A Forum Post Told Them Not to Bother With a Written Agreement

The situation

The letter arrived on a Tuesday, addressed to the family trust that held the shares of both companies, and it took Saskia most of the evening to understand what it was actually asking for. The CRA had opened a review of two related corporations, one that operated a small clinical staffing service where Saskia worked as a registered nurse and picked up management duties on the side, and one that did equipment maintenance and repair, the business Takeshi had built around his trade as a millwright. Both companies were owned by the same family trust, set up years earlier with their son Haruto as a beneficiary, and from early on the two businesses had shared a bookkeeper, a small back office, and a handful of overhead costs — rent on a shared unit, a portion of the bookkeeper's salary, insurance, and some equipment the maintenance company occasionally lent to the staffing side for site visits.

Neither Saskia nor Takeshi had ever thought of this as anything unusual. They split the shared costs roughly by revenue each year, adjusted informally when one business had a slow stretch, and moved on. When they first set the two companies up on the trust structure, Saskia had looked into whether they needed anything formal in place to allocate costs between the companies and found a long forum thread where several posters, none of them tax professionals as far as she could tell, said that as long as the numbers were reasonable and consistent, informal allocation between related companies was fine and a written agreement was optional paperwork most small operators skipped. That matched what she wanted to hear, and it became the plan.

For nearly four years the two companies filed their returns with each one claiming its share of the shared costs as a deduction, based on Saskia's year-end spreadsheet allocating rent, wages, and other overhead between them. There was no written cost-sharing agreement, no board resolution authorizing the allocation method, and no contemporaneous record explaining why the split was calculated the way it was in any given year. The bookkeeper simply entered the numbers Saskia gave her.

The review letter asked for exactly the documentation that did not exist.

Where it went wrong

Related companies are allowed to share costs and allocate them between themselves for tax purposes, but the CRA does not simply accept a year-end spreadsheet as proof that a deduction claimed by one company for a cost actually incurred by, or properly allocable to, another was reasonable. Without a written cost-sharing agreement setting out how costs would be allocated, on what basis, and why, the CRA treats the allocation with considerable skepticism, particularly when the companies are related and controlled by the same trust — a relationship that removes the natural check an arm's length negotiation would otherwise provide.

The reviewer's position was that in the absence of a contemporaneous agreement, there was no reliable basis to conclude the costs claimed by each company matched costs actually incurred for that company's benefit in the years in question. Rather than working through the underlying invoices and trying to reconstruct a fair allocation, the reviewer proposed disallowing the shared cost deductions in both companies for the two most recent years under review, on the basis that neither company had established its entitlement to the amounts claimed. That position, if it stood, would have reassessed both companies for additional tax, interest, and penalties on the disallowed amounts, with the combined exposure across both corporations landing in the range of $50,000 to $150,000.

The forum advice Saskia had relied on was not entirely wrong about the underlying principle — related companies genuinely can share costs, and a formal agreement is not a strict legal precondition to the deduction being valid in principle. What the advice missed, and what cost the family the most, was the practical reality of a review: without documentation created at the time the costs were incurred, the taxpayer is left trying to prove after the fact what the arrangement was and why it was reasonable, against a reviewer who has no obligation to take the taxpayer's word for it. A properly drafted cost-sharing agreement, even a simple one, does two things a spreadsheet cannot: it fixes the allocation method before the fact, removing any suggestion the split was chosen retroactively to minimize tax, and it gives the CRA something contemporaneous to test the numbers against.

There was also a smaller compounding problem. In one of the years under review, the allocation had shifted noticeably — the staffing company absorbed a larger share of the shared rent than in prior years, coinciding with a year when the maintenance company had unusually low income. There was a legitimate operational reason for the shift, but with no written record explaining it at the time, the shift itself looked, on paper, like the allocation had been adjusted to move a deduction toward the company that could use it most.

What we did

1. Gathered every underlying invoice and receipt behind the shared costs. Before responding to the reviewer, we needed to know whether the numbers themselves were defensible, so we assembled four years of invoices for rent, the bookkeeper's wages, insurance, and equipment costs to confirm the total shared pool was accurate. A reviewer who doubts the underlying totals will not seriously engage with an argument about how they were divided, so this groundwork came first and ruled out any suggestion the companies had overstated the costs being shared.
2. Reconstructed the allocation methodology each company had actually used. We worked with Saskia to document, year by year, what basis she had used to split each cost category — largely revenue share, with some costs like the shared equipment allocated by actual usage logs Takeshi had kept informally on paper. This gave us a coherent, explainable method rather than a bare set of numbers.
3. Drafted a cost-sharing agreement reflecting the actual historical practice. Rather than inventing a new methodology going forward, we documented the allocation approach the companies had genuinely followed, including the revenue-share formula and the usage-based method for equipment, so the agreement matched what had actually happened. This mattered because a reviewer can usually tell when a methodology has been reverse-engineered to fit the numbers, and an agreement describing real past practice was far more credible than one that looked freshly designed to win the argument.
4. Explained the rent shift with the operational record. We pulled the maintenance company's own records showing a slow stretch caused by a large client's temporary shutdown that year, unrelated to any tax planning, and presented that alongside the allocation shift so the reviewer could see a business reason rather than a retroactive adjustment. Finding that record before the reviewer flagged the shift let us present it as background context rather than a defensive reaction, which carried more weight in how the file was assessed overall.
5. Negotiated with the reviewer over which years' deductions could be substantiated. Armed with invoices and a reconstructed methodology, we were able to support most of the shared cost deductions for both companies, but conceded that a portion of one year's allocation, the year with the unexplained shift before we found the shutdown record, could not be fully substantiated to the reviewer's satisfaction even with the additional evidence.
6. Put a written cost-sharing agreement in place going forward. To prevent the same problem recurring, we drafted a formal agreement, approved by resolution of both companies' directors, setting out the allocation method for future years so that going forward the documentation exists before the costs are even incurred. The agreement also set an annual review date, so any change in how costs are split gets recorded when it happens rather than reconstructed years later under review — exactly what turned an ordinary allocation into an expensive dispute this time.

The outcome

The CRA accepted the reconstructed documentation and invoices for most of the shared costs across the review period, which meant the bulk of the originally proposed reassessment was withdrawn. The one year with the unexplained rent shift was not fully saved — the reviewer accepted the operational explanation as plausible but not conclusive, and the companies agreed to a partial disallowance for that year rather than pursuing a formal objection over an amount that was, on the numbers, genuinely uncertain.

The final reassessment landed at roughly $38,000 in additional tax, interest, and penalties combined across both companies, against an original proposed exposure closer to $110,000. That is a real reduction, but it is also a real cost the family would not have faced at all had a written agreement existed from the start — the $38,000 reflects genuine uncertainty in one year's numbers that better contemporaneous records would likely have avoided.

Saskia and Takeshi paid the reassessed amount and put the new agreement in place for both companies going forward. Haruto, old enough now to be involved in some of the family's business discussions, sat in on the meeting where the new agreement was signed. Saskia has since stopped looking to online forums for anything involving the trust's companies, and the family now reviews the cost-sharing agreement with our office annually rather than letting years pass on an informal understanding.

What you can learn from this