Theft, mysterious disappearance, and what remains to be proved

Three events look perfectly alike from outside: the object is no longer there. Theft supposes somebody took it. Mysterious disappearance records that it vanished with nobody knowing how. An inventory shortage is found by counting, often months later. A policy in this line treats those three situations very differently, and the difference never bears on the loss, which is the same, but entirely on proof. It is the one part of the contract where the insured loses not because cover is missing, but because they cannot say what happened.

Theft requires establishing a taking. Signs of forced entry, a police report, a chronology, sometimes a camera. The burden sits with the insured, and it is carried well when the object was in a public or commercial place. It is carried badly in a private home with no forced entry, where nothing materially distinguishes a theft from a misplaced object. The file then slides into the second category, and that slide is not neutral: it changes the applicable clause, and often the outcome.

Mysterious disappearance is covered by many wordings in the line, which is a real and rare breadth. But it is frequently covered on condition that the loss be identifiable to a particular time and place. One has to measure what that condition produces: it is impossible to satisfy in exactly the cases the cover exists for, since if the loss could be dated and located it would not be mysterious. This is not a drafting trap, the clause exists so that insurance does not absorb the management failures of a collection. The effect is nonetheless that the cover which reads as the broadest is among the narrowest in use.

An inventory shortage raises a problem of another order, that of the interval. It does not date the loss, it brackets it: between the last record showing the object and the count declaring it absent. That interval can span two policy years, two successive insurers and a loan. Most wordings therefore attach this head to DISCOVERY rather than to occurrence, precisely because occurrence is unknowable. That discovery basis has its own consequence, which is forgotten: cover must be in force when you discover, so letting a policy lapse extinguishes losses that have already happened.

A fourth situation hides behind the first and is not a variant of it. Misappropriation by someone the object was entrusted to is not theft within the meaning of the policy: it is infidelity, belonging to another section or another contract, with its own limits and its own notice periods. Yet the people closest to the object are precisely those: the dealer's employee, the registrar, the restorer, the family member. The theft cover therefore leaves outside its scope the circle where the risk is highest, and the insured discovers this only on notification.

One feature sets this family of losses apart from all the others: the object is not destroyed, it still exists. An insurer settling a total loss therefore acquires property that may resurface, in ten years or in thirty, often on a market where it has appreciated. That is where the clause that matters sits, the one allowing the insured to buy the recovered object back at the indemnity received. It looks like a detail at placement and it is, on a sought-after object, the only thing separating a collector from a former collector.

The discipline that makes all of this defensible is the same as for depreciation, which makes it easy to remember: a dated inventory, with images, and a record of every movement. It prevents no theft. It turns 'the object has disappeared' into 'it was here on this date and not on that one', which is the only fact the three categories are argued on. A collector who inventories once a year narrows the interval to twelve months; one who inventories every three years is preparing an argument no clause will settle in their favor.