Association for Public Policy Analysis & Management

This paper explores the effects of Foreign Direct Investment (FDI) received from China on a country’s mean voting similarity with China in the United Nations General Assembly (UNGA) for a group of 136 countries using a fixed effects model. Recent developments with China’s Belt and Road Initiative (BRI) have led to concerns that China may be using the BRI as a means of purchasing influence. See, for example, a recent New York Times article with the headline “China Renews Its ‘Belt and Road’ Push for Global Sway.” Due to the recency of the initiative, these claims are not supported by empirical evidence. This paper aims to provide some perspective on the concern of influence buying. One possible way this influence-buying could manifest is through increasing policy alignment with China in the UNGA. A similar paper finds that an increase in loans of one percentage point of GDP is associated with a 26 percent increase in voting similarity between a recipient country and the G7 countries (Dreher, Axel, & Sturm, 2012). Ellis found that China’s post-2000 but pre-Belt-and-Road investments in Sub-Saharan Africa resulted in just an 0.00012 increase in similarity, seeming to suggest that vote buying was not taking place during this period (Ellis, 2013). These findings do not rule out vote buying with the increased investment of the BRI.

By using a panel dataset spanning from provided by the American Enterprise Institute on the FDI from China received by all countries in the period between 2005 and 2018, and variation within each country, I estimate the effect of FDI received on my outcome measure. I use GDP per capita, the impex rate, current account balance, total FDI inflow, Official Development Assistance received per capita, and population from the World Bank as controls. This builds upon previous literature by including additional years after the introduction of the BRI. I attempt to use inflation as an instrument to address endogeneity, but it is invalidated by a weak first stage. I find a small, statistically significant positive effect using OLS, but it is not economically significant. These results suggest less than a 1% increase in voting similarity given a 1 percentage point increase in the proportion of FDI to GDP, relative to similar countries that did not receive FDI, or that received less FDI. Although statistically significant, these results are not causal because FDI is endogenously determined during this period.