Ten-year liability, and what exists nowhere else

An engineer who has run projects on three continents discovers in France a regime with no equivalent anywhere else, and generally discovers it too late. The Act of January 4, 1978 imposes two distinct insurances on almost every building work: ten-year constructor liability, taken out by those who build, and inherent defects insurance, taken out by the party having the work built. This is not a market practice one can negotiate, it is a statutory obligation, and understanding it means first seeing the problem it solves.

That problem is lost time. Without this regime, an owner whose building cracks six years after handover must establish someone's liability before being compensated: identify the company at fault among a dozen, prove the fault, withstand cross-claims, and wait for the outcome. Meanwhile the building deteriorates and nobody repairs it, since paying before judgment would amount to admitting liability. The French regime breaks that sequence in two, and therein lies its originality: it separates who pays now from who will pay in the end.

Ten-year liability answers the second question. It weighs on constructors for ten years from acceptance, for disorders that compromise the soundness of the works or render them unfit for their purpose. The decisive point is that it rests on a presumption: the owner need not prove fault, it need only establish the disorder, its gravity and its connection to the works. The constructor escapes only by demonstrating an external cause, which is narrow. That presumption shifts the entire burden of proof, and it, far more than the ten-year duration, is what makes the regime severe.

Inherent defects insurance answers the first question. Taken out by the owner before works start, it prefinances repair of those same disorders, without investigating liability, within deadlines the statute fixes: sixty days to rule on cover, ninety to offer an indemnity. The insurer that has paid then turns, by subrogation, against the constructors and their liability insurers. The liability battle does take place, but it unfolds between insurance professionals, after repair, and not between the owner and a chain of companies while the disorder worsens.

One must still know who counts as a constructor, because the statute reaches further than intuition. It covers the contractor, the architect, the engineering firm, but also the developer, the seller of a building to be built, the technical inspector within its mission, and the maker of an equipment item designed to meet precise installation requirements, which practice designates by an acronym. A foreign equipment supplier can therefore find itself a constructor in the French sense without ever having contemplated it, simply because its equipment goes into works located in France. The rule follows the works and their location, never the nationality of the builder nor the law chosen by the contract.

One must also see what ten-year liability does not cover, because the confusion is constant. Disorders affecting neither soundness nor purpose fall under other, shorter warranties: the one-year completion warranty, obliging the contractor to make good everything listed as a reservation, and the two-year proper-functioning warranty, covering separable equipment items. A purely aesthetic disorder falls under none of the three. On an industrial work the question becomes frankly hard, since unfitness for purpose is assessed against the intended use: is a plant producing at eighty percent of nameplate capacity unfit for its purpose? The answer depends on what the contract defined as the purpose, and that definition is then reread word by word.

That leaves the trap of large international projects, and it is twofold. A foreign turnkey contractor signing for works located in France is subject to the regime, whatever the law of its contract, and its usual liability insurance does not cover it: ten-year liability is a French-law cover that international policies do not carry as a matter of course. It discovers this at the first disorder, ten years later, when its insurer explains that the French presumption does not fit the definition of a loss in its policy. The second half of the trap is the calendar: the two regimes do not overlap but follow one another, the construction policy stopping at acceptance where ten-year liability begins. A disorder appearing in the days following acceptance attaches to one or the other according to a date the acceptance module has already shown to be disputed, and this time the gap is not between two insurers but between two entire legal regimes.