Association for Public Policy Analysis & Management

In 2016, the total amount of foreign aid has reached a new peak of USD 142.6 million. As an emerging donor, from 2000 to 2014, China has given out a total of $354.3 billion official finance, compared to $394.6 billion from the United States during the same period. A decade ago, a New York Times columnist commented on the “rogue aid” from China: “it is development assistance that is nondemocratic in origin and nontransparent in practice, and its effect is typically to stifle real progress while hurting ordinary citizens”. With the improved availability of data, there is an emerging literature about the impact of Chinese aid, and my research question is whether Chinese aid affects conflicts in Africa.

There are competing mechanisms on the causal paths of aid influencing conflict. Aid can either reduce conflicts by strengthening the state military power or fuel conflicts by providing more generous rewards of rebellion victory, but it is not clear which one may be more relevant in certain contexts. Motivated by the debates on Chinese aid and aid-conflict nexus, I examined the average net effect of Chinese aid on African conflicts.

I do not follow the common method of studying the impact of aid at country level. On the one hand, conflicts often happened at subnational level and are related to region-specific conditions. On the other hand, many existing aid datasets, for instance, OECD’s CRS data and AidData, are project-based, so aggregating aid information at country level eliminates the project-specific information, especially the information of location. Researchers have shown that disaggregating aid and conflict data can help identify competing causal mechanisms in aid-conflict nexus. Therefore, with two geo-referenced datasets of aid and conflict from AidData and the Armed Conflict Location & Event Data Project (ACLED), I research the aid-conflict nexus at the disaggregated level. I constructed 3,210 grid cells which cover the entire African continent and obtained a cell-year panel by spatial data manipulation.

In the search for causality, based on Dreher et al.’s paper (2016), I instrument Chinese aid by the interaction of logged amount of China’s steel production and the probability of receiving Chinese aid in a given cell. Steel is considered strategically important in China, and its production is usually not affected by aid or conflict in Africa. Part of China’s steel production surplus was used in the construction of Chinese aid projects. For this instrument, both relevance and exogeneity assumptions seem to hold. I have done the spatial data manipulation and cleaning part. My preliminary results show that, contrary to the view of “rogue aid”, Chinese aid may be in fact associated with fewer conflicts. But this may not be robust to future robustness checks.