When is there enough data to create a global statistic?

To monitor progress towards global goals such as the Sustainable Development Goals, statistics are needed. Yet cross-country datasets rarely truly global, creating a trade-off for producers of statistics: lower data coverage threshold disseminating statistics, more can be made available, but accuracy they will have. We quantify this availability-accuracy by running than 10 million simulations on World Indicators. show that if fraction world’s population which one lacks is x, then should expect to 0.37 *x standard deviations off true value, and risk being much x off. robustness result various assumptions give recommendations when there enough create statistics. Though decision context specific, in baseline scenario we suggest not less half population.