An Understated EBITDA Line Was Worth Seven Times More Than It Recovered

The situation

A strategic buyer acquired 100% of the shares of an Ontario specialty-services business, priced at 7.0x trailing twelve-month adjusted EBITDA of $2,800,000, for a purchase price of $19,600,000. The financial statements representation was standard, and the disclosure schedule included the seller’s EBITDA bridge showing every add-back used to get from reported net income to the adjusted EBITDA figure the price was built on. The indemnity article defined “Losses” broadly for what it covered — direct losses, costs and expenses arising from a breach — and narrowly for what it excluded: no consequential, indirect, special or punitive damages, and, in the same sentence, no amounts “calculated by reference to a multiple of revenue, earnings or any other financial metric.”

The problem

Six months after closing, the buyer's controller was reconciling the first full fiscal year against the pre-closing EBITDA bridge and found a $180,000 annual add-back for what the bridge described as “non-recurring personal vehicle expense.” It was not non-recurring, and on inquiry it was not entirely personal either: the vehicle was used by the seller's former general manager, who remained with the business post-closing, and the lease continued as an ordinary operating cost the buyer was now paying every year going forward. The add-back had inflated the EBITDA figure the price was built on by $180,000 — a misrepresentation of the financial statements the seller had certified as accurate.

The buyer's first calculation of its claim was not $180,000. It was $180,000 multiplied by the 7.0x multiple the deal was actually priced at — $1,260,000 — on the reasoning that the buyer had not merely lost $180,000 of cash flow, it had overpaid by $1,260,000 for a company it valued using a multiple applied to a number that was $180,000 too high.

The numbers

$180,000 a year, understated indefinitely into the business's future cash flows, dollar-for-dollar. $1,260,000, the same error priced at the actual 7.0x multiple the buyer paid for every real dollar of EBITDA the business generated. The two numbers are both honest measures of the same misstatement — one measures the annual cash-flow cost, the other measures what the buyer overpaid at closing to acquire a company that was never as profitable as represented. They differ by a factor of seven because that is the multiple the deal was priced at, and a multiple, once applied to an error in the number it is multiplying, mechanically multiplies the error too.

The rule that decided it

No statute measures indemnity damages in a Canadian share purchase agreement. The measure of “Losses” recoverable for a breach of representation is entirely a creature of how the agreement defines the term — and this is the one indemnity mechanic in this hub where the definition does not just cap what is recoverable, it decides which of two legitimate ways of quantifying the same loss the buyer is even permitted to use.

The exclusion for amounts “calculated by reference to a multiple of revenue, earnings or any other financial metric” is standard boilerplate, and sellers negotiate it for a defensible reason: without it, every small misstatement in the earnings base would expose the seller to damages inflated by the deal's own pricing multiple, which can turn a modest error into a disproportionate liability untethered from the actual cash-flow harm. Buyers accept it too: most breaches (an unpaid invoice, an unrecorded liability, a misclassified employee) cause a loss that is genuinely dollar-for-dollar, and a multiple-based measure would over-compensate for those.

A recurring misstatement in the earnings base is the one breach type where that logic runs out. Every other representation, if false, causes a loss roughly equal to the dollar value of what was misrepresented. A false statement about earnings is different in kind: the buyer did not just receive a company with a $180,000 hidden cost, it paid a price calculated by multiplying a number that included that $180,000 error, so the true measure of what the buyer overpaid is mechanically tied to the multiple, not independent of it. A dollar-for-dollar measure captures the ongoing cash-flow cost, correctly, but it does not capture the separate fact that the purchase price itself was built on the wrong number.

The clause did not distinguish between these two situations. It excluded multiple-based damages across the board, for every representation, because the drafting was generic rather than tailored to the financial-statements representation specifically — treated as boilerplate risk allocation rather than as a clause that would, on this exact fact pattern, cut the buyer's real recovery to roughly a seventh of its actual loss.

The outcome

The buyer recovered $180,000 — the dollar-for-dollar annual cost of the misstated add-back, capped by the definition of Losses at that measure and nothing more. The seller did not dispute the breach; there was nothing to dispute once the general manager's continued use of the vehicle was confirmed. The dispute, such as it was, ended at the definition of Losses, not at the facts.

What it would have cost otherwise

Had the exclusion carved out an exception for breaches of the financial-statements or EBITDA-bridge representations specifically — a common refinement precisely because sellers and buyers both understand why earnings misstatements are different — the buyer's recovery would have tracked the $1,260,000 overpayment, subject to the ordinary basket and cap. Instead the generic exclusion, negotiated once for the whole agreement, silently resolved a $1,080,000 question that neither side appears to have specifically discussed.

The tell

The tell was that the exclusion clause and the financial-statements representation were negotiated by different people on different calls, weeks apart, with no one asking how the two interacted. A multiple-based damages exclusion is a general risk-allocation clause; an EBITDA-bridge representation is a specific, numerical one. Whoever negotiates the general clause needs to know what the specific one is protecting, or the two will be drafted as if they never meet.

Takeaways

• • A recurring misstatement in the earnings base is the one breach type where a dollar-for-dollar measure and a multiple-based measure diverge by an order of magnitude — because the purchase price itself was calculated using the misstated number.
• • A generic “Losses” definition excluding multiple-based damages, negotiated once for the whole agreement, applies to the financial-statements representation exactly as it applies to every other one — unless a specific carve-out says otherwise.
• • If a deal is priced on a multiple, the buyer’s counsel negotiating the indemnity definitions needs to sit in the same conversation as whoever is negotiating the EBITDA-bridge representation, or the two clauses will be drafted without reference to each other.
• • A carve-out preserving multiple-based recovery specifically for breaches of the financial-statements or EBITDA representations is a targeted, negotiable fix — far narrower than removing the exclusion altogether, which most sellers will resist.

Sources

• • No statute measures indemnity damages, or defines “Losses,” in a Canadian share purchase agreement. The measure of recoverable loss is set entirely by how the agreement defines the term, which is why this file turns on drafting rather than on any external rule.
• • Glossary: adjusted EBITDA.
• • Glossary: multiple of earnings.
• • Treadstone Law — the indemnity cap in an Ontario business sale, and how much risk it actually limits — adjacent, not on point: it covers the ceiling on recovery, not the definition of what counts as a recoverable loss in the first place, which is this file’s subject.