The commonly stated axiom is that injury to third-party property is covered but damage to the contractor’s own work is not. I have no quarrel with this general notion as a statement of underwriting aspiration, but the question is whether this limitation is a result of express language in the insurance policy or is it something inherent in the nature of insurance?

The South Carolina court ruled that CGL policies afford coverage where “faulty workmanship causes a third party bodily injury or damage to other property, [but] not in cases where faulty workmanship damages the work product alone.” Slip op. at 23 n.4 (emphasis in original). The result is not so controversial as is its ratio decidendi: the court finds this formula inheres in the very idea of what an occurrence is. But why?

In the South Carolina case, the contractor’s work – a road – began to develop cracks, which resulted from errors in constructing the road. The cracking clearly is damage to property. Once title passed to the road (really, the materials the road comprised), the physical material became third-party property. On its facts, the claim involved damage to property of a third party.

Whether there is coverage for that kind of injury turns on whether that property damage was caused by an occurrence. The court finds that “faulty workmanship is not something that is typically caused by an accident or by exposure to the same general harmful conditions.” Id. Yet, the court concedes that the same conduct – were it to have resulted in some other injury or damage – would be considered to have arisen by accident, i.e., would constitute an occurrence.

In some ways, the court’s distinction is reminiscent of arguments about inherent vice and similar “off the contract” premises about why something is by its nature uninsurable. The problem with these arguments always is that these constitute sub silentio exclusions, ones that are created through argument at the time of claim and then apply necessarily to all claims.

Usually, courts apply the plain meaning of contract terms, and the ordinary meaning of "accident" or "occurrence" is broad enough to pick up unintentional results of even conscious poor workmanship. See David Mellinkoff, The Language of the Law 377 (1963) ("'accident' remains a 'blob of jelly' as a legal construct). The court’s decision might rest on stronger ground were it to have held that the damage on the particular facts was expected/intended by the insured and therefore excluded. Alternatively, the court’s exposition should have waxed more broadly on where the burden of proof should lie and why proof of unintended injury is a sine qua non to coverage vel non such that it should be part of the insured’s prima facie case. But none of this is the essence of the case, for the decision turns on what the court thinks of as first principles going to the heart of insurance.

But at the level of first premises there is nothing illegal about insuring against the risk of faulty workmanship and subsequent breach-of-contract and warranty claims. This risk can be insured because there is uncertainty as to whether, when, or how much liability will be found. See generally Bob Hedges, Back-Dated Coverage as Insurance, 34 CPCU J. 181, 181 (1981).

Nevertheless, one can understand that insurance companies do not want to create economic incentives for the insured not to do a good job in the first place, knowing that insurance will come to the rescue if the insured later has to pay the piper. (Whether such “moral hazard” exists at all in liability policies is doubtful even in theory let alone in reality.) But the issue is how should we set up systems of contract so companies can transfer the risk of loss without insurers picking up costs they do not wish to absorb.

If one were to embrace the objective of not insuring against the risk of faulty workmanship that results in losses to the worked-on property, how should that be expressed and enforced?

One answer lies in the underwriting and selection of insureds to which to sell policies: insurance companies should not sell policies to companies that deliberately would perform poorly. An insurance company has the right to continue to inspect the insured’s operations and, if it believes the insured routinely is engaging in shoddy workmanship, the insurer can cancel the policy (prospectively). (Given that the trial court and the Court of Appeal both sided with the insured, one can be somewhat confident that it is not a total fly-by-night operation.) If the insurer sold the policy nonetheless, then the insurer should own up to its commitment of coverage (and presumably it priced its policy to do so). The South Carolina decision thus undermines the economic incentives of insurers to police the quality of their insureds’ operations.

The manner to circumscribe the risks transferred in the contract is via its terms – not via conceptualist arguments about the nature of insurance, about liabilities that arise only ex contractu versus ex delicto (contract versus tort), or economic loss. None of these is a persuasive, predictable foundation for applying insurance policies by claims people in the field (or by courts). These off-the-contract limitations are not made clear to policyholders – who are admonished to read their policies to know what’s insured and not. Accordingly, the better-reasoned decisions hold that the terms of the contract govern, and if insurers do not want to accept these risks they are required to spell out in the contract the restrictive gloss they would place on the policy terms. See American Fam. Mut. Ins. Co. v. American Girl Inc., 2004 WI 2 (Wis. Jan. 9, 2004).

Besides promoting clarity and administrability of the coverage – and advising insureds ex ante about what is and is not covered – requiring insurance companies to state limitations such as that desired by the South Carolina Supreme Court allows state insurance regulators the opportunity to scrutinize the terms of the policy and the consistency and fairness of an insurer’s handling of claims. Requiring insurers to state these limitations moreover allows policyholders to choose not to do business with them, but rather to shop their premiums to other insurance companies willing to insure the valid, lawful risk that working on a project (where they perhaps might render substandard performance) may end up causing damage to the project.

It is clear that insurance companies do not wish to pick up the risk of damage to the property worked on by the insured. But shouldn’t that be written down in the policy -- or at least more clearly and prominently than making the word on which coverage pivots be one that has long been derided as a "blob of jelly"? The now-reversed South Carolina Court of Appeal cogently reasoned: “'The [insurance] industry has now taken to arguing that whenever a claim of defective construction is alleged against an insured, the claim is automatically barred from coverage as not constituting an ‘occurrence.’ This position is nothing more than a rehash of the ‘business risk’’ doctrine, whose success depends entirely on courts ignoring the actual language of the CGL policy.'” L-J, Inc. v. Bituminous Fire & Marine Ins. Co., 567 S.E.2d 489, 492 n.8 (S.C. Ct. App. 2002) (citation omitted); accord Erie Ins. Exch. v. Colony Dev. Corp., 736 N.E.2d 941, 947 (Ohio Ct. App. 1999). It is unfortunate that the South Carolina Supreme Court has excused insurance companies forever from saying what they mean on this point in the insurance policy or even requiring them to try.