Is Uncertainty Over the Meaning of “Occurrence” Susceptible to a Drafting Solution?

Can astute contract drafting can forestall all contract disputes? No, it cannot. Most contract disputes, sure. But not all.

For example, vagueness is an essential tool for the contract drafter, as often the future is too uncertain to allow you to be precise. But being vague leaves room for future dispute.

And often parties rationally elect to punt on a given issue—they elect not to address it in the contract, on the grounds that the likelihood of that issue arising post-closing isn’t great enough to warrant expending the resources and goodwill that would be required to address it during negotiations. If the issue does ultimately arise, you may well have a dispute on your hands.

The extent to which a drafter can or cannot forestall dispute came to mind when recently I encountered a court opinion addressing whether a given misfortune constituted one “occurrence,” or more than one, for purposes of a given insurance policy. This issue crops up all the time, I thought to myself. And indeed it does. By way of background, here are some extracts from Philip L. Bruner and Patrick J. O’Connor, Jr., Bruner and O’Connor on Construction Law § 11:78:

A separate, but somewhat related, issue arises in situations involving multiple injuries or injury that takes [sic] place over a number of years; namely, is one occurrence or multiple occurrences at work in causing the loss? This issue not only affects the amount of coverage available to an insured, but also the number of deductibles to be applied. There are three basic approaches to answering this question: (1) the “cause” test; (2) the “unfortunate events” test; and (3) the “effects” test.

The majority approach is represented by the “cause” test. Under this analysis, the issue is resolved by reference to the underlying cause or causes of the injury or damage, rather than by reference to the number of resulting injuries or damage claims. Under this test, generally speaking, when a single proximate, continuing, and uninterrupted cause results in injury to persons or property over a period of years, there is only one occurrence.

…

The “effects” tests has been adopted by a minority of jurisdictions and holds that where several injuries result from a single underlying cause, each injury to person or property is a separate occurrence.

…

When the World Trade Centers were destroyed by terrorists, the issue arose as to whether this loss constituted one “occurrence” (i.e., the result of one coordinated act of terrorism) or two occurrences (i.e., each destroyed building constituted a separate occurrence). In S. R. International Business Ins. Co., Ltd. v. World Trade Center, the court determined that the destruction of the buildings was as single occurrence.

So here’s my question: Couldn’t insurance companies draft policies—and an insurance policy is just another kind of contract—in such a way as to bring clarity to the meaning of “occurrence”?

Normally I think through such issues on my own. But I know next to nothing about insurance, and I’m not inclined to give myself a crash course in the subject, so for my own edification I consulted two people who have experience with this issue. I offer you their thoughts, in case this is an issue of any interest to you.

One person I asked is Adam Scales. He’s an associate professor of law at Washington and Lee University School of Law, and he’s written extensively about insurance law. Here’s what he had to say:

Let me first suggest that you double-check the actual outcome in the World Trade Center cases. In my view, the decision you cite is correct—I think it’s really one occurrence. But if I recall correctly, the combination of jury verdicts and settlement pressure resulted in Silverstein getting a lot of benefit of the doubt—something like 1.5 occurrences worth of coverage. I could be wrong, but that’s worth checking.

It is true that in Silverstein, the occurrence provision had not been specified. I don’t think that changes the correct answer, though it did give Silverstein some room for maneuver. There are three basic problems that I think make it unlikely this issue can be satisfactorily resolved through better drafting.

First, liability contracts are obligations that are performed across a very long time. So long, in fact, that the meaning and scope of covered events (what constitute “damages”, for example) will change during the decades intervening between drafting and claiming. In fact, the term “occurrence” itself was an effort to draft around the limitations of the word “accident,” which had been used for over a century before the 1966 introduction of “occurrence-based” policies. Introduction of “occurrence” was successful, in that it broadened the scope of coverage, just as had been intended. But from another perspective, it failed, as it did not adequately (from the insurers’ view) discipline that coverage. In 1966, toxic torts like asbestos and DES were motes on the horizon, and few foresaw the significant impact they would have on insurers’ losses.

The other limitation involves the issue of foreseeing specific types of accidents in advance. Really, this area of the law has seen everything under the sun, and I think insurers might be reticent to adopt a very clear, limiting provision that might drive some insureds away. To give an example, if a parking garage dents your car, chances are it’s only going to happen now and then. A policy that clearly separates Monday’s dents from Tuesday’s dents will not alarm the policyholder. On the other hand, there have been some spectacular cases involving snowplows that damaged hundreds of cars in a single stroke. A plow operator might, or might not, be alive to the possibility that if it gets into trouble, it could be by the handful.

And note just how difficult it would be to provide clear-cut guidelines as to what constitutes an occurrence even in the WTC cases—where now we know what happened. The fact remains that different elements of that transaction can seem quite relevant, though they point in different directions. I think of it as a single occurrence, because it was one “plan.” But, really—would it matter if each plane was operated by independent “cells” acting at the behest of a terrorist leader somewhere? Would anyone think differently about what happened if a single plane managed to hit both towers? Even if it were possible to describe these different possibilities, one would have to know a lot about the world to predict whether that description would be “pro-policyholder” or “pro-insurer”. Or even, “clearer.”

My suspicion is that insurers are basically comfortable with the bargain they have struck, and have no assurance that future courts would not undo any more finely-wrought revisions they might introduce. This is actually quite common in insurance, as insurers are the ultimate repeat players: they may lose in court sometimes, but they can usually get it back come renewal time. And, by the way, insurers are doing very well in court these days, so that is another reason to let sleeping dogs lie.

I also consulted Scott Godes, counsel in the insurance-coverage practice group of Dickstein Shapiro LLP, working out of their Washington, D.C. office. Here’s what he had to say:

Ken, you raise some excellent points regarding the drafting of insurance policies. The meaning of “occurrence” is a question that has been contested for some time in courts across the United States, with questions of whether potential or actual underlying liability against a policyholder is considered an occurrence, and, if so, just how many occurrences are there under one or multiple insurance policies. Adding to the complexity, the question has been answered in multiple ways by state and federal courts (not to mention arbitrators) across the country.

The term “occurrence” replaced the operative term “accident” that had been in place in the 1955 and earlier standard form comprehensive general liability (CGL) policy form. That change was made in the 1966 CGL policy form; some policyholders had their policies revised to include “occurrence” rather than “accident” prior to the 1966 revision.

Although the term was designed to be a clarification of coverage, it comes as no surprise to someone who represents policyholders when claims have been denied that insurance companies would have courts believe that instead, “occurrence” was designed to support coverage denials or limitations. Insurance companies also are happy to argue conflicting interpretations of “occurrence,” depending on which interpretation will mean less coverage for the policyholder in the dispute at issue.

Let me provide one example. In Appalachian Insurance Co. v. General Electric Co. in New York State court, multiple insurers were involved, including CNA insurers and a One Beacon insurer. Many of the insurers argued in General Electric that under decades of CGL insurance policies, the asbestos claims against General Electric (GE) consisted of multiple occurrences. Other insurers did not join the briefing, but were happy to obtain the result sought by the moving parties. The insurers argued that because there was an extensive time frame in which there was exposure to asbestos, and there were disparate locations and circumstances of exposure to asbestos, including the unique design and manufacture of each of the GE turbines to which exposure was alleged, “occurrence” meant each claim pending against GE. Such a finding would lead to a convenient result for the insurers. The insurers’ policies did not incept in an amount less than $5 million per occurrence in the underlying policies, and there were no aggregate limits for those lower layer policies. Accordingly, if every claim was a separate occurrence, the likelihood of GE’s reaching any of their policies was very low. GE and the insurers took the case all the way up to the New York Court of Appeals, and the insurers won their argument for multiple occurrences.

At about the same time, another asbestos insurance coverage case was pending in New York State court while the General Electric appeal was winding its way up through the appellate courts. That case, Continental Casualty Co. v. Employers Insurance Co. of Wausau (aka “Keasbey”), involved CNA and One Beacon as insurers who had sold primary policies and a site-specific wrap-up policy. CNA asked for rulings as to the application of decades of CGL insurance for thousands of asbestos claims. Sound familiar? The arguments sounded familiar as well, but made in reverse. The insurers argued in Keasbey for a single occurrence result, notwithstanding decades of alleged exposure to asbestos by the former insulation contractor at multiple locations and in varying fashions.

Why would the insurers take a completely contrary position? Perhaps they recognized that the language they drafted was not so clear as to require just a single consistent result? Or perhaps it was because the situation was the mirror image of General Electric … but with the insurers in this instance having issued primary insurance coverage that had no aggregate limit for certain asbestos claims? CNA, the plaintiff in Keasbey, had sold Keasbey decades of primary coverage. One Beacon had sold a primary wrap-up policy for a location where it was alleged that Keasbey performed operations. If every claim was a separate occurrence under Keasbey’s policies, without an aggregate limit, there would be a lot of insurance available for the claims against Keasbey. The trial court’s opinion was an interesting read, as it found that under CNA’s policies, every claim was a separate occurrence, and under One Beacon’s policies, all claims at a particular location consisted of a single occurrence. (Keasbey since has been reversed on other grounds, and efforts to seek leave for appeal to the Court of Appeals may still be pending at this time.)

To me, those facts suggest two things. First, “occurrence” is a term that is not plain on its face and is subject to more than one interpretation when the insurers who use the term argue that it means contradictory things. Second, that even if “occurrence” were susceptible to a better drafting solution, it is not clear that the insurance companies would want a solution that could foreclose later arguments to support their coverage positions.

I thank Adam and Scott for their input. I’ve concluded that this issue isn’t conducive to a quick fix, for two reasons. One is that insurance companies are satisfied with the current state of affairs. The second is that the number of possible fact patterns for any given event are innumerable, so deciding which facts are consistent with one occurrence and which are consistent with more than one occurrence can only be an exercise in vagueness.