On July 24, our Center held a symposium in Washington, D.C., titled Sustainable Health Spending and the U.S. Federal Budget (full video here). Our sincere thanks to the Robert Wood Johnson Foundation for their support and to our amazing line-up of participants that included Ceci Connoly, Alice Rivlin, Joe Antos, Len Nichols, Arnie Milstein, Ziad Haydar, Donald Marron, Joanne Kenan, Karen Ignagni, and David Lansky. I was privileged to be among this group and to be paired with Donald Marron, who elevated our session with his terrific insights and commentary. This blog is the fifth in a series on sustainable health spending and was influenced by what I learned from Donald, Alice and others at the symposium.
In my previous blog, I posed the question as to whether “GDP+0” was a sustainable rate of growth in health spending. This refers to a scenario in which national health expenditures grow at the same rate as gross domestic product (GDP) and, therefore, the relentless rise in the health spending share of GDP is halted. I noted, “While there is much debate about the upper limit to health spending as a share of GDP, I haven’t seen anyone suggest that it must remain at [its current value near] 18 percent.” Therefore, one might expect GDP+0 to be easily sustainable. I went on to show that, in order to sustain this rate of growth, it would actually be necessary to increase tax revenues to record high levels and reduce defense and other non-health spending (not counting social security) to record low levels. Surprisingly, even this very low health spending growth rate will require lots of sacrifice to sustain.
I revisit GDP+0 here for two reasons. First, I would like to modify some assumptions I used in my previous analysis based on what I learned at the symposium. Second, Massachusetts has just adopted sustainable health spending growth targets for their state that, if applied nationally, would be quite similar to GDP+0 (news story here; bill summary here). I begin with a short description of the Massachusetts targets and how they would translate nationally. Then I discuss the (minor) modifications to my modeling assumptions and incorporate them into a new diagram that I call the “triangle of painful choices.” This triangle lays out the balanced-budget tradeoffs between federal tax revenues, federal non-health spending, and the rate of growth in national health expenditures. Finally, I apply the triangle to determine the U.S. budget implications if Massachusetts health spending targets were adopted nationally.
Massachusetts Health Spending Goal
As shown in the excerpt that follows, in order to “[Set] health care cost growth on a sustainable long- term path,” Massachusetts has adopted a growth goal over the next 10 years that is at, or slightly below, the “potential growth rate in the state’s gross state product.” This potential growth rate refers to the growth in the state’s “potential GDP” (PGDP), which is defined as what GDP would be at full employment. (For a discussion of the reasons for tracking health spending relative to PGDP rather than GDP, go here.) Thus, the Massachusetts goal can be summarized as holding state health spending to state PGDP+0. At the national level, this would mean holding health spending to national PGDP+0. GDP is currently well below PGDP, but we assume that in 2035 we will be at full employment, and GDP will therefore equal PGDP. For this to happen, the average annual GDP growth rate must be about 0.3 percentage points above that of PGDP. Thus, holding health spending growth to PGDP+0 is equivalent to GDP - 0.3 – a bit lower than GDP+0.
Modified Assumptions in Calculating Sustainable Health Spending
A detailed explanation of my methodology can be found here. In brief, I begin with a projection of what the public will be willing to provide in federal tax revenues in 2035 and how much of these revenues will be set aside for non-health outlays. The remainder is what the government can afford to spend on health. I compute the underlying health care cost trend that must be achieved in order for government health spending to be within what it can afford. Finally, I apply that cost trend systemwide to determine the resulting growth rate in national health expenditures.
In previous calculations, I did not set aside any government dollars to pay down the debt, i.e., I assumed a balanced “primary budget”. However, both Alice Rivlin and Donald Marron emphasized that a small primary surplus would likely be needed to stabilize the debt to GDP ratio at an acceptable level (to provide a cushion for future downturns). In my calculations here, I have assumed that the government will spend 0.5 percent of GDP on interest payments. For defense and other non-health spending, I previously used 7.0 percent of GDP. This was taken from the extended baseline scenario of the long-term budget outlook from the Congressional Budget Office. It is their low estimate, and I wanted to be very conservative on what was set aside for non-health. But Donald Marron noted that the historically low level is 8 percent, so I have decided to use that figure in these calculations.
The “Triangle of Painful Choices”
I call the following chart the Triangle of Painful Choices. Each point in the triangle represents a combination of tax revenues (the horizontal axis), government spending for defense and other non-health items (the vertical axis), and health spending growth rates (the diagonals), that result in a balanced budget in 2035. The share of GDP available for federal health spending is constant along the diagonals and is equal to: tax revenues minus the spending on defense and other non-health items, minus the 6.6 percent of GDP set aside for spending on Social Security and interest payments (the category of defense and other non-health spending does not include Social Security or interest payments and I assume 6.1 percent of GDP for Social Security and 0.5 percent for interest payments in all scenarios). In order to balance the budget, health spending growth must be consistent with the share of GDP available for federal health spending. Thus, at each diagonal in the triangle there is a specific rate of growth in health spending that will exactly balance the budget. I call it the triangle of painful choices because – to balance the budget and hence achieve a sustainable long-run fiscal future – we must choose a point in the triangle and none of the choices are comfortable.
For example, point A represents tax revenues at 18.5 percent of GDP and defense and other non-health spending at 9.3 percent of GDP. These are 2007 values and quite comfortable. However, at point A we only have 2.6 percent of GDP available for federal health spending and, using my model, health spending growth of PGDP- 4, i.e., 4 percentage points below the growth rate in PGDP(!), would be required for a balanced budget in 2035. This is not merely uncomfortable, it is not remotely feasible. Shifting from point A to point B gets to a more feasible health spending growth path (PGDP+1) but raises tax revenues to an uncomfortable 24.6 percent of GDP. Moving down the green line from point B to point C keeps health spending growth at PGDP+1 and brings tax revenues down toward the comfortable 18.5 percent, but spending on defense and other non-health becomes absurdly low.
Implications for U.S. Budget under a National Massachusetts Health Spending Target Through 2035
The red line running through point D represents a balanced budget with health spending held to PGDP+0, which is analogous to the Massachusetts goal of state PGDP+0. At point D, the balanced budget is achieved by setting spending on defense and other non-health items to 8.0 percent of GDP – matching the historical low – and increasing taxes to a record high 21.4 percent of GDP (chart below shows the historical record). There is nothing in this “solution” that is particularly comfortable. Spending for defense and other non-health items is historically low, tax revenues are above historical highs, and health spending growth is extremely low, particularly given the aging of the population. It is possible to move in a direction that is more comfortable in terms of one of the variables (e.g., move down and left along the red line from point D to reduce tax revenues), but this will increase discomfort in one of the other variables, e.g., spending on defense and other non-health items falls below the historical low of 8 percent of GDP.
This analysis suggests that even a health spending growth target as low as PGDP+0 (analogous to the target just approved for Massachusetts) is not sustainable without historically high tax revenues and historically low spending on defense and other health items. There is simply no way around these painful choices.