Who proves what, and what the excavator carries away

A construction policy is an all risks policy, and that label is not a marketing one: it decides who must prove what. Under a named perils policy, the insured must tie its loss to a peril on a list, and until it does it has nothing. Under an all risks policy, it must establish three things only: that the damage exists, that it occurred within the insured period and scope, and that it is sudden and fortuitous. It does not have to establish the cause. It is then for the insurer, if it wishes to escape cover, to show that an exclusion applies, and to show it in all its elements.

That allocation explains the structure of site disputes far better than the vocabulary the parties use. When an insurer writes that it cannot see the cause of the defect and awaits further material, it appears to be asking for information; in reality it is shifting the burden of proof onto the party that does not bear it. An insured who sets out to find the cause in order to persuade is working against itself: it is supplying the material for an exclusion the other side had to establish. The right question at that moment is not what the cause is, but which exclusion is being invoked and on what material it rests.

The design defect exclusion is where these mechanics produce their sharpest effects, and the module devoted to it takes on a different light here. An insurer invoking it must prove two distinct things, and usually only one is remembered: that the design was defective, and that the defect played a causal role in the damage observed. A badly designed structure that collapses for a reason unrelated to its defect remains covered. That second limb is the one most neglected in correspondence, although it is often the harder to establish, because on a site several causes overlap and shares have to be attributed to them.

Fortuity deserves a word, since it is the only one of the three conditions on the insured that invites argument. Fortuitous does not mean unforeseeable but not certain: an insurance contract bears on an event whose occurrence is uncertain, and what was certain is not insurable. The practical consequence is that inevitable damage, wear, expected progressive deterioration, settlement announced by the studies, needs no exclusion to fall outside cover: it falls for want of fortuity, which is a shorter door and one that closes without the insurer having to prove any exclusion. It is the only place where the allocation of the burden works against the insured, and that is why it is worth knowing.

One must then leave the law for the physical world, because the legal burden of proof runs into a reality that empties it of meaning: on a site, evidence is destroyed. Damaged works are cleared to reopen access, the collapsed part is rebuilt to hold a milestone, broken components leave in the skip with the rubble, and all of it is done for good reasons, under real schedule pressure, by people not thinking about an insurance file. Twenty days later the adjuster arrives in front of rebuilt works. The burden of proof has not moved an inch, and it has become impossible for both parties to discharge at once.

That symmetrical destruction does not, however, have symmetrical effects, and this is the least intuitive point in the subject. If nobody can prove anything any more, the insurer invoking an exclusion fails, which ought to benefit the insured. But the insured must still establish the extent of its loss, and it can no longer do that either: repair costs are justified by joint inspections that never took place. It wins the principle and loses the amount. And above all, most policies carry an obligation to preserve the scene, or to obtain agreement before any alteration, whose breach is sanctioned in its own right, independently of cover. The file is then lost not on the cause of the loss but on a conduct clause, which is a particularly bitter way to lose.

The last point concerns the loss adjuster, whose status is regularly misunderstood. The adjuster is not a judge, the report is a technical opinion, and nothing obliges anyone to follow it. But a signed adjustment protocol, an accepted quantum, a written acknowledgment of cause bind their signatory well beyond their procedural appearance, and they are often signed early, to release a payment on account the site needs immediately. The tension is real and has no comfortable solution: cash flow pushes toward signing fast, evidence pushes toward freezing nothing. What is decided knowingly is a trade-off; what is signed unread is a surrender, and at the time the two look alike.