Case Study 2: The Additional-Insured Battle — How a Routine Endorsement Drained a Subcontractor's Limit

This is a labeled composite. Unlike Case Study 1, which describes a documented public episode, this case is a constructed teaching scenario assembled from real, well-established patterns in construction liability — the additional-insured endorsement, the upstream indemnity, the "your work" exclusion, and the litigation that names everyone in the contracting chain. No real company, claim, or figure is depicted. All numbers are illustrative. The patterns, however, are exactly those a casualty underwriter meets on contractor and fabricator accounts every week.

Background

A mid-size structural subcontractor — call it Cardinal Steel Erectors, a constructed stand-in much like Harbor Steel in its exposure profile — fabricated and erected structural components on commercial construction projects. To win and keep work, Cardinal signed the standard subcontract its general contractors required. That subcontract, like nearly every commercial construction agreement, contained two provisions Cardinal's owner barely read and Cardinal's broker treated as routine:

  1. an additional-insured requirement: Cardinal had to name the general contractor (and often the project owner and the lender) as additional insureds on Cardinal's commercial general liability policy; and
  2. an indemnification (hold-harmless) clause: Cardinal agreed to indemnify and hold harmless the general contractor and owner against claims "arising out of" Cardinal's work.

Cardinal carried a \$1 million per-occurrence / \$2 million aggregate CGL — a normal limit for a subcontractor of its size — and a modest umbrella above it. The broker had placed a blanket additional-insured endorsement so that Cardinal didn't have to request a new certificate for each job; an older, broad edition of the endorsement was on the policy, because that's what had been on it for years and no one had revisited it.

Then, on one project, a worker employed by a different trade was seriously injured in an incident on a floor where Cardinal had erected steel. The plaintiff's lawyers did what plaintiff's lawyers do on a construction site: they named everyone — the general contractor, the owner, several subcontractors including Cardinal, and the manufacturers of various components.

The insurance and underwriting issue

Here is where the routine paperwork became a real underwriting event, and where every concept from §21.5 came alive at once.

Because Cardinal had named the general contractor and owner as additional insureds, Cardinal's CGL was now obligated to defend and indemnify those upstream parties for the claim — not just Cardinal itself. The single \$1 million per-occurrence limit was suddenly being shared among Cardinal and the parties it had agreed to insure. The chapter's first warning (§21.5) had arrived in full: more parties can tap the same limit.

Worse, the edition of Cardinal's blanket additional-insured endorsement was a broad, older form. Under its wording, the additional-insured coverage reached liability arising out of Cardinal's ongoing operations generally — and, on the facts, the general contractor argued that the grant extended to the general contractor's own supervision and site-safety responsibilities, not merely to Cardinal's direct fault. The chapter's second warning (§21.5) had arrived: the coverage can extend to the additional insured's own negligence, depending on the endorsement edition. Cardinal's policy was now potentially funding the general contractor's share of the liability.

Layered on top was the indemnity clause. The general contractor tendered the claim to Cardinal not only as an additional insured but as a contractual indemnitor, demanding that Cardinal's coverage respond to the contractual hold-harmless promise. Whether and how far Cardinal's CGL covered that assumed liability turned on the "insured contract" provisions (§21.5) and the precise scope of the indemnity Cardinal had signed — which, it emerged, was broader than the standard insured-contract definition comfortably supported.

And underneath all of it sat the "your work" exclusion (§21.1): to the extent the dispute was about repairing Cardinal's own allegedly deficient erection work, the CGL would not pay for that — only the resulting third-party bodily injury. Cardinal's owner, who had assumed "we have a million in coverage," was discovering that the million was being claimed by several parties at once, possibly for another party's negligence, with the boundaries set by endorsement editions and contract language he had never examined.

What it shows

The composite makes vivid what an underwriter must see before binding, not after a claim:

  • An additional-insured grant is an expansion of exposure the base premium did not contemplate. A \$1 million limit shared among the named insured and several upstream parties is not the same risk as a \$1 million limit protecting one insured. The "diversification" of a clean premises record means nothing when one loss pulls in multiple insureds at once (the limit-sharing problem, §21.5).

  • The endorsement edition is a coverage decision, not a clerical one. The difference between a tightly drafted modern additional-insured edition and an old broad one is the difference between covering liability arising from Cardinal's fault and underwriting the general contractor's site-safety program for free (§21.5).

  • An account that signs a lot of upstream contracts is "underwriting its contracts as much as its operations." Cardinal's real exposure lived in the subcontracts it routinely signed — the additional-insured and indemnity language — far more than in its slip-and-fall record (§21.5).

  • The insured rarely understands any of this, which makes it part of the underwriter's and broker's duty (the social-function and honesty themes) to control the endorsement edition, confirm the indemnity is within the form's support, and price for the multiplication of insureds.

Outcome (illustrative)

In the composite, the claim was eventually resolved — but the resolution consumed most of Cardinal's per-occurrence limit, dragged Cardinal's policy into the defense of parties Cardinal hadn't expected to insure, and triggered a coverage dispute between Cardinal's carrier and the general contractor's carrier over whose coverage was primary and how far the additional-insured grant reached. Cardinal's renewal came back with a higher rate, a tightened (current-edition) additional-insured endorsement, a requirement to provide sample subcontracts, and explicit underwriting attention to the volume of upstream contracts Cardinal signed.

The general contractor, for its part, learned that relying on a downstream subcontractor's broad old endorsement is its own risk. And the carrier that had placed the blanket broad endorsement years earlier "because it had always been there" learned the price of treating an additional-insured grant as paperwork: it had been giving away, unpriced, coverage to strangers across an entire book of contractor accounts.

Lesson for the underwriter

  • Read the additional-insured request as a rated feature. Know the edition, know what it covers, and on any account whose business model involves signing customer contracts (every contractor and fabricator, including Harbor Steel), treat the volume and breadth of those contracts as part of the exposure you are pricing (§21.5).

  • Hold contractual liability and tort liability apart, and check the indemnity against the form. An indemnity broader than the "insured contract" definition is an assumption the insured is carrying naked — and a coverage dispute waiting to happen (§21.5).

  • Remember the limit is shared. A single per-occurrence limit can be claimed by the named insured and every additional insured at once; underwrite (and possibly structure limits and retentions, Chapter 12) with that multiplication in mind.

For Harbor Steel, this composite is the direct preview of a live issue in its file: as a structural fabricator supplying general contractors, Harbor Steel will face exactly these additional-insured and indemnity demands (§21.5, The Underwriting File). The lesson is to control the endorsement edition, require a sample of Harbor Steel's standard customer contract, confirm the indemnity is within what the form supports, and price the account knowing its CGL is funding upstream parties' risk on top of its own — long before any claim arrives to teach the lesson the hard way.

Discussion questions

  1. List the three ways an additional-insured grant expanded Cardinal's exposure in this case, tying each to the chapter's §21.5 warnings. (§21.5)
  2. Why did the edition of the blanket additional-insured endorsement matter so much to the outcome? What would you, as the underwriter, have done differently at binding? (§21.5)
  3. Distinguish the two ways the general contractor tried to reach Cardinal's coverage — as an additional insured and as a contractual indemnitee. How does the "insured contract" exception bear on the second? (§21.5)
  4. The "your work" exclusion meant the CGL would not pay to redo Cardinal's own deficient erection, only the resulting third-party injury. Explain why this is consistent with the principle that "the CGL is not a warranty." (§21.1)
  5. Harbor Steel will face the same demands. Draft the two or three subjectivities or file notes you would attach to Harbor Steel's CGL to control the additional-insured and contractual exposure before binding. (§21.5; Ch. 13, The Underwriting File)