The Value of Prediction in Identifying the Worst-Off
It addresses the challenge for policymakers in designing equitable assistance systems, though it appears incremental by providing analytical frameworks rather than a breakthrough.
The paper tackles the problem of using machine learning to identify vulnerable individuals in government programs, such as long-term unemployment in Germany, and finds that prediction can be effectively compared to other policy levers like expanding bureaucratic capacity.
Machine learning is increasingly used in government programs to identify and support the most vulnerable individuals, prioritizing assistance for those at greatest risk over optimizing aggregate outcomes. This paper examines the welfare impacts of prediction in equity-driven contexts, and how they compare to other policy levers, such as expanding bureaucratic capacity. Through mathematical models and a real-world case study on long-term unemployment amongst German residents, we develop a comprehensive understanding of the relative effectiveness of prediction in surfacing the worst-off. Our findings provide clear analytical frameworks and practical, data-driven tools that empower policymakers to make principled decisions when designing these systems.