42 External debt and sovereign risk

Chapter 42 of 60 in the Bangladesh 2036 research base. Contents of the series.

The debt book is affordable on every ratio that matters, and the decade's job is to stop the interest clock from outrunning the export clock

Bangladesh enters FY26 with an external debt stock of 104,486.9 million USD at end-2024, up 82.9 percent from 57,125.5 million USD at end-2018, yet equal to only 22.25 percent of gross national income [WB IDS 2024]. The thesis of this chapter is that for 2026 to 2036, the window the chapter 15 scenarios assume, the sovereign risk question is not a stock question but a clock question: the stock is moderate by every cross country benchmark, the present value of external debt stands at 57,653.7 million USD, 12.28 percent of GNI [WB IDS 2024], and a fifth of the total book remains IDA concessional at its core. What moved, and moved fast, is the flow: total external debt service reached 3,371.6 million USD in FY24, 111.7 percent above FY19 [ERD 2024], the interest component nearly tripled in two years to 1,349.8 million USD [ERD 2024], all three rating agencies cut the sovereign into the single B category in 2023 and 2024 [SP 2024] [Moody 2024] [Fitch 2024], and the IMF programme signed in January 2023 converts the country from a grant and IDA borrower into a borrower with amortisation dates for the first time since the 1990s [BB AR 2024]. Chapter 03 documented the external repair and chapter 04 the fiscal constraint; this chapter goes one level deeper into the debt book itself, who holds it, what it costs, when it falls due, and what would make it unpayable. The claim it defends: Bangladesh can service this book through FY36, the horizon the chapter 15 scenarios assume, if export and remittance growth stays above the interest growth rate that the post graduation borrowing terms will set, and cannot if it does not.

Where the debt stands: a book that doubled in six years, still anchored by IDA, with the public external share of GDP at 15.5 percent

The stock record has three layers that must not be confused. The widest is the World Bank debtor system measure of total external debt, public and private, long and short term: 57,125.5 million USD at end-2018, 91,477.6 million at end-2021, 104,486.9 million at end-2024 [WB IDS 2024]. Of the 2024 total, public and publicly guaranteed debt is 77,581.7 million USD, private nonguaranteed debt 9,292.9 million USD, short term debt 12,945.2 million USD and use of IMF credit 4,667.1 million USD [WB IDS 2024]; the four components sum exactly to the total. The public and publicly guaranteed share is 74.25 percent and the short term share 12.39 percent, the lowest since 2019, down from a 19.77 percent short term peak in 2021 when trade finance piled up during the reserve defence [WB IDS 2024]. The narrowest layer is the one the ERD publishes on a fiscal year basis: public external debt at 15.5 percent of GDP in FY23, up from 12.8 percent in FY18 [ERD 2023]. Chapter 04 adds the top layer, the IMF measure of general government gross debt at 41.8 percent of GDP in the latest snapshot [IMF WEO 2026], most of which is domestic. The external book is the small half of sovereign risk in stock terms; the mechanism section shows why it is the fast half in flow terms.

Debt carrying capacity is a flow story and the flow indicators sit inside every threshold the World Bank and IMF use. The present value of external debt is 105.73 percent of exports of goods, services and primary income and 12.28 percent of GNI at end-2024 [WB IDS 2024], against the joint IMF-World Bank Low-Income Country Debt Sustainability Framework thresholds for the strong debt-carrying-capacity category, 240 percent of exports and 55 percent of GDP [IMF Art IV 2025], so the stock passes comfortably by a wide margin on the exports ratio and by an even wider margin on the GDP-basis ratio. Short term debt to reserves, the liquidity stress gauge, peaked at 65.10 percent in 2023 and eased to 60.51 percent in 2024 as the reserve rebuild chapter 03 recorded took hold [WB IDS 2024]. The ERD service ratios tell the same pressure story from the budget side: debt service absorbed 12.76 percent of revenue and grants in FY23, up from 8.3 percent in FY18, and 5.8 percent of exports plus remittances, up from 3.9 percent in FY18 [ERD 2023]. Total service ran 1,592.5 million USD in FY19, 2,017.0 million in FY22, 2,670.2 million in FY23 and 3,371.6 million in FY24, split in FY24 between 2,021.8 million USD of principal and 1,349.8 million of interest [ERD 2024]. Chapter 03 and chapter 04 carry these flow figures; they are restated here because the creditor structure below is what explains them.

The creditor core is still official and still concessional. IDA credits alone are 20,730.7 million USD of the end-2024 stock, 19.84 percent of the total and more than a quarter of the public and publicly guaranteed book [WB IDS 2024]. The ERD development partner tracker records an active official financing portfolio of 25,399.0 million USD as of July 2026, of which the World Bank holds 14,080.0 million USD and the Asian Development Bank 10,800.0 million USD, so the two multilateral banks carry 98 percent of the active portfolio; cumulative commitments across 32 tracked partners stand at 58,318.7 million USD over 135 verified financing events [ERD 2026]. Behind the active portfolio sits an undisbursed pipeline of 42,849.7 million USD at end-FY24 [ERD 2024], the contractual tail of signed but unspent loans that will become debt service on a schedule set years before disbursement. New commitments keep the pipeline full: 10,720.0 million USD committed and 9,892.0 million USD disbursed in FY24 [ERD 2024].

The bilateral and commercial layers are where the outline's creditor names live, and where the local record is thinnest. Japan's JICA and China's Export Import Bank are the two large bilateral loan channels the ERD tracker monitors, and the tracker records mid 2026 meetings on Panda bond issuance and cross border payment integration with China Exim [ERD 2026], but no country by country creditor stock table is held in the data lake, so bilateral shares of the public external stock by creditor country could not be confirmed and the ERD external debt bulletin is the resolving source. What the record does show is the project expression of the two bilateral relationships. The Rooppur nuclear plant, financed by a Russian state credit whose terms are not stored here and could not be confirmed, is the largest single project in the development programme at 138,685.77 crore BDT, with 94,689.26 crore BDT spent to the start of FY26 and a 10,011.78 crore BDT allocation in FY26 [Planning Commission ADP 2025]. Bangladesh Bank's FY24 accounts record an escrow account created for Rooppur loan repayment that nearly doubled in one year, from 3,496.84 crore BDT at end-June 2023 to 6,929.78 crore BDT at end-June 2024, and foreign exchange settlements for the project of 879.1 million USD in FY24 [BB AR 2024]. The commercial layer is measurable: private nonguaranteed debt of 9,292.9 million USD [WB IDS 2024], short term trade related debt of 12,945.2 million USD [WB IDS 2024], and no international sovereign bond line on the record available to this chapter, so the state prices nothing off a market yield curve of its own; the bond record could not be confirmed and the ERD debt bulletin the resolving source.

The domestic interest bill is the other half of sovereign risk and it now dominates the payment arithmetic. Actual interest service in FY25 was 134,430 crore BDT, of which domestic interest was 116,617 crore BDT, 86.8 percent, leaving a derived 17,813 crore BDT of foreign currency interest; foreign debt amortisation was a further 30,835 crore BDT in the same year [MoF iBAS 2025]. Chapter 05 traces why the domestic number is compounding, the move from the lending cap to market rates and the state's short maturity bank held borrowing that chapter 06 measures; the point here is distributional: for every taka of external interest in FY25 the state paid roughly 6.5 taka of domestic interest, and the domestic bill re-prices within one refinancing cycle while the external bill is contractually fixed in foreign currency.

Mechanism: three borrowing waves built the book, the third is still being drawn, and all three now carry floating or post concessional cost

Wave one is the concessional base. Through the 2010s the external book was an IDA, ADB and JICA portfolio at near zero interest, and the record shows it: interest was 391.77 million USD of the 1,592.5 million USD FY19 service total, under a quarter of the flow [ERD 2024], which is what made the stock ratios of the era close to costless. Wave two is the megaproject semi concessional build out of FY17 to FY22, the borrowing wave behind chapters 10 and 24: the stock jumped from 62,468.3 million USD at end-2019 to 91,477.6 million at end-2021 [WB IDS 2024], as Chinese buyer credit, Russian project credit and stiffer multilateral blend entered the book and grace periods deferred the service. Wave three is the crisis and programme wave from FY23: interest paid nearly tripled from 492.1 million USD in FY22 to 935.7 million in FY23 and 1,349.8 million in FY24 [ERD 2024], driven jointly by the arrival of IMF credit priced at market linked charge rates, 102.21 million SDR of charges paid in FY24 alone [BB AR 2024], by the end of the zero interest era on new commitments as graduation approaches, the terms shift chapter 02 documents, and by the taka's 2024 float, which raised the taka cost of every foreign currency coupon as chapter 05's exchange rate record shows.

The programme is now the largest single creditor relationship and its arithmetic sets the decade's repayment path. The IMF Executive Board approved 1,645.64 million SDR under the Extended Fund Facility, 822.82 million SDR under the Extended Credit Facility and 1,000.00 million SDR under the Resilience and Sustainability Facility on 30 January 2023; by June 2024 Bangladesh had drawn 939.60, 469.80 and 333.35 million SDR respectively, a derived 1,742.75 million SDR, and total outstanding principal to the Fund stood at 2,153.14 million SDR [BB AR 2024]. The same record targets the start of repayment at 29 July 2027 for the EFF, 1 August 2028 for the ECF and 14 June 2034 for the RSF, on top of legacy facility repayments already running at 241.75 million SDR in FY24 [BB AR 2024]. This new Fund amortisation schedule lands in fiscal years FY28 and FY29, the same window in which the Rooppur construction programme targets completion on 30 June 2028 [Planning Commission ADP 2025] and in which chapter 02's graduation clock removes LDC borrowing terms. The programme's review by review quantitative conditions, reserve floors and deficit ceilings, are not stored in the data lake and could not be confirmed; the IMF review documents are the resolving source [IMF Art IV 2025].

The market's verdict moved before the fundamentals finished moving, and it is on the record. S&P held the sovereign at BB minus from 2010, moved the outlook to negative in July 2023 and cut to B plus in April 2024; Moody's cut from Ba3 to B1 in May 2023 and to B2 in May 2024; Fitch held BB minus from 2014 and cut to B plus in May 2024 [SP 2024] [Moody 2024] [Fitch 2024]. The five year credit default swap spread, the closest thing the sovereign has to a market price, widened from 285 basis points in 2021 to 605 basis points in 2023, with the implied yield near 9.5 percent, before easing to 480 basis points in 2024 [IMF AIV 2024]. The sequence matters for the decade because it prices the post graduation borrowing menu: at 605 basis points over the dollar curve, voluntary market borrowing is closed for a sovereign whose exports must also finance an import bill chapter 03 sizes, so the realistic financing set for FY26 to FY36 is official semi concessional credit, the multilateral banks, and bilateral project credit, with any market issuance a signal decision rather than a financing need.

The rating action of 2023 to 2024 was not a shock event but a lagging confirmation of the same three mechanisms the debt data show: the interest line tripling, the reserve drawdown chapter 03 documented from 46.2 billion USD at end-2021 to 21.4 billion at end-2024 [IMF IIP 2024], and the political transition shock of 2024. The recovery leg since, BPM6 reserves of 26,603.7 million USD at end-September 2025 [BB Econ 2025], has not yet been confirmed by a rating upgrade in the record available to this chapter, which ends at the May 2024 actions; later actions could not be confirmed and the agency publication pages are the resolving source.

The decade ahead: two repayment humps, a graduation terms shift, and a financing identity that must move from official to mixed

The IMF projects general government gross debt at 40.2 percent of GDP in 2026, rising through 41.2 percent in 2027 and 42.5 percent in 2028 to 44.2 percent in 2030 [IMF WEO 2026], a gentle upward drift on the calendar year basis, driven by the domestic primary deficit and the external interest arithmetic above, not by any stock event. Layer the scheduled events on that path and the decade's financing identity becomes concrete. First hump: IMF amortisation from the July 2027 date the programme record targets [BB AR 2024], adding a creditor that must be repaid in scarce foreign currency to the annual service bill, on top of the 3,371.6 million USD FY24 base [ERD 2024]. Second hump: the Rooppur credit, where the escrow pre funding chapter shows the state is already paying ahead of schedule, 6,929.78 crore BDT banked by end-FY24 [BB AR 2024], against a project 68.3 percent complete on the cumulative spend ratio [Planning Commission ADP 2025]. Third force: the terms shift of graduation, which chapter 02 dates to the 2026 to 2029 window and which the IMF and World Bank assessments of 2025 project will raise average interest and shorten maturities on new external borrowing [IMF Art IV 2025] [WB BDU 2025].

Three decisions decide whether the identity holds. The financing mix decision, taken by the ERD and the Finance Division in the next programme cycle, is how much of the 42,849.7 million USD pipeline [ERD 2024] is restructured toward grants and highly concessional tranches before the graduation terms lock, and whether any market instrument, the Panda bond exploration the partner tracker records [ERD 2026] or a debut sovereign issuance, is used to term out rather than to spend. The contingency recognition decision, whether the SOE arrears machine chapters 04 and 41 document, with consolidated SOE liabilities already at 3.46 trillion BDT in FY23 [MoF Budget 2023], is recognised as sovereign debt or kept outside the perimeter, decides the true stock. The escrow institutionalisation decision, whether the Rooppur pre funding model becomes the standard for all non concessional project credit, converts a risk mechanism into a routine; the FY26 development budget still carries 60,290.6 crore BDT of project aid [Planning Commission ADP 2025], much of it on the semi concessional terms whose service this chapter has costed.

Risks compound in the service lines; the upside is a concessional core, a floating currency and a pre funding habit

Risks. First, the interest outrun: if the interest line keeps compounding at the FY22 to FY24 pace while exports grow at the 1.3 to 8.6 percent range chapter 03 recorded for FY23 to FY25, the service ratio passes the 12.76 percent of revenue level [ERD 2023] toward a quarter of revenue by the decade's end, and the revealing indicator is the annual ERD interest line against EPB export growth. Second, the liquidity squeeze of 2027 to 2029 that the repayment schedule targets: IMF amortisation, Rooppur principal and the graduation terms shift arrive together, and a reserve buffer that chapter 03's watch list puts at 26,603.7 million USD BPM6 [BB Econ 2025] would erode fast under a simultaneous import rebound; the revealing indicators are the BPM6 reserve floor compliance and the ERD disbursement series, which falls in financing squeezes before any default signal appears. Third, contingent liability crystallisation: the 3.46 trillion BDT SOE liability stock [MoF Budget 2023], the power sector arrears chapters 10 and 41 trace, and the guaranteed project debt inside the 77,581.7 million USD public and publicly guaranteed book [WB IDS 2024] could move onto the sovereign balance in one accounting decision, and the revealing indicator is the consolidated SOE liabilities series against nominal GDP growth.

Upside. First, the concessional core buys time that peers at this rating do not have: IDA at 19.84 percent of the stock [WB IDS 2024] and the two multilateral banks at 98 percent of the active portfolio [ERD 2026] mean the marginal cost of external financing is a policy variable, not a market one, for most of the book. Second, the float converts solvency risk into price risk: a currency that adjusts, as chapter 03's 2024 to 2025 record shows, services a fixed foreign currency coupon out of a growing taka revenue base automatically, and the 26.8 percent remittance surge of FY25 [BB Econ 2025] is the proof that the adjustment channel works. Third, the pre funding habit: the Rooppur escrow doubling to 6,929.78 crore BDT in one year [BB AR 2024] is exactly the reserve against the first repayment hump that a B plus rated sovereign is not expected to hold, and institutionalising it would move the sovereign risk debate from affordability to optimisation.

What to watch: five indicators whose thresholds mark the sovereign risk regime

1. Total external debt service. Current value 3,371.6 million USD in FY24 [ERD 2024]. Threshold: a year past 5,000 million USD before the graduation terms lock, the level chapter 04 flags, signals the service flow has outrun the financing programme; a print below 3,500 million USD in the fiscal years before the amortisation dates the programme targets signals the hump is being absorbed.
2. External interest paid. Current value 1,349.8 million USD in FY24, nearly tripled from FY22 [ERD 2024]. Threshold: a first year above 1,800 million USD confirms the post graduation pricing regime; a fall back under 1,200 million USD signals new commitments are staying concessional.
3. Short term debt to reserves. Current value 60.51 percent at end-2024, down from a 65.10 percent 2023 peak [WB IDS 2024]. Threshold: a move above 80 percent is the FY22 stress mechanism returning; a sustained print under 40 percent marks a genuine liquidity regime.
4. General government gross debt. Current value 41.8 percent of GDP in the latest IMF snapshot [IMF WEO 2026], on a projected path to 44.2 percent in 2030 [IMF WEO 2026]. Threshold: a print above the IMF projected path by more than 3 percentage points, or any SOE recognition event that jumps the stock, marks the contingent liability regime arriving.
5. The market price of the sovereign. Current values: five year credit default swap spread of 480 basis points in 2024 [IMF AIV 2024] and a B plus stable rating from Fitch and S&P with B2 stable from Moody's as of May 2024, the last actions in the record available to this chapter [Fitch 2024] [SP 2024] [Moody 2024]. Threshold: any second agency moving to single B flat, or a spread back above 600 basis points, closes the market borrowing option and marks the financing regime change; a return to BB category confirms the repair chapter 03 documented.

A note on the set: four of the five print annually or better from published official series and the fifth from agency actions; none requires new statistical capacity; and together they answer the decade's sovereign question, whether the clocks can be kept in the order exports first, interest second, stock last.

Sources used

[WB IDS 2024] World Bank International Debt Statistics via the finance/wb_external_debt_bd parquet, series: DT.DOD.DECT.CD, DT.DOD.DECT.GN.ZS, DT.DOD.DPPG.CD, DT.DOD.DPNG.CD, DT.DOD.DSTC.CD, DT.DOD.DSTC.ZS, DT.DOD.DSTC.IR.ZS, DT.DOD.DIMF.CD, DT.DOD.MIDA.CD, DT.DOD.PVLX.CD, DT.DOD.PVLX.EX.ZS, DT.DOD.PVLX.GN.ZS, calendar years, end-2024 latest. [ERD 2024] Economic Relations Division, Ministry of Finance, debt service and aid series via bdpolicy.db, series: erd_debt_service_total_usd_mn, erd_debt_service_principal_usd_mn, erd_debt_service_interest_usd_mn, erd_foreign_aid_commitment_usd_mn, erd_foreign_aid_disbursement_usd_mn, erd_cumulative_foreign_aid_pipeline_mn_usd. [ERD 2023] Economic Relations Division, Ministry of Finance, external debt and ratio series via bdpolicy.db, series: erd_foreign_debt_gdp_pct, erd_debt_service_revenue_pct, erd_debt_service_export_remittance_pct. [ERD 2026] Economic Relations Division development partner tracker via bdpolicy.db, snapshot of 5 July 2026, series: donor_tracker_total_active_portfolio_usd, donor_tracker_total_cumulative_commitments_usd, donor_tracker_partners_monitored_nos, donor_tracker_verified_events_nos, and per partner records for the World Bank, the Asian Development Bank and the Export Import Bank of China. [MoF iBAS 2025] Finance Division iBAS++ actuals via bdpolicy.db, series: mof_sovereign_debt_interest_service, mof_domestic_debt_interest_service, mof_foreign_debt_amortization, fiscal year actuals to FY25. [IMF WEO 2026] IMF World Economic Outlook database: imf_gross_debt_pct_gdp via bdpolicy.db indicators snapshot updated 2026-09-04; imf_weo_gross_debt_general_government_percent_of_gdp via data/curated/tradeweave_imf/imf_weo_bd.parquet (the identically named bdpolicy.db data_series row holds no data_points and is not the resolving path), calendar years with projections to 2030. [BB AR 2024] Bangladesh Bank Annual Report 2023-24, chapter 12 transactions with the IMF and the notes to the financial statements: programme approvals and drawings in SDR, repayment timetable, escrow account for Rooppur loan repayment, and the foreign exchange settlement record. [SP 2024] S&P Global Ratings sovereign rating actions for Bangladesh via the finance/sovereign_ratings_bd parquet, actions of 2010 to 26 April 2024. [Moody 2024] Moody's Ratings sovereign rating actions for Bangladesh via the finance/sovereign_ratings_bd parquet, actions of 2010 to 29 May 2024. [Fitch 2024] Fitch Ratings sovereign rating actions for Bangladesh via the finance/sovereign_ratings_bd parquet, actions of 2014 to 21 May 2024. [IMF AIV 2024] IMF Article IV market risk indicators via the finance/bd_sovereign_spreads parquet, series: bd_sovereign_spreads, five year credit default swap estimates compiled from Bloomberg, 2017 to 2024. [Planning Commission ADP 2025] Programming Division, Planning Commission, RADP FY2025-26 project table via bdpolicy.db adp_projects, Rooppur project cost, cumulative expenditure and FY26 allocation, and project aid totals. [MoF Budget 2023] Finance Division consolidated SOE evaluation, SPFMS Monitoring Cell via bdpolicy.db, series: soe_consolidated_total_liabilities_bdt_m. [BB Econ 2025] Bangladesh Bank Monthly Economic Trends, October 2025, foreign exchange reserves on BPM6 basis and workers' remittances. [IMF IIP 2024] IMF International Investment Position via bdpolicy.db, series: imf_iip_reserve_assets_total_usd_m. [IMF Art IV 2025] IMF Article IV and ECF, EFF and RSF review documents for Bangladesh, resolving source for review level quantitative conditions and the projected borrowing terms shift; also the joint IMF-World Bank Low-Income Country Debt Sustainability Framework debt burden threshold table by debt-carrying-capacity category published on imf.org, accessed 2026-09-06, source for the strong-category thresholds of 240 percent of exports and 55 percent of GDP. [WB BDU 2025] World Bank Bangladesh Development Update, external financing outlook.

---

Verified line by line against primary sources: 68 claims checked, 2 corrected.

Previous: 41 State-owned enterprises and subsidies

Next: 43 Exchange rate, reserves and capital account

All sixty chapters and the ten narrative chapters.