A construction policy covers a place. It names a site boundary, sometimes a few meters beyond the fence, and everything inside it is covered from the moment materials are received on site. Yet on a large project a considerable share of the value spends months outside that boundary: in a prefabrication shop three hundred kilometers away, on a quay waiting to load, in a hold, on an abnormal load convoy, in a warehouse rented because the site was not ready to receive. That value has been bought, paid for, is often one of a kind, and it is as essential to the schedule as if it were already erected. The question is therefore not whether it is exposed, it is, but which contract holds it at each moment, and experience shows the answer has gaps.
Three contracts share that journey, and none of the three was written with the other two in mind. The cargo policy covers the goods in transit, with its own valuation rules, its own particular average and its own termination, which is often delivery at the named destination. The construction policy takes over from receipt on site. Between the two remains everything that is neither moving nor on site: intermediate storage, waiting at the port, an extended stay at the supplier's after completion. Each contract projects an extension into that space, and the extensions do not coincide: a cargo policy's storage clause is usually counted in days after discharge, a construction policy's in maximum value per declared location. Two different dimensions never overlap perfectly.
The termination of cargo cover is the first place one falls. A classic clause ends cover on delivery at the final warehouse, or a set number of days after discharge from the vessel, whichever comes first. On a project whose site is running late, the equipment arrives on time and waits: the day count runs down at the port while the site finishes its foundations, and cargo cover expires before the construction policy is concerned at all, since nothing has been delivered to site. Delay, which is a scheduling fact, thus produces a break in cover that nobody decided and that appears in neither contract read on its own.
A construction policy's off-site storage extension has boundaries of its own, and they are of three kinds. A boundary of place first: the storage location must be declared, which presupposes knowing that it exists, and a warehouse rented in a hurry by a supplier to absorb a slippage is declared by nobody. A boundary of amount next: a maximum value per location, which bites precisely when a slippage concentrates several deliveries in the same place. A boundary of quality last: requirements as to guarding, fencing or weatherproofing that are reasonable for an organized depot and illusory for an improvised laydown area. All three tighten together at the moment of delay, that is, at the moment off-site storage increases.
These boundaries become far costlier when delay cover is attached to the program. Its mechanics are those set out in the module on anticipated loss: it indemnifies the commissioning slippage caused by covered physical damage. The question that decides everything is therefore whether the physical damage must be covered by this policy, or by any policy on the project. Many wordings require the former, and the consequence is brutal: a transformer destroyed on the quay, paid without argument by the cargo insurer, can push commissioning back six months without opening a single day of delay cover, because the damage was not covered by the construction policy to which that cover is attached. The material indemnity is intact, the financial loss is untouched, and the contrast between the two is what makes the file incomprehensible to whoever discovers it.
To that must be added an asymmetry that reading the wordings does not give and that practice imposes. The value of a project item is almost never its price: it is its price plus the time needed to make it again. A series part is replaced in a few weeks; a power transformer, a turbine or a large-diameter shell is remanufactured in twelve to twenty-four months because few shops can do it and they work to an order book. The same loss therefore carries a comparable material cost and an incomparable calendar cost depending on which item was hit, and it is that asymmetry, rather than the sum insured, that explains why attention concentrates on a few packages and not on tonnage.
That leaves the order in which the questions arise, which is the reverse of the order in which they are discovered. One discovers damage, looks for cover, argues an amount. The useful reading runs the other way: follow one given item from the shop to its final position, and name, for each segment, the contract that holds it, the value applied, the deductible and the fate of delay cover. That exercise is done before the loss or it is not done, because it bears on facts, dates and places that are no longer available afterwards. It is nothing like a formality: it is the only moment when the gaps between contracts can still be repaired, and it lasts only as long as the item has not left.
A grid connection project orders a power transformer made in Asia, worth 4.2 million euros, with a remanufacturing lead time quoted by the shop at eighteen months. The cargo policy covers the goods until delivery at destination or sixty days after discharge from the vessel, whichever comes first. The construction policy attaches from receipt on site and carries an off-site storage extension capped at 2 million euros per declared location. Delay in start-up cover of 46,000 euros per day, with a forty-five-day deductible, is attached to the construction policy and requires physical damage covered by that policy. The transformer is discharged on March 14, 2025. The substation is not ready: it is stored on an undeclared port laydown area. On June 2, 2025, a fire in neighboring pallets damages it severely. What follows?
The sixty-day count expired on May 13, twenty days before the fire, and that expiry is the direct consequence of the substation's delay. The construction policy does not take over: nothing was received on site, the port area is not a declared location, and the amount alone would have ruled it out anyway since 4.2 million exceeds the 2 million per location boundary. So there is no argument to be had between two insurers, there is an absence of cover, which is a different and far worse file. The reading to avoid is believing that the subject is the uninsured value: 4.2 million lands on the project owner's balance sheet, which is heavy but measurable. The subject is the calendar. Eighteen months of remanufacture, at 46,000 euros a day beyond forty-five days, is worth more than twenty million euros of lost margin, and that share would not be owed even had the construction policy paid for the transformer: delay cover requires damage covered BY IT, and the port laydown area falls under none of its extensions. The decisive fact dates back to May 13, when a count written into a contract nobody had reopened reached its end with no decision being taken.
- 01A construction policy covers a site boundary, and a major share of a large project's value spends months outside it: shop, quay, hold, convoy, rented warehouse.
- 02Cargo cover often ends at the first of two terms, one being a day count after discharge: a late site lets that count run down at the port before any delivery to site.
- 03The off-site storage extension is bounded by declared location, by amount per location and by guarding requirements, and all three tighten at the very moment storage increases.
- 04Where delay cover requires damage covered by ITS policy, a loss paid by the cargo insurer opens no delay days: the material indemnity is intact, the financial loss untouched.
- 05An item's value is its price plus its remanufacturing lead time: the same sum insured carries an incomparable calendar cost depending on whether it is a series part or equipment built to an order book.