The Urban Institute Report Did Not Score Medicare for All


Share  Tweet

The Urban Institute and The Commonwealth Fund put out a report today that provided cost estimates for 8 different health care reforms. One of those eight reforms, which is called “Single Payer Enhanced,” describes a comprehensive, no-cost-sharing, national health plan similar to Medicare for All (M4A). However, the Urban plan uses hospital reimbursement rates that are 15 percent higher than the rates in the M4A legislation, meaning that its cost estimates are much higher than the actual costs of M4A.

Urban is certainly welcome to put forward any health plans they can think of and score them to the best of their ability. But it is important for media to understand that this is Urban’s single payer plan, not the Medicare for All plan supported by Bernie Sanders, Elizabeth Warren, and many other Congressional Democrats.

There are reports that have attempted to score Medicare for All as it is actually written, with the most prominent being the one produced by Charles Blahous at the right-wing Mercatus Center. That report found that M4A would reduce national health expenditures by $2 trillion in its first decade.

Urban’s Unserious Approach

Urban’s approach to scoring its own single payer plan is also separately not credible. There are three issues I want to highlight here.

First, Urban’s headline estimate assumes insurance administrative costs will be six percent even though Medicare administrative costs are currently 1.3 percent. The report includes an alternative estimate where they assume the administrative costs will be three percent (rather than six percent) and find an additional $77.2 billion in annual savings. Based on traditional Medicare’s performance, this number is much more plausible.

Second, Urban sets the hospital reimbursement rates at 115 percent of the current Medicare rates, claiming that rates lower than that would lead to a massive contraction in the supply of hospital services. Urban can set the rates wherever they want, of course. But the claim that the supply of hospital services will collapse if Medicare reimbursement rates are used is not credible.

The only support Urban provides for this claim is a citation to a two-page fact sheet produced by the American Hospital Association, a lobbying organization. The fact sheet says that Medicare reimbursement rates currently only cover 87 percent of the cost of providing services to Medicare patients. From there, Urban apparently divided 13 by 87 to conclude that single payer reimbursement rates need to be 15 percent higher than Medicare currently pays.

The problem with this simplistic approach is that it assumes that hospitals will have the exact same costs under single payer as they have under the current system. But they won’t. Most notably, the administrative expenses of hospitals will go down. Hospitals in the US currently spend 25 percent of their revenue on provider-side administration. In single-payer nations like Canada and Scotland, that same figure is only 12 percent. A 13 percent reduction in provider-side costs would bring hospitals costs exactly in line with current Medicare reimbursement rates.

Separate from the administrative savings, any claim that the supply of hospital services will shrink dramatically if unit prices are pushed down needs to actually be proved, not just assumed. For this assumption to be true, you’d have to believe that hospitals are currently receiving no rents and that American doctors (the highest paid in the world) are somehow currently receiving their reservation wage. This is insane of course. The hospital sector is chock full of rents and doctors faced with modest pay cuts would not quit in droves because doctoring will still pay them more than any other job they could conceivably get.

Lastly, Urban’s utilization estimates are ridiculous. When scoring a comprehensive single payer plan, one of the key questions is what will happen when 30+ million more people get insurance and cost-sharing is eliminated for those who currently have insurance. How much more will they go to the doctor? Researchers aiming to answer this question generally scour various micro literature trying to show how much health care utilization goes up when cost-sharing goes down.

It’s unclear from the Urban report exactly how they came up with the utilization elasticities they use for most people. For instance, the report says that utilization for current Medicaid enrollees is “adjusted upward to account for the assumption that the single payer plan uses higher provider payment rates than the Medicaid program” without further elaboration. For the non-elderly, non-poor population, the report says they are treated initially as if they have an Obamacare plan and then they “adjust each person or family‚Äôs spending to account for any reduced cost sharing or additional benefits the single-payer approach may provide,” also without further elaboration.

Urban does provide clarity on the utilization elasticities they use for the elderly population (who will still have Medicare but no longer any cost-sharing). They simply make them up out of thin air:

We assume the following price elasticities of demand for health care by service type: -0.1 for Part A services, -0.2 for Part B services, -0.3 for Part D services, -0.3 for dental care, and -0.35 for vision or hearing care.

This kind of makes you wonder how the other unspecified utilization assumptions are made, but we will leave that aside for now.

Even when you have credible micro estimates of how much utilization goes up when cost-sharing goes down, there is still a separate question of whether those micro estimates scale to the macro level. For example, just because a select group of people used more health care when their deductibles were lowered, that does not mean aggregate health care utilization increased across the entire society. It could instead mean that health care utilization was redistributed: they got more of it and someone else got less.

This redistributive theory is supported by research into prior large expansions of public healthcare in the US. That research shows that after the expansions, health care utilization increased for beneficiaries of the new program but was offset by lower health care utilization elsewhere, suggesting that providers themselves manage demand increases by cutting unnecessary care elsewhere.

Urban’s report makes no effort to address the question of how micro utilization changes translate into aggregate utilization and, even more bizarrely, assumes that there are no capacity constraints on how many health care services can be performed. Specifically they say:

We do not assume limits on utilization of care because of supply constraints because our estimates assume a long-run equilibrium. That is, provider capacity expands to meet the increased demand for services that result from universal coverage, benefit expansion, and the elimination of cost-sharing requirements.

This is a comical assumption. Even if you discount the idea that utilization is largely redistributed rather than increased when health insurance is expanded, there is still a hard limit to just how much health care can be performed because there are only so many doctors and only so many facilities. But not for Urban. All of the increased demand from lower cost-sharing will somehow be met with a corresponding increase in health care supply, without limitation.

Taking the second and third points together, Urban’s position appears to be that reducing provider payment rates to the current Medicare rates will collapse health care supply, but also that there is infinite health care supply if you use provider payment rates equal to 115 percent of the current Medicare rates. This is a very strange-looking supply curve to say the least.

Ultimately the unseriousness of Urban’s scoring approach is beside the point. What matters in the immediate situation is that Urban’s report is not about Medicare for All, as they themselves acknowledge. Reporters implying otherwise are doing a disservice to their readers.