Court of Appeal Says Breach of Covenant to Insure Gives Rise to Remedy In Damages

In Papapetrou v. 1054422 Ontario Ltd., the Court of Appeal dealt with a question that has hitherto received practically no attention from the courts: what are the consequences of a named insured party breaching a covenant to have another party made an additional insured under the named insured’s insurance policy? The Court held that the party in breach (the named insured) became liable in damages to the party to whom the promise was given. The measure of damages was found to be the promisee’s cost of defending an action brought against it by a third party.

In my view, there are a number of problems with the Court of Appeal’s decision. I do agree with its finding, that the breach of the covenant to insure is to render the promisor liable in damages to the promisee. If the Court meant to limit those damages to the promisee’s defence costs, then I disagree. What seems more likely is that it expected that issue to be dealt with in another court, at a later date.

The second and probably more significant difficulty that I have with this decision is the Court having construed the scope of the covenant to insure by referring to an indemnity or “hold harmless” covenant contained within the same contract. It seems to me that this approach is simply wrong and is likely to lead to a number of problems in other cases.


This dispute arose out of an accident in which the plaintiff, Papapetrou, had slipped and fallen at a building managed by The Cora Group, a defendant in the action. The owner of the building, 1054422 Ontario Limited, was also sued.

(Somewhat confusingly, the Court of Appeal’s reasons speak only of The Cora Group, even though the motion from which the appeal was taken had been brought by both The Cora Group and 1054422 Ontario Limited.)

Collingwood Landscape Inc. had contracted with either 1054422 or Cora Group (the Court of Appeal said it was the latter but the motions judge’s reasons indicate that it was the former) to provide winter maintenance and snow removal services at the building. Under the terms of the contract, Collingwood was required to “include the Owners as an additional insured [sic]” in its commercial general liability insurance policy. (The wording of the actual covenant to insure is not clear. The passage quoted in the previous sentence appears in quotations in the Court of Appeal’s reasons. How the property manager, The Cora Group, qualifies as “the Owners” is not explained, but presumably, the word was defined in the contract. The motions judge’s reasons say that the contract “also required them [Collingwood] to name the owners (1054422 Ontario Limited and The Cora Group Inc.) as additional insureds under the policy.”)In any event, both 1054422 and Cora Group were represented by the same counsel and no issue seems to have been taken as to whether the covenant to insure applied to Cora Group as well as 1054422. In the balance of this post, I will refer to the two collectively as “the owners”.

The CGL policy was also supposed to have liability limits of $2 million.

In addition, the contract contained an indemnity or “hold harmless” provision in favour of “the Owner”:

The Contractor shall indemnify and save harmless the Owner, their directors, employees, agents and representatives against all claims, losses, liabilities, demands, suits and expenses from whatever source, nature and kind in any manner based upon, incidental to or arising out of the performance or non-performance of the contract by the Contractor or any of its employees, agents and representatives.

Collingwood failed to honour the insurance covenant in the contract. Its insurance policy limits were only $1 million and neither 1054422 nor Cora Group was named as an additional insured under that policy.

The motions judge’s decision

The owners then moved for summary judgment before Madam Justice Jane A. Milanetti. They asked that the claims against them be dismissed or, in the alternative, that Collingwood be ordered to assume their defence of Papapetrou’s action.

In her reasons, Justice Milanetti found that in the statement of claim, “there are some allegations that fall within the general occupiers’ liability basket”. However, she concluded that overall, “the true nature of the claim is that the defendants were negligent in failing to maintain an ice free pedestrian stairway, and as a result the plaintiff fell and sustained injury”. As a result, Her Honour held that:

…based on the contract, and the context of the pleadings framing the action, the snow removal contractor is obligated to defend and indemnify the property owners. This remains the case despite the fact that the contractor failed to name the property owner as an insured contrary to the contract between the parties. They should not escape responsibility to defend/indemnify merely because they failed to meet their contractual responsibility.

(The order that Collingwood not only assume the owners’ defence but also indemnify them in relation to the Papapetrou claim seems to have been something that the motions judge came up with on her own. Her reasons indicate that the moving parties had only asked, as their alternative relief, that Collingwood assume their defence.)

The Court of Appeal’s reasons

Finding of duty to indemnify held to be premature

Collingwood appealed. In the Court of Appeal, it was conceded that Justice Milanetti’s order that Collingwood indemnify the owners was premature and should be set aside. (A similar order was recently made by Justice Stanley Kershman in Minto Developments Inc. v. Carlsbad Paving et al., 2012 ONSC 1574 (CanLII), a case that I discussed in a recent post on additional insureds.)

The duty to indemnify is ordinarily determined by comparing the coverage in the insurance policy with the facts that are proved at trial. But that approach is not so easily followed in the Papapetrou type of case, where the promisor (the named insured) has failed to have the promisee added to its insurance policy as an additional insured. After the trial, when the findings of fact have been made, what policy language are those factual findings to be compared to? If the promisee was never made an additional insured, what grant of coverage should the court that is considering the duty to indemnify be assuming? In my opinion, in this situation, where there is an actual policy in existence, the court should assume that the additional insured would have had the same coverage as the named insured unless the named insured can prove otherwise. Any other assumption as to the scope of the coverage for the additional insured would necessarily be speculative and in this situation, that is the fault of the named insured, the promisor who failed to have the promisee added to its policy.

So, while I do not disagree with the finding that it was premature for the motions judge to find that there was a duty to indemnify, I am concerned that the Court of Appeal has kicked this question down the road to some extent and that when the issue does fall to be considered by another court, following the trial of the underlying action, the Court’s reasons here will afford little guidance (or worse, will provide erroneous instruction) as to how the scope of the notional coverage that never came into existence should be interpreted.

The scope of the notional coverage

Having found that the order requiring Collingwood to indemnify the owners must be set aside, the Court of Appeal then considered the validity of Justice Milanetti’s order, that Collingwood assume the defence of the owner.

The respondent argued that Collingwood owed a duty to defend because of its failure to discharge the contractual obligation to have the owners named as an additional insured under its CGL policy. But the Court of Appeal rejected this argument, saying, “Collingwood’s breach of this contractual obligation does not create a duty to defend; rather, it gives rise to a remedy in damages” and that “[t]he quantum of such damages is the amount The Cora Group will be required to pay for a defence of the claims Collingwood’s insurer would have been obliged to defend on The Cora Group’s behalf had Collingwood fulfilled its contractual obligations”.

The Court went on to say, quite sensibly, that “[o]rdinarily, the scope of this obligation would be determined by the terms of the insurance contract (in particular, the additional insured endorsement)” but that that could not be done in this case because of the very fact that the coverage for the owners had not been included in the insurance contract. In this situation, it said, “the terms of the intended insurance coverage must be discerned from the insurance obligation and the indemnity provision in the service contract.”

It is the latter finding with which I most strongly take issue. I will summarize how the Court applied the principle that I have quoted in the previous paragraph and then come back to discuss its implications.

The Court’s analysis proceeded by comparing the indemnity provision in the underlying contract (quoted above) and the insurance covenant in the same contract. It held that since the indemnity or “hold harmless” obligation was limited to “claims…based upon, incidental to or arising out of the performance or non-performance of the contract by the Contractor [Collingwood]…the quantum of damages is the amount The Cora Group must pay to defend claims for bodily injury arising out of the manner in which Collingwood performed or failed to perform the service contract”. It added that “these costs will include all costs of The Cora Group’s defence of the Papapetrou action, save for any costs incurred exclusively to defend claims that do not arise from Collingwood’s performance or non-performance of the service contract”.

In order to determine the measure of damages on the part of the promisee (the party that was supposed to have been made an additional insured), the Court of Appeal’s approach was to examine the scope of the separate protection afforded to the promisee by the contract’s indemnity (“hold harmless”) provision and conclude that what has been lost by the failure to name the promisee as an additional insured was insurance liability coverage that was limited by the language of the indemnity provision.

However, it is my opinion that on this point, the Court of Appeal’s reasoning is flawed. Contracts in which covenants to insure appear typically also contain indemnity or “hold harmless” provisions, but the two are quite distinct mechanisms for contractual risk transfer. They exist independently of each other and while it might well be the intention of at least one of the contracting parties that the provisions of the two types of risk transfer mesh with each other, this is by no means necessarily the case.

In the standard reference work in this area, The Additional Insured Book (7th ed.) by Gibson, Legeros and Malecki, the following passage makes this clear:

As discussed in Chapter 4, additional insured status and indemnification under a contract’s hold harmless clause operate independently of each other. And since contractual liability insurance responds only to the named insured’s obligations under the indemnity agreement, additional insured status and contractual liability coverage are likewise independent of each other. As a result, the coverage provided to an additional insured often does not correspond with the scope of liability transferred under the indemnity agreement. In some respects, an additional insured enjoys broader coverage as an additional insured than the protection intended under the contract’s indemnity clause. [Emphasis added]

Thus, the Court of Appeal has (mistakenly, in my view), turned to the contractual indemnity provision in the parties’ contract in order to answer an entirely separate question: what the scope of additional insured protection would have been, had it been obtained. The problem is that the latter is not necessarily informed by the former.

The problem in these cases, involving breach of a covenant to insure, is that it is usually not possible to determine exactly what the coverage would have been, for the very reason that the coverage never came into existence. And frequently (or even usually), the wording of the insurance covenant in the underlying contract is too non-specific to allow a reliable inference to be drawn as to what the scope of the coverage would have been, had it been obtained.

The Court of Appeal went on in its analysis (correctly, in my view) to say that the scope of the duty to defend must be made by comparing the “true nature” of the pleaded claim with “the specific limitations of the insurance coverage at issue”. But its mistaken (in my opinion) conflation of the underlying contract’s indemnity or “hold harmless” provision and the covenant to insure led it to conclude that the “specific limitations of the insurance coverage at issue” in this case were that “the coverage would be limited to the matters relating to Collingwood’s performance or non-performance of the contract”. In my estimation, inferring that the coverage, had it been given, would have contained this limitation just because the indemnity or “hold harmless” provision in the contract contained it is entirely unfounded.

Here, the Court looked at the pleaded claims and concluded that that their “true nature” required that they be classified into three categories:

  1. Negligent maintenance due to Collingwood’s performance or non-performance of its contract;
  2. Negligent conduct on the part of The Cora Group [presumably, this should also include 1054422 Ontario Limited] extending beyond Collingwood’s obligations under the contract; and
  3. A statutory cause of action under the Occupiers’ Liability Act extending beyond those obligations delegated to Collingwood under the contract.

It held that “the duty to defend only extends to allegations that can be classified as falling under the first category of claims”. In making that finding, the Court based its conclusion on its ruling, that the scope of the coverage that would have been obtained was limited by the language of the indemnity provision in the underlying contract. For the reasons set forth above, I believe that the Court was wrong to link the two parts of the underlying contract and therefore, was wrong to find that it was only the first category of allegations in the Papapetrou claim to which the duty to defend would have applied.

For example, what if there had been no indemnity or “hold harmless” provision in the underlying contract? Does that mean that there would have been no restrictions at all on the coverage afforded by the additional insured endorsement? Or would the opposite be true, that a covenant to insure would have conferred no benefit and would have been unenforceable?

In practice, that is not at all the experience, as discussed in my previous post on this subject.

Or take another common situation: a covenant to insure and a “hold harmless” provision in the same contract, just as we have in this case. Assume that the promisee is made an additional insured under the named insured’ insurance policy (unlike this case), but that the extent of the coverage that undeniably exists is inconsistent with the protection afforded by the “hold harmless” clause. Do the two contractual provisions have to be reconciled? The Papapetrou decision strongly suggests that they do. But how? Which provision is to override the other?

It is interesting that the Court of Appeal cited no authority in support of its finding that the scope of the coverage that would have been is to be determined by reference to the indemnity or “hold harmless” provision in the underlying contract. I have not such a finding elsewhere.

The Court went on to hold that “[i]n this case, as Collingwood failed to satisfy its insurance obligation under the service contract, it is unable to demonstrate that it should escape responsibility for paying for The Cora Group’s costs of defending the action save for those costs incurred exclusively to defend uncovered claims.” I agree with this conclusion but I don’t agree with the Court’s approach to determining what are “uncovered claims”.

The Court of Appeal also ruled that in this situation, it would not be appropriate for Collingwood to simply assume the owner’s defence. Rather, it held that Collingwood must pay for the owner’s separate counsel to continue its defence, subject to the proviso that Collingwood would not be liable for “any costs incurred exclusively to defend claims that do not arise from Collingwood’s performance or non-performance of the service contract”.

Its ruling on this point flows too, from its finding that the scope of the duty to defend was a function of the wording of the “hold harmless” provision in the contract. For the reasons outlined above, I think that this conclusion was incorrect.

Borough of Wilkinsburg v. Trumbull-Denton Joint Venture

Nigel Kent of the B.C. firm Clark Wilson LLP has written about this issue (breach of covenants to insure in an additional insured context) in a 2006 paper entitled, “The CGL Policy and the Additional Insured Endorsement in Canada“. At p. 12 of that paper, under the heading “Failure to Obtain the Additional Insured Endorsement”, he considered the very situation that arose in Papapetrou and had this to say:

There are no Canadian cases dealing with a Named Insured who failed to obtain an Additional Insured Endorsement; it is thus necessary to turn to US case law to examine this issue. In Borough of Wilkinsburg v. Trumbull-Denton Joint Venture, 568 A.2d 1325 (Pa. Super. Ct. 1990), Borough and the general contractor had entered into a contract requiring the general contractor to add Borough as an Additional Insured on the general contractor’s insurance policy. After a law suit was brought against Borough, it was discovered that the general contractor had failed to name Borough as an Additional Insured. Borough looked to its own insurer, the one with which it had a policy as a Named Insured, for a defence and that insurer later brought an action against the general contractor for reimbursement. The Court held that a breach of the contract between Borough and the general contractor had occurred and the general contractor was liable for the entire amount that was awarded against Borough in the lawsuit against it. The Court did not inquire into whether Borough would have been entitled to coverage had the general contractor named Borough as an Additional Insured. This may have been because the possibilities for coverage are endless as a result of the availability of several different standard form AIEs and the option of adopting custom wording.

I have posted the Borough of Wilkinsburg case with this post. Like Mr. Kent, I too could not turn up any other case that addressed the type of situation that we have in Papapetrou.

(Until recently, that is. In one of my own files, a contractor covenanted to insure for the benefit of various other parties on a construction job. One of them was a design engineer, whom I represent. The contractor breached the covenant to insure. When the contractor was sued, it brought third party proceedings against a number of other parties, including the engineer. Our office moved to have the third party dismissed as an abuse of process but the motion was dismissed by Madam Justice Giovanna Toscano Roccamo. Her Honour’s decision was handed down prior to the Papapetrou ruling: Castonguay Construction (2000) Ltd. v. Commonwealth Plywood Co. Ltd., 2012 ONSC 3487 (CanLII). The facts were somewhat different in that case, since it was a builder’s risk policy that was the subject of the covenant to insure. However, some of the same issues arise: how does one ever determine what the scope of coverage would have been? The question was even more difficult in the Castonguay case because no policy at all ever came into existence.

Mr. Kent seems to have favoured the Borough of Wilkinsburg approach:

Where a Named Insured breaches a contract by failing to add another party as an Additional Insured to the Named Insured’s insurance policy, the Named Insured can expect to pay any judgment that is made against the person or organisation who should have been named as an Additional Insured that would have otherwise been covered by the Named Insured’s insurance policy. This result is likely to be the case unless it is clear that the insurance coverage that the Additional Insured was to have received would not have covered the claim in question, and that fact can only be clear if the contract specifies the scope of coverage that the Named Insured must obtain for the Additional Insured.

I assume that he would also agree with the Court of Appeal in the Papapetrou case, that the named insured who breaches the covenant to insure should also be required to reimburse the promisee for the cost of having to defend third party claims that would have been covered by the named insured’s policy, had the promisee been made an additional insured.

I would agree with all of this. I think that the Borough of Wilkinsburg approach is a sensible one (assuming that it also includes defence costs). But in my view, requiring the promisee to establish what the coverage would have been, had the covenant to insure been honoured, places an unreasonable burden on that party. It seems to me that the promisor (the named insured) should be responsible for the entire amount of the judgment against the promisee, together with all defence costs, except to the extent that it can prove, on a balance of probabilities, that even if the covenant to insure had been satisfied, there would have been no coverage for some or all of the claim against the promisee.

But to resolve the issue of the notional coverage as the Court of Appeal has done in Papapetrou, by looking to the language of an entirely separate covenant, the “hold harmless” clause, that may or may not even appear in the contract and that may or may not have been intended to relate to the covenant to insure strikes me as entirely the wrong approach.

However, it is the law.

This entry was posted in Uncategorized. Bookmark the permalink.