Counterfactual Analysis with Bayesian Models
summary
Uses PyMC to fit hierarchical Bayesian regressions to Global Burden of Disease and World in Data mortality rates and runs posterior simulations to answer concrete counterfactuals. Finds top contributors to the women‑men life expectancy gap are road traffic, drug overdoses, homicide, suicide, lung and liver disease; model-level sensitivities are ~0.43 years per SD and three US causes (traffic, drugs, suicide) each reduce the gap by roughly 0.75 years if matched to the best observed outcomes (e.g. Lithuania >12 years, Iceland 1.9 years).
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:54Life expectancy gap report
- 5:26Life expectancy trends in OECD
- 7:57Bayesian regression for life expectancy
- 11:05Model coefficients and counterfactuals
- 14:55US death causes over time
- 16:20Why Bayesian regression
- 20:33Advantages and limitations of Bayesian methods
- 21:37Gender differences in traffic deaths
- 24:05Public health policy impact
- 25:52Summary and policy implications
- 26:55Q&A Bayesian trade-offs
- 28:20Q&A AI tools for Bayesian
- 29:05Q&A Priors and skepticism
