Corrosion Exclusion - McCague Borlack LLP

Factual background

A leak of heavy water coolant containing a radioactive element occurred at the Nuclear Research Universal Reactor (referred to in the reasons for judgment as the NRU) in Chalk River, Ontario. The leak was discovered during the course of a generalized power shutdown in Eastern Ontario and Quebec which automatically triggered a shutdown of the NRU. Although the leak was small, the NRU was shut down, ultimately for a period of 15 months, for defueling, drainage of the heavy water, identification, investigation, and repair of the leak location and other pre-leak corroded areas, and to satisfy the regulator's protocols. The plaintiffs were insureds under an all-risk policy and claimed to have lost approximately $121 million, mostly due to loss of sales of medical isotopes and other material that could not be produced during that period of shutdown. There was coverage under the policy for this type of claim in the amount of $25 million.

...claimed to have lost approximately $121 million, mostly due to loss of sales...

The leak consisted of a single, isolated, small penetration in the wall of the calandria, a vessel or tank which contained the heavy water, resulting in the water entering what was known as the J-rod annulus. The leak was localized, unexpected, and was the result of an accelerated process. There were a further nine sites of similar localized, unexpected, and accelerated corrosion which had not yet penetrated the calandria wall. A finding of fact was made that the leak, and the other nine corrosion sites which had not yet progressed to the stage where there was a leak, were due to corrosion, which in turn had been caused by the unknown and unanticipated presence of a contaminant/aggressive agent, probably chlorine.

Interestingly, there was generalized corrosion of the outside walls of the calandria, but that corrosion had been ongoing for several decades (1974 to 2009) and was the result of nitric acid in the J-rod annulus that was created by water leaking from the reflector. The operator of the NRU (AECL) had known about the presence of that generalized corrosion for many years, although it was not considered to be dangerous or yet to require remedial action.

The court drew a clear distinction between the two types of corrosion.

The localized corrosion which caused the penetration of the calandria wall (and the nine other pre-leak sites), had been quick-acting and was not known to be present. This was classified and treated as fortuitous corrosion. It was found to be unanticipated/unexpected, unpredictable, and accidental corrosion.²
The generalized corrosion, on the other hand, was classified as non-fortuitous corrosion, because not only was it anticipated/expected and normal, it was actually known to be present and was monitored.

The insurer acknowledged that the corrosion which caused the leak was fortuitous, in that it was unique, unexplained, unexpected, and unknown.

The corrosion exclusion

The exclusion provided as follows:

This Policy excludes the following, but, if physical damage not excluded by this Policy results, then only that resulting damage is insured:

3) deterioration, depletion, rust, corrosion...

The court's interpretation of the exclusion

The court started from the premise that “[The insurer] must show that an exclusion clearly and unambiguously excludes coverage.”³

The word “corrosion” was not defined in the policy.

The word “corrosion” was not defined in the policy. The parties agreed that it was to be interpreted in its normal, ordinary, everyday meaning, rather than in accordance with any scientific definition. The court arrived at the following definition:

I conclude that the definition of corrosion for the Policy in this case, having regard to the Policy considered as a whole, the factual context and the reasonable expectation of the parties is: “The anticipated and predictable process of corroding, esp. of a rusting metal.” “Corrode is defined as ‘wear away, esp. by chemical action.'”⁴

Is corrosion no longer corrosion when it was unanticipated and unpredictable? Did it become some different type of chemical process, something other than corrosion, because it occurred in that manner? Obviously not. Corrosion is not “the anticipated and predictable process of corroding”; it is “the process of corroding”, which is precisely what occurred in MDS.

All parties, including the court, were in agreement that the factual cause of the loss was corrosion, whether or not that corrosion was anticipated, expected, or predictable. This is reflected in the following findings of fact:

The leak was caused by pitting corrosion.⁵

Classic pitting corrosion, precipitated by the presence of a contaminant/aggressive agent, probably chlorine, caused the leak. The leak was not caused by nitric acid. It was a case of an unanticipated fortuitous pitting corrosion.⁶

While there was no doubt that corrosion was the factual cause of the leak, the exclusion was held to be inapplicable because its language was found to be ambiguous:

Simply because the exclusion clause says “corrosion”... it does not mean all forms of corrosion, however caused, are excluded.⁷

The court rejected the insurer's position that the exclusion applied to all corrosion, whether fortuitous or non-fortuitous in nature, and held that the exclusion applied only to loss or damage caused by non-fortuitous corrosion:⁸

In the context of this Policy, the evidence in this case and the reasonable expectations of the parties, I find that the exclusion applies only to non-fortuitous anticipated corrosion.⁹

The court held that the leak had been due to “fortuitous, unanticipated and unpredictable corrosion”, with the result that “the corrosion exclusion does not apply”.¹⁰

Was there truly an ambiguity in the exclusion?

As stated in an oft-quoted decision,

The ambiguity principle...resolves conflicts between two reasonable but differing interpretations of the policy. The court should not strain to create an ambiguity where none exists.¹¹

The court should not strain to create an ambiguity where none exists.

In similar comments, ambiguity is present where words are capable of two or more meanings,¹² or where there are two equally reasonable interpretations of the word(s) in question.¹³

Just as there was no ambiguity in the words “identified driver”,¹⁴ there is no ambiguity in the word “corrosion”. While the physical state or condition known as corrosion may have different causes, the end result in each case is that physical state or condition. That end result – corrosion – does not have two reasonable but differing interpretations, is not capable of two or more meanings, and is not amenable to two equally reasonable interpretations.

Applying the language of Brissette Estate, “there is nothing ambiguous about the wording of the [exclusion].”¹⁵ Coverage is plainly excluded where the loss or damage is caused by corrosion. Both the parties and the court had no difficulty acknowledging that the loss had been caused by corrosion. In Chilton, the driver was “undeniably unidentified”;¹⁶in MDS, the loss was undeniably caused by corrosion. The insurer, in other words, clearly and unambiguously showed that the loss had been due to corrosion.

It was held that the insurer ought to have shown, but did not, that the loss was due to a non-fortuitous cause; more particularly, it failed to show that the loss was due to non-fortuitous corrosion. Upon what rationale or basis, however, did the insurer have an obligation to show that not only was the loss caused by corrosion, but it was caused by non-fortuitous corrosion?

The exclusion refers simply, and only, to “corrosion”; it does not contain the additional qualification or limitation that the corrosion must have been a non-fortuitous corrosion. Adding that qualification involved the impermissible reading-in of language that was not there, thereby artificially creating the ambiguity upon which the decision rested.

The policy provided coverage for the peril of “riot”...

The decision in Ford Motor Co. of Canada Ltd. v Prudential Assurance Co. Ltd.¹⁷ is both illuminating and governing on this point. That case involved a riot at the insured's plant. The employees left and the plant was shut down. Power to the plant was shut off and winter weather caused serious damage to the machinery. The policy provided coverage for the peril of “riot”, but it excluded coverage for cessation of work, business interruption, and change in temperature.

The following comments were made:

I turn now to the meaning to be given to the causes enumerated in cl. 6(c) [the exclusions clause]. The learned trial judge held that there was no “cessation of work or interruption to process or business operations” as contemplated by the clause, because, in his opinion, these conditions were brought about by the compulsion of the riot. It is quite clear that the plant protection men were available for work and would have entered the plant and saved the damage, had it not been for the display of force by mass picketing that prevented their entry. It seems to me that this limitation upon the application of these causes is to be rejected for two reasons. The first is that as a matter of construction it is impossible to read into the exclusionary clause any such limitation...

The learned trial judge also imposed a limitation upon the operation of the clause “change in temperature”. He rejected the appellant's submission that the cause was too vague and uncertain to have any operation, but he adopted the same principle in dealing with this cause as with the two other causes in 6(c). Anything attributable to the riot was not within this cause. He therefore limited its operation to change in atmospheric temperatures. I think that there is the same error here as there is in the limitation of the other two causes and that the Court of Appeal was correct in rejecting the limitations imposed by the learned trial judge upon any of these 6(c) causes.¹⁸

Had the MDS approach been taken in Ford Motor, the “change in temperature” exclusion would have been found to be ambiguous because “change in temperature” could mean either change in atmospheric temperatures or change in temperature due to other causes. Unlike what was done in MDS, however, the SCC did not look behind the plain meaning of the language of the exclusion. They did not circumvent or abandon that plain meaning, as was done in MDS, by finding ambiguity through the fact that a change in temperature can be due to different causes (just as corrosion can be the result of different causes). Instead, the court applied the clear and unqualified meaning of the actual language of the exclusion and held that the approach taken by the trial judge constituted an impermissible judicial imposition of a limitation on the plain meaning of that language. Each of the two other exclusions in that case (cessation of work and business interruption) could equally have been said to have more than one underlying cause, and therefore be ambiguous, but again, the court did not take that approach.

The trial judge in Ford Motor imposed limitations on the broad but plain language of the exclusions so as to make them accord with, and not limit the scope of, the language of the insuring provisions. The trial judge in MDS did much the same thing. She placed a limitation on the broad but plain language of the exclusion by making it apply only in those cases where the corrosion was non-fortuitous in nature, apparently so as to accord with the insuring requirement that the loss or damage be fortuitously caused.

Regardless, however, of the reason why the trial judge in MDS imposed a limitation on the corrosion exclusion, the highlighted statement reproduced above from the Ford Motor decision is a complete answer: “[A]s a matter of construction it is impossible to read into the exclusionary clause any such limitation”.

A similar refusal to unilaterally insert limiting words into the language of an exclusion clause was made in an earlier (also appellate) decision:¹⁹

I am unable to adopt the construction of the contract that the learned trial judge placed upon it, and am of the opinion that the exclusion clause (g) should be given effect according to its terms and without the limiting words that the trial judge would read into it.

The following passages from two other appellate decisions apply the same principle:

The focus of insurance policy interpretation should first and foremost be on the language of the policy at issue...I would construe the definition of “property damage”, according to the plain language of the definition, to include damage to any tangible property. I do not agree with Lombard that the damage must be to third-party property. There is no such restriction in the definition.²⁰

Reading words into a provision that is unambiguous in its scope is not justified.²¹

The exclusion in this case similarly contained “no such restriction” – i.e. that the corrosion be non-fortuitous in nature. Reading those words into the exclusion, which is “unambiguous in its scope” (in that it unambiguously applies to loss or damage caused by corrosion), is “not justified”. Ford Motor held that the judicial imposition of a limitation on the applicability of clear and unqualified language in an exclusion clause is impermissible. In MDS, the trial judge's approach involved circular reasoning through the creation of, and subsequent reliance on, an ambiguity that was not otherwise there.

The addition of that second precondition constitutes a rewrite of the policy.

The appropriate question is whether the loss was caused by the excluded peril of corrosion (which indisputably was the case) not whether the corrosion which caused the loss was itself fortuitous or non-fortuitous in nature or origin. The exclusion requires only that the loss be caused by corrosion, not that the corrosion, in turn, was non-fortuitous. The addition of that second precondition constitutes a rewrite of the policy.²² It adds words by making “corrosion” to read “non-fortuitous corrosion”.

Stated differently, and as described above, the state or condition known as corrosion is the end result of an underlying process which can be either fortuitous or non-fortuitous in nature/origin. Regardless of the nature or origin of the underlying process, that end result (i.e. corrosion) is a standalone peril. The exclusion in this policy expressly referenced the end result – and the end result only – not the underlying cause of that end result. The loss was caused by that end result – the peril of corrosion – not by the process that led to the corrosion. Ambiguity as to the underlying cause is not ambiguity as to the end result. The court conflated the issues of cause of corrosion and corrosion as cause of the loss.

The bottom line is that the court doubled the number of conditions for the applicability of the exclusion. Not only was it necessary for the insurer to show that the loss was caused by the peril of corrosion, but it was also required to show that the corrosion was non-fortuitous in origin/nature. There is no basis in the language of the exclusion for the addition of that second condition.

The court's interpretation rendered the exclusion meaningless

One of the arguments made by the insurer's counsel was summarized as follows:

The Insurer also argues that the statement that the corrosion exclusion applies only to non-fortuitous events is nonsensical. All-risk policies only insure fortuitous events of physical loss or damage. If physical loss or damage is not fortuitous, it does not fall within the basic coverage grant. There is, therefore, no need to exclude it and the exclusion would be redundant.²³

While that argument was explicitly rejected,²⁴ no explanation or rationale was given for the rejection.²⁵

The insurer's argument that the limitation read into the corrosion exclusion would have the effect of making the exclusion redundant reflects the following comment made in Ford Motor:

These are very serious limitations and come close to destroying any efficacious power in the exclusionary clause.²⁶

Furthermore, the decision would have far-reaching consequences. Following are some commonly excluded perils in homeowner's policies:

War or insurrection; nuclear incident; water, flood, windstorm or hail; earthquake or landslide; vandalism or malicious acts; inherent vice or latent defect; wet, dry rot, fungi or spores; contamination or pollution; settling, expansion, contraction, bulging, buckling or cracking; the cost of making good faulty materials, workmanship or design. Rust and corrosion are on this list as well.

... perils almost always arise or materialize, from the insured's standpoint, fortuitously...

All of those perils almost always arise or materialize, from the insured's standpoint, fortuitously, although there could be circumstances in which they might or would be anticipated or expected and therefore would be considered to have arisen non-fortuitously. Were the MDS decision to stand, these exclusions would almost never be applicable because the insurer would be unable to show that they had been anticipated or expected by the insured and therefore were non-fortuitous. That result would hold true in other types of policies as well.

Reasonable expectations

There are numerous references in the reasons for reasonable expectations. For example:

The interpretation of the Policy should be consistent with the reasonable expectations of the parties, guided by the words of the Policy.²⁷

It is, however, only where the policy language is ambiguous that general rules of contract construction, including the rule regarding reasonable expectations, are employed for the purpose of resolving that ambiguity.²⁸ If I am correct in saying that there is no ambiguity in the corrosion exclusion, then there can be no recourse to reasonable expectations as a prop to support the position adopted by the court in MDS.

The comment may also be made that it is difficult to imagine how it could have been within the reasonable expectation not only of the insurer, but of the insured as well, that the corrosion exclusion would apply only in those circumstances where there would not be coverage to begin with.

A complication

There is a complicating factor in the MDS decision. The insurer, perhaps for business relationship purposes, apparently conceded, to some extent at least, that the exclusion would not apply in circumstances where the corrosion was fortuitous in nature,²⁹although the insurer's final position at trial was that the exclusion applied to loss caused by corrosion of any nature or origin, including fortuitous corrosion.³⁰ While any such concession would be relevant and perhaps even controlling in the circumstances of a particular case (subject to issues such as whether the insured provided consideration in return for a binding agreement, and whether waiver and/or estoppel would preclude a withdrawal of any concession that was made), it has no impact on the more general issues of law that are discussed above, and no comment will, therefore, be made on a factor which was a complication only in the circumstances of this particular case.

See the follow up article - Corrosion Exclusion II - Resulting Physical Damage: An exception to the exclusion in case: MDS Inc. v Factory Mutual Insurance

2020 ONSC 1924
"I conclude that the unanticipated pitting corrosion was caused by the presence of an aggressive agent, probably chlorine, that was not supposed to be in the water that leaked from the reflector to the J-rod annulus”: Reasons for Judgment, para. 417.
Quoting from Progressive Homes Ltd. v Lombard General Insurance Co. of Canada 2010 SCC 33 at para. 51.
MDS, at para. 320.
MDS, at para. 399.
MDS, at paras. 403-04 and 413.
MDS, at para. 273.
MDS, at para. 316.
MDS, at para. 317.
MDS, at para. 417.
Chilton v Co-operators Insurance Co. (1997) 32 O.R. (3d) 161, C.A. at para. 26.
Brissette Estate v Westbury Life Insurance Co. [1992] 3 S.C.R. 87 at para. 4.
Lombard Canada Ltd. v Zurich Insurance Co. 2010 ONCA 292 at para. 33.
Chilton, at para. 26.
Brissette Estate,at para. 6.
Chilton, at para. 27.
[1959] S.C.R. 539.
Ford Motor, at paras. 15-16 (emphasis added).
J.R. Mooney & Company v Pearl Assurance Co. Ltd. [1943] O.R. 634, C.A. at para. 10.
Progressive Homes, at paras. 35-36 (emphasis added).
Sam's Auto Wrecking Co. Ltd. v Lombard General Insurance Co. of Canada 2013 ONCA 186 at para. 43 (emphasis added).
“These terms cannot simply be rewritten under the guise of interpretation”: Brissette Estate,at para. 6.
MDS, at para. 284.
MDS, at para. 285,
There is no shortage of authority for the principle that a contract provision should not be interpreted so as to (a) render it meaningless: Bruinsma v Cresswell 2013 ONCA 111 at para. 47; Rankin Construction Inc. v Ontario 2014 ONCA 636 at para. 58 (b) effectively read it out of the contract: RBC Travel Insurance v Aviva Canada Ltd. (2006) 82 O.R. (3d) 490, C.A. at paras. 11 and 41, or (c) render it devoid of content or superfluous: Venture Capital USA Inc. v Yorkton Securities Inc. (2000) 75 O.R. (3d) 325, C.A. at para. 28; Kentucky Fried Chicken Canada v Scotts Food Services Inc. (1998) 41 B.L.R. (2d) 42, Ont. C.A. at para.35.
Ford Motor, at para. 7.
MDS, at para. 406.
Ledcor Construction Ltd. v Northbridge Indemnity Insurance Co. 2016 SCC 37 at para. 50. This is the very authority cited in MDS in support of the statement reproduced directly above.
MDS, at paras. 198, 255, 257-70.
MDS, at para. 316.