Bad Bets on Additional Insured Obligations To Defend and Indemnify

Examining the consequences of failing to properly analyze coverage

March 02, 2024 Photo

Additional insured (AI) coverage often determines who is responsible for paying tort claims. However, as compelling as AI coverage is, confusion abounds among parties, insurers, coverage attorneys and courts.  

As measured by the vast number of citations nationally to the leading AI case of Burlington v. NYC Transit Auth. 29 N.Y.3d 313 (2017), the most common AI endorsements may generate the most coverage disputes. As of the time of this writing, Burlington has been cited over 700 times in reported cases and commentary across numerous state and federal jurisdictions in the last seven years.

Moreover, most coverage disputes stem from just a short phrase found in most standard insurance industry Insurance Services Office (ISO) AI endorsements.  

For example, many construction and real estate agreements’ insurance procurement provisions require downstream tenants and contractors to buy the ISO CG 20 38 04 13 AI endorsement.  

The CG 20 38 04 13 provides AI status to owners, lessees, and contractors for ongoing operations when required by contract. It states, in pertinent part, “Who is an insured is amended to include as an AI…any…person or organization you are required to add as an AI on your policy…only with respect to liability for ‘bodily injury’ and ‘property damage’…caused in whole or in part by your acts or omissions.”   

In fact, most ISO AI endorsements for the last 20 years beginning with the 07 04 forms contain similar “caused by acts or omissions” wording.

This discussion will address four of the most frequently asked questions in the order in which the wording appears in this AI endorsement, and then examine the potential consequences to claims professionals and coverage attorneys when they fail to properly analyze AI coverage.

To Whom Does “Liability” in the Endorsement Refer?  

The answer must be that the liability referenced is that of the putative additional-insured party seeking AI status. This is because the purpose of the endorsement is to extend coverage to putative AIs for liability to third party claimants albeit conditioned on a link between the loss and the named insured. Coverage for named insureds is found elsewhere in the policy, namely in the first sentence of the coverage grant of the ISO common conditions.  Accordingly, coverage for the named insured’s liability to third parties is neither granted nor referenced to in these AI endorsements.

Does the Word “Liability” Connote a Role for Negligence in Considering AI Status.  

While liability is widely associated with negligence, the endorsement does not explicitly mention the word negligence. Moreover, as Burlington and its progeny demonstrate, there is at most a miniscule role of negligence in AI analysis.

As to the AI duty to defend, negligence is irrelevant. If the pleadings allege that there is a sufficient causal nexus between the loss and the named insured, then the duty to defend the putative AI is triggered. Moreover, even if the pleadings do not reference any such nexus, in most states extrinsic evidence and facts known beyond the complaint may also be considered. 

Accordingly, where the insurer knows or should know of facts establishing a sufficient causal nexus, then the duty to defend the AI is triggered. This is because the courts have long held that the CGL policy is “litigation insurance,” meaning the policy is intended to defend against allegations whether proven or not. [See Servidone Constr. Corp. v. Security Ins. Co., 64 N.Y.2d 419 (1985)]. In addition, where the complaint alleges that the named insured was negligent, then AI should be similarly triggered since negligence is inclusive of causation and therefore, more than enough to trigger AI.     

As to the AI duty to indemnify, negligence has a minuscule role in most cases. As referenced in Burlington, ISO issued a circular when it introduced the “caused by acts or omissions” trigger indicating that their intent was to prevent AI cover where the putative AI is solely negligent. However, AI will still cover up to 99% of the putative AI’s negligence. 

Case law on AI coverage litigation overwhelming stems from construction-site accidents, and in such claims putative AI owners usually have no active role in the work and thus are seldom solely negligent. In addition, putative AI general contractors are usually able to sufficiently implicate some trade contractor to avoid sole negligence and obtain full AI defense and indemnity. Furthermore, insurers usually have enough information from the initial investigation to know there is no realistic chance of the putative AI being found solely negligent well before any judicial finding deciding liability.   In fact, as discussed below, there can be serious financial consequences when claims professionals or coverage counsel choose to ignore information in their own file to delay accepting AI cover.

What Is the Nature of the Causal Nexus Between the Named Insured and the Loss?  

While causation has vexed courts since the seminal case of Palsgraf v. Long Island R. Co., 248 N.Y. 339 (1928) nearly 100 years ago, as a practical matter an adequate causal nexus is readily known in most cases without getting to any profound legal debate on the types or nature of causation.

Typical Case: As reflected by the body of reported case law, AI often presents where an injured employee sues the owner and general contractor, which seeks AI coverage from the employer’s insurer. The putative AI’s status as the claimant’s employer alone is not enough to establish a sufficient causal nexus between the loss and the named insured. However, as a practical matter, it is often clear at the time of the tender requesting AI cover from the initial investigation that the named-insured employer was the proximate cause of the loss. 

For example, injured workers usually allege that they made prior complaints to their supervisor of the dangerous condition or methods involved in their accident. This prior notice establishes foreseeability to the named-insured employer, which is a decisive component of proximate cause both for plaintiff’s prima facie liability case and AI coverage. 

A widely used definition of “proximate cause” nationally is found in the IRMI (International Risk Management Institute) glossary which states: “As a principle of tort law, proximate cause refers to a doctrine in which a plaintiff must prove that the defendant’s motions set in motion a relatively short chain of events that could have reasonably been anticipated to lead to the plaintiff’s damages.  If the defendants’ actions were ‘proximate’ or close enough in the chain of causation to have foreseeably led to the plaintiff’s damages, then courts will impose liability. Otherwise, if the defendant’s actions set in motion a bizarre chain of events that could not have reasonably been foreseen to lead to plaintiff’s damages, courts will not impose liability.” 

Further support for the notion that foreseeability is essential to the nature of proximate cause can be found case law. In Hain v. Jamison, 28 N.Y. 524, 529 (2016), New York’s highest court stated that “the overarching principle governing proximate cause is whether a party was a substantial cause of events which produced injury, the determination of which turns on questions of foreseeability.” Accordingly, injured employees’ prior complaints to foremen often have the dual effect of establishing foreseeability, notice, and thus causation for plaintiff’s prima facie case and AI from employers insurers. 

Moreover, other actionable violations short of negligence by the named insured may be sufficient to trigger AI coverage. As repeatedly noted in Burlington, AI coverage can be triggered by the named insured’s negligence orother actionable deed.” This suggests that rebuttable adverse findings of the named insured’s violation of OSHA, rules, codes, regulations, and other laws usually issued long before any negligence finding can trigger AI.

What Effect Does the Phrase, “In Whole or in Part” Have on AI Coverage Analysis?  

While the majority and dissent in Burlington disagreed on whether there is a difference between proximate cause and “but for” causation, again, as a practical matter any difference may be academic. Arguably “in whole or in part” has no impact on causation at all.  

In many states the jury charge for proximate cause directs jurors to consider whether the defendant’s conduct was a substantial factor in bringing about the loss.  However, even in states that adopt the Restatement of Torts 2d definition in proximate cause jury instructions including California, there can be multiple substantial factors and multiple actors causing harm.  

Moreover, and counterintuitively, “substantial” need not amount to much.  For example, NY PJI 2:70 states in part, “you may find that a cause is substantial even if you assign a relatively small percentage to it.”   So, if proximate cause is established where the loss is foreseeable to a defendant, then slight foreseeability is foreseeability, proximate cause is established, and AI is triggered.

The next question may be why ISO would include superfluous language in the endorsement. It is possible that ISO attempted to change as little wording as possible when updating forms, thus failed to remove legacy wording from prior versions of endorsements.


Two of the most common mistakes downstream named insurers make are failing to appreciate that the putative AI can insure away up to 99% of its negligence and how readily evident causation is. Too often downstream insurer claims professionals and their counsel either ignore AI tenders until threatened with a declaratory judgment (DJ) coverage lawsuit or they offer an AI defense only under an ROR when they should accept AI full defense and indemnity.  

However, accepting only AI defense under an ROR is effectively making a bet unlikely to pay off that discovery and investigation will ultimately prove that the upstream owner or general contractor is solely negligent and the named insure did not cause the loss.   

Moreover, too often these bets are waged by seasoned adjusters and counsel who ignore their own experience that injured workers make prior complaints to their foremen thus, establishing foreseeability, proximate cause and full AI defense and indemnity. Often, information about prior complaints is readily available or even already in the claim file even before the AI tender is received which could prove embarrassing when revealed in a DJ action. Moreover, when named insured employers are third party defendants, juries are likely to find that they failed to properly supervise, instruct, train, or control their employees and their work. 

Moreover, the costs of improperly handling AI tenders can be substantial not just to insurers but with rising deductibles to policyholders also.  

For example, when downstream insurers fail to properly accept defense and indemnity without reservation, they often don’t just lose declaratory judgment coverage actions.  They also can lose control of costs in the underlying tort claim because they must pay multiple defense firms to litigate cross claims against each other as well as coverage counsel.  Similarly, they lose control of the litigation since finger pointing among the defendants may inure to the plaintiff’s benefit and drive up the settlement value.  And, too often, they lose sight of opportunities to resolve the underlying claim with the plaintiff’s attorney.

The concepts of causation and negligence are nuanced.  However, as this discussion has highlighted, confusion can often be avoided if the stakeholders remember that putative AI’s can contract away up to 99% of their negligence, and they fully consider the available facts and jury proclivities.  

About The Authors
Julian D. Ehrlich

Julian D. Ehrlich, Esq. is national managing director, claims for Aon Global Risk Consulting.

Sponsored Content
Daily Claims News
  Powered by Claims Pages
About The Community

CLM’s Construction Community provides a forum for construction-related claims and litigation professionals to exchange ideas and share best practices. The community identifies trends and creates needed resources to meet the needs of the industry.

Community Events
No community events