The policy analyst spreads three state implementation plans across her desk. Georgia counts every hour equally: employment, education, job search, volunteering all accumulate toward the same 80-hour threshold. Ohio proposes weighting activities differently, with workforce training counting 1.25 hours for every hour completed. Arkansas wants to adjust the threshold itself, reducing requirements for members facing documented barriers.
Same federal mandate. Radically different implementations. The One Big Beautiful Bill Act specifies 80 hours of qualifying activities monthly, but Congress left states extraordinary discretion in how to structure those hours. Her governor wants a recommendation by Friday.
She knows the tradeoffs. Georgia’s equal-hour model offers administrative simplicity but treats job searching while homeless the same as job searching with a car, childcare, and internet access. Ohio’s weighted model acknowledges that some activities generate more long-term value but requires verifying not just whether someone participated but what kind of participation counts. Arkansas’s barrier-adjusted model recognizes human complexity but opens questions about who qualifies for reduced requirements and how to prevent gaming.
Her spreadsheet shows the math. Under Georgia’s model, a single mother attending community college full-time while caring for children cannot reach 80 hours without also working. Under Ohio’s model, her 15 credit hours of coursework might count as 56 weighted hours, leaving a manageable gap. Under Arkansas’s model, her caregiving responsibilities might reduce her threshold to 60 hours, making compliance achievable.
Three philosophies. Three definitions of what work requirements are meant to accomplish. Her recommendation will shape whether her state’s policy promotes workforce attachment, human capital investment, or coverage stability. None of these goals is wrong. But they cannot all be maximized simultaneously.
The Design Space#
OBBBA establishes the federal floor: 80 hours monthly of qualifying activities for expansion adults not otherwise exempt. But the legislation leaves states substantial room to define what qualifies, how activities are measured, and how partial compliance is treated. This design space creates a natural experiment across 50 states, each making choices that reflect different theories about what work requirements should accomplish.
The qualifying activities specified in federal law include employment, job search, education, vocational training, community service, and caregiving for non-dependents. States can add activities but cannot subtract from the federal list. They can define what counts as legitimate participation within each category. They can establish documentation requirements that effectively narrow or broaden access. And critically, they can decide whether all hours count equally or whether some activities generate more credit than others.
The simplest implementation treats the requirement as a straightforward accumulation problem. A member needs 80 hours. Any qualifying activity contributes hour-for-hour. Verification confirms participation occurred. Compliance is binary: either you reached 80 or you did not.
This simplicity has appeal. It minimizes administrative complexity, reduces verification disputes, and creates clear expectations. Members know exactly what they need to do. Caseworkers can assess compliance without subjective judgment. Systems can be automated around simple counting.
But simplicity comes at a cost. An hour of employment differs meaningfully from an hour of job searching. Employment generates income, builds work history, and demonstrates labor market attachment. Job searching is necessary but produces nothing until it succeeds. Treating them identically ignores the policy’s underlying purpose of promoting workforce participation.
Similarly, an hour of community college coursework differs from an hour of watching job training videos online. The coursework leads to credentials that improve long-term employment prospects. The videos might satisfy a requirement without building human capital. Equal treatment creates no incentive to choose more valuable activities.
The design space extends beyond activity weighting to the treatment of partial compliance. What happens to someone who logs 75 hours? Under binary models, they lose coverage entirely, the same consequence as logging zero hours. Under graduated models, they might receive prorated benefits or extended grace periods. The choice between cliff effects and gradual consequences shapes member incentives and coverage stability.
States must also decide how to handle measurement periods. Monthly verification creates twelve compliance checkpoints annually, each an opportunity for administrative failure. Quarterly measurement reduces verification burden but allows longer periods of potential non-compliance. Some states are exploring rolling averages that smooth month-to-month variation.
Finally, states can adjust requirements based on documented circumstances. A member facing housing instability, serious mental illness, or caregiving responsibilities might receive a reduced threshold rather than exemption. This creates a middle ground between full 80-hour requirements and complete exemption, acknowledging that barriers exist on a spectrum rather than as binary conditions.
Activity Weighting Models#
Four distinct approaches to activity valuation have emerged from state planning documents and policy debates. Each reflects different assumptions about what work requirements should accomplish.
The Equal Hour Model
Georgia’s implementation exemplifies the equal hour approach. One hour of qualifying activity equals one hour toward the 80-hour threshold regardless of activity type. Employment, job search, education, training, and volunteering all accumulate identically.
The model’s logic is straightforward: the requirement exists to ensure members are doing something productive, and the state should not privilege one form of productivity over another. A member volunteering at a food bank contributes to their community just as a member working a retail job contributes to the economy. Both demonstrate the reciprocal obligation that justifies public benefit.
Administrative simplicity is the model’s primary advantage. Verification requires confirming that an activity occurred, not categorizing it for differential treatment. Disputes about whether a particular hour counts as training versus job search become irrelevant. Systems can count hours without complex classification logic.
The disadvantage is that equal treatment creates perverse incentives. If a member can satisfy requirements through low-effort job searching as easily as through employment, rational actors choose the easier path. The model does nothing to encourage actual workforce attachment, which undermines the policy’s stated purpose.
Research on the original Arkansas implementation found that members who lost coverage were overwhelmingly those struggling with employment, not those refusing to work. Equal-hour models do not distinguish between inability and unwillingness, applying the same consequence to both.
The Productivity-Weighted Model
An alternative approach assigns differential credit based on labor market productivity. Employment hours count at full value. Training and education count at reduced rates, perhaps 0.75 hours of credit per hour completed. Job search and volunteering count at further reduced rates, perhaps 0.5 hours per hour.
The logic is that employment represents the policy’s ultimate goal, and other activities are valuable only insofar as they lead to employment. A member working 40 hours weekly is demonstrating exactly the behavior the requirement seeks to encourage. A member spending 40 hours in training is making progress but has not yet achieved workforce attachment.
This model creates clear incentives for employment. A member can satisfy 80-hour requirements through 80 hours of work, 107 hours of training, or 160 hours of job searching. The math pushes toward employment as the path of least resistance.
The disadvantage is that productivity weighting penalizes human capital investment. A member choosing community college over immediate low-wage employment is making a rational long-term decision that the model punishes in the short term. It privileges any job over preparation for a better job, potentially trapping members in low-wage work that satisfies requirements but offers no advancement path.
The Investment-Weighted Model
Inverting the productivity logic, some states are considering investment-weighted models that credit education and training at higher rates than employment. A member in an accredited degree program might receive 1.5 hours of credit per hour of coursework. Workforce training through approved providers might count at 1.25 hours.
The underlying theory is that work requirements should promote long-term self-sufficiency, not just immediate workforce attachment. A member who completes a nursing degree will never need Medicaid again. A member cycling through minimum wage jobs might remain Medicaid-eligible indefinitely. Investment weighting encourages the education and training that produce permanent exits from public assistance.
Implementation requires defining which educational activities qualify for enhanced credit. Accredited degree programs at recognized institutions are straightforward. Certificate programs, online courses, and informal training create classification challenges. States must build capacity to verify legitimate educational enrollment and distinguish it from nominal participation that generates credit without building skills.
The disadvantage is delayed workforce attachment. A member pursuing a four-year degree satisfies requirements without working, potentially for years. Critics argue this inverts the policy’s purpose, using work requirements to subsidize higher education rather than promote work.
The Barrier-Adjusted Model
Rather than weighting activities, Arkansas and several other states are exploring models that adjust the threshold itself based on documented barriers. A member with stable housing, transportation, and childcare faces the standard 80-hour requirement. A member experiencing homelessness might face a 40-hour requirement. A member with serious mental illness in active treatment might face 20 hours.
The logic acknowledges that capacity varies. Holding a homeless individual to the same standard as a stably housed individual ignores the reality that finding and maintaining 80 hours of qualifying activities requires baseline stability that not everyone possesses. Barrier adjustment preserves the principle of mutual obligation while calibrating the obligation to realistic capacity.
Implementation is substantially more complex than activity weighting. States must define which barriers qualify for adjustment, how much adjustment each barrier warrants, and how barriers are documented and verified. The system requires ongoing assessment as circumstances change. A member whose housing stabilizes should see requirements increase; a member who becomes homeless should see them decrease.
The approach also requires addressing barrier stacking. A member experiencing homelessness, serious mental illness, and limited English proficiency faces compounding challenges. Should barriers stack additively, with each reducing requirements by a set amount? Or should states establish minimum thresholds below which requirements are effectively waived?
Series 11 of this analysis documented populations for whom single-barrier accommodations fail at intersections. The barrier-adjusted model forces states to operationalize insights about compounding disadvantage, translating conceptual understanding into administrative categories with numerical consequences.
Verification Implications#
Each weighting model creates distinct verification requirements. The choice of model is simultaneously a choice about administrative burden, error rates, and the accuracy of compliance determination. States cannot select a weighting model without considering whether they can verify the activities that model privileges.
Equal-hour models require verifying that qualifying activities occurred but not distinguishing among them. A member claims 80 hours; the verification system confirms participation in something qualifying. The binary question, did this happen, is simpler than categorical classification. But the simplicity is somewhat illusory. Verifying job search activities remains difficult. Confirming volunteer hours requires attestation from supervising organizations. Employment verification depends on employer cooperation that is not always forthcoming, particularly for informal work arrangements and small businesses without HR infrastructure.
The infrastructure for employment verification exists in most states through connections to wage databases maintained for unemployment insurance purposes. But these databases have lag times, typically reporting quarterly rather than monthly, and miss cash employment entirely. Real-time employment verification requires employer cooperation that simple data matching cannot compel.
Productivity-weighted models add a classification layer. The system must determine not only whether activity occurred but what type of activity it was. A member claiming training hours must prove the training qualifies for training credit rather than mere job search credit. Disputes arise at category boundaries. Is watching online videos training or job search? Is an unpaid internship employment or volunteering? Is caring for an elderly neighbor caregiving that counts or informal help that does not?
Classification disputes require resolution mechanisms. Someone must decide whether a particular activity falls into one category or another. This creates administrative burden and introduces subjectivity that equal-hour models avoid. Different caseworkers might classify identical activities differently. Appeals challenging classifications add workload and delay final determinations.
Investment-weighted models require institutional verification. Credit hours must be confirmed with educational institutions. Training programs must be validated against approved provider lists. The administrative apparatus expands to include relationships with colleges, universities, and workforce development boards. Verification becomes a multi-party problem requiring data sharing agreements and reconciliation processes.
States without existing data connections to educational institutions face substantial infrastructure investment. Building these connections takes time and money. States rushing to meet federal deadlines may find investment-weighted models infeasible regardless of their policy appeal.
Barrier-adjusted models require assessment infrastructure that other models can avoid. Someone must evaluate each member’s circumstances, document qualifying barriers, and assign appropriate requirement levels. This assessment cannot be fully automated. It requires trained staff making judgment calls about complex situations. It creates appeals opportunities when members disagree with assessments. And it must be repeated periodically as circumstances change.
The assessment burden falls disproportionately on the populations Series 11 documented: those experiencing serious mental illness, substance use disorders, homelessness, and other conditions that make standard administrative interactions difficult. The very people who need barrier adjustments are often least equipped to navigate assessment processes.
The verification burden also differs by stakeholder. Equal-hour models place burden primarily on members and employers. Weighted models spread burden to educational institutions and training providers. Barrier-adjusted models create significant burden for MCO care coordinators and state eligibility workers who must conduct assessments.
States with limited administrative capacity may find equal-hour models more feasible regardless of their policy preferences. States with robust care coordination infrastructure may be positioned for barrier-adjusted approaches. The optimal model depends not only on policy goals but on implementation capacity.
The Cliff Problem#
All models must address what happens at the boundary. A member who logs 79 hours has demonstrated substantial compliance. Under binary enforcement, they lose coverage entirely, the same consequence as logging zero hours. This cliff effect creates arbitrary outcomes disproportionate to the marginal failure.
The economics of cliff effects deserve attention. A member at 79 hours has contributed 99% of the required reciprocal obligation. Treating them identically to someone who contributed nothing violates basic proportionality principles. It also creates perverse incentives: if falling one hour short produces the same consequence as not trying at all, members who realize mid-month they cannot reach 80 hours have no incentive to continue attempting.
Research from the Arkansas implementation found that most members who lost coverage were close to compliance. They were not refusing to participate; they were failing to document sufficient participation. The cliff transformed minor administrative gaps into complete coverage loss. Many had worked the required hours but could not prove it. Others faced temporary circumstances, a sick child, a car breakdown, that reduced one month’s hours without indicating chronic non-compliance.
The consequences of cliff enforcement extend beyond individual members. Providers lose patients mid-treatment. MCOs lose members whose risk scores were generating revenue. Care management investments evaporate when the member they were managing disappears. The healthcare system absorbs the costs of coverage gaps through emergency department visits and uncompensated care.
Several alternatives to cliff enforcement have emerged in state planning.
Prorated benefits would scale coverage to compliance level. A member at 75% compliance might retain 75% of benefits, losing some covered services while maintaining others. This preserves incentives while avoiding all-or-nothing consequences. Implementation is complex, requiring states to define which benefits can be prorated and how partial coverage interacts with provider networks and MCO contracts.
Grace period approaches give members time to cure deficiencies before coverage terminates. A member below threshold in January might have until March to demonstrate compliance before losing coverage. This acknowledges that month-to-month variation is normal and allows catch-up. The risk is that grace periods become de facto exemptions if members can perpetually cure deficiencies without ever achieving sustained compliance.
Hour banking allows members to accumulate credit during high-activity months that can offset low-activity months. A member working 100 hours in January banks 20 hours against future shortfalls. This smooths the seasonal variation common in low-wage employment and acknowledges that consistent 80-hour months may be unrealistic for workers in variable-hour jobs.
Quarterly measurement replaces monthly cliffs with quarterly assessment. A member needs 240 hours over three months rather than 80 hours each month. This provides flexibility for members whose circumstances fluctuate while maintaining annual hour requirements. Administrative burden decreases with fewer verification checkpoints, though quarterly gaps before detection may concern those worried about program integrity.
Each alternative creates different incentives and different risks. States must choose based on how they weigh coverage stability against consistent enforcement, administrative simplicity against equitable outcomes.
State Design Choices in Practice#
Georgia’s experience with the Pathways program illustrates what happens when states prioritize simplicity. The program requires 80 hours monthly of qualifying activities verified through a state portal. Activities count equally. Compliance is binary. The result has been substantial enrollment barriers: fewer than 7,500 enrollees against projections of 50,000 or more eligible individuals. The state spent over $90 million implementing the program, yielding a per-enrollee cost exceeding $13,000 annually before any healthcare services were delivered.
Analysis suggests the Georgia model’s strictness deters enrollment rather than promoting compliance. Potential members assess the likelihood of sustained compliance against the effort required and conclude the risk of coverage loss is too high. The simple model does not accommodate the complex realities of low-wage labor markets where hours fluctuate, jobs are temporary, and documentation is inconsistent.
Georgia has recently modified its approach, reducing reporting frequency from monthly to annual and adding caregiver exemptions for parents of young children. These modifications acknowledge that the original design created unsustainable burden without achieving enrollment goals. But the core equal-hour model remains, and enrollment remains far below projections.
The Georgia experience raises a fundamental question: does a program that costs more to administer than it saves in coverage actually serve the reciprocity goals it was designed to advance? If the answer is achieving coverage reduction rather than promoting work, simpler mechanisms exist. If the answer is promoting work, the evidence does not support the model’s effectiveness.
Ohio’s pending waiver takes a different approach. The state proposes weighting workforce training activities above employment hours, creating incentives for human capital investment. Ohio has significant community college and vocational training infrastructure it hopes to leverage. The theory is that members who complete training programs will achieve more stable employment than those who cycle through low-wage jobs satisfying hour requirements without building skills.
Arkansas is reconsidering its approach after the 2018-2019 failure. The state’s new “Pathway to Prosperity” proposal includes success coaching and personal development plans that create individualized pathways to compliance. Members work with coaches to identify achievable activity combinations given their circumstances. The model is more resource-intensive than Georgia’s but attempts to address the documentation failures that caused the earlier debacle.
These three approaches reflect different theories of change. Georgia assumes members will respond to clear requirements by meeting them; failure indicates unwillingness that should have coverage consequences. Ohio assumes members need incentives to make optimal choices about workforce investment; weighting shapes those choices. Arkansas assumes members need support to navigate complex systems; coaching fills capacity gaps that requirements alone cannot address.
Each theory has evidence in its favor. Behavioral economics supports the idea that clear simple rules generate compliance. Human capital theory supports the idea that investment in skills produces better outcomes than immediate low-wage employment. Implementation science supports the idea that navigation assistance improves outcomes for complex programs.
The federal mandate forces all states to implement something, but it does not force them to implement the same thing. The variation that emerges will generate evidence about which approaches work for which populations under which conditions. States that choose wrong will learn through coverage losses and political backlash. States that choose right will become models for future policy.
The Counterargument: Complexity Creates Gaming#
Critics of weighted and barrier-adjusted models argue that complexity invites manipulation. If training hours count more than employment hours, members will enroll in training programs to generate credit without building skills. If barriers reduce requirements, members will claim barriers they do not genuinely face. The more accommodations a system offers, the more opportunities for gaming.
This concern is not hypothetical. Disability determination systems struggle with claims that are difficult to verify. Educational benefits have faced fraud through diploma mills and phantom enrollment. Any system that offers differential treatment based on member characteristics creates incentives to present those characteristics strategically.
The gaming argument has particular force in discussions of barrier-adjusted models. If a member claiming homelessness receives a 40-hour requirement instead of 80, incentives exist to claim homelessness whether or not it is genuine. Verification of housing status is difficult. A member between stable housing situations might strategically present their circumstances to maximize accommodation. Multiply this across millions of members and the integrity concern becomes substantial.
Proponents of complex models respond that the alternative is worse. A system that prevents all gaming by offering no accommodations also excludes everyone who genuinely needs accommodation. The fraud prevention is achieved by failing to serve the hardest cases. If 10% of accommodation claims are fraudulent but 90% are genuine, eliminating accommodations to prevent the 10% harms nine legitimate claimants for every fraudster stopped.
The question is whether fraud risk justifies designs that exclude legitimate claimants. Research consistently shows that Medicaid work requirement enforcement errors fall disproportionately on members who qualify for exemptions but cannot navigate documentation requirements. The Arkansas experience found that most coverage losses occurred among members who were working, exempt, or both. The members who fell through cracks were not gaming the system; they were failing to prove compliance the system should have recognized.
States can mitigate gaming through verification design. Training programs that generate enhanced credit might require institutional enrollment verification rather than self-attestation. Barrier documentation might require clinical assessment rather than member reporting. These safeguards add administrative cost but preserve accommodation benefits while reducing fraud opportunity.
The empirical question remains contested. How much gaming actually occurs when accommodations are offered? Evidence from other benefit programs suggests gaming is real but often overstated in policy debates. Most people do not lie about their circumstances to gain marginal benefit. The transaction costs of deception, including maintaining false narratives and risking exposure, deter all but determined fraudsters.
The honest assessment is that all designs involve tradeoffs between access and integrity. Simple systems are harder to game but exclude legitimate claimants. Complex systems serve more people appropriately but create exploitation opportunities. States cannot eliminate this tradeoff; they can only choose how to balance it based on their values and their assessment of their population’s characteristics.
Conclusion#
The policy analyst finishes her recommendation. She proposes a hybrid model: equal-hour baseline with barrier adjustments for members with documented circumstances, plus quarterly measurement to smooth month-to-month variation. It is not the simplest approach, but her state’s population includes substantial numbers of members facing barriers that make monthly 80-hour compliance unrealistic.
She includes an honest assessment of the risks. Barrier adjustment creates classification disputes her agency will need to resolve. Quarterly measurement delays detection of non-compliance. The hybrid approach requires systems and staff her state does not currently have. She estimates 18 months to build the necessary infrastructure, which means her state will need the federal extension to implement properly.
She also includes the alternative: Georgia’s simple equal-hour monthly model that her state could implement with existing infrastructure. This approach will cost less to administer and will exclude members her staff would otherwise struggle to categorize. The exclusion is a cost. Whether it is a cost worth paying depends on values her analysis cannot resolve.
Her recommendation acknowledges uncertainty. No state has operated weighted or barrier-adjusted models at scale. The evidence base consists of Georgia’s troubled equal-hour experience, Arkansas’s failed 2018-2019 experiment, and policy documents from states that have not yet implemented anything. She is recommending based on theory and limited analogy, not proven results.
The federal mandate creates a floor, not a ceiling. States will build different structures on that floor, reflecting different philosophies about what work requirements are meant to accomplish. Some will emphasize simplicity and accept exclusion. Others will emphasize inclusion and accept complexity. Still others will experiment with novel approaches no one has yet attempted.
This variation is by design. Congress chose to mandate work requirements while leaving implementation details to states, creating the conditions for policy learning. Over the coming years, evidence will accumulate about which approaches maintain coverage, which promote employment, and which simply generate administrative burden without achieving either goal.
The analyst knows her recommendation will shape her state’s outcomes. She also knows that learning from other states will enable refinement over time. The first implementation need not be the final implementation. What matters is building the capacity to measure results and adjust based on evidence.
Weighted hours and activity credits are technical policy instruments. But they embody deeper questions about what society expects from the people it helps and how much variation in circumstances those expectations should accommodate. The answers states provide will determine whether work requirements function as pathways to self-sufficiency or as barriers that exclude people who cannot clear arbitrary thresholds. The technical choices matter because they determine who keeps healthcare coverage and who loses it.