Aid effectiveness in sustainable development: A multidimensional approach
通过将细分的援助流数据与可持续发展目标指标关联，并利用人工智能模型模拟因果过程，评估21世纪头十年国际援助对146个受援国各维度的影响。
What is the impact of international aid? We answer this question by linking disaggregated aid-flows data to a large set of indicators classified into the Sustainable Development Goals (SDGs). Since such linkage is not perfect (due to the nature of the data), we deploy an artificial intelligence model of the causal process through which changes in aid flows contribute to the dynamics of individual indicators. The model accounts for salient features of real-world development such as multidimensionality, complex interconnections between indicators, heterogeneous aid-to-expenditure ratios, rationally-bounded bureaucracies, fungibility, and the temporal structure of contemporary aid flows across development dimensions. The model does not require cross-country pooled data, so we calibrate its parameters for each of the 146 aid-recipient countries in our sample, preserving important contextual information of each nation. By producing counterfactual simulations where aid is removed, we obtain nuanced estimates of the impact of international assistance during the first decade of the 21st century, at the level of each country, SDG, and indicator. We validate our results using a sector-specific study with similar–but more aggregate–findings. Such a large and detailed picture of the multidimensional impact of aid has not been documented before.