Do Human Rights Treaties Work? Causal Re-Analysis for Peace Science

This project is a re-analysis of Hill (2010) who sought to estimate the impact of three human rights treaties - the International Covenant on Civil and Political Rights (ICCPR), the CAT (Convention Against Torture), and the CEDAW (Convention on the Elimination of All Forms of Discrimination Against Women) - on physical violence, torture, and womens' rights, respectivley. This project improves on Hill's (2010) design by 1) introducing a formal graphical causal model, 2) defining an estimand, 3) implementing an estimator that accounts for the biasing properties of the employed panel data, and 4) executes a sensitivity analysis to examine the potential impact of unobserved confounding. Overall, the results suggest that these human rights treaties generally fail to demonstrate a causal effect on their associated outcomes of interest and, where an non-zero "effect" was estimated, these estimates were highly sensitive to varying levels of unobserved confounding. Overall, the results suggest that this area of research requires increased theory to sufficiently account for confounders and increased attention to alternative identification strategies that do not rely on selection on observables.