A country should build its export credit agency for speed, because it does not set the terms

Export credit agencies compete on two things, and only one of them belongs to them. A country building or reforming an agency should build it for speed: the date it can commit, and the date it can refuse. That is my position, and it means treating the term sheet as a floor to be published rather than a bid to be improved.

The half nobody in the agency sets

Terms are priced off a base rate no agency sets. On the ten-year Treasury that base ran 0.89 percent in 2020, 1.45 in 2021, 2.95 in 2022, 3.96 in 2023, 4.21 in 2024 and 4.29 in 2025. The move across the window is 3.40 points, the largest single-year move is 1.50 points between 2021 and 2022, and the 2025 reading is 4.8 times the 2020 reading.

Short rates moved faster still. The federal funds effective rate averaged 0.08 percent in 2021 and 5.14 percent in 2024, a climb of 5.06 points across three years, and 3.34 of those points arrived in one step, from 1.68 percent in 2022 to 5.02 in 2023. In 2025 it averaged 4.21 percent, 0.93 points below the peak. A concession on margin is measured in fractions of a point at every agency I have seen. The base underneath it moved by more than three points from one year to the next.

Currency does the same work from the other side. The broad dollar index, with January 2006 at 100, read 117.8 in 2020, 113.1 in 2021, 120.7 in 2022, 120.3 in 2023, 122.9 in 2024 and 122.7 in 2025. From the 2021 low to the 2024 high that is a rise of 9.8 index points, or 8.7 percent; measured across the whole window it is 4.9 points, or 4.2 percent. A borrower servicing a dollar loan out of local earnings met that movement whatever the coupon said. That index is not a bilateral rate and no single country's currency sits inside this paragraph, but the direction of the pressure is in the published series, and it is not a lender's to offer or withhold.

What two years in a queue cost

Delay is the other half, and it is priced as well. Producer prices for all commodities, on a 1982 base of 100, ran 194.4 in 2020 and 264.5 in 2022. That is 70.1 index points, or 36.1 percent, in two years. A buyer whose financing closed two years later than it might have done re-entered a market more than a third dearer on that measure. No extension of tenor gives that back, because the sum borrowed is larger and a longer loan only spreads it.

The index did not keep climbing. It fell to 255.7 in 2023 and 254.6 in 2024 before reaching 260.3 in 2025, which leaves it 4.2 points below the 2022 peak. That is the honest version of the point. Escalation is not a law, and a project delayed into a falling market is delayed into a cheaper one. What the series shows is that two years of drift can carry a price change larger than anything a lender writes into a term sheet, in either direction, and that whoever holds the timetable is holding the larger variable.

Where the objection is strongest

A serious reply runs the other way: speed is a property of process and terms are the substance. An agency that commits quickly to the wrong exposure has done its taxpayer a faster disservice. Tenor and grace are what make a project bankable for a borrower with no long domestic market, and for that borrower a shorter loan is not a slower loan, it is no loan. The reply rests in part on a premise no source on this site can settle: that established agencies mostly work inside an agreed framework on maximum tenors and minimum premiums. Grant it, and terms compete exactly where that framework does not reach, and speed is a tiebreak between lenders already offering the same thing.

Half of that is right, and it is the half about thresholds. Terms have to clear a level at which the project finances at all, and below that level nothing else gets discussed. A threshold is a floor, though, not a race. Publish the standard, in a document that does not change per transaction: the tenor, the cover ratio, the premium grid, the local-content rule. Then compete on the date. An agency that will not publish its terms is not competing on terms at all, it is negotiating, and negotiation is the slowest thing it does.

Refusal belongs on the same clock. An agency slow to decline costs the exporter the delay and costs the borrower the option of going elsewhere inside it. A published date by which a refusal will arrive is cheaper to commit to than a published date for approval, and it is worth more, because almost nobody offers one.

What choosing speed costs

Four things get worse, and they should be said plainly. Decision quality falls at the margin, because time is where diligence lives, so the loss rate rises and has to be defended rather than explained away. Staff cost rises, since speed is bought with people who can read a credit and not with a committee calendar. The agency gives up the concession it could have extracted from a borrower under time pressure, which is a fee foregone. And a fast approval that later fails is legible to a legislature in a way that a slow failure is not, so the political exposure lands on the institution that chose the clock.

Those are prices worth paying, because the alternative is paying for a variable somebody else moves. A country that funds a terms war commits itself to subsidising against a base rate set by a central bank it does not sit on, in a currency it may not issue. The bill arrives largest in the years the base rate is highest, and those are the years the budget is tightest.

One thing these series do not say. Exports of goods and services ran 2.57 trillion dollars in 2021, 3.04 in 2022, 3.07 in 2023, 3.22 in 2024 and 3.32 in 2025, a rise of 0.75 trillion across the window, of which 0.28 trillion came after 2022. Nothing in that series records how a single shipment was financed. It is not evidence for the argument on this page and I am not offering it as any. It is here because an export credit agency is pointed at that aggregate, and the aggregate cannot tell anyone whether the agency mattered.

Put one thing to an agency that calls itself competitive. Ask for the date by which it will give an answer, in writing, before it has seen the file. An agency competing on terms will reply with a term sheet. An agency competing on speed will reply with a date, and the date is the only part of the offer it can keep on its own.