The value of an asset one can no longer visit

The principle of indemnity requires that a payment restore the insured to its former position without enriching it, which presupposes knowing the value of the asset on the day of loss. On an immobilised asset taken by the state hosting it, that requirement meets an obstacle no other risk presents so sharply: the day of loss is precisely the day from which the insured can no longer enter the plant, no longer holds the accounting records, no longer appoints the auditors and no longer gets an answer from local management. Proof of value is demanded at the very moment access to all proof has been removed, and by the very party whose measure removed it.

Two contractual regimes answer that difficulty, and they must be held apart before being discussed. Value at the date of loss is the ordinary rule: the insured establishes what its asset was worth, the insurer argues, an expert decides. Agreed value instead fixes a sum at inception, payable without having to establish actual value at the date of loss, subject to the policy's conditions. The first is faithful to the principle of indemnity and often unworkable here; the second is workable and departs from the principle, which is why insurers do not grant it without something in return.

What they take in return deserves to be understood, because it goes by the misleading name of routine administration. An agreed value shifts a risk onto the insurer: if the asset has deteriorated since inception, it is worth less than the agreed sum, and expropriation becomes the most remunerative outcome for its holder. No insurer accepts that configuration as it stands. It frames it with a periodic revaluation obligation, almost always annual, supported by certified accounts or by an independent valuation filed on a date. Breach of that obligation is not sanctioned by a penalty: it tips the contract back into the value at date of loss regime, that is, into the one the insured cannot satisfy. The loss then reveals that protection had ceased long before.

An intermediate figure is very widespread and is misread by those who buy it. The policy takes book value at the date of loss, capped at the agreed sum, which means it pays the smaller of the two numbers. The insured therefore bears the fall and does not receive the rise, which is the exact opposite of what a figure written at the head of a contract suggests. There is nothing unfair in the drafting, and the quick reading is what costs: an asset whose value has doubled since inception is covered at the agreed sum, and the same asset having lost half its value is covered at that half. The stated number is a ceiling and not a promise.

There remains the question the brief poses and which decides most files: what to do when neither regime can work because the last certified balance sheet is old? The answer lies in moving the burden of proof through time. A value is not proved after the taking, it is built before. The file that holds is the one containing, on dated documents, certified accounts, independent valuation reports, fixed asset registers, physical inventories, the site's property insurance surveys, and comparable transactions on the same market. Each of those items is ordinary, none is expensive, and they become impossible to assemble the day after the measure.

A method the ground imposes must be added, one wrongly thought reserved for spectacular disputes. Where the accounts are inaccessible, one reconstructs from outside: the group's consolidated financial statements, which carry the subsidiary on an equity basis; local tax returns filed before the measure; deliveries and invoices from suppliers and customers, which live outside the country; dated satellite imagery, which establishes the existence and condition of the installations. That reconstruction never reaches the precision of an audit, and precision is not what is asked of it: it must reach the level of conviction an arbitrator or an insurer accepts, which is a markedly more modest and often attainable objective.

One final asymmetry deserves naming because it runs against intuition. The more specific the asset, the less establishable its value at the date of loss, since no comparison market exists for a refinery configured for a particular crude or for a concession whose perimeter is unique. Yet those are exactly the assets a state has an interest in taking, because they do not move and they have value where they stand. Evidential difficulty is therefore correlated with the probability of loss, not independent of it, which makes the agreed value question less incidental than it looks: it is the regime matching the assets actually exposed.

The conduct that follows from all this is short and rarely observed. Know which regime one is in, which means reading the valuation clause and not the figure on display. Keep the revaluation calendar as one keeps a premium due date, since forgetting it produces the same effect as non payment without looking like it. And build the valuation file continuously, in a place outside the country concerned, because a file kept at the subsidiary's registered office is on the wrong side of the door the day it closes.