The developing world's debt service climbed from 8.1% to 12.4% of export earnings in a decade — and across the 114 economies the World Bank scores, the heaviest loads sit on middle-income borrowers, not the poorest

The United Nations gauges whether a country's external borrowing is sustainable with a single ratio — Sustainable Development Goal indicator 17.4.1, debt service as a share of exports. It sets no numeric target; it is a pressure gauge, not a pledge. The World Bank operationalizes it as DT.TDS.DECT.EX.ZS: total debt service — principal and interest actually paid on external debt — as a percentage of exports of goods, services and primary income. We pulled that series, updated 13 July 2026, and kept every economy with a 2023 value: 114 of them. Almost all are low- or middle-income; only three high-income economies report at all, because the underlying figures come from the World Bank's Debtor Reporting System, which the rich world largely sits outside. This is, in effect, a developing-world ledger. The median economy paid 12.4% of its export earnings to service external debt in 2023. A decade earlier, in 2013, the median was 8.1%. Among the 111 economies reporting in both years, the median climbed from 8.3% to 12.6%, and 81 rose against 30 that fell. The World Bank's low- and middle-income aggregate moved from 9.2% to 15.4% over the same span; its poorest, IDA-only group from 8.1% to 11.9%. Twenty-eight of the 114 now spend 20% or more of export earnings servicing external debt; 15 spend 30% or more; 7 spend 40% or more.

The following is analysis, not fact. The frame a single-country desk reaches for is that debt distress tracks poverty — that the deepest holes are the poorest countries. Read all 114 filings at once and that is not what the ledger shows. Sort by income and the low-income median is 10.5%, below the lower-middle (13.3%) and upper-middle (13.4%) tiers. The extreme top of the list is entirely upper-middle-income: El Salvador at 83.1%, Brazil at 53.5%, Argentina at 51.4%, Kazakhstan at 50.3%. The only low-income country in the top six is Mozambique, at 46.2%. The reading this is consistent with — not one these data isolate — is that the poorest borrow concessionally: IDA credits and bilateral loans at low rates and long maturities, some under active debt relief, which keep the export ratio down. Middle-income economies borrow more at market terms — Eurobonds and commercial credit that repay harder and faster. We did not test it: no creditor-composition or interest-rate variable was joined, so treat it as a hypothesis the pattern fits, not a mechanism. ^sspeculative One caveat the whole-record read makes visible: single-year ratios are lumpy, because principal repayments spike when big bonds mature. El Salvador swung from 32.2% in 2022 to 83.1% in 2023 as debt came due. The robust signal here is the aggregate decade climb and the income-tier ordering — not any one country's rank in any one year.

The number depends on which slice of debt you count. Our 12.4% is total debt service — public and private, principal and interest. UNCTAD's widely cited figure, that half of developing countries pay at least 6.5% of export revenues, measures external public debt service only, a narrower slice, which is why its median runs lower; the World Bank's roughly-6% headline is IDA interest payments alone. Three different concepts, three different numbers — each correct about its own cut, none directly comparable. And a ratio can rise two ways: debt service climbing, or exports falling — a commodity bust flatters no borrower. A high reading is also not automatically a crisis. A country rolling cheap debt with deep reserves is not the country that has lost market access, even at the same 30%. The gauge flags pressure; it does not diagnose solvency, and it should not be read as if it did.

**View from a middle-income finance ministry.** A ministry could read part of the burden as a consequence of graduation and market access. Graduate out of the poorest tier and the concessional windows narrow; you fund yourself in the bond market instead, at market rates rather than the low, long concessional terms the poorest still get — so the same access that signals you have arrived is the access that puts you near the top of a list of who spends the most of their exports paying creditors back. This is a plausible reading of the ordering, not a mechanism these data isolate; the analysis did not test it. ^sspeculative