There is evidence of ICHRA growth, but no complete national census in the sources reviewed. The useful question is which part of the market each measurement can establish.
The measures are different
| Source measure | Reported value | Boundary |
|---|---|---|
| HRA Council: known participating employees | 261,000 | January 2026; page 8 |
| HRA Council: covered lives | 402,000 | Derived using a 1.54 dependent factor; not a direct count of unique member records |
| HRA Council: broader covered-life estimate | 500,000+ | Includes an allowance for coverage outside its contributors |
| HealthSherpa: national market estimate | Above 1 million members | Company estimate in a mid-year update; not its own enrollment count |
| CMS: 2026 Marketplace selections | 23.1 million | Wider Marketplace, not an ICHRA count |
The HRA Council figures come from its report; HealthSherpa’s estimate comes from its company-channel post. CMS measures a different insurance population. These figures must not be added together or averaged. source source source
A stable comparison changes the growth story
HRA Council’s page 9 reports ALE growth of 108% in its expanded aggregate and 39% in the repeated prior-year contributor cohort. The corresponding non-ALE figures are 97% and 33%. New contributors affect the aggregate comparison. source
Our analytical preference is to use a repeated contributor group when describing change within an observed population. Use the expanded total when describing the research program’s broader reach. Even a repeated group is not a random national sample: changes within the group can include transfers from administrators outside it.
For example, an administrator gaining an employer from another administrator increases its own book but does not necessarily add an employer to the category. A national adoption measure would need to distinguish those transitions from first-time ICHRA adoption.
What the newer channel update adds
HealthSherpa reports 58% year-to-date growth in ICHRA enrollment flowing through its channel and 83% of that enrollment off-Exchange. Its exact post date was not displayed. Channel mix and an estimated national population remain different measures. source
A larger estimate later in the year could reflect real adoption, a broader definition, more complete observation or a different extrapolation. The reviewed evidence does not quantify each explanation. Selecting the largest number would conceal that uncertainty rather than resolve it.
The definitions a credible market table needs
Store the observation date, release date, unit, eligible population, collection method, contributors and treatment of duplicates with every number. “Employer” should specify active or historical accounts. “Member” should specify whether dependents are included. “Enrollment” should distinguish a selection from coverage that actually started.
CMS maintains a separate series for effectuated coverage. We located that series but did not extract a new numeric paid-enrollment total in this review. Its existence is another reason to preserve the “plan selections” label on the Marketplace figure. source
Does 10.5% direct-purchase coverage mean 10.5% use ICHRA?
No. Census's September 15, 2026 release, Health Insurance Coverage in the United States: 2025, reports direct-purchase coverage for 10.5% of the population during 2025. That category describes individual insurance obtained from an insurer or an exchange. It is not a count of people receiving ICHRA reimbursements. source source
Keep two questions separate: where did the insurance policy come from? and how was the premium financed? An ICHRA can help pay for an individual policy, but the broad direct-purchase category does not isolate that financing arrangement. It includes people whose premiums are funded in other ways. The aggregate percentage does not establish ICHRA enrollment, employer contributions or ICHRA growth.
The report measures coverage during a calendar year. People can appear in more than one coverage category, so categories must not be added as if they were mutually exclusive market shares. Nor should an annual coverage measure be compared directly with a single month's enrollment. These are national coverage statistics, not local premium or carrier-availability data. source
Does the government ask employers specifically about ICHRA?
Yes. Question 21 of the 2025 MEPS Insurance Component establishment questionnaire separately identifies an ICHRA offer, a QSEHRA offer, neither, or an unknown answer. Its instructions distinguish these arrangements from traditional HRAs. That is evidence that the instrument asks the relevant offer question; it is not an adoption estimate. source
The corresponding plan questionnaire's Question 22 asks about traditional HRAs and excludes ICHRA and QSEHRA. This distinction matters when reading a study labeled simply “HRA”: establish which arrangement it measured before applying its findings to ICHRA. See our HRA foundations for the relationship between the categories. source
To publish an employer-offer rate, we would need a verified weighted result, its survey population, denominator and uncertainty. Establishments and firms are not interchangeable; the form generally asks about the specified location. An offer also does not show how many eligible workers enrolled, received reimbursements or covered dependents. source
Current boundary: we verified the questions but have not verified a published weighted ICHRA/QSEHRA estimate from them. AHRQ states that MEPS-IC data files are not public because of confidentiality; approved research access is a separate process. Published tables must be checked before proposing any independent estimate. This is a research path, not a dataset we have already analyzed. source
Which number answers your question?
| Question | Required measure | Common substitution to avoid |
|---|---|---|
| How many businesses offer ICHRA? | Firms or establishments offering it, with the unit explicit. | Administrator accounts, survey interest or all HRA offers. |
| How many workers participate? | Enrolled workers for a defined period, with duplicates addressed. | Workers eligible for an offer. |
| How many people are covered? | Workers plus covered dependents; identify counts versus estimates. | A dependent multiplier presented as an observed headcount. |
| How large is the underlying insurance market? | Individual coverage or Marketplace measures with dates and definitions. | Direct-purchase coverage presented as ICHRA enrollment. |
What would change this conclusion?
A reconciled dataset with shared definitions, documented deduplication and comparable time periods could support a stronger national estimate. Until then, we publish the figures with their boundaries. This is ICHRA Report’s analysis of external research; it is not a survey or market census we conducted.
Related reading: employer interest versus implementation and what contribution data can establish.
Sources and evidence
Review dates are recorded for each source above. Company pages are useful for confirming how a product is described, but they do not prove service quality or customer results.