Search Results
Working Paper
One Threshold Doesn’t Fit All: Tailoring Machine Learning Predictions of Consumer Default for Lower-Income Areas
Modeling advances create credit scores that predict default better overall, but raise concerns about their effect on protected groups. Focusing on low- and moderate-income (LMI) areas, we use an approach from the Fairness in Machine Learning literature — fairness constraints via group-specific prediction thresholds — and show that gaps in true positive rates (% of non-defaulters identified by the model as such) can be significantly reduced if separate thresholds can be chosen for non-LMI and LMI tracts. However, the reduction isn’t free as more defaulters are classified as good risks, ...
Discussion Paper
Combining AI and Established Methods for Historical Document Analysis
This paper examines methodological approaches for extracting structured data from large-scale historical document archives, comparing “hyperspecialized” versus “adaptive modular” strategies. Using 56 years of Philadelphia property deeds as a case study, we show the benefits of the adaptive modular approach leveraging optical character recognition (OCR), full-text search, and frontier large language models (LLMs) to identify deeds containing specific restrictive use language— achieving 98% precision and 90–98% recall. Our adaptive modular methodology enables analysis of ...