Does the international public deem global governance (il)legitimate – and how so? A machine-learning-based content analysis of 22 million tweets

In recent years, scholars and world leaders have increasingly focused on the perceived legitimacy of international organizations. Despite significant advances, extant studies are limited by their focus on surveys or restricted samples of observational data. We provide the most comprehensive study yet of real-world debates on the legitimacy of global governance institutions (GGIs). We fine-tuned a transformer-based language model to code more than 22 million tweets from 2019 on 16 diverse GGIs. Our principal findings are: First, the international public evaluates GGIs positively on average, which is associated more with output than input considerations. Second, global, multi-purpose, governmental, and/or high-authority GGIs are considered less legitimate than regional, task-specific, non-governmental, and/or low-authority institutions. Third, a perceived lack of input virtues explains legitimacy deficiencies. Our study provides fresh evidence concerning public attitudes toward present-day GGIs, bearing significant implications for the study and practice of global governance.