chevron-down Created with Sketch Beta.


A Policyholder’s Self-Insured Retention Obligations under a Pollution Legal Liability Policy

Adrienne Nicole Kitchen and Bevin A. Carroll


  • Arguments made by policyholders and insurers in the context of environmental claims may vary widely.
  • Whether there is an SIR, and, if so, the amount of that SIR contributes to that.
  • The effect of the number of pollution Incidents on the number of SIRs.
A Policyholder’s Self-Insured Retention Obligations under a Pollution Legal Liability Policy
Bilanol via Getty Images

Pollution policies protect against onsite and offsite liability for bodily injury or property damage, including cleanup costs, triggered by a pollution incident.

Pollution legal liability (PLL) policies typically define “pollution incidents” as the physical release of pollutants into the environment. But where the U.S. Environmental Protection Agency (EPA) or a state environmental authority initiates an investigation, that investigation may reveal numerous separate pollution incidents, one pollution incident, or no pollution incidents. Often there may be no clarity on whether or how pollution was released until the investigation is complete and a determination is made about a legal requirement for the policyholder to remediate. Such investigations can take years.

Under a PLL policy, the insurer may have a duty to defend the policyholder against environmental claims. Typically, the policyholder is responsible for satisfying self-insured retentions (SIRs) that apply on a “per pollution incident” basis before the insurer’s duties are triggered under the policy. However, the cost of the policyholder’s remedial investigation alone frequently exhausts the policy’s per incident limits. This creates a conundrum. The policyholder and insurer both need to know the number of pollution incidents to determine both the policyholder’s SIR obligations and the available limits of coverage. But how can they make this determination when the number of incidents is unknown until the investigation is complete? Is it based on the number of alleged pollution incidents under investigation? Or the number of allegations in the administrative order on consent, if one exists? Or a ruling entered by a court in a related coverage action? Also, what if the investigation finds no pollution incidents? Does that mean there is no coverage and the policyholder has to reimburse defense costs provided by the insurer?

In this article, we initially explain how courts determine the number of occurrences under a commercial general liability (CGL) policy because it may be instructive to our understanding of how courts determine the number of pollution incidents under a PLL policy. We next consider the effect that the number of pollution incidents has on the availability of coverage and, more specifically, on the number of SIRs that must be satisfied by the policyholder. Finally, we consider some practical issues that may arise between policyholders and insurers in the environmental pollution context.

How Courts Determine the Number of Occurrences in CGL Policies

A CGL policy’s occurrence language may be helpful in analyzing the number of pollution incidents under a PLL policy. Extensive case law addresses the issue of the number of occurrences, and the analysis can be complex. “Occurrence” is defined by earlier CGL policies as “one happening or series of happenings arising out of or due to one event taking place during the term of this contract.” Later CGL policies define “occurrence” as follows:

(1) an accident, or (2) event or continuous or repeated exposure to conditions which result in bodily injury, personal injury, death or physical damage to or destruction of tangible property, including the loss of use. All damages arising out of such exposure to substantially the same general conditions shall be considered as arising out of one occurrence.

Courts have interpreted the meaning of “occurrence” to include a series of exposures, none of which is sufficient in itself to cause injury or damage but which, when aggregated over a sufficient length of time, culminate in an injurious result said to be caused by accident. Both definitions, specifically “a series of happenings” and “continuous or repeated exposure,” lead courts to make a causation inquiry. Some courts look at proximate cause, some at immediate cause, others at the cause that sets everything in motion. The common theme is that courts must determine where to cut off the chain of events leading up to the occurrence.

Generally, occurrence language is broad—more so than that defining the number of pollution incidents under a PLL policy.

Courts have come to differing conclusions regarding the number of occurrences in very similar sets of circumstances. Case law construing the number of occurrences is fact-specific and result-oriented and is inconsistent even within individual states. The majority of courts use the “cause” test (or the closely related “unfortunate events” test) to determine the number of occurrences. The cause test looks to the cause of the damage or injury.

The Cause Test and Its Application

A majority of jurisdictions apply the “cause” test and look to the cause of the damage or injury rather than the number of injuries or claims. Under the “cause” test, “where a single event, process or condition results in injuries, it will be deemed a single occurrence even though the injuries may be widespread in both time and place and may affect a multitude of individuals.”

How courts apply the cause test varies. There are two basic variations: the proximate cause theory and the liability event theory.

The proximate cause theory and the cause test. The proximate cause theory “focus[es] . . . on the uninterrupted nature and closeness in time and location between the event and its consequent injuries. Timing and location are therefore critical factors in assessing whether there is a single occurrence or multiple occurrences.” For that reason, courts have found a single occurrence when “there was but one proximate, uninterrupted, and continuing cause which resulted in all of the injuries and damage.” For example, the court in Plastics Engineering Co. v. Liberty Mutual Insurance Co. found that, where the injury was caused by “repeated exposure to asbestos-containing products,” there were multiple occurrences “because each individual’s injury stem[med] from his or her repeated exposure to asbestos-containing products.”

The liability event theory and the cause test. The “liability event theory” looks to the immediate cause—the event or events that gave rise to liability—to determine how many occurrences there were. For example, Michigan Chemical Corp. v. Home Assurance Co. involved multiple shipments of toxic flame retardant mislabeled as animal feed supplement. The court remanded the case for determination of the number of shipments to find the number of occurrences. The court reasoned that “[t]he shipment of the substance constituted the act from which liability arose. Other shipments, if any took place, created additional exposure to liability and therefore were separate occurrences. In such a situation, there would not have been one uninterrupted and continuing cause, but several distinct acts from which liability would have resulted.” Under the proximate cause theory of the “cause” test, this case could have come out the other way because there was a single “cause” of all liability—the mislabeling.

How Courts Determine the Number of Pollution Incidents in PLL Policies

In contrast to occurrence language, a PLL policy’s language addressing the number of pollution incidents focuses on the number of discharges, dispersals, or release events that may have resulted in pollution. “Pollution incident” is frequently defined as “the same, related or continuous pollution condition.” “Pollution condition” is defined as follows:

The discharge, dispersal, release or escape . . . of any solid, liquid, gaseous or thermal irritant or contaminant, including, but not limited to, smoke, vapors, soot, fumes, acids, alkalis, toxic chemicals, [or] hazardous substances . . . into or upon land, or any structure on the land, the atmosphere or any watercourse or body of water, including groundwater.

“Discharge, dispersal, release or escape,” in turn, generally refer “to damage or injury caused by improper disposal or containment of hazardous waste.” In interpreting the meaning of “discharge, dispersal, release or escape,” courts have found as follows:

A “discharge” is defined as “a flowing or issuing out.” To “disperse” is defined as “to cause to breakup and go in different ways”; “to cause to become spread widely.” A “release” is defined as “the act of liberating or freeing: discharge from restraint.” An “escape” is defined as an “evasion of or deliverance from what confines, limits, or holds.”

Courts consider the number of causes. As in the occurrence analysis, courts have used the “cause” test to determine the number of pollution incidents in PLL policies. Courts look to various factors to determine the number of causes, such as whether there were separate spills, the types of pollutants, the number of operating units, whether there were discrete areas of contamination, disposal of specific chemicals, shipments between sites, EPA treatment, and whether there was common ownership or operation. Courts also look to the mechanism leading to the discharge. For example, dumping potentially hazardous materials at a landfill or lagoon is not the same mechanism as a failure of the landfill or lagoon to contain those materials.

In Pennzoil-Quaker State Co. v. American International Specialty Lines Insurance Co., the court stated that the CGL policy’s “substantially the same conditions” language compares with the PLL policy’s “same, related or continuous” language. Pennzoil relied on various cases analyzing the number of occurrences under a CGL policy in order to determine whether there was more than one pollution incident under the PLL policy at issue. Pennzoil involved multiple lawsuits stemming from pollution at one site. In five underlying suits, the plaintiffs alleged Pennzoil released various pollutants into the air and groundwater surrounding its refinery. The insurer alleged there were four separate pollution incidents at issue in the suits, including two separate fires and explosions, long-term air emissions, and long-term groundwater contamination. The parties agreed that the pollution conditions alleged in the suits were neither the same nor continuous. Thus, the case turned on whether the pollution conditions were related.

Pennzoil argued that all five underlying lawsuits involved losses from the same, related, or continuous pollution conditions. Pennzoil claimed the pollution conditions had a logical or causal connection because “they focus[ed] on the alleged release of pollutants over time from Pennzoil’s refinery in Shreveport, Louisiana into the surrounding air and groundwater.” Pennzoil further argued that the releases originated from the same refinery, affected the same group of people, contained benzene, caused exposure to airborne and waterborne contaminants, and were caused by a failure to maintain equipment or properly train employees. The insurer countered that the allegation that the damages were caused by a failure to inspect, test, maintain, and repair is a legal theory and, as such, could not serve as the basis for concluding that the pollution conditions were related.

The court applied the cause test to determine whether the insured satisfied its deductible obligations relating to a single pollution incident, thus triggering the insurer’s duty to defend. It relied on cases involving general liability policies that defined “occurrence” to include “continuous or repeated exposure to substantially the same conditions” or “related conditions.” The court concluded that, “[w]hen there is more than one immediate cause of events giving rise to an insured’s liability in an underlying lawsuit, courts have rejected the argument that there is a single ‘occurrence’ based on continuous ‘exposure’ to the insured’s alleged negligence.”

Ultimately, the court held that Pennzoil failed to meet its deductible obligations. A single deductible “applie[d] to losses or costs arising from ‘the same, related, or continuous’ Pollution Conditions.” Because the policy did not define “related,” the court used the ordinary meaning of the word, a logical or causal connection. The underlying suits alleged distinct emissions, releases, and causes of different pollutants spanning nearly two years and were not related. Hence, the court held that Pennzoil had to satisfy multiple deductibles necessary to trigger the duty to defend.

Court determinations under CGL policies may help anticipate determinations under PPL policies. The reasoning of court decisions discussing the number of occurrences within the context of a CGL policy can assist our understanding of how courts would determine the number of pollution incidents under a PLL policy. In Endicott Johnson Corp. v. Liberty Mutual Insurance Co., the court looked at both the number of sites and the number of sources to determine the number of occurrences. There, the policyholder manufacturer sent waste products to two different sites—a landfill and a barrel recycling facility. The court found there were two occurrences because repeated dumping at two sites was “repeated exposure to substantially the same general conditions.” However, all of the deliveries to each site constituted one occurrence in part because the plaintiff “was not ordered by EPA to clean up the waste sites based on the specific number of deliveries it made or based on differences in the substances dumped. Rather, [the policyholder’s] liability ha[d] one cause: it disposed of tons of waste at [the two waste sites] over a twenty-year-period.” The court reasoned that while many instances of property damage resulted, “they all derived from one ‘occurrence’ at each site.”

Similarly, in Nicor, Inc. v. Associated Electric & Gas Insurance Services Ltd., 195 homes suffered mercury spills between 1961 and 1978, when Nicor employees removed old regulators containing mercury. Individually, investigation and remediation costs for each home were less than the applicable SIR. Nicor asserted that its liability at the 195 homes arose from one occurrence, with one applicable SIR. The insurers contended that Nicor’s liability arose from 195 separate occurrences, subject to 195 separate SIRs. The court reasoned that “liability was incurred only when mercury happened to spill,” and found there were separate occurrences because there was no common cause, there was no overall scheme linking the spills, and different workers spilled the mercury at different times. The court held that “the 195 spills were not ‘one happening’” or one event and did not involve continuous or repeated exposure, but instead were “sporadic, not part of an interrupted process, and in no instance . . . was any customer’s home ever subject to more than one exposure to mercury contamination” and, for that reason, could not have been repeated. The court recognized that this decision meant that Nicor could collect nothing from the insurers, but it noted that insurance policies do not guarantee “indemnification for every dollar of every loss.” It pointed out that SIRs are a way “a policyholder may elect to absorb part of the risk for any covered loss in exchange for reduced premiums” and nothing in the policies or circumstances surrounding their purchase suggests that the parties contemplated situations like this one would be consolidated to determine the applicable SIR.

Keyspan New England v. Hanover Insurance Co. involved a manufactured gas plant complex that consisted of six different industrial operations designed to produce coke by-products as well as two plants designed to further refine the coke by-products. The court found that claims arising out of contamination of waters and lands surrounding the complex arose from a single occurrence, because the policyholder “conducted interrelated operations throughout the complex.” While certain facilities generated different types of waste in different parts of the complex, “all of the contaminants resulted ultimately from the processing of tars generated by the coke works, and all of them contributed to the pollution of the nearby waters. . . . and [the policyholder] failed to remove the contamination. That single pattern of conduct is the basis of its liability and, under the policy definition, constitutes one occurrence.” Essentially, the court found that there was one “cause”—one single pattern of conduct—resulting in contamination.

In Certain Underwriters at Lloyd’s v. Southern Natural Gas, the policyholder operated 14,000 miles of pipeline running from Texas to Georgia. The pipeline operations involved various features, including compressor stations. Environmental testing revealed polychlorinated biphenyl (PCB) contamination at various compressor stations due to the use of a synthetic lubricant. The court held that the policyholder’s claim for indemnification of the cost to remediate the PCB contamination arose from a single occurrence. The court reasoned that the policyholder “expressly asserted that there was one continuous exposure over the course of the policy periods to the same PCBs emanating from the same synthetic lubricant” and “undertook a single project to investigate and remediate the damage it discovered throughout its integrated system of compressor stations.” “In short, [the policyholder] has asserted the unitary nature of its pipeline operation, including the unitary nature of the type and location of contamination at all of the compressor stations and has expressly asserted in the trial court that its remedial activities at its compressor stations were likewise unitary.”

The Effect of the Number of Pollution Incidents on the Number of SIRs

Insurers and policyholders alike face issues determining the number of SIRs. One common point of contention is whether the issue of the number of pollution incidents is ripe for adjudication prior to a final decision by the EPA. Insurers may assert that the number of pollution incidents cannot be determined until the investigation is complete, while policyholders may respond that the investigation need not be complete to determine the number of alleged pollution incidents under investigation. Apart from the issue of ripeness, a determination of the number of pollution incidents may be a point of contention because it will affect the number of SIRs that the policyholder is responsible for before the insurer’s duties are triggered under the policy.

What policyholders and insurers typically want may be turned upside-down when SIRs are involved. Policyholders generally advocate for multiple pollution incidents in order to maximize the available limits of coverage. But if the policy includes an SIR, the policyholder is more likely to argue that diverse losses arose out of one single pollution incident. Whether an insurer advocates for one or more pollution incidents may be determined by the facts of the case, the law at issue, and where the insurer sits in the policyholder’s coverage profile. Insurers may advocate for one pollution incident in order to minimize the available limits of coverage. If a policy includes an SIR, then insurers could argue that the loss arose out of multiple pollution incidents.

1. Just the facts. Policyholders and insurers alike should look to the facts. Several questions may play an important role in the determination of the number of pollution incidents.

  • Were multiple products produced?
  • Were multiple manufacturing processes involved?
  • Did the manufacturing process change over time?
  • Did the contamination arise from the same cause or the same location?
  • Are different types of pollutants involved?
  • Did the contamination occur at different areas?
  • Does the EPA investigation involve multiple operable units—distinct areas of the site separated by geography, specific site problems, or areas requiring a specific action?
  • Did contamination occur at different times?
  • Was the pollution ongoing or periodic?
  • If there were separate injuries, how closely were they linked in time and space?

2. Solely attempting to maximize or minimize coverage is not always the best strategy. If multiple insurers are involved, due to multiple policy periods, excess, umbrella, or reinsurance policies, the various insurers may want different outcomes regarding the number of pollution incidents and applicable SIRs. This is particularly true in jurisdictions using vertical exhaustion, where umbrella and excess insurers may prefer a single pollution incident to lower limits and the primary insurers may prefer multiple incidents to force the policyholder to foot more of the bill due to multiple SIRs.

Whether a policyholder wants one or multiple pollution incidents may depend not only on the applicable SIRs but also on several other factors. One factor, for a high-cost remediation, is that it may be more prudent to pay a second SIR in order to trigger additional available limits than to hope the total site costs stay low and incur significant uninsured costs by failing to trigger the additional available limits. Another factor is the availability of umbrella and excess coverage in the policyholder’s coverage profile and the potential for exhaustion of underlying limits, including whether the jurisdiction uses vertical or horizontal exhaustion.

The insurer’s duty to defend may become a duty to pay investigation costs as incurred. PLL policies frequently define “Loss” as “costs, charges and expenses incurred in the defense, investigation or adjustment of Claims” where “Claims” include “a written demand received by the Insured alleging liability or responsibility and seeking a remedy on the part of the Insured for Loss” covered under the policy. With respect to these “Claims,” “the Company shall pay all reasonable expenses incurred by the Insured at the Company’s request to assist it in the investigation or defense of the Claim.”

When an insurer cannot defend due to a conflict of interests, its duty to defend converts to a duty to pay defense costs. In such cases, the duty to pay defense costs requires payment as incurred, not at the end of the case.

In the environmental context, an insurer may argue that a policyholder’s investigation costs do not constitute defense costs and, thus, it has no obligation to advance any investigation costs until the EPA’s investigation concludes. Conversely, policyholders may argue that the insurer should pay investigation costs as incurred because investigation costs do constitute defense costs, and allegations trigger the duty to pay defense costs as incurred. This issue is often an area of contention in environmental coverage actions, as there is case law supporting both arguments.


The number of pollution incidents is a complex issue that is fact-dependent. Arguments made by policyholders and insurers in the context of environmental claims may vary widely based on the facts, the amount and type (primary, excess, umbrella) of coverage available, whether there is an SIR, and, if so, the amount of that SIR. Because the available limits and SIR obligations are inextricably linked to the number of pollution incidents, policyholders and insurers frequently must undertake an investigation of a claim. In some instances, they may, as a practical matter, reach an agreement regarding these issues at least until the underlying case is resolved.

Disclaimer: The views and opinions expressed in this article do not necessarily reflect those of the authors or the companies for which they work.