Political violence, and the loss nothing broke

Political violence cover pays for damage, and the word must be taken strictly: something was destroyed, burned, damaged. It is a cover for things. Yet in a political crisis, most of the money is lost without anything being broken. A riot that burns a warehouse costs a few hundred thousand euros; a six-month curfew that idles the same plant costs several million, and it breaks nothing at all. The cover is therefore built around the small loss, while the insured's real exposure sits beside the trigger. This is not a drafting oversight, it is the founding choice of this line, and it has to be understood before opening a file.

The mechanism that follows is consequential business interruption, and it reads the opposite way from intuition. In these policies, business interruption does not attach to the political event but to the covered physical damage: it indemnifies the interruption that RESULTS from destruction of insured property. The recoverable amount therefore does not depend on the size of the economic loss, it depends on what was broken and on what that breakage actually caused. Two companies losing the same margin, one with a destroyed workshop and one with its gates closed by the authorities, do not recover sums of the same order. Physical damage functions as a key rather than as a loss, and the causal chain is checked link by link: the fact that a covered peril occurred, and that covered damage exists, does not mean the stoppage flows from the latter.

The market saw the gap and fills it with extensions, whose shape is worth knowing because it is always the same. Denial of access covers the loss suffered when the authorities bar access to the premises; threat covers an evacuation decided on a credible threat, with no attack. These extensions are useful and they are narrow, almost always bounded three ways: a radius around the site, a maximum duration counted in days, and frequently the condition that physical damage occurred somewhere within that radius. In other words, the extension re-imports the physical trigger one step away. Reading the radius and the cap in days is worth more, in money, than arguing about the principle of the cover.

Then comes the boundary that decides the large files, and it has the peculiarity of being drawn on a continuum. Political violence covers riot, civil commotion, insurrection and terrorism, and it excludes war. Yet those words do not name separate categories but rising intensities of the same phenomenon. A riot that lasts becomes civil commotion, organized civil commotion becomes insurrection, an insurrection holding territory becomes civil war. The facts do not change nature along the way, they change name, and that name is given afterward, by historians, by governments and by counsel who each have an interest. The cover and its exclusion are therefore drawn by the same facts read at two different scales, which is the definition of a dispute.

One drafting nuance weighs very heavily here and is easily missed. Some war exclusions are tied to a formal act: a declaration of war, recognition of a state of war. Others rest on a factual test: organized armed forces, sustained hostilities, control of territory. The difference is considerable, because states have essentially stopped declaring war since the middle of the twentieth century. An exclusion tied to a formal declaration therefore almost never bites, whereas a factual test catches most contemporary conflicts. Two policies that appear to exclude the same thing do not exclude the same thing at all, and nothing flags it on a quick reading.

Finally, escalation makes a loss travel from one contract to another. Most property programs carry an extension for strikes, riots and civil commotion while excluding terrorism and war; the standalone political violence policy takes over beyond that. One and the same event can therefore fall under the property program at the low end of the intensity scale, drop out as it rises, and be picked up by the standalone cover, or not. The two contracts come from different markets and their definitions of terrorism do not coincide. Insureds almost never fall because a cover was missing: they fall into the seam between two covers that both existed.

All of this commands the same discipline of proof, and it is a peculiar one because the useful facts are ordinary at the moment they occur. What will decide the file is the classification used by the local authorities at the time, the curfew orders and their exact perimeter, the fact that police and not the army intervened, the fact that the courts kept sitting, the date the zone reopened. These are public, datable elements, and they vanish beneath a unified narrative as soon as the crisis has passed. Six months later everyone will say the country was in civil war, and the only thing allowing anyone to say otherwise will be a record built while it was still unremarkable.