Medicaid block grants (Tennessee Edition)

Late last week, Politico reported that Tennessee is about ready to jump head first into asking for a Medicaid block grant for the traditional federal matching rate populations.

More than 1.3 million people are enrolled in Tennessee’s Medicaid program, known as TennCare, and the state is one of 14 that has not joined Obamacare’s Medicaid expansion for low-income adults. The federal government on average covers about two-thirds of Medicaid costs in Tennessee…
State Republican lawmakers on Thursday, emboldened by the Trump administration’s promise to provide states with more flexibility to run their Medicaid programs, approved legislation requiring Tennessee to submit a Medicaid block grant plan to the federal government within six months. The legislation now goes to Republican Gov. Bill Lee, who will sign the bill

The logic of a block grant is that Tennessee has weak incentives to save money. Right now, if Tennessee spends a dollar, the Feds kick in two bucks. If Tennessee works hard to save three dollars, they only get to keep a buck. A block grant means that Tennessee is on the hook for every marginal dollar spent and Tennessee can keep all the savings from its efforts to reduce spending.

There are several major problems with block grants. The first and the biggest is that block grants are not responsive to shocks. Shocks can be anything; they can be a major storm like Hurricane Maria in Puerto Rico, they can be mass unemployment like 2008-2009, they can be a new technology like Hep-C cures or they can be a new disease like Zika.

December 2016 I wrote about Zika and block grants:

We also know that locally transmitted Zika infections will not be uniformly distributed. Alaska will have far fewer proportional Zika infections than Florida…
The Medicaid block grant procedures would give states a fixed head payment for each enrollee. It could vary by category of assistance and a few other criteria but the fee would be flat within subgroups by the number of enrollees….
here is where there is a problem. The capitated payments would be based on average expected costs in year 1 and then get weaker. States with disproportionate clustering of high cost conditions will be significantly worse off. Long run Zika neurological impairments will hit warmer states’ Medicaid budgets much harder and more disproportionally than Zika will hit cold weather states’ Medicaid budgets. This could be adjusted for by having a Zika bump in the block grant calculation much like there could be a diabetes bump or a maternity bump or any other number of risk adjusted bumps to capitation payments. But what happens when there is a new high cost and very concentrated disease that will have major impact on a few states’ Medicaid budgets? The block grant system fails unless there is a side payment of new federal funds…

Block grants are not designed to respond to negative shocks.

Secondly, even if the block grants are designed reasonably well and are risk adjusted, nasty incentives to selectively deny care will pop up. Risk adjustment will frequently fail at the individual level even if it is good at the group level. For example, hemophilia is risk adjusted in the ACA at between 61 and 62 times average state monthly premium for the 2020 individual market. That translates to about $455,600 dollars using 2019 Open Enrollment average premiums. We know hemophilia has an extreme tail. Some boys with this coagulation disorder can be treated for significantly under the mean cost of care. There are a few predictable, multi-million dollar claims as we have discussed before. A block grant structure (even if risk adjusted) would strongly encourage a state to place treatment limitations and caps on care to limit the outlier expense. The incentive for the state would be stronger to drive extreme and persistent cost outliers to move out of state and become someone else’s problem.

I’m using hemophilia as the cleanest example of something that is both expensive and highly variant. Less expensive conditions with high variance also invite risk adjustment games where the state is happy to pay for folks who have predictable below average costs while running off the folks who have above average costs but who also provide a good amount of the risk adjustment value.

The other major problem with block grants is that they are often linked to “flexibility” which means easier service cuts. Expensive services that help few people and are not seen as randomly distributed and randomly occurring would be the logical target of a state trying to spend as little as possible while not angering too many people.

It is possible to design a block grant program that is responsive to shocks, risk adjusts well at the group level with appropriate secondary payments to handle outliers and provide greater flexibility while improving care. It is possible. It is not the priority of any of the block grant programs that have been submitted to Congress in the past three years.