RAND researcher presents state tool to estimate costs if states expand coverage to undocumented immigrants

Preeti Rao, a health economist at the RAND Corporation, told the Nevada Senate Committee on Health and Human Services that undocumented immigrants are excluded from most federally funded insurance programs and that states considering filling that gap need better data and realistic assumptions.

Rao summarized RAND’s research on state-funded options—Medicaid-like coverage or marketplace-style subsidies—and demonstrated a web-based tool RAND developed to produce high-level estimates of enrollment, state spending and out-of-pocket costs for selected eligibility groups and income cutoffs.

The tool, Rao said, uses a national microsimulation model reweighted to state populations and relies on imputation where immigration-status data are missing. "Undocumented immigrants are not eligible for most forms of health insurance in The United States," Rao said, adding that RAND uses Migration Policy Institute population estimates and American Community Survey cross-tabs to build state-level synthetic populations.

Why it matters: decisions to use state funds to cover immigrants carry budgetary implications entirely borne by the state, and policymakers need credible cost ranges to plan. Rao said RAND’s tool is intended to give policymakers a ballpark estimate rather than a precise budget figure because key inputs are uncertain.

Key findings and limitations

Rao noted three practical takeaways from RAND’s state work and Connecticut case study: covering children and young adults tends to be less costly than covering older adults; early enrollment and costs in some states (Illinois was cited) have exceeded initial expectations; and assumptions about lower health spending for immigrants (the "healthy immigrant effect") may not hold once previously uninsured people gain coverage and use pent-up care.

Rao described methodological choices: RAND excludes ACS respondents who are clearly citizens, permanent residents or hold occupations/programs unavailable to undocumented immigrants, then reweights remaining observations to match MPI state totals (she cited an example of roughly 110,000 estimated undocumented people in Connecticut). She said the tool updates population and policy maps (current as of September 2024) and allows users to override RAND’s population estimate with a jurisdiction's preferred figure.

On cost-offsets and take-up

When senators asked whether the tool accounts for savings from reduced uncompensated care or fewer emergency-room visits, Rao said it does not automatically include such offsets. She explained RAND performed back-of-the-envelope estimates of possible uncompensated-care savings in Connecticut but cautioned such estimates rely on large assumptions about what share of uncompensated care is attributable to undocumented immigrants.

On take-up and spending assumptions, Rao said RAND initially assumed lower take-up and lower spending for immigrants, but observed state programs produced higher-than-expected enrollment, and presumed pent-up demand after gaining coverage can raise early spending relative to baseline.

Committee response and follow-up

Vice Chair Taylor and other senators asked about the "healthy immigrant effect," the microsimulation inputs, and how ACS and MPI data are combined. Rao said RAND imputes undocumented status and pregnancy using natality statistics, then reweights to MPI totals while borrowing ACS cross-tabs for joint distributions. She offered to provide committee members with RAND’s reports and the underlying tool documentation.

Ending

Rao emphasized that the RAND web tool is intended to inform discussion by providing flexible, scenario-based estimates rather than final budget figures. She reiterated that the tool should be treated as giving a plausible range rather than a precise prediction and said RAND would share reports and answer follow-up questions.