This whole course has dealt with the date on which construction cover stops: acceptance, the testing window, the maintenance period. There remains the symmetrical date, less discussed and just as costly, the one on which it starts. It is assumed obvious, believed to be fixed by the contract, and indeed it is; what is not obvious is that the works themselves start on another calendar, decided by other people, for reasons having nothing to do with it. Between the two dates opens an interval where something is already at risk and nothing is yet covered.
It must first be seen that the inception date of a construction policy is not a single date but a combination of conditions. Most wordings start cover at the later of a date set in the contract and the actual commencement of works, or at the earlier of those two events depending on the case, and some further make inception conditional on payment of the first premium. Those three elements almost never occur on the same day on a large project, where negotiating the insurance program follows its own logic, where the notice to proceed depends on administrative permits, and where payment follows an accounting circuit.
What starts before, and that nobody counts as works, deserves naming. Geotechnical boreholes and ground investigations, which really do dig. Clearing, stripping and the removal of existing services. Setting up the site accommodation, erecting a fence, bringing in the first machines. Preliminary demolition, often the most dangerous operation of the whole project, which happens before the construction site opens. Each of those operations can damage an adjoining property, injure somebody or destroy equipment, and each sits, on most projects, before the inception date of the construction policy.
The other end of that interval is early ordering, and it carries the largest amounts. On a project whose heavy equipment has manufacturing lead times of twelve to twenty-four months, orders go out well before the site opens, and shop fabrication starts at once. A transformer being wound at its maker's, paid by staged instalments, already represents several million committed, and the construction policy does not cover it, since it has not incepted and the equipment is not on site anyway. The module on value outside the fence described that boundary in space; here it is the same boundary in time, and it closes just as badly.
The rule governing all of this is an old one and is worth stating plainly: an insurance contract bears on an event whose occurrence is uncertain, and it cannot retroactively cover a loss that has already happened and is known. A policy bought in March will not cover February damage the insured was aware of, and a retroactive inception date, where accepted, is worth only what nobody knew about. It is the same requirement of fortuity the module on proof met from the other end. The practical consequence is that a gap in cover noticed at the start cannot be closed once a loss has fallen into it: it is closed beforehand, or not at all.
There is an asymmetry of attention that explains why this subject is neglected. At the end of a project everyone watches the dates: acceptance is prepared, minutes are signed, statutory warranties start, and the stakes are visible. At the start, attention is elsewhere, on permits, financing and schedule, and insurance sits on a list of things to do among thirty others. Yet both ends carry the same type of risk, and the start additionally carries demolition and boreholes, which are statistically high-loss operations. The imbalance of attention is exactly the inverse of the imbalance of risk.
The useful reading therefore consists of asking a single question, but early: what is the project's first physical operation, and on what date. Not the first operation of the main contract, but the first operation of any kind, including a borehole ordered by the design office eight months earlier. Then comparing that date with the policy's real inception date, allowing for the premium payment condition if there is one. The difference between the two, expressed in days, is the measure of the gap, and it is a number that can be known before it costs anything. That comparison is almost never made, because it assumes somebody looks at both calendars at once, and they belong to two different people.
A factory project plans to open the site on September 1. The contractor's all risks policy has an inception date of September 1, conditional on payment of the first premium, actually paid on September 12. The design office had additional geotechnical boreholes carried out from August 7 to 22. A demolition firm took down an old building on the plot from August 18 to 29. A transformer worth 3.1 million euros, ordered in February, has been in manufacture since April and will be delivered in January. On August 21 a demolition excavator severs a drinking water main and floods the neighboring building's cellar: 340,000 euros of damage. On September 5 a machine knocks over a retained existing boundary wall: 90,000 euros. What does the construction policy cover?
None of what is described, and three distinct reasons are needed to establish that, which is the real subject of the case. The August 21 loss pre-dates the contractual inception date: it falls in the interval between the project's first physical operation, which goes back to August 7 with the boreholes, and September 1. The September 5 loss post-dates that inception date, and the reflex is to think it covered; it is not, because inception was conditional on payment of the premium, made on the 12th. A condition precedent of payment shifts the real inception date while the contract carries no date other than September 1, and it is the quietest drafting trap in the class. The transformer, finally, has been committed since February and in manufacture since April, off site and outside the period, therefore outside on two grounds. The non-obvious point is that all three gaps would have been measured by a subtraction done in July: the first physical operation is the August 7 borehole, real cover starts on September 12, the interval is thirty-six days, and it contained a demolition, that is, the most loss-prone operation of the project. That number was knowable, and nobody computed it because the two calendars belonged to two different people.
- 01Inception is not a date but a combination: a date set in the contract, actual commencement of works, and sometimes payment of the first premium, which shifts inception without appearing as a date.
- 02Operations preceding the opening of the site are counted by nobody as works: boreholes, stripping, removal of services, site accommodation, and above all preliminary demolition.
- 03Early orders carry the largest amounts: equipment in manufacture for months is committed, off site and outside the period, therefore outside on two grounds.
- 04An insurance contract does not retroactively cover a loss that has already happened and is known: a gap noticed at the start is closed before a loss falls into it, or not at all.
- 05Attention goes to the end dates, where everything is visible, while the start additionally carries demolition and boreholes: the imbalance of attention is the inverse of the imbalance of risk.