Insurance: the four cases where paying for it actually pays

Insurance decisions go wrong when they are made as a single yes-or-no question. The useful version is four branches, and the branch is decided by what your parcel is rather than by what it is worth: declared value under the destination threshold, electronics carrying a battery, a single high-value item, or a mixed parcel of several kilograms. Each branch has a different failure mode, and cover only helps against some of them.

Two facts about the product shape the whole decision, and both are worth stating before any branch. First, the charge is normally based on declared value, and our own snapshot series records the insurance charge basis as Not verified for one quarter because the basis appeared to vary by category and we could not establish a single rule. Second, cover is defined by an exclusion list published by the line or the agent, and what is excluded matters more than what is included.

What cover does not do is worth as much space as what it does. It does not compensate delay, it does not pay out on a parcel seized because the contents were prohibited or mis-declared, and it does not fix a packaging decision you declined. Buying cover against one of those three is buying a document rather than a protection, and the branch questions below exist to separate the three.

A short glossary helps before the branches, because these words get used loosely. Declared value is the figure you state on the shipping instruction, and it functions as a ceiling for any payout rather than as a description of the item. Insured value is what the cover is written against, usually the same figure. Assessed charge is what the destination levies, which is a customs matter rather than an insurance one, and it is the reason people under-declare in the first place. Keeping those three apart removes most of the confusion surrounding the decision, and it makes each branch below answerable in about a minute.

When a consignment sits under the destination's published threshold, the assessed-charge risk is largely off the table, and that removes one of the reasons people reach for cover. What remains is loss and damage in transit. Against those, the decision is arithmetic of an unflattering kind: a premium is calculated against a declared value, while the expected loss is a small probability multiplied by that same value, and on low-value parcels the two figures can be uncomfortably close.

Our own practice is to self-insure below a few hundred in the destination currency and to buy cover above it. That figure is a rule we apply to our own parcels rather than a measured rate, and we are explicit about it because the honest version of this advice is "pick a line and hold it consistently", not "copy our line". A consistent rule beats an optimised one here, because the whole point of self-insuring is that the losses average out across parcels over a year.

One condition attaches to self-insuring: you have to be able to absorb the loss. A parcel whose disappearance would derail a month is not a candidate for self-insurance regardless of its value, because the arithmetic of expected loss assumes you can carry the realisation of it. That is a budget question rather than a shipping question, and it is the one place where we would override the threshold rule.

Setting that line for yourself is easier than it sounds. List the parcels you shipped in the last year, count the events, and divide nothing: the count is the input, not a rate. If your own count is zero, resist the conclusion that cover is therefore pointless, because zero events in a small sample is not evidence that the probability is zero, only that it did not land on you yet. The practical form of self-insurance is a reserve rather than a per-parcel calculation. Set an amount aside that would absorb one bad parcel, let it sit, and top it up when it is used.

Here the binding constraint is not value, it is acceptance. Services differ on whether they take lithium batteries installed in a device, whether they take them loose, and whether they take power banks at all, and the restrictions are usually published per service rather than per carrier. A parcel that was never eligible for the service is not made eligible by buying cover, and a claim on a prohibited item is the most reliably refused claim in the category.

The order of operations follows from that. Verify acceptance first, on the service documentation for the specific service you intend to use. Declare the contents truthfully, including the battery and where it sits. Then and only then decide whether the value justifies cover. Reversing those steps is the most common way buyers in this branch pay a premium for nothing, and it is also the way parcels get held at screening rather than at customs, which is a slower and less explicable delay.

If the parcel passes acceptance, cover does have a real function here, but a narrower one than people expect. It addresses partial loss and damage to a device, not confiscation and not screening delay. That makes the evidence habit more important than the premium: photographs at inspection showing the unit and its serial number, and a note of what was in the box. Without that record a damage claim becomes a conversation about what used to be inside the parcel.

A single-item parcel has a clean failure mode: the item arrives or it does not, and there is no ambiguity about which line of the manifest was affected. That clarity is worth something at claim time. What decides the branch is the declared-value ceiling on the service and how the declared figure relates to what you would actually lose, because cover is capped by the declaration rather than by the market value.

This is where under-declaring does double damage. Declaring below the true figure to reduce an assessed charge also caps the cover, and it creates a mismatch between the declaration and the proof you will be asked to produce if the parcel disappears. The mismatch is the part that costs: a claim on a parcel valued at one figure and declared at another is a claim with an internal contradiction, and it tends to be settled slowly if at all.

The practical preparation for this branch is cheap and belongs at inspection. Photograph the item and any authenticity or serial marking, keep the purchase record from the seller, and note the value you are declaring so that the three agree. On a single high-value item those three documents are the whole claim file, and assembling them before shipping takes minutes rather than the weeks it takes to reconstruct them afterwards.

A mixed parcel fails differently from a single-item parcel. Nothing dramatic usually happens to the parcel as a whole; instead, one item inside it arrives damaged, or one item is missing against the manifest, and the claim becomes a per-item argument that requires you to know what was in the box and what it looked like before packing. Incidence is the driver: the more items in a parcel, the more chances there are for a single-item problem, even when the per-item risk is unchanged.

Our own log points the same way, with the caveat that the sample is small. Across forty-one parcels it holds three loss or damage events, and all three were mixed parcels under seven kilograms. None of them was a single high-value item. Three events is not a rate and we do not present it as one, but it is enough to have changed our default: on mixed parcels in this weight class we now buy cover as a matter of course rather than deciding case by case.

Consolidation itself is a risk factor worth naming. Before the international leg the goods are unpacked, checked, sometimes re-bagged, and packed again, and every one of those handovers is a chance for an item to be mislaid inside a warehouse rather than lost in transit. Cover applies to the parcel as a whole, so a single missing item is a partial-loss claim, which is why the inspection photographs and a written item manifest matter more here than anywhere else in this list.

Four mis-costs, each of which we have seen in claim threads and each of which is avoidable. Buying cover that excludes your actual failure mode, which is the delay or seizure case: the premium is paid, the exclusion is quoted, and the loss is total. Skipping cover on the branch with the highest incidence, which is the mixed parcel, where the event is a partial loss that a small premium would have covered in full.

The third is under-declaring to save on an assessed charge and then claiming at the real value, which converts a small saving into a capped or contested claim. The fourth is insuring a parcel while declining the packaging that the cover assumes, since a claim on a box with no corner protection and no void fill invites exactly the argument about adequate packing that the exclusion list is written to allow.

The branch test takes a minute: what is inside, what is it worth, does the service accept it, and how many items could go wrong separately. Two of those four questions have nothing to do with money, which is the reason a pure cost comparison keeps producing the wrong answer. Where the terms themselves are unclear, the agent or carrier documentation is the authority and we link to the provider pages rather than paraphrasing a clause we cannot date.