Pushing the Development Finance Envelope: AidData Improves Upon Its China Data

Employing these supplementary data collection methods, we were able to generate a more reliable picture of Chinese development finance in Africa by adding 114 new projects and updating 131 existing records in our database from 2000 to 2011.

January 14, 2014
Ellie Kaufman

AidData launched the 1.0 version of its “Chinese Development Finance to Africa” database at china.aiddata.org in April 2013. Publication of this large data store uncovered approximately 1,670 Chinese official finance activities between 2000 and 2011 worth over $75 billion. Setting the stage for our updated China in Africa data and methodology to be released next week, we’d like to catch you up on what’s been going on behind the scenes since last April.

The initial version of the Chinese development finance data and methodology provoked varied responses from the media, academy and policy community. Journalists from Reuters, China Daily, and Ghana Business News published stories with new insights mined from the database. Scholars debated the merits and shortcomings of open source data collection methods. Policymakers used the data to revisit popular assumptions about China’s political influence, commercial interests, and development impact in Africa.

Since April, AidData has joined forces with faculty, staff, and students from the University of Heidelberg, University of Cape Town, and Zhejiang University to improve upon our Chinese development finance data in several ways: integrating 2012 data, refining the data collection methodology, and making it easier for users to access our data and contribute useful content  (e.g., project documents, photos, videos, media reports) to individual project pages.

Committed to continuous learning, AidData has sought to address concerns that our Chinese development finance relied too heavily on media reports. We have revised and enhanced project records with information from official sources gleaned through annual reports from recipient government ministries and Chinese embassy websites, as well as a systematic search of African government websites. For projects with the fewest corroborating sources, AidData also performed additional English and Mandarin searches—using Google and Baidu, respectively. Employing these supplementary data collection methods, we were able to generate a more reliable picture of Chinese development finance in Africa by adding 114 new projects and updating 131 existing records in our database from 2000 to 2011.

Recognizing that the term “media-based data collection” led some data users to conclude that our database draws information exclusively from news reports, we have renamed our methodology to better represent our data collection approach. In fact, media reports are used only as a departure point for project records, which are supplemented with case studies undertaken by scholars and NGOs, project inventories from Chinese embassy websites, as well as grants and loan data published by recipient governments. Going forward, we will refer to our open source approach to data collection and triangulation as the Tracking Under-Reported Financial Flows (TUFF) methodology.

Increasing the accessibility of our data on china.aiddata.org and eliciting crowd-sourced content has been another important area of focus since April. A multimedia uploader added to each project page enables a wide range of stakeholders to contribute and download images, videos, or documentation related to a specific project. We hope this also makes it easier for “auditors” to suggest corrections and improvements to individual project records.

Users of china.aiddata.org can now visit a project page, such as this entry for a Chinese investment in South Africa’s platinum industry, to quickly scan project details and upload their own content. Source: china.aiddata.org

Listening to end users of our data, AidData has upgraded how we store and present informational resources in china.aiddata.org to include information on the publisher, author, and date published, in addition to the source URL. Rather than saving sources as a single field of a project entry, the database now stores each independently as a “resource” which allows researchers to analyze relationships between sources more easily. This allows us to link a single resource to all relevant projects, examine the “health of project records,” and evaluate which types of sources yield the most useful project information.

Stay tuned for the release of our updated 2000-2012 database of Chinese development finance activities next week. It will cover more than 1900 non-investment projects worth approximately $83 billion. The database includes 268 pledged projects; however, the reported total of $83 billion only captures official commitments, including projects in implementation and completed projects. As we continue to improve china.aiddata.org, we want to hear your feedback. Drop us a line at china@aiddata.org as you have ideas on additional features or information you’d like to see.

*Post updated to reflect current financial information available at china.aiddata.org.

Brad Parks is AidData’s Co-Executive Director, and Charles Perla is an AidData Research Associate based at the College of William & Mary. AidData is a partnership of the College of William & Mary, Development Gateway, and Brigham Young University.