Abigail Kai sits down with Simon Glairy, a veteran of insurance coverage and Insurtech who has spent decades untangling legacy umbrella forms, aggregate limits, and environmental liabilities. In this conversation, Simon unpacks what the Ninth Circuit’s April 23, 2026 decision really does to negotiating leverage, how to dissect “where applicable” in 1960s–70s forms, and why internal insurer memos can make or break a case. He walks through step-by-step tests for determining whether a general aggregate applies when property damage isn’t named, maps the difference between a single occurrence and multiple occurrences across decades of airport contamination, and frames arguments around “separately” and “where applicable” without overreaching. From governance controls and auditing legacy files to reserves for exposures that could crest $162 million, Simon offers granular guidance for claims teams, brokers, and coverage counsel facing long-tail environmental cleanup costs and multi-year policy stacks.
What’s the practical significance of the Ninth Circuit’s April 23, 2026 ruling on aggregate limits; how does it change negotiating leverage for policyholders and insurers; and which claims handlers or coverage counsel actions should change first?
Practically, the ruling takes the safety net out from under carriers who assumed a quiet, universal $9 million annual aggregate across all coverages in vintage umbrellas. By holding that the aggregate does not apply to property damage where the policy names only products and occupational disease, the court keeps the door open to exposure reaching $162 million across three policies—versus the $81 million cap the insurer urged. That delta moves leverage to policyholders at mediation: instead of arguing over one $9 million bucket, you’re now discussing per-occurrence towers that can reset annually across 1966–1975. Claims handlers should immediately stop defaulting to a “general aggregate” assumption, re-read Declarations and Limit of Liability provisions for “where applicable,” and preserve every contemporaneous note that shows good-faith interpretive steps. Coverage counsel should also recalibrate reservation letters to acknowledge ambiguity, document extrinsic evidence searches, and avoid categorical statements that a general aggregate applies unless the policy actually says so.
The policies had $9 million per occurrence and a “where applicable” aggregate. How should adjusters parse “where applicable” in legacy forms; what red flags in Declarations or Limit of Liability sections should they document; and what contemporaneous notes should they keep?
Start with the plain sequence: if “where applicable” appears on the Declarations and the Limit of Liability then names only products liability and personal injury by occupational disease, you have a strong textual cue that the aggregate is compartmentalized. Red flags include any aggregate reference that’s not mirrored by a corresponding named coverage category, or that appears only on the Dec page without a matching operative clause. Another red flag is when the Limit of Liability says the aggregate applies “separately in respect of” named categories but never mentions property damage—silence can be dispositive. In your notes, record the exact page and line where “where applicable” appears, capture screenshots, and memorialize discussions with underwriting about whether a general aggregate was ever intended; note dates like 2012, 2014, and any reinsurer communications mentioning no general aggregate. Those notes become your shield if litigation later interrogates why you paid past $9 million on a single occurrence or why you didn’t cap annual payments.
When property damage isn’t expressly named in the aggregate clause, how should claims teams decide whether a general aggregate exists; what tests would you apply step-by-step; and what metrics would you track to quantify potential exposure?
I use a four-step funnel. First, textual naming test: is property damage explicitly tied to the aggregate? If not, presume no general aggregate absent contrary language. Second, structural coherence: does the policy use “separately” to demarcate aggregates for products and occupational disease only, suggesting a limited, not general, aggregate design? Third, extrinsic evidence: pull historical market practice and any contemporaneous memos; 1986 analyses and commentary often confirm that pre-1986 general aggregates were not standard for non-products claims. Fourth, course of performance: if the carrier’s 2012 or 2014 documents say no general aggregate applied, that weighs heavily. For metrics, track the number of distinct occurrences identified per policy year, per-occurrence spend against the $9 million cap, cumulative paid and outstanding across 1966–1975, and sensitivity scenarios that scale from $81 million to $162 million depending on occurrence counts.
The airport contamination spanned decades with multiple industrial activities. How would you analyze “occurrence” versus “occurrences” in that setting; what facts meaningfully separate events; and how do you build a defensible occurrence map that survives scrutiny?
I start with causation clustering: group releases by discrete industrial operations—World War II dismantling and melting, Vietnam-era napalm and bombs production, and separate storage or disposal practices. Temporal separation matters; activities in the 1940s, the 1960s, and early 1970s are natural fault lines. Operational modality also matters: a continuous discharge from a degreasing line is different from episodic drum dumping; separate sources can be separate occurrences. Build the map with layers: a base chronological timeline (e.g., 1966–1975 policy window), a spatial layer showing the 280 soil borings and seventy-five groundwater monitoring wells, and a source layer tying each plume to tenant conduct. Annotate each cluster with sampling dates and cost milestones to show how one “occurrence” drove a specific tranche of spend versus another, so that each identified occurrence can bear its own $9 million per-occurrence limit.
The court weighed “separately” against “where applicable.” How should coverage lawyers argue each term without overreaching; what drafting history matters most; and how can they use examples from products liability and occupational disease aggregates to frame property damage outcomes?
For “separately,” argue it as a term of addition within named categories, not as a stealth gateway to a universal aggregate—frame it as “two separate buckets were created because they needed separate caps.” For “where applicable,” emphasize that it conditions the aggregate’s very existence; if the policy doesn’t say it applies to property damage, it doesn’t apply there. The most relevant drafting history is the market’s pre-1986 practice of attaching aggregates to products/completed operations and occupational disease while leaving other coverage parts uncapped, later flagged by industry commentators and a 1986 GAO report as a driver of open-ended risk. Use those examples to show that the policy intentionally named the only areas where aggregates were “applicable,” so reading a general aggregate into property damage would rewrite the bargain.
Internal insurer documents stated no general aggregate applied. What governance controls should carriers implement to align internal memos, reinsurer reports, and litigation positions; how would you audit legacy files; and what training reduces the risk of damaging admissions?
Carriers need a single source of truth protocol: before any reinsurer report or authority memo goes out, run a pre-clearance checklist confirming policy language, including quotes of “where applicable” and the specific categories named. Require matter-level alignment meetings so claims, legal, and reinsurance are reading the same pages—literally. For audits, inventory all legacy files that mention aggregates, pull 2012–2014-era documents where staff may have written “no general aggregate,” and tag them with a privilege and accuracy review. Training should include case studies from this ruling: show how four internal documents—payment authority, file memo, reinsurer report, and loss-run summary—became Exhibit A for ambiguity. Teach staff to document interpretive uncertainty rather than declaring absolutes, and to preserve context so a court sees diligence, not contradiction.
Discovery included an admission later withdrawn. What’s your playbook for managing and, if needed, retracting mistaken admissions; how do you minimize prejudice; and what documentation helps a court accept withdrawal without undermining credibility?
Move swiftly and transparently. File a motion that acknowledges the gravity of the aggregate issue and explains the precise misunderstanding—cite the policy’s Declarations, the “where applicable” qualifier, and the Limit of Liability phraseology. Offer the court a clean evidentiary record: attach the original admission, the corrected interpretation, and a timeline of discovery showing prompt correction once the error surfaced. To minimize prejudice, tender supplemental discovery, agree to targeted depositions, and, if needed, stipulate to cost-shifting on discrete re-opened issues. Credibility improves when your internal review memos contemporaneously show that multiple interpretations were considered and that your team acted in good faith rather than playing hide-the-ball.
Standard CGL forms in the 1960s–70s often lacked a general aggregate outside products/completed operations. How should litigants use historical market practice and GAO/industry reports; what experts are most persuasive; and how do you keep that evidence tied to the specific wording at issue?
Use market practice to corroborate—not replace—the text. Lay the foundation with industry commentary and a 1986 GAO report explaining why pre-1986 policies frequently had aggregates only for products/completed operations and occupational disease. The best experts are former underwriters or form drafters from that era who can explain, in plain language, how “separately in respect of” was used and why “where applicable” signaled conditional application. Keep the tether tight: every time an expert references market practice, bring it back to these policies’ exact wording—Demonstrate that property damage is not named, while products and occupational disease are, which mirrors the market’s pattern. That blend of history and text resonates with judges who are wary of turning custom into contract.
Garamendi (2005) was distinguished on timing and who benefited. How should practitioners assess whether a later judicial gloss can apply to older policies; what retroactivity factors matter; and how do you argue insured-friendly canons without inviting inconsistent outcomes?
Start with chronology: if the policy predates the decision by decades—as these 1960s–70s policies did relative to 2005—courts are rightly reluctant to retroactively import a gloss the parties never saw. Next, examine the beneficiary shift: if applying the prior gloss would now favor the insurer where it once favored the insured, that’s a red flag under California’s pro-policyholder construction rules. Retroactivity factors include parties’ reasonable expectations at the time of contracting and whether the earlier decision addressed the exact phrases at issue—here, “where applicable” wasn’t considered in the earlier case. To keep outcomes consistent, frame the insured-friendly canon as a tiebreaker once ambiguity is established, not as a shortcut; show your work through text and extrinsic evidence before you ask the court to resolve the tie for the policyholder.
The ruling favors resolving ambiguity for the insured. What concrete tools should insurers use to reduce ambiguity in renewals and endorsements; which phrases would you retire or redefine; and how should brokers counsel clients reviewing vintage excess and umbrella programs?
Insurers should deploy annotated endorsements that explicitly state whether a general aggregate applies to property damage and bodily injury outside products/completed operations, in bold, not buried. Retire or redefine “where applicable” unless the policy then expressly lists every coverage part to which the aggregate applies; otherwise you’re inviting the same fight. Add a matrix in the Declarations mapping each coverage category to “per occurrence,” “aggregate,” or “none,” with the $9 million-style figures clearly slotted. Brokers should tell clients with 1966–1975 stacks to inventory forms and endorsements, flag any policy where only products and occupational disease are named in the aggregate clause, and model two scenarios—$81 million with a general aggregate versus $162 million without—so management sees the swing and pushes for clarifying endorsements at renewal.
Environmental cleanup costs are ongoing and escalating. How do you structure reserves and reinsurance notices when exposure could reach or exceed $162 million; what milestones trigger reserve adjustments; and how do you communicate uncertainty to boards?
Build a tiered reserve: baseline per-occurrence at $9 million, then scenario bands that stack occurrence counts per policy year to reach the $162 million outer edge. Trigger reserve adjustments at factual milestones—new plume confirmation from additional borings, expansion of the seventy-five-well network, regulatory orders like the 1990 directive ratcheting scope, and major spend events such as mass drum removal phases. Issue early and periodic reinsurer notices that quote the exact aggregate language and explicitly state the competing interpretations; mirror that candor in board decks. To boards, use a heat map: show low, medium, and high ranges based on one, several, or many occurrences, and disclose the litigation posture candidly—remember, in 2012 and 2014 the internal documents leaned “no general aggregate,” which is part of today’s risk profile.
Multiple policy periods (1966–1975) are implicated. How would you approach allocation across years and occurrences; what settlement structures (e.g., banded layers, occurrence-by-occurrence caps) work best; and how do you preserve subrogation or contribution rights?
I’d run parallel allocations. Pro rata by time-on-the-risk across 1966–1975 for investigation and monitoring costs that don’t tie cleanly to a single event; and occurrence-based allocation for discrete release clusters tied to a tenant’s operations. In settlement, banded layers work: carve the first $9 million for Occurrence A, the next $9 million for Occurrence B, and so on, with reopener clauses if new occurrences are found. Occurrence-by-occurrence caps let each side manage risk while preserving the argument that additional occurrences could reset limits. Reserve contribution and subrogation rights in the agreement text and exchange tolling agreements with co-carriers and former site operators so future allocations and recoveries aren’t time-barred.
An amicus from a religious organization signaled broader policyholder interest. What other sectors likely hold similar vintage policies; how should they inventory, authenticate, and reconstruct missing wording; and what early steps create negotiating leverage before suing?
Expect municipalities, universities, hospitals, utilities, and legacy manufacturers to hold similar 1960s–70s umbrellas. Start with a forensic inventory: pull binders, loss runs, reinsurer reports, and broker files; the latter two often repeat key language verbatim. Authenticate with contemporaneous evidence—premium audits, endorsements, or even claims correspondence describing the aggregate as “where applicable” and naming only products and occupational disease. Where wording is missing, triangulate using sister policies from adjacent years or the same program. For leverage, prepare a detailed chronology, quantify the $81 million versus $162 million scenarios, and present the carrier with internal memos—if any—acknowledging no general aggregate. Showing you can tell the story crisply often opens settlement talks without filing.
If you were advising a county facing long-tail environmental liabilities today, how would you triage coverage recovery; what documents would you chase in the first 30 days; and how would you position for mediation while discovery on aggregates is still developing?
Triage begins with policy archaeology and occurrence mapping. In the first 30 days, chase Declarations, Limit of Liability pages, endorsements, broker placement files, reinsurer bordereaux, and any 2012–2014-era internal memos or loss runs stating “no general aggregate.” Pull regulatory records—like the 1990 cleanup order—and field logs for the 280 borings and seventy-five wells to anchor your occurrence clusters. For mediation, present two spreadsheets: one allocating costs per occurrence with $9 million caps, the other showing total exposure if multiple occurrences span 1966–1975. Acknowledge ambiguity but emphasize the ruling’s directive to resolve doubts for the insured; that tone—confident but not overreaching—builds credibility and momentum.
What’s the practical significance of the Ninth Circuit’s April 23, 2026 ruling on aggregate limits; how does it change negotiating leverage for policyholders and insurers; and which claims handlers or coverage counsel actions should change first?
The ruling redraws the battlefield. Instead of treating “aggregate” as a catch-all ceiling, parties must read what’s actually named—products and occupational disease—before they cap property damage. Policyholders now credibly model exposure up to $162 million across three policies, not just the $81 million the carrier touted, which shifts tone and tenor in settlement conferences. Claims and coverage teams should stop sending letters that proclaim a general aggregate unless the policy says so; instead, cite the “where applicable” qualifier, acknowledge ambiguity, and invite a joint extrinsic evidence review.
What is your forecast for legacy umbrella policy aggregate limit litigation?
Expect a wave of re-opened files and fresh suits as policyholders revisit 1966–1975 programs and beyond, armed with this ruling’s blueprint. Courts will likely continue to split the baby on “separately” versus “where applicable,” but when silence persists on property damage, ambiguity will be resolved for the insured. I foresee more settlements structured around occurrence-by-occurrence bands, with careful reservations on yet-to-be-mapped sources. For readers wondering what this means for them: if your vintage umbrella names aggregates only for products and occupational disease, pull the file now—your path to expanding recoveries from $81 million toward $162 million may be hiding in plain sight.
