The Experiment Societyexperimentsociety.org⌕Get involved
Public HealthPrice signalMixed

RAND Health Insurance Experiment

RAND Corporation · Multiple US cities · 1971

Summary

The RAND Health Insurance Experiment remains the most influential health economics study in American history. Its central finding is counterintuitive: when people pay more out of pocket, they use significantly less health care — but for the average person, their health is not measurably worse. The experiment answered the key question of 1970s health policy: if we give people free care, will they waste it? Yes, people used more free care, but most of the additional utilization was discretionary and did not improve average health outcomes. The critical exception — which often gets lost in summary — is that for poor people with hypertension, free care meaningfully improved blood pressure control. The experiment's lesson is not that insurance doesn't matter, but that the design of insurance — specifically, where cost-sharing falls and for whom — has large, measurable effects on both utilization and health equity.

Research question

"Does cost-sharing in health insurance reduce utilization, and does reduced utilization harm health?"

Methodology

Intervention

3,958 non-elderly families in six US cities were randomly assigned to health insurance plans ranging from free care (0% cost-sharing) to plans requiring patients to pay up to 95% of costs. Participants remained in their assigned plan for 3–5 years. The experiment, costing $300M in today's dollars, remains the most expensive social science randomized trial ever conducted.

Assignment

Randomized controlled trial; families assigned to insurance plans varying by coinsurance rate (0%, 25%, 50%, 95%) and out-of-pocket maximum; outcomes measured via claims data and annual health surveys

Sample size

3,958 families (~7,700 individuals) across Dayton OH, Seattle WA, Fitchburg MA, Franklin County MA, Charleston SC, and Georgetown County SC

Primary outcome

Healthcare utilization (visits, hospitalizations, prescriptions); health outcomes (blood pressure, cholesterol, vision, dental, mental health, mortality risk)

Effect estimate

Utilization: free care participants used 30–40% more health care than those in high-cost-sharing plans. Health outcomes: almost no significant difference across cost-sharing levels for the average person, except for the poorest and sickest individuals — for whom free care meaningfully improved blood pressure and vision outcomes. Free care improved low-income hypertension control by 10 percentage points.

Decision

RAND HIE became the evidentiary foundation for cost-sharing in US health insurance design; strongly influenced the Affordable Care Act's benefit design debates; copayment policy in Medicare and Medicaid traces to RAND findings; results have been challenged by the Oregon Medicaid Lottery finding that coverage expansion improves depression and financial security even when physical health effects are small; reanalysis by Newhouse (1993) and subsequent work has deepened interpretation without overturning the core findings

Result

Mixed

Utilization: free care participants used 30–40% more health care than those in high-cost-sharing plans. Health outcomes: almost no significant difference across cost-sharing levels for the average person, except for the poorest and sickest individuals — for whom free care meaningfully improved blood pressure and vision outcomes. Free care improved low-income hypertension control by 10 percentage points.

Evidence strength

Strong

Randomized trial, replicated across multiple sites or studies.

Replication status

Replicated

Institution

RAND Corporation

Location

Multiple US cities

Year

1971

Policy area

Public Health

Mechanism

Price signal

More from these institutions

Other trials of this mechanism

Cite this entry

RAND Corporation. (1971). RAND Health Insurance Experiment. The Experiment Society Registry. Retrieved from https://www.experimentsociety.org/registry/rand-health-insurance-experiment (primary report: https://www.rand.org/pubs/research_briefs/RB9174.html)

Related experiments

Mexico (national) · 2014

Mexico Sugar-Sweetened Beverage Tax

Positive

Seattle, WA, United States · 2016

Seattle Minimum Wage Study

Mixed

England, UK · 2015

5p Plastic Bag Charge

Positive

Mentioned in the Dispatch

Issue 11 · May 23, 2026

The Oregon health lottery.

When the state couldn't afford to give Medicaid to everyone who wanted it, a lottery created the most important health insurance experiment in history. What the data showed was inconvenient for everyone.

← Back to registryWhat the evidence shows →