WORKING PAPER - Global Governance under Populism: The Challenge of Information Suppression

Data Missingness We examine rates of data missingness using World Bank data both for comparability with previous work in this area (e.g., Hollyer, Rosendorff, and Vreeland 2011) and because of the substantive importance of the Bank in many scientific domains, including those pertaining to the environment and health-related issues. [...] Further, the Bank is active in the public health arena, where key functions include the identification of disease outbreaks, the measurement of disease incidence, and the communication of effective medical practices.25 In this space, the data contained in the World Development Indicators (WDI) are often initially collected by other IOs that explicitly engage in monitoring. [...] Information related to the environment often requires detailed scientific models and projections, measurements of pollutants and energy use, and estimation of the impact of environmental factors on health and well-being. [...] The strongest results in the table, both in terms of magnitude and statistical significance, are for the variables reliant on raw, state-provided scientific data.30 The entry into office of a populist government is associated with an increase in missingness of 5 to 8 percent of a standard deviation in variables using raw state data. [...] Lower capacity should add random noise to state-reported emissions data, not biasing them in a particular direction.38 To measure the quality of state-provided emissions data, we compute the absolute difference between emissions data reported directly to the UNFCCC and the emissions data contained within the WDI.39 Emissions data within the WDI are based on independent estimates from the Emissions.