The 8.2 Percent Borrowing Rate: What the World Bank's Debt Report Changes for Developing Economies

Data as of November 5, 2026.

The number that matters is not the $9.4 trillion

The World Bank's annual International Debt Report, published today, puts the external debt of low- and middle-income countries at a record $9.4 trillion. That headline figure grew 5.5 percent in 2025, the fastest pace in four years, and the growth came overwhelmingly from private lending rather than from official creditors.

The figure that actually changes behaviour sits lower in the report. Countries eligible for concessional lending paid an average of 8.2 percent on new borrowing, the highest rate since the late 1980s. Preliminary data for 2026 shows lending continuing at a rapid pace at the same elevated cost.

Those two numbers describe different problems. A larger stock is a stock problem. An 8.2 percent price on new money is a flow problem, and it binds immediately, because every refinancing at that rate raises the interest line in the current budget rather than some future one.

The concessional floor stopped holding

Concessional finance worked because it priced below the market instead of alongside it. The report's own examples show how far that floor has moved. Laos borrowed at 11.5 percent, the Republic of Congo at 13.88 percent, and Kenya at 9.2 percent.

Each of those rates sits close enough to a commercial yield that the distinction between concessional and market access has narrowed to a matter of degree. The World Bank's chief statistician, Haishan Fu, framed it directly: access to international finance has improved, but the price of that finance has not, leaving less for long-term development.

The mechanism linking the two is worth stating plainly. When a government refinances at 9 percent instead of 4 percent, the difference is paid out of the same revenue pool that funds teachers, clinics and roads. Debt service becomes the fastest-growing line in a budget that is not itself growing.

The creditor has changed, and so has the negotiation

The 2005 Gleneagles write-off worked because the lenders were governments sitting in the same room as the borrowers. The report's picture is different now: private creditors and bond issuance by public enterprises supplied most of the increase, which is why the UK, holder of next year's G20 chair, has commissioned a review of the debt regime rather than a fresh round of relief.

A bondholder cannot be persuaded to forgive principal the way a finance ministry can. Restructurings become slower, and the collective action problems inside each creditor class consume the time that a country does not have. This is the same constraint visible in Sovereign Debt and Restructuring, which has spent years in process while the arrears accumulate.

There is a second, quieter consequence. Private creditors lend against collateral and cash flow, not against a development plan, so borrowing capacity starts to track commodity receipts and export concentration. That shape is visible in Egypt's debt dynamics, where sovereign borrowing inside the domestic banking system crowds out the private credit that would otherwise fund investment.

Where the crowd-out shows up first

The report draws a link between debt burden and food affordability, noting that global shocks hit food availability harder in heavily indebted countries. That is not a coincidence of geography. Imported food and imported fuel are paid in dollars, and dollars compete with debt service for the same reserves.

Once reserves are earmarked for creditors, a currency depreciation transmits straight into the food basket instead of being absorbed by the central bank. Food inflation then becomes the channel through which external debt reaches household welfare, which is why the report places nutrition security next to debt statistics rather than in a separate section.

The concessional lending architecture was built to interrupt exactly this sequence. Its weakening is the substantive change in this year's report, more than the record total.

What the G20 debt drive can and cannot do

The proposals under discussion involve mobilising balance sheets of the World Bank and other development banks to buy back outstanding debt, extending maturities and lowering coupons. That addresses the price of the stock. It does not address the price of new flows, which is set by markets and by the borrower's own risk premium.

A buyback at a discount also transfers value from creditor to debtor only if the discount is real, and creditors holding a performing bond have little reason to accept one. The 2005 precedent does not carry over, because the parties are not the same.

What the G20 can supply is coordination and a common disclosure standard. What it cannot supply is a lower policy rate in the creditor countries that sets the base for every one of these coupons. For a fuller picture of the institutions involved, the IMF and World Bank relationship remains the starting point.

The three options a finance ministry actually has

Faced with an 8.2 percent refinancing rate, a government can raise revenue, cut spending or borrow more. In a low-income economy the first means taxing consumption that is already thin, and the second in practice means cutting capital projects first, because they have no constituency in the current budget year. The third is the default, and that is the trap: borrowing to pay interest raises the debt stock without adding the productive capacity that would service it.

There is a fourth option that appears in few budget documents, which is delay. Deferring a supplier payment improves this quarter's cash position and creates a larger liability later. Programme design treats arrears accumulation as an early warning for exactly this reason, since a government missing small domestic payments is signalling that the external ones are next.

The domestic transmission rarely gets stated. When a sovereign pays 9 percent, that yield becomes the benchmark against which banks price corporate and household credit, and a sovereign premium does not stay sovereign. Lending to the state at 9 percent with a capital preference beats lending to a small firm at 12 percent with credit risk attached, so credit is rationed away from the borrowers who create jobs. That is crowding-out in its clearest regional form.

Transparency as an operational problem

The report's consistent theme is that debt is undercounted, because state-owned enterprise borrowing, collateralised resource deals and hidden guarantees move liabilities off the central government's books. Both a buyback and a restructuring depend on knowing the full creditor list, and that list is what transparency gaps hide. A country that does not know what it owes cannot negotiate a schedule, and an unidentified creditor cannot be bound by a majority vote.

The consequence is that the headline total will keep rising partly because reporting is improving, which makes the record figure a poor measure of deterioration. The coupon is the cleaner signal.

What to watch

Three markers will show whether the 8.2 percent is a peak or a plateau. First, whether 2026's preliminary figures are revised down once full-year data arrives, since a plateau at this level makes refinancing a permanent budget line rather than a cyclical squeeze. Second, whether the buyback proposals produce a completed transaction, not an announced framework. Third, whether the composition of new lending shifts back toward official creditors, which would signal that private lenders have repriced risk rather than merely repriced price.

The distinction to hold onto is between the dollar total and the coupon. Debt stocks rise in good years and bad. Borrowing costs reveal what lenders actually believe about being repaid, and this year they are charging the poorest borrowers the most since the 1980s.