The blind spot in sovereign debt data
AidData's Chinese PPG Loan Performance Dataset, Version 3.0 is the only source granular enough to examine a single loan and current enough to reflect what has happened to that loan since it was issued.
$804B
in PPG lending commitments, 2000-2023
3,304
PPG loans, each with three repayment schedules
127
borrowing countries
139
Chinese state-owned creditors tracked
Beijing is now the largest official creditor to low- and middle-income countries. As grace periods expire and global interest rates remain elevated, many sovereign borrowers are entering their peak repayment years with Chinese state-owned lenders.What these borrowers owe is largely undocumented.
Recent research finds that about half of China's state lending commitments to PPG borrowers go unreported in the World Bank's International Debt Statistics (IDS), or they are recorded without attribution to Chinese creditors. Beijing publishes nothing on how its loans perform, and most of its contracts bar borrowers from disclosing the terms.
IDS — the source that finance ministries, ratings analysts, and rival lenders treat as the official record — publishes nothing at the loan level, covers a limited set of debt instruments, and underreports public-sector borrowing outside the central government. Nor does any other source show what borrowers owe China loan by loan, or what falls due in any given year. As a result:
- Sovereign exposure is understated.
- Liquidity risk is underestimated.
- Debt sustainability analyses rest on unrealistic assumptions.
- Restructuring outcomes cannot be verified.
- Comparability of treatment cannot be enforced.
Risk exposure that official statistics do not capture
More than double the debt pile
PPG debt stock, end-2024
$313.3B vs. $148B in the IDS
+112% above official statistics
Official statistics understate exposure by a wide margin. Low- and middle-income countries held $313.3 billion in PPG debt to Chinese official-sector creditors at end-2024, against $148 billion in the IDS.
The gap begins with which loans get counted. AidData records $804 billion in commitments against $331 billion in IDS, roughly $2.40 committed for every dollar recorded.
Understated repayment pressure
Projected debt service, 2026
$80B vs. $26B in IDS
More than three times the IDS
Repayment pressure is cresting now. About $268 billion is expected to remain outstanding in 2026, and $80 billing falling due that year alone. Pressure peaks in 2027 as PBOC swap drawdowns come due, assuming they are not rolled over as in the past.
Over one-third of the portfolio is impaired. $286.6 billion of PPG lending commitments show arrears, default, or some other indication of financial distress. Twenty-three borrowers carry a high-risk classification.
China is now a net collector. Net transfers turned negative around 2019: borrowers repaid $185.5 billion more than they received across 2020–2026, and net transfers are projected to stay negative through 2035.
Questions that only loan-level data can answer
Each answer is traceable to the loan records behind it.
Investors, creditors & rating agencies
How much Chinese debt is missing from a sovereign's published figures?
Loan-level stocks and flows for 127 borrowers, reconcilable against IDS, IMF, and government reporting.
Which borrowers are closest to a repayment wall?
Debt service and arrears trajectories compared across 127 borrowers, with peak years identified.
Governments & international institutions
What did a restructuring actually deliver?
Cash-flow relief by year, present value reduction, maturity and grace extension, and the treatment of arrears and penalty fees.
What falls due, to whom, and when?
Forward debt service by year and by creditor, with arrears accumulation tracked as it builds.
Researchers & analysts
What did each loan actually cost?
Loan-year cash flows for 3,304 loans and 9,912 repayment schedules, in nominal and present value terms.
How concessional was the lending, in practice?
Grant element at signing against grant element achieved, by creditor and vintage, at a 5% discount rate.
Repayment paths
Loan Performance 3.0 takes the PPG subset of China's Global Loans and Grants (CLG-Global) Dataset and reconstructs what actually happened to each loan after signing. Military and security loans, loans routed through offshore affiliates of Chinese banks, and People's Bank of China swap drawdowns are all included.
AidData rebuilds the record from the borrower's side, country by country, drawing on public debt registers, audited national accounts, central bank reports, IMF surveillance documents, borrowing-SOE financial statements, and unredacted loan and restructuring contracts. More than 17,000 observed data points document actual disbursements, repayments, arrears, outstanding balances, and restructuring terms.
Every loan carries three parallel repayment schedules. Reading them against one another shows the gap between the loan as signed and the loan as it has performed.
What happens when borrowing terms or loan flows have no information available?
AidData's detailed Loan Performance 3.0 methodology document sets out what we do when loan terms are not disclosed or when observed data points are unavailable. Flags for both scenarios are built directly into the data files, so users can see when information is estimated and to what extent.
How are the data structured?
Four linked worksheets, joined on one loan record id
A user can start from a country total and drill down to an individual repayment behind it, or start from one loan and roll up.
Kenya’s USD-to-RMB Debt Conversion Was Really a Restructuring
When Kenya announced in October 2025 that it had converted its Standard Gauge Railway debts to China Eximbank from U.S. dollars (USD) into Chinese renminbi (RMB), the move was widely framed as a breakthrough for RMB internationalization and a clever way to reduce borrowing costs. Early reporting suggested the switch could save Kenya roughly $215 million a year due to interest rate reductions associated with RMB.
But this narrative does not tell the full story. In a new policy note, AidData researchers Sailor Miao and Oshin Pandey find that the largest source of debt relief did not come from the benchmark rate change alone.
