Skip to content

The Great Reversal

This page describes the methodology and data used for the Great Reversal report produced by ONE Data for the Development Finance Observatory. You can read the report here.

For this research, we use data from:

Data and code to replicate the analysis are available on this GitHub repository.

Inflows

In terms of inflows we focus on:

  • OECD Development Assistance Committee (DAC) data for grants from DAC, non-DAC countries, and multilateral institutions from the Creditor Reporting System (CRS) database.
  • OECD DAC data for loans and other non-grant flows (e.g. private sector instruments) from DAC, non-DAC countries, and multilateral institutions from the CRS. The net flows dataset only includes this data for provider-recipient pairs that are not included in the World Bank International Debt Statistics database.
  • Additional data from the International Forum on Total Official Support for Sustainable Development (TOSSD) for grants, loans, and other non-grant flows from TOSSD reporting parties. The dataset only includes TOSSD data that is not included in the CRS. Since 2024 data is not yet available for TOSSD, when needed we assign the 2020-2023 yearly average (in real terms) to 2024.
  • World Bank International Debt Statistics data for new long-term, external public and publicly guaranteed debt disbursements, including from private sources like bondholders and commercial banks. This does not include foreign direct investment or other forms of private capital flows.

Outflows

In terms of outflows we focus on:

  • Debt service payments (including both principal and interest payments) on PPG long-term external debt (from the World Bank IDS database)

Net flows

For this report, net flows mean PPG debt, grants, and other official inflows minus debt service payments on PPG debt.

All debt is Public and Publicly Guaranteed long-term debt. In other words, “private” debt (for example) is public or publicly guaranteed debt owed to private creditors.

All numbers are presented in constant 2024 US dollars, using currency and price deflators for debtor / recipient countries.

Note

Note Except where noted otherwise, for this analysis we only include low and lower-middle income countries. We include Ethiopia in this group, even though it is not currently classified by income by the World Bank. We exclude Ukraine given the exceptional flows following Russia's invasion.

Additional Notes

On Double-Counting

Because we combine data from multiple sources (DAC/CRS, TOSSD, and World Bank IDS), we apply rules to prevent double-counting flows that appear in more than one database:

  • DAC non-grant flows (loans, Other Official Flows, Private Sector Instruments) are only included for provider-recipient pairs where the counterpart has a DAC code but no World Bank code in our concordance table. This prevents overlap with IDS debt data, which captures the same underlying loans.

  • TOSSD non-grant flows are only included for providers that do not appear in the World Bank IDS database (providers with no wb_code in our concordance).

  • Grants from all sources (DAC and TOSSD) are included without deduplication since grants are not captured in the IDS debt statistics.

Note that we only include the TOSSD data which does not appear on the CRS.

The concordance tables mapping provider codes across systems are available in the repository at source_data/counterpart_concordance.csv and source_data/recipient_concordance.csv.

Coverage of Non-Traditional Providers

China

China does not report to OECD DAC or TOSSD systems. All Chinese development finance data in this analysis comes from the World Bank International Debt Statistics, which captures:

  • Debt disbursements from Chinese bilateral lenders
  • Debt service payments to Chinese creditors

This means our analysis of China data:

  • Captures loans but may miss grants or technical assistance
  • Represents debtor-reported data rather than provider-reported data
  • Does not distinguish between policy banks (e.g., China Development Bank, Export-Import Bank of China), commercial banks, or other Chinese lenders

For more granular data on Chinese development finance commitments, users may consult AidData's Global Chinese Development Finance Dataset, though that dataset tracks commitments rather than disbursements.

Other Non-DAC Providers

Middle Eastern and other emerging providers (including Kuwait, Qatar, Saudi Arabia, Turkey, and UAE) are partially captured through:

  • TOSSD reporting (where they participate as reporting parties)
  • World Bank IDS (debt component, as reported by debtor countries)

Coverage varies by provider and year. These providers have uneven reporting practices, so data completeness cannot be guaranteed.

Price Deflation Methodology

All values are presented in constant 2024 US dollars. Deflation is performed from the debtor/recipient country perspective, which accounts for both price changes and exchange rate movements as experienced by the recipient.

Data Source Deflator Used Library
DAC/CRS data DAC deflators oda_data
TOSSD data DAC deflators oda_data
World Bank IDS data World Bank GDP deflators pydeflate
Net flows aggregations IMF GDP deflators pydeflate

This debtor-perspective deflation means that the purchasing power represented by flows is measured from the recipient country's viewpoint, rather than the provider's.

Country Inclusion and Exclusion Criteria

Income Groups

The analysis focuses on low-income countries (LICs) and lower-middle-income countries (LMICs) based on World Bank income classifications.

Special Cases

Country Treatment Rationale
Ethiopia Included as LIC Ethiopia is not currently classified by the World Bank due to data gaps, but we include it given its significance and historical LIC status
Ukraine Excluded Exceptional inflows following Russia's 2022 invasion would distort trend analysis

Regional Recipients

Data sometimes reports flows to regional programs rather than individual countries (e.g., "West Africa, regional"). These are mapped to a "REGIONAL" entity code and:

  • Excluded from country-level analysis
  • Included in aggregate regional and global totals where appropriate

Provider Classification

Providers are classified as Bilateral, Multilateral, or Private using a concordance table (counterpart_concordance.csv) that maps each provider code to a counterpart type:

Type Definition Examples
Bilateral Sovereign governments providing official finance United States, Germany, Japan, China
Multilateral International organizations including MDBs, UN agencies, and regional development banks World Bank, IMF, African Development Bank, Asian Infrastructure Investment Bank
Private Bondholders, commercial banks, and other private creditors (for PPG debt only) International bondholders, commercial banks