Why These Two Cases Matter

Two fresh Court of Appeal decisions draw a hard boundary for Ontario termination clauses in 2025.

Bertsch v. Datastealth Inc., 2025 ONCA 379^[1] upheld a plain-language ESA-minimums clause.

By contrast, De Castro v. Arista Homes Ltd., 2025 ONCA 260^[2] struck a clause because its “for cause” wording left room to deny notice below the ESA threshold.

Read together, they reconfirm that text—not intention—decides enforceability.

What Bertsch Actually Says

In Bertsch, the employee was terminated without cause after a comparatively short stint with the company. He sued for common-law reasonable notice, arguing that the contract’s termination language was invalid. The motion judge disagreed, and the Court of Appeal affirmed.

The agreement did something unfashionable but lawful. It promised only what the Employment Standards Act, 2000 (“ESA”)^[3] and its regulations require, and nothing more, and it did so in clear, direct language. The clause expressly tied all termination entitlements—notice or termination pay, any severance pay that might be triggered, and benefits continuation—to the ESA framework, and it contained a simple “failsafe” clause ensuring that if any word in the agreement ever drifted below the statutory floor, the ESA would prevail. Because the words tracked the statute and regulations, and because there was no ambiguity inviting a lower-than-ESA outcome, the Courts enforced the clause and shut the claim down at the pleadings stage, as disclosing no tenable cause of action.

The real significance is practical. When the drafting pins entitlements to the ESA architecture with no loose ends, Ontario courts will enforce it. That is not an employer windfall. It is contract law doing exactly what it says on the label.

Where De Castro Went Off the Rails

De Castro presents the mirror image. The employee, with several years of service, was dismissed without cause and paid a limited sum under the contract’s termination provision. She challenged the clause. The problem was not the existence of a without-cause paragraph. It was the companion for-cause language. Instead of confining no-notice termination to the ESA’s narrow carve-out—“wilful misconduct, disobedience or wilful neglect of duty” as reflected in the ESA regime—the contract used open-ended phrasing. It spoke in terms like “shall include” and referenced generic breaches and a broad, common-law-style “cause” concept. Those words permitted no-notice outcomes in scenarios where the ESA would still require notice or pay in lieu.

On that footing, the Court of Appeal applied the now-settled rule that a defective for-cause sub-clause contaminates the entire termination section. The employee was therefore entitled to common-law reasonable notice. The lesson is not subtle. Courts examine what the language allows, not what the employer intended. If the text leaves a path to an unlawful result, the clause falls.

The Doctrinal Anchors: Waksdale and Bardal

Two pillars explain why these outcomes were inevitable.

First, Waksdale v. Swegon North America Inc., 2020 ONCA 391^[4] held that termination provisions are read as a whole. If any piece of the termination scheme violates the ESA—most commonly the “for cause” part—all termination language is invalid, even if the employer never relied on the offending paragraph. Severing doesn’t save it.

Second, once a termination clause fails, damages revert to common-law reasonable notice under Bardal v. The Globe & Mail Ltd., 1960 CanLII 294 (ON SC)^[5], which looks at the character of employment, length of service, age, and the availability of similar work—factors that anchor the negotiation and, at trial, the court’s assessment.

Put simply, Waksdale answers whether the clause stands. Bardal answers how to value the case when it doesn’t.

The line in 2025

Against that backdrop, the line is unmistakable.

Clauses survive when they anchor every termination outcome to ESA minimums, name the regulations (not just the Act), and close escape hatches with a clean failsafe.

Clauses fail when “cause” language is non-exhaustive, mixes common-law cause ideas with the ESA’s narrow wilful-misconduct carve-out, or otherwise leaves a route to an ESA-offside, no-notice dismissal. It is not enough to say “we intended to comply.” Courts are policing effects, not aspirations.

That is why Bertsch stands and De Castro falls.

Drafting in Practice (Keep It Tight)

For employers, use a plain, self-contained termination section. Say that whether employment ends with or without cause, the employee gets only ESA-minimum entitlements (Act and regulations), and list notice/termination pay, severance (if any), and benefits as ESA-governed. Add a short failsafe making the ESA floor decisive. If you mention cause, mirror the regulation and stop—no “shall include,” no generic “breach,” no catch-alls.

For employees, look for open-ended cause wording, muddy cross-references, or any importation of common-law “cause.” De Castro is your path to void the whole termination section and claim common-law notice. If the clause tracks the ESA, cites the regs, and has a failsafe, Bertsch signals a tougher, ESA-bound fight.

How This Might Play Out in Mediation (and Valuation)

On a Bertsch-style agreement, a Rule 21 exit is credible and should steer numbers toward the statutory envelope. On a De Castro-style agreement, the Bardal runway becomes the valuation driver, with mitigation and the employee’s job market realities doing the heavy lifting. The risk conversation is simpler when you price the words on the page, not the policy someone hoped they’d written