Measuring the Financial Performance of China's Sovereign Lending Portfolio

AidData’s Chinese PPG Loan Performance Dataset 2.0 addresses this major evidentiary gap. It is a loan-level dataset that tracks the disbursement, repayment, arrears, restructuring, and outstanding debt trajectories of PPG loans issued by Chinese state-owned creditors to low- and middle-income countries. For all loans formally issued between 2000 and 2022, it captures financial performance information (disbursements, repayments, restructuring events) between January 2000 and March 2025.

The dataset provides dynamic amortization schedules that reflect each loan’s full life cycle, including three parallel amortization schedules:

• T0 Planned Schedule, reflecting the repayment path implied by contractual terms at signing.
• Perfect Compliance Schedule, incorporating actual disbursement details and prevailing benchmark interest rates under the assumption of full and timely repayment.
• Actual Performance Schedule, integrating observed disbursements, repayments, arrears, restructurings, and time-stamped outstanding balances to measure actual performance over time.

For institutions that monitor public debt exposure, evaluate sovereign credit risk, or design debt relief operations, the dataset offers a uniquely detailed and comprehensive source of evidence.

Each loan's three amortization schedules roll up into a standardized set of loan-level summary statistics, reported in both nominal USD and net present value terms (discounted at 5% from the year of commitment). Reporting every metric under all three models gives users a direct view of how borrowing costs and concessional depth shifted from what was promised at signing to what the borrower actually experienced.

The summary statistics cover the headline financial flows for each loan: total disbursements, total principal repayments, total interest paid, and total fees (commitment, management, and insurance). For loans that have encountered repayment difficulties, the Actual Performance summary also captures cumulative principal and interest arrears, interest recapitalized into principal during restructuring events, and any debt forgiveness granted by the creditor. Together, these fields make it possible to quantify, loan by loan, the gap between expected and realized debt service.

Three derived metrics translate the underlying flows into standardized indicators of loan cost and concessionality. The all-in cost of borrowing expresses the sum of interest, fees, and recapitalized interest as a share of total disbursements—a single figure that captures the full economic price of credit, including costs that emerge only after arrears or restructuring events. The OECD grant element and grant equivalent measure the concessionality of each loan against internationally recognized benchmarks. Calculating all three indicators across the three models allows users to observe how restructurings, arrears, and rate movements alter a loan's true cost and concessional value over its lifetime.

The evidentiary base behind these metrics is substantial. In total the dataset covers 3,719 PPG loans across 124 low- and middle-income countries, supported by over 15,000 direct loan performance observations drawn from official government and creditor sources.