In the third quarter of 2025, the Public Development Banks (PDBs) and Development Financing Institutions (DFIs) database team introduced three key enhancements, including converting the database to a panel data structure, providing a brief analysis of institutions with inaccessible total assets data to call for more transparency, and adopting a new classification system to improve data processing efficiency.

Update the Database to Panel Data Format

A key upgrade was undertaken to the PDB and DFI database, migrating its structure to a standard panel data format to enhance user experience as the database has collected and disclosed time-series data for basic indicators including total assets, total equity, total liabilities, net income, profit before tax, net interest income and number of employees over six years (2018-2023).

The new format of the PDB and DFI Database is now available for download on the Data Downloading webpage, alongside the previous version which remains an option.

Call for More Transparency: PDBs and DFIs with Inaccessible Total Assets Data (FY2018–FY2023)

Of the 544 institutions included in the database, total assets data is unavailable for 102 institutions between fiscal years 2018 and 2023. The main reasons include: no access to institutions’ website[1]; access to the website is possible but there is no access to annual report (or similar documents); accessible reports do not display total assets data. The QR code for the list of concerned institutions is presented in the Appendix and a summary of their key characteristics is presented below.

Geographically, these institutions with inaccessible total assets data are spread across all continents. Africa (32%) and Oceania (30%) show the highest proportions of institutions with inaccessible total assets data relative to their regional totals, indicating a potentially higher inaccessible tendency compared with other continents. Notably, many are members of regional DFI networks, with 17 belonging to the Association of African Development Finance Institutions (AADFI) and 16 to the Association of Development Financing Institutions in Asia and the Pacific (ADFIAP).

By income group, low-income countries have the highest inaccessible rate at 47%, followed by lower-middle-income countries at 36%. In contrast, only 13% of institutions in high-income countries and 9% in upper-middle-income countries have their data on total assets unavailable. The accessibility gap reveals a clear correlation between a country's income level and the transparency of its PDBs and DFIs. The high inaccessible rates in low-income and lower-middle-income countries may stem from deeper issues of institutional capacity and regulatory environment.

In terms of ownership structure, inaccessibility is most common among subnational institutions (26%), less so at the national level (18%), and rare among multilateral institutions (5%). The small number of multilateral institutions with inaccessible total assets data are comprised of Bank of Alba, Maghreb Bank of Investment and Foreign Trade, and INFRAMED. Mirroring the overall pattern, the institutions with inaccessible total assets data in the top three countries—Vietnam (11), India (8), and the United States (7)—are exclusively subnational, except for one national PDB in the United States.

To address the data gap, the database team is contemplating stronger collaboration with regional associations of DFIs to engage with those members with inaccessible financial statements to encourage a move towards more transparency.

Summary Tables of Institutions with Inaccessible Total Assets Data

Continent	Number of Institutions with Inaccessible Total Assets Data	Total Number of Institutions	Percentage
Africa	32	100	32%
Oceania	6	20	30%
Asia	37	158	23%
Americas	18	122	15%
Europe	9	134	7%
World	-	10	-
Total	102	544	19%

Income Group	Number of Institutions with Inaccessible Total Assets Data	Total Number of Institutions	Percentage
LIC	9	19	47%
LMIC	53	146	36%
UMIC	12	137	9%
HIC	25	186	13%
NA	3	56	5%
Total	102	544	19%

Levels of Ownership	Number of Institutions with Inaccessible Total Assets Data	Total Number of Institutions	Percentage
SUBNATIONAL	32	121	26%
NATIONAL	67	367	18%
MULTI	3	56	5%
Total	102	544	19%

Forthcoming Steps

To optimize the data processing workflow, the team has implemented a new categorization framework for PDBs and DFIs. Our strategy prioritizes applying human expertise to the more complex institutions while leveraging digital tools to process the approximately 260 institutions with more readily available data. This two-fold approach is designed to allocate resources more efficiently by streamlining workflows and creating synergies between manual data collection and digital technology, thereby maximizing operational efficiency and ensuring data reliability at scale.

Moving forward, the database team will continue to proactively and rigorously update the global list of PDBs and DFIs, including their basic information, key financial indicators, and thematically-focused data modules.  We hope that our ongoing effort will contribute to advancing original research in the field of public development finance and help unlock the full potential of PDBs and DFIs for achieving Sustainable Development Goals.

To learn more, please visit our data visualization website (http://www.dfidatabase.pku.edu.cn/) to register and download the database for free.

For using the information from the database, please cite: Xu, Jiajun, Régis Marodon, Xinshun Ru, Xiaomeng Ren, and Xinyue Wu. 2021. “What are Public Development Banks and Development Financing Institutions? ——Qualification Criteria, Stylized Facts and Development Trends.” China Economic Quarterly International, volume 1, issue 4: 271-294; database DOI: https://doi.org/10.18170/DVN/VLG6SN.

We welcome feedback from academia, policymakers, practitioners from PDBs and DFIs, and other stakeholders to provide constructive suggestions and fill gaps in the database. Please contact us at nsedfi@nsd.pku.edu.cn.

Appendix

List of Institutions with Inaccessible Data on Total Assets (FY2018–FY2023)

Scan the QR code in the Appendix to view the full list of institutions.

---

[1]The unavailability of total assets data for fiscal years 2018-2023 from approximately 25 institutions is a result of website accessibility issues.