Among the regime's perils, ground movement holds a place of its own and is the worst served by the trade's tools. It barely models at portfolio scale, it maps usefully only case by case, and it produces losses whose distribution resembles nothing flood or windstorm have taught underwriters. An underwriter applying to landslide the reflexes acquired on flooding is wrong on three counts at once.
The first is scale. A flood is described in kilometers, a windstorm in hundreds of kilometers, a landslide in tens of meters. Two buildings a hundred meters apart can sit, one on an unstable slope, the other on sound substrate. No grain available in an insurance file, neither commune nor postcode nor even address, carries that information. What carries it is the terrain's topography and the history of its earthworks, two things that cannot be bought as a database.
The second is the shape of the damage. A flood produces many partial losses; a ground movement produces few losses and they are often total, or close to it. A building whose foundations move is not a damaged building, it is a building whose demolition is being discussed. That changes what has to be looked at: on this peril the question is not the expected damage ratio but the probability of being on the wrong slope, and the penalty for an error is the loss of the entire value.
The third is the hardest and it governs almost every file: the cause is rarely purely natural. A slope moves because it rained, and also because an embankment was cut back to gain a platform, because drainage was removed during works, because roof water was discharged at the top of the slope for eight years. The natural hazard is present, it is not alone, and the branch that will pay depends on that split.
One must therefore know how that split is settled, because it is this peril's useful skill. The regime answers for damage whose determining cause is the abnormal intensity of a natural agent. If the expert establishes that the slope would have held without the earthworks, the determining cause is not the rain, and the route is no longer natural catastrophe cover but the liability of whoever designed or executed the works. The two routes have different deadlines, different counterparties and different evidence: opening the wrong one costs months that are not recovered.
One must add what this configuration does to accumulation, because it is counterintuitive. On flooding, an underwriter fears discovering a hundred sites in the same catchment. On ground movement, geographic accumulation is weak: few sites move together. The dangerous accumulation lies elsewhere, it is CONSTRUCTIONAL: a hundred sites built in the same years, by the same earthworking methods, on the same type of slope, carry the same latent defect and declare themselves within one decade. That accumulation shows on no map and is measured only through the attributes of the building stock.
In practice, conduct comes to two steps and they cost almost nothing. Before any site backed against a slope, one asks for the construction's geotechnical studies and the existence of a retaining structure, with its date and its maintenance: those documents almost always exist and nobody asks for them. And one looks at what happens upstream, because a site can be perfectly stable and receive the neighbor's slope, a situation in which the question stops being one of vulnerability and becomes one of recourse.
A 6,200 square meter warehouse was built in 2014 on a platform gained by cutting back a hillside, with a 4-meter retaining wall at the top of the slope. After a very wet winter, the slope moves in March: the wall tips over, the warehouse's north gable cracks over its full height, operations stop. An order recognizes the state of natural catastrophe for ground movement in the commune, for the period from March 2 to 9. The appointed expert notes that the quarter's rainfall is indeed exceptional, and that the retaining wall was designed without allowance for water ingress, the drain specified in the 2013 geotechnical study never having been installed. The insured claims under the regime. What happens, and what must be done first?
The file carries this peril's own tension: the order exists, the hazard is established, and that is not enough. The regime answers for damage whose DETERMINING CAUSE is the abnormal intensity of the natural agent, and the expert's finding identifies a second cause without which the failure would probably not have occurred: a retaining structure designed without water ingress and a drain specified in 2013 that was never installed. If the expert concludes that a properly drained wall would have held this rainfall, the determining cause is not the rain, natural catastrophe cover does not respond, and the route becomes the liability of the builders, with its own deadlines running from the 2014 handover. If the expert concludes that the rainfall would have carried the structure away even properly drained, the regime responds. The error to avoid is choosing one of the two routes at once, because the two close through different mechanisms and at different speeds. The first step is therefore not a notification but a preservation: having the state of the slope, the wall and the absence of a drain formally recorded before any make-safe earthworks, because making safe will destroy the evidence of cause within days. The second is to notify under both headings, without choosing, and to say so: notifying under the regime does not bar a claim against the builders, and waiting for the expert's conclusion before opening the second route may find it shut. The third is to check exactly what the 2013 geotechnical study specified and what the works contract executed, because the gap between the two is the document that decides the file, and it sits in the employer's archives, that is, with the insured itself.
- 01A landslide is described in tens of meters: no insurance file grain, commune, postcode or address, carries that information.
- 02This peril produces few losses and often total ones: the question is not the damage ratio but the probability of being on the wrong slope.
- 03The cause is rarely purely natural, and the split between hazard and structure decides which branch will pay.
- 04The dangerous accumulation is not geographic but constructional: a hundred sites earthworked the same way carry the same latent defect.
- 05Geotechnical studies, the existence and upkeep of a retaining structure almost always exist and nobody asks for them before the loss.