation between an anxious client and an auditor he had already, understandably, rubbed the wrong way. That distance also gave Menachem room to focus on his business and his bank renewal instead of rehearsing his next call with the auditor.

The outcome

The auditor accepted the corrected claim on the compromise date, with the smaller mistaken invoice removed from the calculation entirely. Menachem kept roughly three quarters of the original adjustment, a meaningful recovery given how the file had started, though not the full amount he had originally claimed. The portion tied to the earlier, mistaken invoice was conceded and not recovered, and a modest amount of interest applied to the slice of the claim that the compromise date treated as having effectively been filed late.

The line of credit renewal went through without incident once the audit closed, which had been Menachem's real concern from the very first phone call to our office, well before he ever mentioned the dollar figure itself. The bank saw a resolved file with a documented, agreed outcome rather than an open dispute sitting on the account, and the renewal proceeded on the same terms he had expected before the review letter ever arrived. For a business his size, that mattered more day to day than the exact percentage of the adjustment he ultimately kept.

What the file also showed, plainly, was how much the early tone of an audit shapes what happens later in it. The technical question here was always answerable on the documents, once someone actually organized them properly. What made the file harder than it needed to be was two calls in which a taxpayer, genuinely frightened for his business and his livelihood, came across as combative to an auditor who then treated every subsequent claim with more suspicion than the underlying record actually warranted. Once that dynamic was removed from the exchange, by handing the conversation to someone with no stake in how the earlier calls had gone, the file resolved largely on its facts, at a result both sides could reasonably accept without either one treating it as a defeat.

What you can learn from this