The most widespread wording of the defect clause excludes the defective part and covers consequential damage. It looks balanced and it is: the insurer does not pay for what the contractor should have done correctly, and it pays for what that failure destroyed around it. The whole difficulty lies in the word part, which is almost never defined. On a real loss, the gap between the two possible readings of that word frequently exceeds the gap between two different wordings of the clause itself. The clause is negotiated at length, and the definition is simply endured.
The textbook case is the weld. A pipe fails at a badly executed run. Is the defective part the run, a few centimeters and a few hundred euros, or the whole pipe of which it is part? Both readings are arguable: the defect affects the run, but one does not replace a run, one replaces a section. A third reading exists, the functional unit, which would take the complete line since that is what must be requalified before return to service. Three answers, three orders of magnitude, and a clause that says nothing.
Three tests compete for the definition, and they must be named because they appear as such in discussions. The physical test takes the element materially affected. The repair test takes the smallest set it is technically possible to replace, which is wider and often more honest, since nobody repairs a square centimeter. The functional test takes the set performing a function and requiring requalification, which is widest and is justified on installations where return to service demands a whole-system test. None is wrong; the contract designates one or designates none, and that last case is the most frequent.
The effect of that delimitation on the split is mechanical and worth seeing with figures. The more widely the defective part is defined, the larger the excluded share and the smaller the consequential damage, since the latter is defined by subtraction. Ten million of damage whose defective part is worth a hundred thousand euros leaves 9.9 million of covered consequential damage; the same damage whose defective part is defined as the whole line can leave six million. The insured therefore has an interest in the narrow definition and the insurer in the wide one, and that opposition is stable, which allows it to be handled coolly, at placement, rather than in the heat.
An important exception nonetheless reverses that interest, and it surprises. Where the clause covers full reinstatement, including the defective part, less only the betterment cost, the insured no longer has an interest in shrinking the defective part: what matters then is showing that the repair does not constitute an improvement, that is, that the repaired works are no better than those specified. The argument changes nature and becomes a comparison between the repaired works and the promised works, not between two pieces of the same works. Applying the reflexes of one wording to the other is the most frequent error on this theme.
A case the clause does not provide for and that projects meet constantly must be added: the defective part is sound. Forty welds were made by the same faulty procedure; one failed, the other thirty-nine hold. Are they defective, therefore excluded, when none is damaged? Most wordings assume damage, and an intact item is not damage; the next module treats that question in its own right, since it commands amounts that often exceed the initial loss.
What must be done comes down to a sentence to be written before the loss. Ask that the contract define the defective part by one of the three tests, naming it, rather than leaving the word bare. That precision adds no cover, it costs nothing to negotiate because it does not obviously move the price, and it removes the argument that decides the amount. Failing that, one must at least know, before notifying a loss, which of the three readings will be argued and on what documents, because the first quantification submitted often fixes a position one never leaves.
On a process unit, a high-pressure pipe fails at a weld run executed outside procedure. The run represents 400 euros of labor. The welded section, the smallest technically replaceable set, is worth 62,000 euros. The complete line, which must be requalified by a whole-system test before return to service, is worth 840,000 euros. The rupture sprayed hot fluid that damaged two neighboring exchangers and a control station, for 4.3 million euros, and shutting the unit during the repair costs 1.1 million. The policy excludes the defective part and covers consequential damage; the contract does not define the defective part. What is at stake in that definition?
The three outcomes must be computed before anything is argued. On the physical test, 400 euros are excluded and 5.4 million stay covered. On the repair test, 62,000 are excluded and the rest follows. On the functional test, 840,000 are excluded. The gap between the narrowest and widest reading is therefore 839,600 euros, and that is the first lesson: on this file, the definition of an undefined word is worth more than most clauses negotiated line by line. The second point is subtler and often missed. The insured instinctively favors the physical reading, but it is untenable in practice: nobody replaces an isolated run, and arguing that the defective part is worth 400 euros undermines the credibility of the whole quantification. The defensible position is the repair test, and it has the advantage of matching what the invoice will show. The third point is the one the insurer will raise: requalification of the whole line appears in the quotation, and it will argue that a return to service requiring a whole-system test reveals a functional unit. The answer lies in a factual distinction, to be documented from the start of the repair: is the whole-system test imposed by replacing the section, or by a safety requirement that would have applied anyway. The first quantification submitted will fix the position, and it is prepared before it is sent.
- 01The word part is almost never defined, and the gap between its readings often exceeds the gap between two different wordings of the clause itself.
- 02Three tests compete: physical, the element affected; repair, the smallest replaceable set; functional, the set requiring requalification before return to service.
- 03The effect is mechanical: the wider the defective part, the smaller the consequential damage, since it is defined by subtraction. Opposed, stable interests, therefore negotiable coolly.
- 04The wording covering full reinstatement reverses the interest: it is no longer about shrinking the defective part but about showing the repair is not an improvement.
- 05The most favorable reading is not always the most defensible: arguing that a defective part is worth four hundred euros undermines the credibility of the whole quantification.