Almost every policy in this family leaves the insured a fraction of each loss, ten per cent most often, sometimes more. It is not a deductible, and the confusion is costly because it suggests a threshold when it is a proportion: a deductible subtracts a fixed sum and shrinks in percentage terms as the loss grows, a retained share stays at ten per cent on a five million loss as on a hundred million one. On a subject where losses are rare and enormous, the gap between the two mechanisms is counted in tens of millions.
The reason for that requirement is not accounting, it is behavioural, and it must be said plainly because it is unflattering to the insured. A fully covered investor no longer has a financial interest in the loss not happening. It may negotiate less firmly with the local administration, defer a costly compliance programme, let a concession degrade, launch aggressive proceedings whose political cost will fall on an asset that no longer costs it anything. None of those courses is fraudulent and all of them change the probability of loss. The retained share buys alignment, and its size is calibrated to stay painful.
It also buys something more concrete, which insurers rarely spell out and which decides the quality of claim files: the conduct of the claim. An insured carrying ten per cent of the final loss has a reason of its own to build its chronology, to bring records into being, to take mitigation steps at the right moment and not to settle too quickly with the state. The same insured covered in full delegates everything to its insurer and hands it a thin file. The retained share is therefore an instrument for producing evidence as much as one of alignment, and that is the strongest argument to put to a client that wants rid of it.
Then comes the question this module must settle, and it arises at every renewal in well advised groups: can that retained share be placed elsewhere? Technically, yes. A broker will find a market to cover a ten per cent share, often with a secondary carrier or through a separate facultative placement, and the premium will look modest since the slice is thin. Legally, the answer depends on a clause few buyers read, which frequently forbids the insured to insure its retained share elsewhere on pain of forfeiture or reduction. That clause is no drafting flourish: without it the retained share ceases to exist and the leading insurer has paid for an alignment it did not obtain.
The effect of such a placement deserves precise description, because it is not limited to a contractual breach. An insured that has placed its share no longer has any interest of its own in the loss, so the behavioural mechanism falls away, so the probability the leading insurer rated is no longer the right one. There is an alteration of the risk there, independently of any prohibition clause, with the usual consequences of inaccurate disclosure. And there is worse, an effect nobody anticipates: at the time of loss, two insurers will have to agree on the classification, on the date of the triggering event and on whether to settle, while the insured, carrying nothing, no longer has any reason to arbitrate between them.
There are nonetheless situations where the question arises legitimately, and they must be told apart from a simple wish for total cover. A lender may require the borrower to retain no uncovered exposure, which places the insured between two contradictory obligations; the answer is not to place in silence but to have one of the two amended, and it is the banking clause that most often yields. A group captive may carry the retained share, which many insurers expressly accept because the risk stays within the group and alignment is preserved. The difference between those two constructions is not one of technique but of who ultimately carries the risk, and that is the only test the leading insurer cares about.
One point of arithmetic deserves stating because it changes treasury decisions. The retained share compounds with the deductible and with the gap between the limit and the value of the asset, and those three terms appear on three different lines of the same document. An asset of two hundred million covered for one hundred and fifty, with a two million deductible and a ten per cent retained share, leaves the insured fifty million of underinsurance, two million of deductible and fifteen million of co participation, that is sixty seven million on a total loss. The figure on the front of the policy is one hundred and fifty; the figure that counts is eighty three.
This module therefore closes on a calculation rather than on a rule: add the three terms before discussing the premium, and present the committee with the amount actually transferred rather than with the limit. It is a figure nobody produces spontaneously, because each of the three terms is negotiated by a different person, and it is the only one that answers the question the committee thinks it is asking.
A European group insures a refinery held in a South American country. The policy carries a 180 million euro limit on an asset valued at 240 million, a 3 million deductible, and a 10 per cent retained share stipulated with an express prohibition on insuring it elsewhere. In January 2026 the bank pool refinancing the project imposes a new clause: the borrower must retain no uncovered political exposure. The broker proposes placing the 10 per cent with a secondary carrier for 240,000 euros of annual premium, observing that the leading insurer has no way of knowing. The finance director finds the solution elegant. What should he be told, and what should be done?
One must begin with the sentence the broker uttered, because it is the real subject of the file: the leading insurer has no way of knowing. That is false at the moment that counts. On the day of the loss the presence of a second carrier will surface, if only because it will have to be party to the discussions on classification and settlement, and the insured will then have to explain cover placed in breach of an express prohibition, at the worst possible moment. This is therefore not a question of discretion, it is a question of when discovery occurs. On the merits, two consequences compound and they are independent of one another. The first is contractual: the prohibition is express, its breach opens forfeiture or reduction depending on the drafting, and it would bear on the whole indemnity and not on the disputed 10 per cent. The second is graver because it would survive even without a prohibition clause: placing the retained share removes the insured's own interest in the loss, therefore alters the risk the insurer rated, therefore amounts to an undisclosed increase in risk. One must then address the real problem, which is genuine and is not solved by a refusal: the insured is caught between two contradictory obligations, that of its policy and that of its loan. The way out is not to choose in silence which one to breach, it is to have one of the two amended, and it is the banking clause that most often yields, because lenders accept a named exception as soon as it is explained that the alternative puts the whole policy at risk, which is exactly what they want to avoid. The construction that satisfies everyone exists and is negotiated openly: have the retained share carried by a group captive, which many insurers expressly accept since the risk stays within the group and alignment survives. Finally the committee should be given the figure nobody has produced: on a total loss the insured carries 60 million of underinsurance, 3 million of deductible and 18 million of co participation, that is 81 million. What is transferred is not 180 million, it is 159, and that conversation is worth more than the one about 240,000 euros of premium.
- 01A retained share is not a deductible: a deductible fades in percentage terms as the loss grows, a proportion never fades.
- 02It buys behavioural alignment, and it also buys the conduct of the claim: an insured carrying ten per cent builds its evidence, an insured covered in full hands over a thin file.
- 03Placing it elsewhere often breaches an express clause, and in any event amounts to an undisclosed increase in risk, since the rated probability rested on that interest.
- 04The test the leading insurer cares about is not technique but who ultimately carries the risk: a group captive preserves alignment, a third party carrier removes it.
- 05Underinsurance, deductible and retained share sit on three different lines and compound: the figure on the front of the policy is not the amount transferred.