Block grants are all the rage. Take the latest G.O.P. proposal to repeal and replace the Affordable Care Act: the Graham-Cassidy bill. It proposes to replace the current system and instead give grants to the states, essentially taking the funds the federal government now spends under the ACA for premium subsidies and Medicaid expansion and give those funds to the states as a lump sum with little regulation.
There is a complicated formula by which the bill proposes divvying up this money among the states. Many think the formula is unfair, that it benefits red states over blue states, and that it just flat isn’t enough money. These are incredibly important concerns. But let’s put them to the side for just a moment and consider the theory behind block granting. Is there any world, for instance assuming that the amount and allocation of the funding could be resolved (probably crazy talk), in which switching to block granting may actually improve upon the status quo?
Proponents of block granting health care make two main arguments. First, it will reduce costs. By block granting Medicaid and the ACA subsidies, we end the blank check open entitlement that these programs have become and give states more skin in the game. Second, these cost savings will come from empowering states to innovate. States will become more efficient, improve quality, and solve their own state-specific problems.
These arguments have an understandable appeal. But how will states really react to providing health care coverage on a budget?
There are at least three options. They can find innovative ways to reduce costs and even improve quality, as supporters hope. They could simply cover fewer people or cover fewer services (essentially, cut corners on coverage). Or they could add more of their own state funds to make up for the shortfall in what the government provides, which is more or less what has happened in Canada. As someone from the great state of Illinois, I must state the obvious that many state governments don’t have a lot of spare cash on hand.
Which result is most likely? This is an age old problem in outsourcing a service for a flat fee. How does the principal (here the federal government) control the agent (here the state) to ensure that they innovate and enhance efficiency rather than cut corners? Monitoring is one tool. But we know that monitoring is practically very hard to do and costly to even attempt. And perhaps more to the point, it is not currently in the cards. The whole idea behind Cassidy-Graham is for the federal government to be hands off.
Another tool at the federal government’s disposal is to credibly threaten not to renew the “contract” if the state does a bad job. “Does a bad job” is admittedly a loaded statement. Part of the problem is a difference in value judgments. Most liberals want to see more poor people have insurance coverage, even if through Medicaid. Many conservatives would argue that poor people will be better motivated to find jobs that provide insurance if we don’t give it to them for free. I’ll show my hand that I find the latter approach to be highly flawed and not supported by the data. Notably, if we reduce short-term cost by reducing access to care, particularly that will prevent people from getting sick, it will just serve to increase long-term cost. But back to the point—although it is theoretically possible for the federal government to take back the block grant or get a private entity to come in and run health care for the state, there is no credible threat of that in the bill. States keep getting the grants even if they do a bad job.
The final possibility—at least in terms of traditional economic arguments—is that the federal government could refuse to guarantee a particular level of funding, giving the state an incentive to “do a good job” to get the highest possible level of funding in year two and so on. A state might even risk losing funds in year two if it underperforms (again, I concede the same definitional problem). This is essentially the concept behind Accountable Care Organizations (ACOs). But the bill doesn’t do this either. It commits up front to how much money the states will get each year. They get their fee regardless of how they run their program, through 2026.
In terms of traditional economic arguments, then, there is reason for concern. States might very well take their grants and reduce benefits or cover fewer people. (I must flag that I have written about similar problems arising in the outsourcing of other government services.) Ironically, it is this concern—over how many insureds would lose coverage—that defeated past G.O.P. attempts to repeal and replace Obamacare.
There is one last potential saving grace. State actors may be intrinsically motivated to innovate and not cut corners. A large body of literature from the behavioral sciences suggests that intrinsic motivation is a strong force, likely to be strongest in situations with a strong moral framing (such as donating blood, and one could hope, the provision of quality health care). Perhaps giving the states a grant with few strings attached would signal trust and motivate states to do their best—even as disagreement over end goals persists. But the HMO experiment, here, should give pause. States might be different in lacking a profit motivation, but HMOs were delivering health care and many still cut corners despite negative implications for care.
Perhaps after all of this, the only real answer is that we don’t know the answer. Maybe there is something about states that just make them better suited to do this job than the federal government. (On this, check out Nick Bagely’s recent paper.) At the very least, we need more data before moving the whole system to a block grant. Ironically, a mechanism for state experimentation is already built into the ACA—the system of waivers. And we’ve already seen that waivers can be a genuine source of state-level experimentation on a narrower, and more controlled scale. Arguments of resource allocation may be crippling, but even if we could get past those, there is tremendous risk in moving to a block grant system absent data to support its effectiveness.