Medicaid and the OBBBA

September 29, 2025

Key Findings

The vast majority (84%) of survey respondents disagreed or strongly disagreed that Medicaid work requirements introduced by the One Big Beautiful Bill Act (OBBBA) would substantially increase employment among Medicaid-enrolled, working-age adults.
Most experts (61%) thought that the Congressional Budget Office’s estimate of Medicaid coverage losses related to OBBBA work requirements (5.3 million people by 2034) is about right. Another 26% of respondents felt that CBO’s estimate was too low.
More than three-quarters of experts (77%) agreed or strongly agreed that more people will lose Medicaid coverage because of challenges submitting documentation to comply with OBBBA work requirements than because they are unemployed.
A number of respondents said that they based their assessments on evidence from work requirements introduced by Arkansas in 2018. Several noted that, although this data is relevant, it may also be limited in terms of generalizability.

Survey Questions

Please note: our surveys will consistently use two modifiers to describe the size of an effect:

“Substantial”: when an effect is large enough to meaningfully influence policy decisions, program implementation, or outcomes of interest.
“Measurable”: when the direction of an effect is clear, but the effect may not be sufficiently large to make much of a difference for a given policy, program, or outcome.

The One Big Beautiful Bill Act (OBBBA) introduced several reforms to Medicaid, including work requirements for working-age adults.

Response rate of 95% (61 out of 64 panelists responded)

Question 1: Work requirements introduced by the OBBBA will substantially increase employment rates among Medicaid-enrolled, working-age adults.

a. Strongly Agree
b. Agree
c. Uncertain
d. Disagree
e. Strongly Disagree
f. No Opinion

Question 2: The Congressional Budget Office’s estimate of Medicaid coverage losses related to OBBBA work requirement provisions (5.3 million people by 2034) is likely _____.

a. Too High
b. Too Low
c. About Right
d. No Opinion

Question 3: Most people disenrolled from Medicaid because of the OBBBA’s work requirements will be individuals who lose coverage due to challenges submitting documentation demonstrating that they are compliant or exempt, rather than individuals who lose coverage due to unemployment.

a. Strongly Agree
b. Agree
c. Uncertain
d. Disagree
e. Strongly Disagree
f. No Opinion

Individual Survey Responses

Question One

Work requirements introduced by the OBBBA will substantially increase employment rates among Medicaid-enrolled, working-age adults.

Name	Vote	Confidence	Comments
Margarita Alegria	Disagree	9	Most Medicaid participants already work, so it is unlikely that employment will increase in this target population.
David Asch	Disagree	7
John Ayanian	Disagree	8
Peter Bach	Strongly Disagree	7
Laurence Baker	Disagree	8
David Blumenthal	Strongly Disagree	9
Erin Fuse Brown	Disagree	6
Melinda Buntin	Disagree	8
Michael Cannon	Uncertain	7	The proposal may still be worthwhile if it substantially affects other outcomes of interest.
Lawrence Casalino	Disagree	5
Amitabh Chandra	Disagree	7
Lanhee Chen	Uncertain	5
Michael Chernew	Disagree	7
Janet Currie	Strongly Disagree	9	Work requirements have been tried already in several states. They do result in disenrollment, but since many working age people on Medicaid are either disabled or already working, the requirements do not have much effect on enrollment. Rather, by increasing the administrative burden of being enrolled, requiring paper work, and visits to welfare offices during working hours, they tend to disenroll people who are already working.
Lesley Curtis	Disagree	7
David Cutler	Disagree	4
Julie Donohue	Strongly Disagree	10	Evidence from work requirements 1115 waiver in Arkansas showed no impact on employment.
Joseph Doyle	Disagree	7
David Dranove	Uncertain	4
Stacie Dusetzina	Strongly Disagree	8	The size of the eligible to work but not working population is not large enough to create a substantial change in employment.
Liran Einav	Agree	1
Jose Escarce			Did not answer
Elliott Fisher	Strongly Disagree	9	The best data comes from the evaluation of Arkansas work requirements which found no effect (NEJM dif and dif evaluation of first year). Several other follow up stuides supported this.
Richard Frank	Disagree	8
Craig Garthwaite			Did not answer
Darrell Gaskin	Disagree	8
Martin Gaynor	Strongly Disagree	8	Evidence does not support the notion that a work requirement will lead to increased employment.
Sherry Glied	Strongly Disagree	7
David Grabowski	Disagree	9
Jonathan Gruber	Strongly Disagree	9
Vivian Ho	Uncertain	10	Garthwaite et al 2014 QJE study suggests that there will be a substantial rise in employment, but multiple less rigorous studies argue this won’t happen.
Jason Hockenberry			Did not answer
Haiden Huskamp	Strongly Disagree	9
Benedic Ippolito	Disagree	3
Anupam Jena	Uncertain	7
Nancy Keating	Strongly Disagree	9	Researchers have examined Medicaid work requirements in Arkansas and found a significant reduction in the proportion of individuals covered by Medicaid but no meaningful change in employment rates. This is because most Medicaid recipients who can work are already working.
Aaron Kesselheim	Strongly Disagree	10
Jonathan Kolstad	Strongly Disagree	8
R Tamara Konetzka	Strongly Disagree	7	Limited evidence from states that have implemented work requirements suggests that employment effects will be minimal.
Rick Kronick	Disagree	9
Valerie Lewis	Strongly Disagree	8
Nicole Maestas	Strongly Disagree	10
Tom McGuire	Disagree	6
Ellen Meara	Strongly Disagree	9	Evidence tells us most people who are not exempt from work requirements on Medicaid already work.
Ateev Mehrotra	Strongly Disagree	8
David Meltzer	Uncertain	5
Joseph Newhouse	Strongly Disagree	7	Arkansas evidence is strong about little effect, but it is only one state.
Sean Nicholson	Strongly Disagree	7
Steve Parente	Agree	8	It happened in welfore reform in the 1990s.
Stephen Patrick	Strongly Disagree	10	Many Medicaid enrollees currently work. Evaluations of Medicaid work requirements suggest they do not work.
Harold Pollack	Disagree	6	The existing studies lead me to be pessimistic. I do believe we should have wide confidence intervals, and much depends on implementation.
Daniel Polsky	Disagree	8	The evidence to date on work requirements do not find meaningful changes in employment rates.
Ninez Ponce	Disagree	8
Thomas Rice	Strongly Disagree	9
Meredith Rosenthal	Strongly Disagree	10	Empirical evidence is clear on this point.
Joseph Ross	Strongly Disagree	9
Brendan Saloner	Disagree	7	My opinion is based on experiences from Arkansas
Kosali Simon	Strongly Disagree	8	1) Past research (and CBO report) shows that vast majority already meets the requirement 2) Administrative burdens will mean several who are meeting requirement will be unable to prove and lose coverage (research related to admin burdens, as synthesized by CBO reports including the one this summer)
Jon Skinner	Strongly Disagree	8
Ben Sommers	Disagree	7
Neeraj Sood	Uncertain	6	There is evidence from one state only that work requirements do not expand employment . Evidence from welfare reform suggests possibility of employment expansion.
David Stevenson	Disagree	7	Evidence of implementation to date seems clear, not to mention underlying work demographics of affected populations.
Kevin Volpp	Uncertain	7
Rachel Werner	Strongly Disagree	10

Question Two

The Congressional Budget Office’s estimate of Medicaid coverage losses related to OBBBA work requirement provisions (5.3 million people by 2034) is likely _____.

Name	Vote	Confidence	Comments
Margarita Alegria	Too Low	8
David Asch	No Opinion	0
John Ayanian	About Right	7
Peter Bach	About Right	4
Laurence Baker	About Right	6
David Blumenthal	About Right	4	The implementation of work requirements will be highly variable across states and depend on the interest of states in preserving access to Medicaid. This social experiment has not been tested in the past.
Erin Fuse Brown	About Right	4
Melinda Buntin	About Right	8
Michael Cannon	Too High	10	Medicaid’s matching-grant system determines each state’s grant according to a ratio — the FMAP. Medicaid further gives states substantial flexibility to game that system by tampering with the denominator. My sense is that the CBO does not fully account for the creativity that state policymakers, providers, and consultants bring to that process.
Lawrence Casalino	Too Low	5
Amitabh Chandra	About Right	8
Lanhee Chen	Too Low	8
Michael Chernew	About Right	2	I have not followed the CBO methods, but generally believe they are a good benchmark
Janet Currie	Too Low	8	I’m not sure if CBO included effects due to eligible children becoming disenrolled when their parents lose coverage.
Lesley Curtis	Too Low	8
David Cutler	About Right	2
Julie Donohue	Too Low	6	Very hard to know given the uncertainty about what guidance CMS will give to states and how states will implement but my sense is that these requirements will be more disruptive than CBO estimates
Joseph Doyle	About Right	6
David Dranove	About Right	7
Stacie Dusetzina	About Right	8	Though CBO methods aren’t usually fully explicated, there are prior studies and experiences related to work requirements and administrative burden / enrollment verification that provide strong inputs for models that project coverage losses.
Liran Einav	No Opinion	0
Jose Escarce			Did not answer
Elliott Fisher	Too Low	6	While other studies suggest greater losses, the much more important point is that strong recent evidence shows a strong effect of Medicaid coverage on mortality. Any increase is likely to be harmful.
Richard Frank	About Right	7
Craig Garthwaite			Did not answer
Darrell Gaskin	About Right	7
Martin Gaynor	About Right	6
Sherry Glied	About Right	5
David Grabowski	About Right	8
Jonathan Gruber	About Right	6
Vivian Ho	About Right	8
Jason Hockenberry			Did not answer
Haiden Huskamp	Too Low	7
Benedic Ippolito	Too High	6	Financing structure gives states a strong incentive to help workers satisfy community engagement requirements. I expect states will be nontrivially successful in doing so (particularly blue states with large expansion enrollment, like CA and NY).
Anupam Jena	No Opinion	0
Nancy Keating	About Right	6	Hard to know for sure-a lot will depend on how states implement the work requirement reporting procedures. If these are complex it could be many more who lose Medicaid coverage.
Aaron Kesselheim	Too Low	7
Jonathan Kolstad	About Right	5
R Tamara Konetzka	About Right	7	Good estimate based on limited evidence, but with some uncertainty.
Rick Kronick	No Opinion	0
Valerie Lewis	About Right	5	I don’t know the literature on this that well, so I expect there will be large losses, but I don’t personally have an intuition as to the extent. I generally trust CBO for these type of estimates.
Nicole Maestas	Too Low	5
Tom McGuire	About Right	6
Ellen Meara	About Right	7	The estimate is likely a lower bound, but a reasonable one based on evidence from Arkansas work requirements. The OBBBA has several features that could add to administrative burden, and therefore some estimates of lost Medicaid are higher. The answer is – it depends how states respond…
Ateev Mehrotra	About Right	3	This one is tough. Given experience of Arkansas, I hope that many states will make it logistically easier for Medicaid beneficiaries to demonstrate that they are working or are exempt.
David Meltzer	No Opinion	0
Joseph Newhouse	About Right	5	As a point estimate it is probably near an expected value but the uncertainty is large.
Sean Nicholson	About Right	8
Steve Parente	Too Low	6
Stephen Patrick	About Right	6
Harold Pollack	About Right	6	Same comment–wide confidence interval
Daniel Polsky	About Right	7	The CBO brings together evidence for projections better than any other source. I expect there is a wide confidence band around these projections, but I have no basis for providing an opinion they are either too high or too low.
Ninez Ponce	Too Low	8
Thomas Rice	About Right	6
Meredith Rosenthal	About Right	8	I am not sure how much to trust CBO in the current administration — ordinarily I would have said that they use the best available evidence for these estimates.
Joseph Ross	Too Low	5
Brendan Saloner	About Right	6
Kosali Simon	About Right*	8*
Jon Skinner	About Right	6
Ben Sommers	Too Low	5	Really hard to know what states will be able to do on this timeframe and how aggressively they will aim to use ex parte renewal to protect coverage
Neeraj Sood	Too High	5	Many assumptions and scant evidence base. Depends on how states implement
David Stevenson	About Right	5
Kevin Volpp	About Right	7
Rachel Werner	Too Low	8

Question Three

Most people disenrolled from Medicaid because of the OBBBA’s work requirements will be individuals who lose coverage due to challenges submitting documentation demonstrating that they are compliant or exempt, rather than individuals who lose coverage due to unemployment.

Name	Vote	Confidence	Comments
Margarita Alegria	Agree	8
David Asch	Agree	6
John Ayanian	Strongly Agree	9
Peter Bach	Agree	7
Laurence Baker	Agree	8
David Blumenthal	Agree	8
Erin Fuse Brown	Agree	8
Melinda Buntin	Strongly Agree	8
Michael Cannon	Uncertain*	7*	What is more relevant is that they are likely to be the healthiest of the Medicaid-expansion enrollees, which further suggests that the health impact of those coverage losses is likely to be small or zero.
Lawrence Casalino	Agree	5
Amitabh Chandra	Uncertain	7
Lanhee Chen	Uncertain	5
Michael Chernew	Agree	6
Janet Currie	Agree	8
Lesley Curtis	Strongly Agree	10
David Cutler	Agree	6
Julie Donohue	Strongly Agree	8
Joseph Doyle	Agree	9
David Dranove	No Opinion	0
Stacie Dusetzina	Agree	8	Prior research and the extent of administrative burden and coverage losses during redeterminations will be more impactful.
Liran Einav	Uncertain	1
Jose Escarce			Did not answer
Elliott Fisher	Agree	9
Richard Frank	Agree	8
Craig Garthwaite			Did not answer
Darrell Gaskin	Agree	7
Martin Gaynor	Uncertain	5
Sherry Glied	Agree	6
David Grabowski	Uncertain	5
Jonathan Gruber	Uncertain	1
Vivian Ho	Strongly Agree	10	State management of Medicaid has become a bureaucracy nightmare.
Jason Hockenberry			Did not answer
Haiden Huskamp	Uncertain	3
Benedic Ippolito	Uncertain	1	Agree this is possible, but I am uncertain about how representative existing evidence (e.g., Arkansas) will be of other states.
Anupam Jena	Agree	7
Nancy Keating	Strongly Agree	9	Administrative barriers and confusion about reporting requrements were the biggest driver of uninsurance in the Arkansas study.
Aaron Kesselheim	Agree	10	This is a feature, not a bug.
Jonathan Kolstad	Disagree	7
R Tamara Konetzka	Agree	7	This is what the available evidence suggests.
Rick Kronick	Agree	8
Valerie Lewis	Agree*	5*
Nicole Maestas	Strongly Agree	8
Tom McGuire	Agree	7
Ellen Meara	Strongly Agree	10	again, the evidence is very strong here and most people subject to requirements already work!
Ateev Mehrotra	Strongly Agree	8
David Meltzer	No Opinion	0
Joseph Newhouse	Strongly Agree	7	This overlaps with question 1; again the Arkansas evidence is strong, but it is only one state.
Sean Nicholson	Agree	6
Steve Parente	Strongly Disagree	8	Most of the population in question have government-issued phones or access to a smartphone. This will be an app affirmation process.
Stephen Patrick	Strongly Agree	9	As evidenced by recent Medicaid redetermination, these structural barriers frequently result in lower enrollment and/difficulty enrolling.
Harold Pollack	Agree	5	An unwise policy. Many uncertainties and calculated risks. The only upside to the work requirement is that this may provide a politically dignified path for some Republican officials to support Medicaid.
Daniel Polsky	Agree	5	This statement would be consistent with past experiences with work requirements. It is hard to be highly confident in this future projection given that the administration of this massive change has yet to be set in regulation.
Ninez Ponce	Agree	7
Thomas Rice	Strongly Agree	10
Meredith Rosenthal	Strongly Agree	10	Again, the empirical evidence is clear.
Joseph Ross	Strongly Agree	9
Brendan Saloner	Agree	8
Kosali Simon	Strongly Agree*	8*
Jon Skinner	Agree	5
Ben Sommers	Strongly Agree	8	Prior literature is pretty good on this topic
Neeraj Sood	Uncertain	6	Evidence from 1 state only. Depends on implementation
David Stevenson	Strongly Agree	9	History is a guide to my answer.
Kevin Volpp	Uncertain	6	i think both will be reasons people lose coverage, not sure relative magnitude
Rachel Werner	Strongly Agree	10

* This response was captured after analysis was completed and is not included in the aggregate survey results.

Meet our panel of experts

Cornell Health Policy Center