A whole certification devoted to insuring a risk must end with the question it has put off until now: when is it better not to insure? It is not rhetorical. A group present in eleven countries whose cost of full cover exceeds the maximum plausible loss on the worst of them does not have a negotiation problem, it has an answer in front of it that it dares not state because nobody is paid to give it.
The governing calculation is the maximum plausible loss, and it must be told apart from adding up sums insured. A group holding eleven assets rarely adds eleven simultaneous expropriations, because they do not occur together absent a common factor, and we have seen that common factors exist, political or physical. The relevant figure is the heaviest loss a single event can produce, increased by the exposures genuinely correlated with it. That figure is almost always far below the sum of the limits, and it is the one to compare with the cost of cover.
Then comes the second term of the comparison, and it is wider than a premium. The cost of full cover comprises the premium, but also the retained share on each loss, the deductibles, the gap between limits and real values, management time spent underwriting eleven programmes, and advisers' fees. A group comparing a premium with a maximum loss compares two figures one of which is incomplete, and it often wrongly concludes that it is well covered cheaply.
One must then name what replaces the policy when one decides not to take it, failing which the decision is mere abandonment. Organised retention is an instrument, not an absence: a dedicated cash line, a provision, a captive smoothing over time, or simply an exposure ceiling per country decided and kept. Diversification is another, provided it is real, which means having checked the common factors rather than counting flags. And structuring is a third, often the most effective: holding through an entity covered by a treaty, bringing a multilateral carrier into the financing, obtaining written undertakings from the state, are protections that are not paid for in premium.
One test decides more than all the others and it is not an insurance one: the group's capacity to absorb the loss without its own continuity being in question. A group whose worst plausible loss is three per cent of its own funds does not need insurance on that risk, it needs discipline. A group for which the same loss is forty per cent must insure even if the ratio looks poor, because what it buys is not an expectation of gain but the survival of a going concern assumption. Between the two lies the zone where the decision is arguable, and it is the only one where economic calculation has the last word.
Two external constraints limit that freedom and must be checked before concluding. Lenders frequently require cover, as we have seen, and the undertaking sits in the loan agreement rather than in the policy. Local partners, minority holders and sometimes home country authorities may require it too. A group deciding not to insure without rereading its financing agreements discovers its error at the first covenant test, and the discovery costs more than the premium it was saving.
Finally one must say what a decision not to insure demands of whoever takes it, because it is not a saving but a transfer of burden onto the organisation. It presupposes tracking the countries oneself, with the few public indicators that deteriorate before ratings; maintaining the documentation that would serve a claim when there is no insurer to bring it; and preserving the ownership structure that opens a treaty, since it becomes the only protection. Those three obligations cost time, they cost no premium, and a group deciding not to insure without keeping them has simply decided to do nothing.
This module closes on the sentence this trade utters least and which makes it respectable. There are situations where the best advice is not to buy, and knowing how to recognise them is what separates an adviser from a salesperson. This certification will have failed if it has only taught how to place policies; it will have succeeded if it has taught how to say, to a committee expecting the opposite, that the best protection on this file is not for sale.
A European building materials group holds eleven industrial subsidiaries across eleven countries, in Africa, Southeast Asia and Latin America, for a combined net book value of 840 million euros. The largest asset, a cement works, is worth 190 million; the other ten are worth between 40 and 95 million. None of these countries belongs to one region or one bloc, and none shares an export route, output being sold locally in each case. The annual cost of a full cover programme is put at 14.2 million euros, before deductibles and retained shares. The group's consolidated own funds are 3.1 billion. Two of the eleven subsidiaries are project financed. The finance department asks whether to insure.
The figure to put first is neither 840 million nor 14.2: it is the maximum plausible loss, and it is 190 million here, the cement works. The eleven countries share neither region, nor bloc, nor export route, and output is sold locally, which removes the two common factors this certification has taught one to look for, the political and the physical. That must be verified rather than assumed, but if it holds, the exposures are genuinely independent and there is no ground for adding several together. The second figure is that loss against own funds: 190 against 3,100, that is six per cent. A loss of six per cent of own funds is painful and does not put the group's continuity in question. That is the test that decides, and it leans clearly towards not insuring. The third calculation completes the demonstration. The programme costs 14.2 million a year before deductibles and retained shares; over twelve years it will have cost more than the worst loss it could cover, and one is paying here for a risk the group absorbs without danger. Two checks are required before concluding, and the first is blocking. Two subsidiaries are project financed, and those agreements almost certainly require political risk cover to be maintained: those two are not free, and the agreements must be reread before any decision, on pain of discovering the undertaking at the first covenant test. The second concerns the local partner and minority holders, who may have required the same. The recommendation that follows is therefore neither to insure everything nor to abandon everything, but to cover what is contractually required and the cement works if the committee wants its worst case bounded, and to retain the rest in an organised way. Retaining presupposes three obligations the group will have to keep and which cost no premium: tracking itself the few public indicators that deteriorate before ratings, maintaining in each country the documentation that would serve a claim no insurer will bring on its behalf, and preserving the ownership structures that open a treaty, since they become the only protection. A group that decides not to insure without keeping those three has not taken a decision, it has merely stopped paying.
- 01The figure to compare is not the sum of the sums insured but the maximum plausible loss, increased only by exposures genuinely correlated with it.
- 02The cost of cover is not the premium: it includes retained shares, deductibles, the gap between limits and real values, and management time.
- 03The deciding test is not an insurance one: it is the worst plausible loss against own funds, and what the group's continuity can absorb.
- 04Not insuring is not an absence: organised retention, verified diversification and structuring through a treaty are instruments that cost no premium.
- 05Two external constraints outrank the calculation: loan agreement undertakings and partners' requirements, which sit somewhere other than in the policy.