Reading the policy in light of principles of insurance law and the reasonable expectations of the parties, Montrose is entitled to access available coverage under any excess policy once it has exhausted directly underlying excess policies for the same policy period.  A rule of vertical exhaustion is appropriate.  The insured has access to any excess policy once it has exhausted other directly underlying excess policies with lower attachment points, but an insurer called upon to indemnify the insured’s loss may seek reimbursement from other insurers that issued policies covering relevant policy periods.

IT’S FUNDAMENTAL: THE DUTY TO DEFEND – The Anomaly of D.R. Horton-Tex., Ltd. v. Markel Inter’l Ins. Co.

It is well-established that the duty to defend is broader than the duty to indemnify.  Often touted as being “axiomatic”, if there is a potential for coverage under the CGL policy, there is a duty to defend a suit. The duty to indemnify, on the other hand, is narrower because it depends on the insured proving that a loss is actually covered.

Is it possible to have a duty to indemnify but not a corresponding duty to defend? Is the duty to indemnify without a corresponding duty to defend an anomaly, or at least a bit illogical?  Or is it perfectly logical in the context of an “eight-corners” state, like Texas? 

The analysis of the duty to defend begins with an examination of the facts and the policy. It is necessary to determine, however, what facts an insurer must, or can, review when determining its defense obligation in a particular jurisdiction. Is the examination of the facts limited to those stated in the lawsuit or can information extrinsic to the lawsuit be consulted?

While it is universally accepted that the duty to defend is broader than the duty to indemnify, jurisdictions differ as to what information can be used to determine that duty. For instance, in an “eight-corners” state, the decision to defend must be based on a review of the policy and the lawsuit. Use of information outside of the complaint is generally not permitted. In other states, information extrinsic to the complaint must be considered.  In some cases, this information can be used to defeat coverage as well as afford it. In other cases, the information can be used only to afford coverage.

So, while it appears that the defense duty is broad because only a potential for coverage need be established, an eight-corners state restricts the amount of information that can be examined.  Does this restriction detract from the axiom that the duty to defend is broader than the duty to indemnify when only limited information can be used?

Common rationale for rejecting the eight-corners rule can be persuasive. First, the denial of a defense based on information restricted to the lawsuit would appear to be contrary to the broad duty to defend which is universally recognized; and, second, the decision to defend, or not, would be based on the pleadings which could be limited in its recitation of the facts and be subject to “the pleading strategies, whims, and vagaries of third party claimants to control the rights of parties to an insurance contract.”[1]

On March 21, 2013, a federal district court in Texas ruled that the duty to defend and duty to indemnify are distinct duties under Texas law and that one can exist without the other.  Colony Ins. Co. v. Price, 2013 U.S. Dist. LEXIS 38952 (N.D. Tex. Mar. 21, 2013).

Citing the Supreme Court of Texas in D.R. Horton-Tex., Ltd. v. Markel Inter’l Ins. Co, 300 S.W.3d 740, 743, the Colony court stated that “although the duty to defend is ‘strictly circumscribed by the eight-corners doctrine, it is well settled that the facts actually established in the underlying suit control the duty to indemnify.’” Colony 13-14

Is the duty to defend truly broader than the duty to indemnify if determination of the duty to defend is “strictly circumscribed by the eight-corners doctrine?”  To answer this question, it is helpful to distinguish the “potential v. actual” criterion for determining duties from any constraints on the information to which this criterion can be applied.

When the duty to defend must be determined from the lawsuit only and not from extrinsic information, it is possible that no duty to defend exists given the way the lawsuit is drafted.  On the other hand, there may be a duty to indemnify because it can be determined based on information outside of the complaint.

In D.R. Horton-Texas, Ltd., the Court ruled that the defense and indemnity duties are independent of each other and that an indemnity duty can exist without a defense duty: “We hold that even if Markel has no duty to defend D.R. Horton, it may still have a duty to indemnify… That determination hinges on the facts established and the terms and conditions of the CGL policy.” D.R. Horton 744.

Hence, the anomaly…the duty to defend is broader than the duty to indemnify, but in an eight-corners jurisdiction, the information used to determine the duty to defend is limited. And while the duty to indemnify is narrower, the information used to determine that duty is broad. Since the duties exist independent of each other, it is possible to have a duty to indemnify without a duty to defend.

[1] Randy Maniloff and Jeffrey Stempel quoting from M. Mooney Corp. v. U.S. Fid. & Guar. Co., 618 A.2d 793, 797 (N.H. 1992), General Liability Insurance Coverage/Key Issues in Every State (New York, NY: Oxford University Press, Inc. 2011) 71.


On December 11, 2019, an appellate court in the State of Illinois ruled that Liberty Mutual Fire Insurance Company had no duty to defend or indemnify its insured in the underlying litigation and also had a valid right to be reimbursed for defense expenses paid pursuant to a specific endorsement in the policy. Before discussing this further, let’s go back to 1997.

In February 2013, Dr. Jerry Buss, long-time owner of pro basketball’s Los Angeles Lakers and hall of famer, died at the age of 80. Aside from a legacy of world championships, he also left one for the insurance industry in the form of a California Supreme Court case decided in 1997, Buss v. Superior Court, 16 Cal. 4th 35 (Cal. 1997), probably the leading case supporting the right of an insurer to seek reimbursement for expenses incurred in the defense of causes of action that are not potentially covered under the policy.

It is common that the defense of an insured is conducted under a reservation of rights due to coverage issues. The reservations may include the right to seek reimbursement for defense expenses paid. Because the duty to defend is very broad, and because the insurer must defend the entire action, some jurisdictions provide for the right to reimbursement of defense expenses specific (solely) to causes of action that are not potentially covered in a mixed action.

The Buss case is probably the poster child for this right to reimbursement. In the underlying case against Buss and others, H&H Sports had contracted with Buss to provide advertising and other services. A dispute between the parties evolved into a lawsuit in which 27 causes of action were asserted, including count 23 for defamation. Buss’ insurer defended the entire lawsuit on the premise that only the defamation count was potentially covered.

The central issue was whether the insurer could recover the expenses incurred in the defense of its insured that were specific to the non-covered causes of action. The Supreme Court ruled that the insurer had the right to reimbursement of expenses in the defense of specific causes of action that are not potentially covered. Furthermore, the insurer has the burden to demonstrate that the expenses related solely to the uncovered cause of action. And this burden is by a preponderance of evidence. Buss 3940

Buss, then, provided that in a “mixed” action, or an action in which some of the causes are not potentially covered while others are, the insurer can seek reimbursement for expenses paid solely to defend the causes of action that are not potentially coverage, and the insurer has the burden, by a preponderance of evidence, of making the distinction.

But as a noted commentator points out, theory and practice are significantly different, and the burden on the insurer to make that distinction may not be an easy one:

the Buss rule is more valuable in theory than practice. The court held that “an insurer is only entitled to recover those defense expenses which can be fairly and reasonably allocated solely to non-covered claims for which there never was any potential for coverage.1

1 Randy Maniloff, Reimbursement of Defense Costs, An Insurer’s Duty to Defend or Duty to Lend, FC&S Online (the National Underwriter Company 2006, Ohio) M.28-3.

Of course, while California is a bellwether state in many instances, not every jurisdiction always follows its lead. For example, in 2010, the Supreme Court of Pennsylvania ruled in Am. & Foreign Ins. Co. v. Jerry’s Sport Ctr., Inc., 606 Pa. 584 (Pa. 2010) that the right of reimbursement should be included in the insurance policy if the insurer wishes to preserve this right, as opposed to creating a new contractual obligation by way of an unilateral reservation of rights letter within which the insurer attempts to preserve its right to reimbursement and where the insured accepts the defense.2 The Court also rejected the insurer’s argument that the insured was unjustly enriched for receiving the defense:

…Royal cannot employ a reservation of rights letter to reserve a right it does not have pursuant to the contract…the policy here did not provide for a right of reimbursement of defense costs for non-covered claims. A reservation of rights letter asserts defenses and exclusions that are already set forth in the policy…We are persuaded that permitting reimbursement by reservation of rights, absent an insurance policy provision authorizing the right in the first place, is tantamount to allowing the insurer to extract a unilateral amendment to the insurance contract…A reservation of rights letter “does not relieve the insurer of the costs incurred in defending its insured where the insurer was obligated, in the first instance, to provide such a defense”…Absent the creation of a new contract through the reservation of rights letters, there is no contractual basis upon which to order Insured to reimburse Royal for defense costs. Am. & Foreign Ins. Co. 614-615

Nor are we persuaded that there are any equitable bases upon which to grant a right to reimbursement…

Insured was not unjustly enriched…Royal had not only the duty to defend, but the right to defend under the insurance contract. This arrangement benefited both parties. The duty to defend benefited Insured to protect it from the cost of defense, while the right to defend allowed Royal to control the defense to protect itself against potential indemnity exposure. Am. & Foreign Ins. Co. 615-616

2 Randy Maniloff and Jeffrey Stempel, General Liability Insurance Coverage/Key Issues in Every State (New York, NY, Oxford University Press, Inc. 2012, 2nd ed.) 164.

In Pennsylvania, then, an insurer’s right to recoupment of expenses incurred in the defense of a claim or cause of action that is not potentially covered must be provided for in the insurance policy, and cannot be preserved in a simple, unilateral reservation of rights letter extrinsic to the policy.

In Liberty Mutual Fire v. Ferrara Candy, 2019 IL App (1st) 181385-U Liberty Mutual prevailed on its duty to defend and indemnify its insured in the underlying case as well as in its pursuit of reimbursement of defense costs. Liberty initially agreed to defend subject to reservations which included its right to withdraw from the defense and to seek reimbursement of defense expenses should it be determined later that there was no coverage.

The second amended complaint prompted Liberty to assert that there was no coverage and withdrew from the defense. It subsequently filed a declaratory judgment action. As to the duty to defend, there was no coverage because no bodily injury or property damage, as defined, occurred during the policy periods. Nor was any personal and advertising injury caused by an offense during the policy periods. Any trade dress infringement occurred post-merger and post policy periods, and this was clear from the complaint. There was no duty to defend or indemnify.

The right to recoup defense expenses was also upheld given Liberty’s endorsement that it was entitled to reimbursement if it was determined that there was no duty to defend and subject to the parameters of that endorsement.

The take-away is that perhaps it is best in any state, but certainly in Pennsylvania and Illinois, to include an endorsement in the policy reserving the right to reimbursement as opposed to a unilateral reservation of rights contained in a letter.


Given the nature of latent, progressive, and cumulative injury or damage, and in the context of the definition of occurrence, it can be quite challenging to determine how many occurrences you are dealing with in a construction defect claim, providing a lot of fertile ground for controversy.  This is important because the limits of liability and deductibles or self-insured retentions are impacted by the number of occurrences. In addition, more than one occurrence can complicate an already-challenging allocation scheme (among the carriers and potentially the insured with respect to defense cost and indemnity sharing).

The “cause and effect” paradigm is useful, and, in fact, forms the basis of case law to determine how many occurrences there are in a construction defect claim.   But it must be applied with flexibility given the variety of fact situations in construction defect claims.  

The “cause test” determines the number of occurrences by focusing on the specific cause of property damage, and not on the number of effects of the cause.  A defective roof that is leaking, for example, can cause damage to attic insulation and personal property.  Using the “cause test”, there is one occurrence.  On the other hand, the “effects test” bases the number of occurrences on the number of effects.  In our example, the damaged insulation and personal property may constitute two occurrences.

Most jurisdictions look to the cause to determine number of occurrences but what constitutes a particular cause is highly fact-specific and not without controversy, the controversy existing generally in the application of the definition of occurrence to the underlying facts. (“Cause” can mean a couple of things.  The physical cause, or cause-in-fact, with a nearness, or not, in time or space to its effect, is distinguished from the legal or proximate cause.)

In the CGL policy, “occurrence” is defined as, “…an accident, including continuous or repeated exposure to substantially the same general harmful conditions.”  Some issues to consider:

  • Are the effects, e.g. property damage, the result of the same or substantially the same conditions?
  • Is the exposure of the property to these conditions repeated or continuous, and not unbroken by an intervening cause or causes?
  • Are the exposures, while substantially similar, separated by time or distance to such an extent that to conclude that there is one occurrence pushes the envelope of reason?
  • How many trades are potentially implicated as causes?
  • Doesn’t the determination of number of occurrences for insurance purposes depend on whose insurance you are evaluating, e.g. number of occurrences as to the general contractor whose scope of work is broad as opposed to a subcontractor whose scope his narrow?
  • Nature of the defect and legal theory, e.g. a product defect in windows throughout the development versus various different installation defects throughout the development.

Furthermore, there is a result-oriented consideration, the result being maximizing coverage where the claims are numerous and the damages are within the deductible, if applied on an effect basis.  Finding one occurrence with just one deductible would be advantageous to the insured.


“Trigger” is a term of art used by legal and insurance practitioners that you will not find in the CGL policy.  It is a useful term that simply describes how and when coverage is activated.  In the CGL occurrence form, property damage must occur during the policy period.

In the context of insurance claims, most of the time there is no controversy.  The cause (occurrence) and effect (property damage) are definite in both time and place, and readily apparent. It is simple to pinpoint when the property damage occurred and it is when the property damage occurs that triggers coverage, not when the work is done, for example, in a construction defect claim. 

But in the case of latent (hidden) damage, the effect may not become apparent until sometime after the work is completed so the timing of the property damage is unclear and the controversy begins. 

Furthermore, latent property damage can be characterized by a continuous and progressive process (and practically indivisible if there are multiple causes). This is the rule and not the exception in the construction defect claim world. Even experts cannot agree on when property damage occurs.

A coverage analysis begins with the requirement that property damage happen during the policy period as the result of an occurrence. The insured has the burden of demonstrating at least the potential that an occurrence caused property damage during the policy period.

The trigger of coverage analysis includes an examination of the underlying facts and the law in a particular state. Because of the challenge of determining when property damage and bodily injury occur in continuous damage/injury claims, various trigger theories have developed. The impetus for these theories actually preceded the proliferation of construction defect claims.  Asbestos and environmental claims provided fertile ground for controversies concerning when injury and damage actually occur, so these trigger theories evolved in an attempt to defuse these controversies. 

Brief descriptions of the major trigger theories follow.  There are also variations of these theories.  But, while useful, these theories are not definitive and do not relieve the claim professional from examining the defect/damage process.  Whether a particular trigger theory should apply in the first place depends on the underlying facts and the prevailing case law.

To facilitate an understanding of these theories, our focus is on the insured’s work and one identifiable cause, or faulty work. The typical construction project and the claims that arise, however, are not so neat and clean.  The insured is frequently not alone on the project so an analysis is complicated by the existence of multiple parties and multiple causes.

The Theories

Before tackling the trigger theories, when analyzing the underlying facts, it is useful to create a chronology of construction events.  When property damage occurs may not be known exactly, but a good chronology can help narrow down when property damage may have occurred.  Specific dates and activities can not only define the parameters of when property damage may have occurred, they can also rule out when property damage definitely did not occur.  A chronology generally, however, cannot determine how much property damage occurred at a particular point in time, only when property damage may have occurred.

For example, the date of the insured’s construction contract, the dates of work, including completion and acceptance dates, the certificate of occupancy date, etc., should generally be available.  These dates, and the dates insurance coverage are available to the insured, can aid in the establishment of a chronology that narrows down the triggered periods.  The following simple chronology illustrates the point:

Date                         Activity/Event

5/1/04                      Date of subcontract

8/1/04                      Date insured’s work completed/accepted

11/1/04                    Date of substantial completion

12/1/04                    Certificate of Occupancy issued

6/1/05 – 6/1/06    Insurer A policy

6/1/06                      Insurer B policy begins

12/1/06                    First documented complaints of water intrusion

6/1/07                      Insurer B policy ends

6/1/08                      Lawsuit filed


The manifestation trigger holds that the policy in effect at the time the property damage becomes apparent, either subjectively or objectively, is activated. A subjective manifestation occurs when the property damage actually becomes apparent and is discovered. On the other hand, an objective manifestation is one in which the property damage should have become apparent and discovered.  For example, an inadequately designed foundation on expansive soil, i.e. soil that expands and contracts, will start to fail before the symptoms become apparent and eventually are discovered.  In either case only one policy is triggered. 


The exposure trigger requires that all policies in effect during the period that the property is exposed to the harmful, damage-causing agent be activated. Consequently, more than one policy can be triggered. For example, in the case of a defective foundation, all policies are triggered from the moment of installation through the period during which the office building is exposed to the defective foundation and is damaged. 

Continuous Injury or Damage

The continuous injury trigger, the broadest trigger, begins with the time of the defective work through to when the work manifests or is discovered, and possibly beyond.  More than one policy is triggered.


Injury-in-fact stays true to the policy requirement that only property damage that occurs during the policy period is covered, whether detectable or not.  In the context of latent, continuous and progressive damage, more than one policy can be triggered.  

When Does a Construction Defect Cause Property Damage?

Until property damage is discovered, has it occurred?  A manifestation or discovery jurisdiction would say no.  

The CGL policy requires that property damage occur during the policy period.  It does not say that the property damage has to be seen, heard, or discovered.  So, it must be conceded that property damage can begin before it is discovered or becomes manifest.  But what does the law of a particular state say?  In other words, while it is easy to conceptualize that property damage can begin the moment of the creation of the defect and continue undetected, it is necessary to determine when the law says the property damage occurred.

One cannot forget logic and common sense, however, when analyzing when property damage occurred.  For example, suppose a roof is installed but flashing at the chimney is not.  Three months later water damage is discovered.  The damage occurred before discovery but did it occur during the entire three months?  Suppose it did not rain for the first two months

Care must be taken to avoid being so mired in theories and legal concepts that logic and common sense are trumped by an automatic, mechanistic approach to coverage for a claim.

Certificate of Insurance Disclaimer…Not Bulletproof

In construction defect claims, certificates of insurance (“COI”) routinely accompany additional insured tenders. During the normal course of business, at the time a construction contract is made, the general contractor will typically require a COI from the subcontractor and usually the sub’s insurance agent provides it. The COI includes policy information important to the general contractor (risk management) such as named insured, names of insurers, types of coverage, policy periods, limits, and other basic information.

In other words, the COI provides the general contractor with confirmation that the subcontractor is insured, and what the terms of the policy are. The COI only provides information at the point in time that it is issued. It is a snapshot. But whether this information is accurate is another matter.  It is not infrequent that an insurance agent will provide information on a COI and attachments that does not comport with the policy nor the insurer’s records, a disappointing revelation to the certificate holder.

Furthermore, the COI does not amend or change the policy in any way, explicitly stating so in the form of a disclaimer. The certificate is not the contract, the policy is. However, the disclaimer is proving not to be bulletproof, at least in Washington.

In T-Mobile USA Inc. v. Selective Insurance Co. of America, Slip. Op. No. 96500-5, 2019 WL 5076647 (Wash. Oct. 10, 2019), the Washington Supreme Court, on a certified question from the Ninth Circuit, held that despite the disclaimer, Selective’s agent, under apparent authority, bound Selective to cover T-Mobile USA as an additional insured even though T-Mobile USA was not a signatory to the construction contract that contained the additional insured obligation.

Selective’s named insured contracted with T-Mobile Northeast, not T-Mobile USA, for the construction of a cell tower and the former was required to be an additional insured on the policy.  Initially, T-Mobile USA was named as the defendant in a subsequent construction defect lawsuit. Since T-Mobile was not a signatory to the construction contract and, therefore, not an automatic additional insured, Selective denied coverage. 

Significantly, the COI identified “T-Mobile USA Inc., its Subsidiaries and Affiliates” as the additional insureds. The agent, Selective’s “authorized representative”, repeated this practice over a period of seven years. T-Mobile USA approved the form of the policy and was aware that the COI identified it as the additional insured notwithstanding it not being a signatory to the subcontract (Selective should have been aware of it as well).

While the disclaimer was clear, the representation by the agent with apparent authority from Selective trumped the disclaimer, requiring Selective to cover T-Mobile USA as an additional insured even though it did not make the contract with the named insured which would have qualified it as an automatic additional insured, nor was it specifically scheduled as an additional insured.

The take-away is that a certificate of insurance disclaimer is not bulletproof. Accuracy matters.