A public export credit agency is not an insurer with patient shareholders. It is an instrument a state uses to support its exports, and that nature imposes limits on it the private market does not know. The choice between agency and market is ordinarily presented as a comparison of price and capacity; it is a comparison of freedoms. The agency is bounded by an international framework, by a national content requirement, and by environmental and social rules; the private market is bounded only by its underwriting and its reinsurance. Understanding where those bounds fall explains most of the refusals exporters experience as caprice, and above all it makes it possible to know in advance which file has no chance at which window.
The first bound is tariff-based, and it surprises because it works against what one expects of a support scheme. An international framework sets minimum premiums and maximum repayment terms, precisely to stop states subsidizing their exports through cheap insurance: competition between exporters is meant to turn on the product, not on the generosity of the treasury. It follows that an agency's rate is a regulatory floor and not a risk price. On a well-rated country and a short tenor, the private market can be cheaper than the agency, not because it is more efficient but because the agency is not allowed to go lower. An exporter without that key concludes there is a malfunction where there is only a rule.
The second bound is tenor. The same framework caps repayment periods by country category and by the nature of the transaction, and those caps are not negotiable. An agency can therefore have appetite for a file and be unable to offer the fifteen years the financing plan assumes. The refusal is then not a refusal of risk: it is a legal incapacity, and no commercial argument lifts it. The practical consequence is that a structure is designed knowing these caps, failing which one discovers after six months of processing that the chosen financial structure falls outside the only window that was targeted.
The third bound is the least well understood, and it is marketable risk. A regulation reserves short-term risks on certain countries to the private market, on the assumption that private cover exists there, and correspondingly forbids public agencies from writing them. That classification is not fixed: when a crisis withdraws private capacity, it can be suspended, and public cover reopens until the market returns. An exporter's access to cover therefore depends on an administrative classification that moves with the state of the market, generally in the right direction, but with a lag, and that closes again afterwards. An exporter can lose and regain access to cover without its risk having moved an inch, and it is a parameter to watch the way one watches a country limit.
The fourth bound concerns national content, and its peculiarity is to be a continuing obligation disguised as an entry condition. Eligibility is measured on the contract as presented at underwriting, but an industrial contract lives: a subassembly is bought elsewhere, a subcontractor changes, a production line is moved. The national share then falls inside a department that has no reason to know a cover depends on it, and verification happens at the loss, that is, when nothing can be corrected. To this are added, on large projects, environmental and social requirements bearing on the effects of the financed works, whose review adds months to the schedule.
What the private market does follows from what the agency cannot do. It covers the transaction the framework excludes, the tenor the cap forbids, the buyer whose national content falls short, and the country a classification closes to the public window. It does so more expensively, and the premium is not an abuse: it pays for capacity with no sovereign guarantee behind it, for reinsurance that must be bought, and for a risk nobody has made eligible by political decision. It also does so with different contractual material: narrower loss definitions, more precise evidentiary requirements, and a cancellability a state commitment does not have.
The operational reading amounts to reversing the order of the questions. One does not first ask what the cover costs; one asks whether the transaction falls within the public window's scope, and the answer comes from three checks that cost nothing: the tenor sought against the cap for the country's category, the real national share of the contract as it will be performed rather than as it was priced, and whether the country is classified as marketable at the date of commitment. If any of the three shuts the door, the price comparison never had a reason to exist, and one saves the six months a review destined to fail would have consumed.
A French manufacturer is negotiating the supply of a 60 million euro industrial plant in a Central European country that is a member of the European Union. The buyer's financing plan assumes repayment over eight years. The manufacturer approaches its national public agency, which redirects it to the private market without reviewing the file. Three months later it obtains a private offer, appreciably more expensive than the public rate it had anticipated, and carrying a narrower definition of the triggering event. Meanwhile, the purchasing department has moved the manufacture of two subassemblies to a group plant outside the Union, cutting national content from 62 to 44 percent. How should this sequence be read?
The agency's refusal is not a refusal of risk and it is not arguable: on a Union member country, short-term risk is in principle classified as marketable, which forbids the agency from writing it while the classification holds. The manufacturer thus lost three months seeking a judgment call where there was only a prohibition, having failed to ask the classification question before the price question. The rate gap, next, does not measure private market inefficiency but the absence of a regulatory floor on one side and its presence on the other: the anticipated public rate was not a market price, it was a floor, and it would not have been available in any event. The gravest point is the one the sequence conceals. The fall in national content from 62 to 44 percent was decided in a purchasing department with no connection to the cover, and it would have made the file ineligible at the public window had that window been open: eligibility is verified at the loss, not at underwriting. The manufacturer must therefore treat national content as a continuing obligation and track it, and it must reread the triggering-event definition in the private offer, which is the real consideration for the price rather than a drafting detail.
- 01A public agency's rate is a regulatory floor, not a risk price: the private market can be cheaper without being more efficient.
- 02A tenor cap is a legal incapacity, not an underwriting judgment: no commercial argument lifts it.
- 03The marketable-risk classification closes the public window on certain countries, and it is suspended then restored as the market moves.
- 04National content is a continuing obligation verified at the loss, while it erodes inside departments unaware that a cover depends on it.
- 05Check tenor, national share and classification before comparing prices: three free checks avoid six months of pointless processing.