Stress-Test Bangladesh's China-Financed Infrastructure Debt Before the Repayment Cliff Hits

Diagnosis

The note names three large China-linked infrastructure obligations now moving from construction into repayment: Padma Rail, the Karnaphuli tunnel, and Payra port debt service. These are not abstract macro risks. They are dated contractual obligations with grace periods that end, interest that accrues, and revenue assumptions that were made years before the assets opened. The label frames this as a debt-trap scenario, which is the right lens: the danger is not the borrowing itself but the combination of concentrated single-creditor exposure, foreign-currency repayment, and assets whose actual earnings may not cover their own debt service.

The context records no current indicator value for this risk, which is itself the most important finding. If the lead body cannot put a single number on consolidated China-linked debt service over the next several years, it cannot manage it. The window to act is now, before grace periods on these projects lapse and several repayment streams stack into the same fiscal years. The Ministry of Commerce (MoC) is the named lead responsible body, with supporting roles for the Bangladesh Investment Development Authority (BIDA), Bangladesh Standards and Testing Institution (BSTI), Bangladesh Trade and Tariff Commission (BTTC), and Chittagong Port Authority (CPA).

Recommended actions

1. Build a consolidated obligation register. Owner: MoC, coordinating with CPA for the Payra and port-related streams. Mechanism: a standing project-by-project ledger for Padma Rail, Karnaphuli tunnel, and Payra port debt service, recording principal, grace-period end, repayment schedule, and currency of repayment. Signal it is working: a single living document that the leadership can query, replacing scattered project files.
2. Run a forward debt-service stress test. Owner: MoC, with BTTC providing the trade and tariff revenue assumptions. Mechanism: a scenario model that projects combined annual repayments under a base case and an adverse case (currency depreciation, revenue shortfall on the financed asset). Signal: a published year-by-year repayment profile showing which fiscal years carry the heaviest stacked obligations.
3. Open structured renegotiation on the highest-risk stream. Owner: MoC as lead negotiator. Mechanism: a formal request to the creditor to reprofile the obligation whose stress-test result is worst, seeking extended tenor or revised grace terms before the repayment cliff. Signal: a signed amendment or a documented creditor response, not an informal assurance.
4. Tie each asset's revenue to its own debt service. Owner: CPA for Payra and port assets, MoC for coordination. Mechanism: ring-fence the operating revenue of each financed asset and report it against that asset's repayment line. Signal: a per-asset coverage ratio reported each cycle, so a shortfall is visible before it becomes a default.
5. Set a single-creditor exposure ceiling for future deals. Owner: MoC, with BIDA screening new investment proposals. Mechanism: a circular requiring any new large infrastructure financing to disclose creditor concentration and pass a debt-service test before approval. Signal: new project approvals that carry an attached affordability assessment.

Sequencing (first 12 months)

Start with the obligation register (Action 1): nothing else is credible without it, and it is the cheapest step. Once the register exists, the stress test (Action 2) becomes a mechanical exercise rather than a guess. The stress test then tells MoC which single stream to renegotiate first (Action 3), so negotiating capital is spent where it matters most. Per-asset coverage reporting (Action 4) and the exposure ceiling (Action 5) institutionalize the discipline so the next cycle of projects does not recreate the same blind spot.

Risks and constraints

The binding constraint is negotiating leverage: a single dominant creditor holding multiple obligations can resist reprofiling, and Bangladesh's request signals stress. Foreign-currency repayment means a depreciation can worsen the burden faster than any domestic action can offset. Politically, projects already opened are sources of prestige, so acknowledging that an asset cannot service its own debt is uncomfortable. Fiscally, if the named streams stack into the same years, even an accurate register cannot conjure repayment capacity it does not have; the register only ensures the shortfall is seen in time to manage it.

Bottom line

Bangladesh's China-financed infrastructure obligations on Padma Rail, the Karnaphuli tunnel, and Payra port are entering repayment with no consolidated number attached to them, which is the core management failure. MoC should build the obligation register, stress-test the combined repayment profile, and renegotiate the worst stream before grace periods lapse and the repayment cliff arrives.