Anonymised, illustrative composite. The decision not to file was made in ten minutes on the basis that no filing was required. Nobody costed the two extra years it bought.
At a glance
A sponsor built a specialty-bakery platform out of three family businesses supplying grocery in-store bakeries in Ontario and Quebec. The acquisitions closed over nineteen months at $46,000,000, $18,500,000 and $11,200,000. None of them came near the notification thresholds: the party-size test in s.109(1) requires assets in Canada or gross revenues from sales in, from or into Canada exceeding “four hundred million dollars” on a combined basis, and the platform never reached a quarter of that.
So no filing was made, correctly. The decision was recorded in a single line of a closing checklist on each deal: not notifiable, no filing required. That line was right and it answered a different question from the one that mattered.
Twenty-six months after the last closing the sponsor ran a sale process. The buyer was a national baked-goods business, and its counsel's regulatory memo did something the sponsor's advisers never had: it asked when the Commissioner's ability to challenge the three build-up acquisitions expired.
The answer was that it had not. The most recent acquisition had ten months left on a three-year clock, and the second had six. Had any of them been notified, all three would have been beyond challenge long since. The platform's competitive position — roughly a third of in-store bakery supply in two provincial grocery channels, assembled from three businesses that had previously competed — was exactly the kind of fact pattern a buyer's counsel flags.
The sponsor's reaction was that this made no sense: surely filing exposes you to review and not filing keeps you out of it. That intuition is the trap, and s.97 runs the other way.
Both notification thresholds were missed by a wide margin. The transaction-size test in s.110 turns on an amount that is re-set each year: s.110(7) fixed it, in the year the subsection came into force, at “$70,000,000”, and s.110(8) then indexes it by nominal GDP, with s.110(9) requiring that “As soon as possible after determining the amount for any particular year, the Minister shall publish the amount in the Canada Gazette”. Anyone applying the test must take the current year's published figure rather than a number remembered from a previous deal.
The cost of the open window was priced by the buyer, not argued about. It required a $6,000,000 specific indemnity, escrowed for the balance of the s.97(b) periods, against the cost of a Commissioner's application — a divestiture, the associated loss on a forced sale, and the legal spend. On a $118,000,000 enterprise value that is 5.1% of the price sitting in escrow for reasons that had nothing to do with the business's trading.
The alternative had been cheap. An advance ruling certificate request or a voluntary notification on the largest of the three acquisitions would have cost the filing fee and perhaps six weeks of process. It would have closed the window at the twelve-month mark and there would have been no escrow.
s.97 is short and it is a limitation period on the Commissioner, not on the parties. No application may be made under s.92 “in respect of a merger that was the subject of a request for a certificate under section 102 or a notification under section 114, more than one year after the merger has been substantially completed”, or “in respect of any other merger, more than three years after the merger has been substantially completed”.
Three units have to be read correctly. The trigger for the shorter period is a request for an advance ruling certificate under s.102 or a notification under s.114 — not the granting of a certificate, and not clearance. s.102(1) provides that “Where the Commissioner is satisfied by a party or parties to a proposed transaction that he would not have sufficient grounds on which to apply to the Tribunal under section 92, the Commissioner may issue a certificate to the effect that he is so satisfied.” Making the request starts the shorter clock whatever the answer.
Second, the clock runs from the merger being “substantially completed”. That is not signing, not announcement and not the date the Commissioner learns of it. On a build-up it is measured deal by deal, so each acquisition carries its own window.
Third, s.97 limits applications under s.92. It says nothing about Part IX obligations, which is why a genuinely notifiable deal cannot be improved by staying quiet — the s.123.1 exposure for closing without notice is separate and survives. The interaction with an advance ruling certificate is worth noting too: under s.113(b), a transaction is exempt from Part IX where the Commissioner has issued “a transaction in respect of which the Commissioner has issued a certificate under section 102”, so a certificate both removes the filing obligation and shortens the challenge window.
The sale completed with the $6,000,000 indemnity escrowed and released in two tranches as each remaining window closed. No application was ever made and the money came back, which is the usual outcome and does not make the escrow a waste — the buyer was pricing a real, uninsurable contingency.
The sponsor changed its standing practice on the next platform. Where a build-up produces a materially concentrated position in any market it actually competes in, it now makes a s.102 request even on a plainly non-notifiable deal, treating the filing decision as a risk-allocation decision rather than a compliance one. The concentration test that tells it when to bother is set out in a third acquisition that flipped the burden of proof; on the exit itself, see exiting a roll-up to a larger acquirer and the survival period glossary entry on how these tails are usually papered.
The escrow is the visible cost and the smaller one. The real cost of an open s.97(b) window is that it narrows the buyer universe. A strategic acquirer with its own competition counsel will price it; a smaller financial buyer may simply decline, because it cannot carry a divestiture risk it has no way to quantify. The sponsor's process produced four indicative offers and two of them fell away at exactly this point.
There is also an asymmetry in who bears it. A vendor who filed and cleared has a clean answer to the question; a vendor who did not has to fund an escrow out of proceeds. Two years earlier, the same protection was available for a filing fee. That is the whole of the lesson: the decision not to file is a decision to keep a contingent liability on the business for an extra two years, and it should be taken with that on the page.
The tell was a checklist line that answered a narrower question than it appeared to. “Not notifiable” establishes that no filing is required. It does not establish that no filing is useful, and on a roll-up those are different questions with different answers.
The generalisable form: whenever a rule gives you an option you are not obliged to take, ask what the option buys as well as what it costs. Here the whole benefit sits in a different section from the obligation — the duty is in s.114 and the reward for volunteering is in s.97 — which is why a compliance-shaped analysis never finds it.
A 30-minute call is enough to tell you whether AI pays for itself here.
Canadian small-business transaction data is not published anywhere, so most valuations in this country quote an American benchmark. The Deavo–Treadstone Acquisition Index is a daily record of Canadian listings built to replace that: asking-price distributions by province and city are published now, and days on market, departure rates and asking-to-sale spreads follow as the series lengthens. Leave an email and we will tell you as each measure lands.
No pitch, no listings. One email when the first report lands.
