A political risk underwriter does not decide alone what it accepts. Behind it stands a reinsurance structure that has fixed, before any file arrives, which countries are covered, up to what amount, and on what conditions. That structure is invisible to the insured, which never sees it and is not party to it, and it explains most of the market behaviour it finds incomprehensible: a sudden refusal on a country where nothing has changed, capacity evaporating in the first quarter, an information requirement appearing from one year to the next.
Two forms must first be distinguished, with different effects on underwriting. The treaty covers in advance a category of business defined by its terms, and an underwriter writing within that definition is reinsured automatically, without asking anyone. Facultative cover is negotiated case by case, which lets anything through but slowly and at greater cost. An underwriter therefore works inside a perimeter written elsewhere, and stepping outside it is not forbidden, it costs a delay and a margin. That is why a routine submission takes ten days and one on a country outside the treaty takes ten weeks.
Treaty renewal concentrates the effects and it has a date, generally the first of January. A reinsurer deciding to remove a country from its cover does not warn its cedant's insureds, it does not know them; it amends a schedule. On the morning of 1 January the underwriter discovers it is no longer reinsured on three countries, and its appetite contracts instantly on files it would have accepted in December. For the insured, the country is identical, the asset is identical, and the market's reception has changed entirely. It concludes there is arbitrariness; it is a decision taken elsewhere, on another horizon, by someone looking at a global portfolio.
Then comes the point that gives the brief its interest, and it rests on an asymmetry of duration. Treaties renew every year, investment policies often run three years or more. When a country leaves the treaty, policies already issued on that country remain perfectly valid: the contract binds the direct insurer to its insured, and reinsurance is a separate contract to which the insured is not party. The insurer therefore finds itself carrying net, to their expiry, commitments it believed ceded. That situation is uncomfortable and perfectly lawful, and it must be named because it produces behaviour the insured undergoes without understanding it.
That behaviour is predictable and better anticipated. An insurer that has become a net carrier on a country grows far more attentive to the letter of its policy: it investigates longer, it asks for more documents, it raises grounds it would not have taken when it carried only ten per cent of the risk. It also looks for legitimate exits, an inaccurate disclosure, an undisclosed increase in risk, a defect in the ownership chain. None of that is disloyal, and the effect on the insured is real: the rigour of investigation depends on who carries the risk, and that appears nowhere in the contract.
One pricing consequence deserves stating because it explains increases no event justifies. The price an insured pays contains the cost of its insurer's reinsurance, and that cost follows a global cycle unrelated to its country. A costly natural catastrophe season raises the cost of reinsurers' capital, which they pass on across all classes, including political portfolios that have had no losses at all. An insured negotiating its premium by showing that its risk has not moved argues correctly and beside the point: it is not its risk that has been repriced, it is the capital carrying it.
Finally one must say what an insured can actually do with this knowledge, without which the module would remain an explanation. Three things. Ask its broker whether its policy is written into a treaty or facultatively, information that exists and is not volunteered, because it predicts the stability of its renewal. Avoid placing long cover with an insurer it senses will cede everything, since the net carrier will be the more rigorous. And place its renewals other than at the first of January where possible, so as not to negotiate at the exact moment the market discovers its own constraints.
What this module teaches is therefore not a reinsurance technique, it is a reading of a market's behaviour. Most of the decisions an insured attributes to its own file were taken before it existed, by people who do not know it, on portfolio considerations. The professional who knows this stops pleading its file to someone who has no say, and goes looking for capacity where it still is.
A European insurer writes a political risk portfolio of 1.4 billion euros of commitments, including 210 million on a Central African country spread over eleven policies of which seven run to 2028. Its excess of loss treaty, renewed on 1 January 2026, now excludes that country at its reinsurers' request, following a coup in October 2025. An insured holding a three year policy expiring in June 2028 notifies in March 2026 a series of administrative measures and seeks the classification of creeping expropriation. It observes that the investigation is markedly heavier than at the previous claim settled by the same insurer in 2023: three requests for further documents, a challenge to the ownership chain, a question about a disclosure made in 2025. It sees bad faith in this. Is it right?
No as to intent, yes as to observation, and it should be given both answers because the second changes its conduct. What it observes is real: its insurer is investigating more harshly than in 2023, and it is not imagining it. The cause is not a decision taken against it but a change in a financial position it is no part of. In 2023 the insurer carried a fraction of the risk and ceded the rest; since 1 January 2026 the country is outside the treaty and it carries net, until 2028, the eleven policies already issued. An insurer discovering it carries 210 million instead of a fraction becomes attentive to the letter of its policy, asks for more documents and raises grounds it would not previously have bothered with. That is lawful, reinsurance being a separate contract to which the insured is not party, and its own policy remains perfectly valid until June 2028. Calling it bad faith would be an analytical error and above all a tactical one, because it would lead to arguing intent instead of dealing with the three grounds. Useful conduct follows from the diagnosis. The document requests are met quickly and completely, because a net carrier is looking for legitimate exits and an incomplete file hands it one for free. The challenge to the ownership chain is the most dangerous of the three and must be handled first, by producing the history of internal restructurings and checking that none escaped disclosure. The question about the 2025 disclosure must receive a written, dated and documented answer before it becomes a constituted ground. Finally this insured must be told what it does not want to hear: its policy is good until 2028 and there will probably be no renewal with this carrier, nor with many others. The search for replacement capacity, notably from public or multilateral carriers whose appetite does not follow this cycle, starts now and not in 2028.
- 01An underwriter works inside a perimeter written elsewhere: the treaty covers automatically, facultative cover lets anything through but slowly and at greater cost.
- 02A country removed from a treaty at renewal changes nothing for policies already issued: the insurer carries them net to expiry, and the insured is not party to the reinsurance.
- 03A carrier that has gone net investigates harder and looks for legitimate exits: the rigour of investigation depends on who carries the risk, and that appears nowhere in the contract.
- 04A premium increase may have nothing to do with the country: it is not the risk that has been repriced, it is the capital carrying it.
- 05Three useful steps: know whether your policy sits in a treaty or is facultative, beware long cover with a carrier that will cede everything, and avoid renewing at the first of January.