---
type: source
title: "Saved by Medicaid: New Evidence on Health Insurance and Mortality from the Universe of Low-Income Adults (NBER)"
author: "NBER Working Paper 33719"
url: https://www.nber.org/papers/w33719
date: 2025-01-01
domain: health
secondary_domains: []
format: article
status: unprocessed
priority: high
tags: [Medicaid, mortality, NBER, causal-evidence, coverage, insurance, low-income, quasi-experiment]
intake_tier: research-task
---

## Content

**NBER Working Paper 33719: "Saved by Medicaid: New Evidence on Health Insurance and Mortality from the Universe of Low-Income Adults"**

**Key finding:**
- Medicaid expansion increased Medicaid enrollment by **12 percentage points**
- Reduced mortality of the low-income adult population by **2.5%**
- Translates to a **21% reduction in the mortality hazard** of new enrollees specifically

**Methodology:**
- Uses "universe of low-income adults" (full population, not sample)
- Quasi-experimental design exploiting state Medicaid expansion variation
- This is among the most rigorous causal designs available in health insurance research

**Implication for OBBBA:**
- If Medicaid expansion reduced mortality hazard by 21% for new enrollees
- Then coverage loss symmetrically should increase mortality hazard
- The Lancet Regional Health Americas modeling (7,049-9,252 excess deaths/year) is consistent with this causal magnitude
- The directionality of the mortality effect is well-established from the expansion side

**Context within insurance-mortality literature:**
- Previous NEJM study found loss of drug subsidy coverage associated with higher mortality among low-income Medicare beneficiaries (drug subsidy loss → mortality)
- Oregon Health Insurance Experiment (OHIE): found mental health improvements and financial protection but inconclusive mortality signal at small sample size
- NBER WP 33719 is more powerful than OHIE because it uses full population data

## Agent Notes

**Why this matters:** This is the causal foundation for the mortality modeling studies. The Lancet Regional Health Americas study builds on the established Medicaid expansion mortality effect and applies it in reverse to project coverage loss deaths. If the NBER estimate is correct, 12 percentage points of Medicaid gain = 21% lower mortality hazard — meaning 12 percentage points of coverage loss should produce a symmetric mortality increase.

**What surprised me:** The 21% mortality hazard reduction for new enrollees is larger than I expected for an insurance intervention. The magnitude suggests that low-income adults newly gaining Medicaid coverage face substantial pre-enrollment mortality risk (uncontrolled chronic conditions, preventable hospitalizations avoided, preventive care accessed). The coverage → survival pathway is remarkably strong.

**What I expected but didn't find:** A published version (the paper appears to be a working paper as of this date). NBER working papers are not peer-reviewed. However, the authors are credible and the methodology is sound.

**KB connections:**
- [[medical care explains only 10-20 percent of health outcomes because behavioral social and genetic factors dominate as four independent methodologies confirm]] — coverage has a mortality effect, suggesting that the 10-20% clinical care slice is critically important for specific high-risk subpopulations even if it's less important at the population level
- [[social isolation costs Medicare 7 billion annually and carries mortality risk equivalent to smoking 15 cigarettes per day making loneliness a clinical condition not a personal problem]] — comparable mortality magnitude evidence for a social determinant

**Extraction hints:**
- This NBER paper is the grounding evidence for the mortality claims in the OBBBA context. It's the mechanism: Medicaid → coverage → managed chronic conditions → mortality reduction.
- Important nuance: the 21% figure applies to NEW enrollees (who were previously uninsured with unmanaged conditions). It's not the mortality risk for the stable Medicaid population. The OBBBA disenrollees include many who have been enrolled for years — the mortality effect for this group may be smaller (conditions managed) or larger (sudden care disruption).
- Do NOT conflate the 21% new-enrollee effect with a population-level mortality claim. The extractor should scope this carefully.

**Context:** NBER working papers are pre-publication but use full-population administrative data. The causal design (state variation in expansion timing) is among the strongest in observational health economics.

## Curator Notes (structured handoff for extractor)
PRIMARY CONNECTION: [[medical care explains only 10-20 percent of health outcomes because behavioral social and genetic factors dominate as four independent methodologies confirm]]
WHY ARCHIVED: This is the strongest causal evidence for the mortality impact of Medicaid coverage. Used as the foundation for the OBBBA mortality projections. Important for KB grounding of the "coverage loss → deaths" claim chain.
EXTRACTION HINT: Scope carefully: 21% mortality hazard reduction applies to NEW enrollees (previously uninsured). The population of OBBBA disenrollees is different (long-term enrollees losing established coverage). Both mechanisms cause mortality — but through different pathways (sudden care disruption vs. absence of care initiation). Extract two separate claims if possible.