Ranganathan, ShyamNicolis, Stamatios C.Swain, Ranjula BaliSumpter, David J. T.2018-07-312018-07-312017-02-27e0171560http://hdl.handle.net/10919/84446The Millennium Development Goals (MDG) programme was an ambitious attempt to encourage a globalised solution to important but often-overlooked development problems. The programme led to wide-ranging development but it has also been criticised for unrealistic and arbitrary targets. In this paper, we show how country-specific development targets can be set using stochastic, dynamical system models built from historical data. In particular, we show that the MDG target of two-thirds reduction of child mortality from 1990 levels was infeasible for most countries, especially in sub-Saharan Africa. At the same time, the MDG targets were not ambitious enough for fast-developing countries such as Brazil and China. We suggest that model-based setting of country-specific targets is essential for the success of global development programmes such as the Sustainable Development Goals (SDG). This approach should provide clear, quantifiable targets for policymakers.application/pdfenCreative Commons Attribution 4.0 InternationalSetting development goals using stochastic dynamical system modelsArticle - RefereedPLOS ONEhttps://doi.org/10.1371/journal.pone.0171560122282410571932-6203