The Court of Appeal decision in Marvelous Mario’s Inc. v. St. Paul Fire and Marine Insurance Co. is a noteworthy addition to “rolling limitation period” jurisprudence.
The appellants sought coverage under the business interruption cover of a commercial insurance policy issued by the respondent. The policy contained a contractual limitation period:
ACTION: Every action or proceeding against the insurer for the recovery of any claim under or by virtue of this contract is absolutely barred unless commenced within one year next after the loss or damage occurs.
The trial judge held that the claim for business interruption losses was an ongoing claim and subject to a rolling limitation period. The appellants discovered it on a date more than a year before commencing the coverage action. However, it was subject to a rolling limitation period, a new claim accrued each day the appellant sustained a business interruption loss. Those days within a year of the action give rise to timely claims.
The Court of Appeal rejected this reasoning. First, it considered the nature of a rolling limitation period:
 The jurisprudence suggests that a rolling limitation period may apply in a breach-of-contract case in circumstances where the defendant has a recurring contractual obligation. The question is not whether the plaintiff is continuing to suffer a loss or damage, but whether the defendant has engaged in another breach of contract beyond the original breach by failing to comply with an ongoing obligation. In cases where there have been multiple breaches of ongoing obligations, it is equitable to impose a rolling limitation period.
Second, it found that the trial judge erred by considering whether the appellants were continuing to suffer damages rather than whether the respondent was breaching a recurring contractual obligation:
 In my view, where the trial judge erred was in focussing her analysis on the question of whether the appellants were continuing to suffer damages rather than on the issue of whether the respondent had a recurring contractual obligation. Unlike Pickering Square, where the tenant had a recurring obligation to occupy the premises every month during the term of the lease, the respondent was not obliged to make recurring payments. Rather, the policy covered business interruption losses and the respondent was obliged to pay those losses in their totality, subject to any limits in the policy. The fact that there was a 24-month cap on the business interruption losses does not convert the respondent’s obligation to indemnify into a recurring contractual obligation. Therefore, this was not a proper case for the application of a rolling limitation period.
 The appellants knew as of the closing date for the sale of the businesses that they no longer had the assets under their control and that consequently they would thereafter suffer business interruption losses. While the precise amount of the damages was unknown, the appellants knew at that point that they had suffered loss or damage and under the policy they were obliged to commence a claim within one year. The fact that the extent of damages may not be known with precision does not stop the commencement of the limitation period: Peixeiro v. Haberman, 1997 CanLII 325 (SCC),  3 S.C.R. 549, at para. 18. The Second Action was consequently time-barred in its entirety.
This reasoning raises some interesting issues (which, I note sheepishly, I missed entirely when writing about the trial decision).
First, a word on the “rolling limitation period”. The court’s language suggests that it’s an equitable principle: “In cases where there have been multiple breaches of ongoing obligations, it is equitable to impose a rolling limitation period”. This isn’t quite right.
Statutory limitation periods apply to causes of action. Subject to certain exceptions, one limitation period applies to one cause of action (in the case of the Limitations Act, one basic and one ultimate limitation period apply to one “claim”, which derives from a cause of action).
When there are circumstances that cause multiple causes of action to accrue periodically, one limitation period applies to each cause of action. In the case of a contractual obligation that recurs daily (the court’s example), each day the defendant fails to perform the obligation a discrete breach of contract occurs that gives rise to a discrete cause of action. Each discrete cause of action has its own limitation period.
Because these limitation periods have the same length, it will appear as if one limitation period commences anew each day, The limitation period appears to roll. This is the “rolling limitation period”.
A rolling limitation period is accordingly just convenient way to describe multiple limitation periods of the same length commencing regularly and consecutively. It’s not a special kind of limitation period that exists independently of the statutory limitation scheme to be invoked as equity requires.
On my reading, the Court’s analysis is correct. The Court found there was one cause of action that accrued on a particular date. There were no multiple causes of action, and there were accordingly no multiple limitation periods that could appear to roll.
All of this is to say that I question the value of the “rolling limitation period” in limitations analyses. What purpose does it actually serve? Instead of asking whether there is a rolling limitation period, you might just as easily, and certainly more accurately, ask whether there are recurring causes of action.
Now for the second issue: I’m doubtful cause of action accrual, or causes of action generally, had any place in the parties’ limitations analysis.
The limitation period is contractual. It reflects the parties’ agreement on when the insured can sue the insurer.
The limitation period commences on the occurrence of the “loss or damage” and bars “absolutely” an action “against the insurer for the recovery of any claim” under the policy (that is, a claim for indemnification).
The triggering event is not the accrual of a cause of action, but the occurrence of the “loss or damage” for which the insured claims indemnification. I suspect the policy provided coverage for some business interruptions that don’t arise from actionable misconduct. This means the insured could seek recovery for a claim for indemnification of a loss that has nothing to do with a cause of action.
This makes it bewildering that the parties appear to have agreed that common law discovery applied to this limitation period
As I read the provision, its language excludes discoverability of any kind. The limitation period commences on the date of an event—the occurrence of loss or damage—and expires absolutely one year later. The word “absolutely” would seem an explicit bar to discoverability or the suspension of time. If the parties intended for discoverability to apply, they could have used language like “…unless commenced within one year next after the insured could first reasonably have learned of the loss or damage”.
It’s also impossible to apply common law discovery to a limitation period that commences on the date of an event rather than on the accrual of a cause of action. Discovery is of a cause of action: the rule provides that a cause of action accrues when a plaintiff can first reasonably discover its material facts. This is why the rule doesn’t apply to statutory limitation periods that commence on the date of an event (like s. 38 of the Trustee Act) .
Further, discovery is always in relation to the cause of action that gives rise to the proceeding. In this case, the cause of action giving rise to the proceeding was the breach of the insurance agreement, and this is the cause of action to which the contractual limitation period applies. The discovery of some of other cause of action that may have resulted in the event that is the trigger of the limitation period wouldn’t in the normal course be material to the commencement of time. It’s so unusual for the limitation of one cause of action to be determined by the accrual of another cause of action that if the parties intended this to be so I’d expect explicit language to that effect in their agreement (I’ve never seen such language).
More generally, because this is a contractual limitation period, the parties ought to have grounded their positions in the language of the policy. The material question strikes me as being whether the “loss or damage” occurred within the meaning of the policy each day the business interruption lasted, or on the date of the event that caused the business interruption. This is a question of policy interpretation.
Without having read the policy, I see arguments on both sides. I think the strongest argument is the one the court accepted. The day of the peril that caused the business interruption—the transaction—caused the appellants to suffer the “loss or damage” for which they sought indemnification. That the interruption lasted days goes to quantifying the loss, not when it occurred. I would think the appellants’ notification would have been helpful; it’s easy to imagine that the appellants filed one notification in regards of the business interruption, which might have indicated they considered there to be one loss. I’m not sure why this didn’t appear to be part of the record.
The counterargument is that each day of business interruption was a new occurrence of loss. This is effectively the rolling limitation period argument, except not one based on cause of action accrual.
I welcome your thoughts on this!