Approaching belts and roads through better research design
I’ve written previously about my concerns relating to the quality of BRI-related MSc dissertations.(1) To summarize, these projects tend to suffer from significant problems. These tend to stem from at least one of the following:
- Inability or unwillingness to separate hype/propaganda from actual policy and activities
- Poor research design or a lack of quality data
These, individually and in combination, lead to poor performance on this topic relative to other topics. While my strategy of getting students to ignore the BRI component and focus on research design at the initial stages of their work (typically through the implementation of a regression discontinuity design), projects still run into a number of issues specific to this topic.
Separating noise from the signal
Most coverage of BRI amounts to noise: propaganda from the Chinese government (or inspired by it) and criticism from its opponents. This isn’t to say that these sources of information are entirely useless, but they are primarily distractions from the core quantities of interest for most research projects.(2) For empirical projects in particular, a common problem is extensive reliance on these sources as an indicator of the related academic literature, leading to a lack of foundation in relevant theoretical and empirical work, and a tone that reads more like the position pieces referenced.
One approach that avoids this trap is to focus the project’s framing and contribution to the literature in reference to broader debates, either with reference to China or more generally. Focusing primarily on high-quality academic journals will also help avoid the propaganda trap. If the public debate over BRI is central to the research question, then pieces that make up the debate should not be extensively referenced in the literature review; instead, the focus should be on broader issues of policy framing, power in international relations, or other related academic debates. Sources that are part and parcel of the debate over BRI itself should instead be primarily referenced within the body of the research activity itself.
Ultimately, students need to be very aware of the diversity of opinion relating to the BRI, particularly if they’ve only been exposed to Chinese media coverage and official proclamations (an interesting counterpoint to all of this can be found in the Heritage Foundation’s analysis of the project). For a balanced account, Peter Frankopan’s The New Silk Roads offers an insightful and easy-to-read discussion of the BRI and the debate surrounding it (focusing on Central Asian countries), from the perspective of an academic historian writing for a general audience.(3)
Disentangling belts and roads
Even with improved focus on research-design issues, problems with data still present difficulties.
FDI vs. Foreign Aid
Funding for BRI projects often consists of a combination of direct investments and foreign lending (typically with high interest rates attached). Because the two flows are related and interdependent, any study that aims to discuss the effects of one without accounting for the other will yield biased inferences.
Announced vs. Actual Flows
Announcements of investment and aid flows are not necessarily followed by the distribution of those funds. Furthermore, an announced mixture of investment and loans may differ significantly from the actual composition of direct investment and lending disbursed for any particular project. Depending on the research question, announcements may be more important than actual distribution or sourcing, but very careful attention needs to be paid to both the data source and construct reported.
Unobserved Outcomes
Many interesting outcomes remain unobserved from an empirical perspective. A popular student question relates to BRI’s effect on economic growth in partner countries. Measurement issues aside (many of which extend beyond those mentioned here), infrastructure projects often take years to be realized; their effects on the economies around them may take even longer to materialize. Other potential topics, like the effects of a reliance on Chinese labor on local-economy unemployment, require a more creative (and potentially expensive) approach to data collection than typically observed at the BSc/MSc level.
(1) I should note that these concerns are not limited to the MSc level; they apply equally to BSc- and PhD-level projects, and beyond.
(2) Potential exceptions to this are computational text analysis or discourse analysis of these rhetorical stances, but these sorts of studies would be of limited interest unless they are explicitly tied to broader theoretical and empirical debates beyond BRI.
(3) I’m happy to take recommendations of other books that provide balanced overviews of the subject.