A buy-sell clause that says shares transfer at “fair market value” without naming a method has not really named a price. Three formulas recur in Canadian shareholder agreements, and run on the same company's numbers, they do not land anywhere close to the same answer.
Key takeaways
A shareholder agreement can be otherwise complete — governance, information rights, a clean exit trigger — and still leave the single most contested number in the whole document unresolved: what the shares are actually worth when the trigger fires. Treadstone Law's advice is to fix this “in advance when relations are still good,” because the three formulas below produce materially different numbers, and nobody wants to be arguing about which one applies at the exact moment a shareholder is trying to leave.
Book value prices the shares off the balance sheet — assets minus liabilities, divided by the shares outstanding. It is fast, objective, and requires no outside professional. It is also the formula least likely to reflect what a buyer would actually pay for a profitable, going concern, because it captures none of the business's goodwill, customer relationships, or earning power beyond what happens to be sitting on the balance sheet at the measurement date. A company with modest net assets and strong recurring earnings can be worth several times its book value to an actual buyer, and a book-value buy-sell clause hands that entire gap to whoever is doing the buying, not the person being bought out.
A capitalized-earnings formula applies an agreed multiple to an agreed earnings figure. It comes far closer to a market price than book value, but only if the clause is specific about two things the parties otherwise fight over later: which earnings line the multiple applies to — net income, EBITDA, some adjusted or normalized figure with defined add-backs — and over what period, since a single bad or unusually good year can swing the answer sharply if the clause does not specify an average. A multiple with no defined base is not really a formula; it is an argument deferred to the day it matters most.
The third route names a professional valuator in advance rather than fighting over who values the business after a dispute has already started. Since January 1, 2026, an independent valuator's conclusion is governed by the CBV Institute's Valuation Practice Standards — Practice Standard 100 (Valuation Conclusions and Valuation Reports), 110 (Report Disclosure), 120 (Scope of Work) and 130 (File Documentation) — which CBV Institute states “set the minimum requirements for a valuator to establish a credible and properly supported conclusion of value.” It is also the slowest and most expensive of the three, and the agreement still has to specify who pays the valuator's fee and what happens if the parties cannot agree on which firm to retain.
Some agreements instead let the shareholders set a share value by resolution once a year, avoiding both the balance-sheet distortion of book value and the cost of an independent appraisal. It only works if the group actually does it — an agreed value that was set once at formation and never revisited is functionally the same trap as no formula at all, just with extra false confidence attached to it.
Whichever formula is chosen ends up feeding the buyout obligation the group's life or disability insurance is supposed to fund. A book-value formula understates what the insurance needs to cover for a growing business; an independent-valuation formula can produce a number the group only discovers well after the coverage was last purchased. Reviewing the valuation formula and the insurance coverage on the same schedule — rather than treating them as two unrelated clauses — is the only way the funding actually matches the price the formula produces.
Whichever formula is chosen also has to feed the forfeiture price on an unvested share repurchase and every other exit trigger the agreement defines, so it is worth choosing once, deliberately, rather than letting each clause default to its own unstated assumption about what “fair value” means.
The same company, three formulas, one exiting shareholder
To show how far the answers can diverge, not to state a market benchmark: a private company has a balance-sheet book value of $1,200,000, trailing annual EBITDA of $600,000, and an independent valuator's fair market value opinion of $2,100,000, reflecting real property appreciation and customer relationships the balance sheet does not capture. All three figures are illustrative parameters chosen only to make the arithmetic legible. A shareholder holding 25% is exiting under each formula in turn, with the agreed earnings multiple set, for illustration only, at 3×:
The same 25% stake is worth $300,000 or $525,000 — a 75% swing — depending entirely on which formula the clause happens to name. Nothing about the business changed between the three numbers; only the method did.
There is no single best formula -- book value is fastest and least accurate for a going concern, an earnings multiple is closer to market value but needs a precisely defined earnings base, and independent valuation is the most defensible but the slowest and most expensive. The 'best' choice depends on how much the group is willing to spend and how precisely fair value needs to track the real business.
It can, but doing so defers the real question rather than answering it -- the parties will still need to agree on a method, or fight over one, at the exact moment a shareholder is trying to exit. Naming a specific formula, or naming the valuator in advance under an independent-valuation clause, avoids that fight happening at the worst possible time.
The CBV Institute's Valuation Practice Standards 100/110/120/130 took effect January 1, 2026 and set the minimum requirements a Chartered Business Valuator must meet for a credible, supported conclusion of value. They matter directly to a buy-sell clause that names an independent valuation as its pricing method, since that is the standard the named valuator's report will be measured against.
Not necessarily, but using a different formula for each trigger multiplies the number of assumptions the group has to maintain and revisit. Many agreements deliberately use one formula across every trigger for simplicity, and reserve a different method only for a genuinely different situation, such as a forced buyout under an oppression finding where a court sets the price instead.
A 30-minute call is enough to tell you whether AI pays for itself here.
Nobody publishes Canadian transaction data, so every valuation in this country quotes an American benchmark. We are building the Canadian one — multiples, asking-to-sale spreads and days on market, by sector and by city. Leave an email and you will see it first.
No pitch, no listings. One email when the first report lands.
