How to Use Global Chinese Official Finance Data

Carefully consider which projects are relevant to answer your question, and include only those projects in your analysis. For most types of analysis that require aggregate estimates of Chinese official financing, we recommend only including records that have been marked "TRUE" in the "recommended_for_research" field. This field marks projects that are in the commitment, implementation, or completion stages of the project cycle, and are not umbrella projects. Using projects marked "recommended for research" will ensure that: (1) financial values are not double-counted; (2) all selected projects have moved beyond the pledge stage; and (3) all suspended or cancelled projects are excluded. We have included umbrella projects as well as pledged, cancelled, or suspended projects in the full dataset for those who may have relevant research questions regarding these types of projects.

When comparing China to other donors, keep in mind the flow classification of projects in our dataset. Most statistics reported as "aid" only include flows classified as Official Development Assistance (ODA), as defined by the OECD Development Assistance Committee (DAC). AidData’s Global Chinese Official Finance Dataset (Version 1.0), however, includes a broader set of state financing activities, including both ODA and Other Official Flows (OOF).  We have applied the OECD’s criteria (as set out in the DAC Directives for 2013-2015) to classify projects into ODA-like and OOF-like categories to enable more direct and accurate comparisons to other donors. We also use a third residual category (called "Vague Official Finance") to capture officially-financed Chinese projects where there is insufficient information to make an ODA-like or OOF-like determination. These are projects that are known to be officially-financed, so they can be included in comparisons of Chinese and Western official finance.

Generally, each record in this dataset represents a unique project. However, if disaggregated financial data is available for individual activities nested within a project, then each activity will have its own record. The field "title" indicates the specific activity that each record represents, as well as the project IDs of any linked projects related to that record.  

The project records in this dataset are most useful for estimating total financial commitment amounts. In order to generate useful estimates of total ODA commitments (and/or other forms of state financing) from China to specific sectors, recipients, and/or other groupings, we recommend (a) summing all projects that reached the official commitment stage, projects in implementation, and completed projects; and (b) excluding all pledges that did not meet the official commitment stage, canceled projects, suspended projects, and “umbrella" projects. The logic of this summation procedure to derive estimates of Chinese official finance (ODA and OOF) commitments is that if a project has already reached the implementation or completion stage, then it must have previously reached the official commitment stage.  All financial values in the dataset have been deflated to USD 2014 constant dollars using AidData’s deflation methodology detailed in Appendix F of AidData's TUFF Methodology, Version 1.3.

Users should pay extra attention to debt forgiveness/rescheduling projects.  We have included debt forgiveness/rescheduling projects in the ‘recommended_for_research’ field, but there may be some overlap if the original loan is captured elsewhere in the dataset. Users may choose to exclude these projects if they so desire, as there is insufficient information to identify whether forgiven debts are already captured by other loan project rows.

Familiarize yourself with the key fields in the dataset and the various categorizations they offer. These fields include: donor intent, flow, flow class, status, umbrella, sector, and recommended for research.  See the dataset README file for detailed descriptions, as well as the TUFF Methodology, Version 1.3 document for examples and explanations to responsibly use the data for analysis.  

We recommend using the most recent research release version of the dataset for analysis.  AidData makes its Chinese official finance dataset available through two mediums: (1) a static, cleaned, and carefully curated research release; and (2) a dynamic, searchable interface (via that reflects the current state of the database at any given point in time (including projects that are in the process of being assembled and vetted). The research release dataset provides a reliable, replicable, and stable source of data for analysis, whereas the online platform that houses the dynamic data is subject to change based on our own internal data collection and quality assurance cycles. To conduct analysis, we recommend users use one of our static "research release" datasets that contain content that has been standardized and carefully vetted.  

We strongly recommend that researchers exclude North Korean projects from statistical models. Our dataset uncovered 20 projects to North Korea totaling $272.65 million (including pledges), but we have reason to believe that this is a substantial underestimate of total Chinese official financing to North Korea. China reportedly contributes significant official finance to North Korea, but at least some of these flows are kept secret by the Chinese and North Korean governments. Given the secrecy of these financial flows, it is not possible the for TUFF methodology to produce close-to-complete coverage of concessional and non-concessional Chinese government financing to North Korea. Researchers interested in flows to North Korea should therefore treat these data with significant caution and consider excluding North Korea from their analysis.

The data are only as good as the underlying information. AidData’s Global Chinese Official Finance Dataset (Version 1.0) represents the most comprehensive and detailed source of project-level information on Chinese official finance (and official development finance) that has ever been assembled. However, the dataset was compiled using open-source data collection and triangulation methods, which means that individual project records are only as good as the underlying sources of publicly available information that enabled their construction (e.g., information via English and Chinese-language news reports; Chinese ministries, embassies, and economic and commercial counselor offices; aid and debt information management systems of finance and planning ministries in counterpart countries; and case study and field research undertaken by scholars and NGOs).

There are inevitably some errors and omissions in the dataset. AidData’s TUFF methodology is a systematic, replicable, and transparent methodology that seeks to reduce the likelihood of errors and omissions. Over the last five years, it has been stress-tested, refined, codified, and subjected to scientific peer-review, resulting in dozens of scholarly publications. AidData also has a track record of updating individual project records whenever errors, omissions, or new sources of information are identified. We have taken great care to learn from the past mistakes of previous open source data collection efforts that have run into major challenges such as: heavy reliance on individual sources (particularly English language news sources); insufficient attention to duplicate projects; over-counting as a result of not following projects from announcement to implementation; and opaque methods and sources.

Our approach is a second-best solution in the absence of transparent reporting. We would prefer to use official, project-level data from the Chinese Government that is both accurate and complete, but this is simply not possible at present. The Chinese Government does not yet disclose comprehensive or detailed information about its overseas development program, nor does it publish a bilateral breakdown of its international development finance activities. It has also opted out of international reporting systems, including the International Aid Transparency Initiative (IATI) and the OECD’s Creditor Reporting System (CRS). Therefore, the the open source data collected through the TUFF methodology is the best available solution to gain a reasonably comprehensive and detailed picture of China’s global development footprint.  China plays an increasingly central role in the global development finance regime, and its overseas investments are simply too important to ignore.