G7 partner vote changes partly offset the U.S. agreement loss on 54 matched UN resolution titles in 2016 and 2018

We measured which side changed its votes when agreement with the United States fell. We used a fixed public data copy of United Nations General Assembly votes. The main sample had 54 resolution titles present in 2016 and 2018. All seven G7 states cast a yes, abstain, or no vote on each record. A score gave half agreement to abstain versus yes or no. Mean agreement between the United States and its six G7 partners fell from 65.90% to 62.96%. An exact split that averages both change orders gave a U.S. vote component of -5.09 percentage points and a partner component of +2.16 points. Partner switches thus offset 42.42% of the negative U.S. component. Every partner had the same component sign pattern. Exact vote equality, strict title strings, and a smaller provider-link sample kept this pattern. Removing three largest-loss titles reduced the net decline to 0.16 points. The drop across all eligible records was 13.49 points, but sample differences do not identify an agenda cause. Equal titles can have different text. This is a descriptive account of selected recorded votes. It does not establish changes in state preferences or an effect of a president. The package preserves source data, all matching decisions, code, three figures, and an independent exact audit. A fresh source rebuild matched all 19 result files and all three figures.

United Nations · voting agreement · repeated resolutions · G7 · vote switches · reference state · reproducible analysis