Case Study 1: The Long Tail That Rewrote the Form — How Asbestos and Pollution Drove the 1980s Liability Crisis and the Claims-Made Trigger

DataField.Dev

Case Study 1: The Long Tail That Rewrote the Form — How Asbestos and Pollution Drove the 1980s Liability Crisis and the Claims-Made Trigger

A real, public episode in insurance history. The events — the mid-1980s liability insurance crisis, the long-tail asbestos and environmental losses, and the 1986 revision of the standard commercial general liability form — are matters of public record. Consistent with this book's rules, this study stays qualitative: it describes what happened and why it mattered for underwriting, and it does not attach any invented loss figure, combined-ratio number, or precise statistic to these real events.

Background

For most of the twentieth century, the commercial general liability policy was an occurrence form. A business bought coverage; that coverage attached to injuries and damage that occurred while the policy was in force; and it stayed attached no matter how long afterward a claim arrived. For the everyday liability loss — the customer who falls, the product that fails and is reported within months — this worked well, and it was generous to insureds: coverage followed the moment of harm and could be relied on years later.

The occurrence form rested on a quiet assumption that the industry had not been forced to examine: that the gap between the occurrence of harm and the reporting of the claim was manageable — months, a few years at most. For a whole category of harms, that assumption turned out to be catastrophically wrong.

Two long-tail exposures broke it. The first was asbestos. Workers exposed to asbestos fibers across the middle decades of the century developed asbestos-related diseases — including cancers with latency periods that can run decades — that did not manifest, and could not be claimed, until long after the exposure. The second was environmental pollution: contamination that seeped into soil and groundwater over years or decades, discovered and litigated long after the polluting acts, and given enormous new legal force when the United States enacted its federal environmental "Superfund" liability regime (the Comprehensive Environmental Response, Compensation, and Liability Act) in 1980, which imposed broad, retroactive cleanup liability.

Both exposures shared the feature that occurrence coverage handles worst: a long, slow latency between the act and the harm, and therefore between the policy that was in force at the time of the act and the claim that arrived a generation later.

The insurance and underwriting issue

The structural problem was this. Under an occurrence form, an insurer that wrote a policy in, say, the 1950s or 1960s had accepted responsibility for harms occurring in those years — including asbestos exposures whose diseases would not manifest until the 1980s and beyond. The premium for that policy had been set with no knowledge of the coming wave; the actuaries of the day could not have reserved for losses no one anticipated. When the claims finally arrived decades later, they landed on policies written long before, at prices that never contemplated them.

This created two acute difficulties for underwriting and reserving, both of which this chapter has named:

The tail was open-ended and nearly impossible to reserve. An occurrence insurer's exposure for a given policy year does not close when the year ends; it stays open for as long as claims from that year's occurrences can still be reported. For asbestos and pollution, "as long as" meant decades. The "incurred but not reported" liability (the long gray tail an actuary must estimate, Chapter 10) was not a modest add-on; for these exposures it dwarfed the reported losses, and the uncertainty around it was enormous.
The losses were correlated across the whole industry and the whole economy. This was not a diversifiable book of independent risks; it was a systemic exposure that hit insurers, reinsurers, and whole industries at once — the failure of the independence assumption (Chapter 1) on a grand scale.

The convergence of these long-tail losses with a broader hardening of the market produced the liability insurance crisis of the mid-1980s (the hard-market episode introduced in Chapter 3). Across whole segments of commercial liability, coverage became expensive and, for some classes, hard to obtain at all. Underwriters and carriers, confronting a tail they could neither bound nor price, retrenched.

What it shows

The episode is the clearest demonstration in the book of why the occurrence-versus-claims-made distinction (§21.2) is not a technicality but the structural heart of liability underwriting:

An occurrence form transfers the latency risk to the insurer. Whatever the gap between act and claim, the occurrence insurer owns it — and for slow-developing harms, that gap can outlive the underwriting team, the pricing assumptions, and sometimes the carrier.
Loss runs lie most on exactly the risks where lying matters most. The chapter's image of the loss run as "a photograph of a parade still marching past" (§21.2) is precisely the asbestos lesson: the reported losses at any moment were a fraction of the ultimate losses, because the worst claims hadn't manifested. An underwriter who priced off reported experience was pricing off a systematically incomplete picture.
The trigger is a risk-allocation device. Whoever controls the trigger controls who bears the tail. The industry's response was to change the trigger for the exposures it could no longer absorb on an occurrence basis.

Outcome

The industry's structural response, in the public record, was twofold. First, the standard commercial general liability form was revised — the 1986 revision of the CGL is the one practitioners still reference as the modern baseline — and that revision, among other changes, introduced a claims-made version of the CGL alongside the occurrence version, and tightened the form's treatment of long-tail exposures (the broad pollution exclusion pushing environmental liability out of the standard CGL is a direct descendant of this episode).

Second, much long-tail professional, management, and pollution coverage migrated to claims-made triggers as the market standard. A claims-made form closes the tail: the insurer's exposure for a policy year is fixed once that year's claims-reporting window shuts, which is what made these exposures reservable and pricable again. The retroactive-date and tail-coverage (extended reporting period) machinery that Chapter 24 develops is the apparatus the industry built to make claims-made coverage work without leaving insureds with gaps.

The asbestos and environmental claims themselves, meanwhile, continued to develop for decades — a reminder that changing the form for new business does nothing about the tail already accepted on old occurrence policies. That accepted tail had to be run off, reserved, and paid, long after the underwriters who wrote it had retired.

Lesson for the underwriter

The transferable lessons are exactly the chapter's themes, learned by the industry the expensive way:

On a long-tail line, you are pricing a future you cannot see, and the trigger decides how much of that future you own. Know which trigger your policy uses, every time, and never let an insured or a broker operate on the wrong mental model (§21.2). For the standard CGL you will usually stay on the occurrence form — which means you are accepting the latency risk, and your pricing and reserving must respect it.
Reported losses understate ultimate losses on long-tail business — build that in. Develop the recent years upward; do not take a clean recent loss run on a products or completed-operations account as proof the account is clean (§21.4). The worst claims may simply not have arrived.
Pricing follows risk, and the discipline is hardest exactly where the feedback is slowest. The asbestos generation of underwriters did not underprice out of recklessness; they priced for a tail they did not know was coming. The modern underwriter has no such excuse for known long-tail exposures — and the temptation to underprice them in a soft market, because the losses won't surface for years, is the central discipline of a casualty desk (§21.4; Chapter 11 on rate adequacy).

For Harbor Steel, the connection is direct. Its products-completed operations exposure — load-bearing structural steel installed in other people's buildings, with slow failure modes and a litigious field — is a long-tail exposure on an occurrence form. The asbestos episode is the reason you will not price that tail off the quiet premises record, and the reason the one pending bracket claim goes in the file in red ink.

Discussion questions

Explain, in your own words, why an occurrence trigger left mid-century insurers holding asbestos claims on policies written decades earlier — and why a claims-made trigger would have prevented that specific problem (while creating others). (§21.2)
The chapter says the worst mistake in CGL underwriting is to "price the part you can see and undervalue the part you can't." How does the asbestos episode illustrate that at the level of an entire industry? (§21.1, §21.4)
Changing the CGL form in 1986 did nothing about the tail already accepted on old occurrence policies. What does this tell you about the durability of an underwriting decision on a long-tail line, compared with a short-tail property decision? (§21.2; Ch. 19)
Why is correlated, systemic long-tail loss (asbestos across an entire economy) so much harder for the insurance mechanism to absorb than a book of independent, short-tail liability claims? Connect this to the law of large numbers and the independence assumption. (Ch. 1, §21.4)
A broker today asks you to write a manufacturer's CGL with a product whose failure mode is slow and cumulative, off a three-year clean loss run, at a competitor's low rate. Using this case, write the two sentences you would put in your file explaining why you will not price it off the reported experience. (§21.2, §21.4)