Malesky and Nguyen (2024, Journal of Comments and Replications in Economics, MN) reassess a recent study on the long-term effects of wartime violence on civic engagement in Vietnam, attributing discrepancies in previous findings to coding and historical errors. They argue that prewar party strength, rather than wartime exposure, drives Vietnam’s contemporary engagement and membership in social organizations. However, their reanalysis introduces a significant error by omitting crucial geographic covariates, which are essential for the validity of the original study’s quasi-experimental design and were included in the model specifications by Barceló (2021) and, also, in Miguel and Roland (2011). MN claim that two control variables, South and Latitude are collinear, leading them to drop both from their models. This decision is a fundamental mistake: first, collinearity affects the relationship between control variables, not the treatment of interest (Distance), making it irrelevant for estimating the treatment effect. Second, the appropriate response to collinearity is to remove one of the collinear variables, not both.
This misspecification leads to omitted variable bias and a violation of the exclusion restriction of the instrumental variable model, as the border distance treatment becomes conflated with the South-North distinction due to the asymmetrical distribution of Vietnamese residents across the 17th parallel. Once this omission is corrected and either \textit{South} or \textit{Latitude}, or both, are properly accounted for, the originally reported results withstand all of MN’s other proposed modifications, including controlling for prewar party strength, excluding provinces with greater prewar party strength, applying various clustering structures of standard errors, and controlling for ethnicity, among others.
This manuscript confirms that wartime exposure indeed enhanced long-term civic engagement and participatory values across Vietnam, countering MN's critique. Furthermore, it contributes to legacy studies by highlighting how improper model specifications can produce biased estimates that obscure, rather than clarify, our understanding of the legacy of historical events.
