Automating the US Welfare State: Historical Experiments in Technological Federalism
Talkat Bridging Prediction and Intervention Problems in Social Systems (Simons Institute, 2026) · 2026-01-12..01-18·
summary
Historical research shows U.S. welfare computerization aims less at abstract prediction than at solving political coordination: microsimulation tools like TRIM, MIM and the KGB model (derived from RIM/TRIM/MATH lineages) simulate households to estimate who wins and loses; Project MATCH matches federal payroll to welfare rolls, finds 33,000 matches, 96 prosecutions, and almost no large-scale fraud; statewide systems travel (Alaska → North Dakota → Mississippi) via vendors and consortia; many builds persist, partially work, and produce political momentum rather than purely accurate savings.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:32Automation in US welfare
- 3:54Mississippi’s 1970 budget anomaly
- 7:17Lessons on automated welfare
- 8:04Califano and welfare reform
- 11:49Microsimulation in the 1970s
- 15:56Battles over welfare modeling
- 17:11Computers for fraud detection
- 19:45Project MATCH data unification
- 23:18Momentum in welfare automation
- 25:41Technological federalism
- 28:45How scores are made
- 29:31Q&A
related talks
28mExperience report: implementing a real-world, medium-sized program derived fr...
05mTemplated Analyses within R Packages for Collaborative, Reproducible Research
55mDDD Surrounded by Legacy, Govt Edition
26mMake world less shit. NOW.
30mWhen software and law are the same thing
27m Agile transformation in large public organization
04mGovernment is a system.
46mFacebook, Dynamite, Uber, Bombs, and You