The argument over California's use of Medicaid dollars for social services is usually staged as a partisan fight: a Democratic governor with national ambitions defending a program congressional Republicans call waste, fraud, and abuse.
That framing captures the politics and obscures the substance. Underneath is a genuine, unresolved question that predates this administration and will outlast it: when a medical program pays for non-medical things because they cause medical problems, where does it stop?
What California actually built
CalAIM, California Advancing and Innovating Medi-Cal, launched in 2022 under Biden-era waivers within Medi-Cal, which has a proposed budget of $217 billion and covers more than 14 million residents. The state has been selective about who receives help from 15 types of social services, targeting enrollees with complex needs rather than offering them broadly.
Since 2022, California has spent nearly $12 billion in joint state and federal money, with the hope of reducing long-term Medi-Cal spending by keeping enrollees out of costly institutions including emergency rooms, jails, nursing homes, and mental health crisis centers. As of September 2025, CalAIM had provided social services to more than 528,000 patients, with nearly 453,000 receiving intensive case management.
The case for it
The logic is straightforward and has real clinical grounding. A person with uncontrolled diabetes living in a car will end up in an emergency room repeatedly, and no amount of medication management fixes the underlying instability. An asthmatic child in a moldy apartment will keep coming back. Paying a security deposit or removing mold is, on this view, treating the cause rather than repeatedly treating the consequence, and it is cheaper.
Supporters point to operational results. Health Plan of San Joaquin CEO Lizeth Granados said CalAIM helped place homeless patients who were routinely hospitalized into housing, and that patients with uncontrolled diabetes saw blood sugar drop after nutrition counseling and home-delivered meals. She reported inpatient hospitalizations dropping to 44 per 1,000 members since it launched in 2022, down from 61 per 1,000 before CalAIM.
That is a substantial reduction if it holds up. It is also worth reading carefully: it is a single health plan's self-reported figure, without a control group, over a period in which many other things changed, including the end of pandemic disruptions to normal care. It is genuinely suggestive, and it is not the same as rigorous evidence that the program pays for itself. The honest state of the literature on health-related social needs interventions is that the clinical benefits are better established than the cost savings.
The stronger version of the objection
"Waste, fraud, and abuse" is a political characterization, and it does not describe what is actually contested. Nobody serious argues that paying an asthmatic's mold remediation is fraud in the legal sense. The substantive objections are different and worth stating properly, because they are more defensible than the slogan.
The first is definitional. Medicaid is health insurance, established to pay for medical care. If it can pay rent because housing affects health, the boundary becomes hard to locate, since almost everything affects health. Income affects health. Education affects health. A program with an unbounded definition of its own scope is difficult to budget and difficult to constrain, and dollars spent on rent are dollars not spent on medical services for other enrollees.
The second is about fiscal federalism, and it is the strongest technical objection. Medicaid is jointly funded, with the federal government matching state spending. Housing and nutrition programs exist elsewhere in government, funded through fixed appropriations. Routing those services through Medicaid instead lets a state draw federal matching dollars for activities it would otherwise fund entirely on its own. Whatever the merits of the services, that is a real shift of state costs onto the federal government, accomplished through waiver rather than through Congress appropriating money for housing. Someone can support housing assistance and still object to funding it this way.
The third is procedural. These programs operate under Section 1115 demonstration waivers, which are supposed to be experiments testing whether an approach works. Waivers granted by one administration and revoked by the next make for unstable policy, and the fact that CalAIM's authority ends at the close of 2026 illustrates the problem. Building durable benefits on discretionary federal permission means benefits that can vanish with an election.
The complication that undercuts both sides
Here is the fact the partisan framing tends to leave out, and it should temper both camps.
California is already cutting these programs itself, for budget reasons unrelated to Washington. The CalAIM service providing up to six months of temporary housing and ongoing care is ending at the close of this year, and the state is cutting recuperative care benefits, no longer paying for beds for patients to recover from illness or injury.
That complicates the Democratic position: a state genuinely convinced these services save money would be expanding them, not trimming them under fiscal pressure. Programs that pay for themselves do not usually get cut first.
It also complicates the Republican one: if California is voluntarily scaling back, the picture of a state recklessly expanding Medicaid into unlimited social spending is less accurate than the rhetoric suggests. The state has been selective about eligibility from the start, and is now narrowing further.
The likeliest explanation is the unglamorous one. The savings are real for some patients and slower or absent for others, the up-front costs are immediate while the offsets are diffuse and delayed, and a state facing budget pressure cuts the spending whose return is hardest to demonstrate on a two-year horizon, regardless of whether it works over ten.
What is actually at stake this year
The practical question is not whether Medicaid should ever fund social services. It is what happens when California's waiver expires at the end of 2026, and the answer will be decided by an administration that has been expected to reverse the waiver policies that allowed Medicaid money to be used for these experiments.
If the waiver is not renewed, roughly half a million people lose services they currently receive, and the state absorbs whatever costs follow. If it is renewed, the underlying questions about scope and federal matching remain unsettled, waiting for the next administration.
That instability is arguably the real problem, and it is one both parties have contributed to. If health-related social services are worth funding, the durable answer is Congress deciding so explicitly, with defined scope and appropriations. What is happening instead is a four-year cycle in which the same programs are good policy or fraud depending on who won the last election, while the people receiving mold remediation and rental deposits find out which it is when their coverage changes.