This is my fourth blog on the topic of sustainable health spending (others are here: 1, 2, 3).  I have argued that under the Patient Protection and Affordable Care Act (ACA), long-run sustainability in health spending is largely determined by the federal government’s ability to meet its Medicare, Medicaid, exchange subsidy and other health obligations under a balanced budget. Cost growth beyond what the government can afford will not be sustainable. 

Sustainable cost growth for the government will also be affordable to individuals because of the subsidies in ACA designed for that purpose. I have developed a model, described in earlier blogs, that computes the growth rate that is affordable by the government (hence sustainable) under alternative assumptions about what the public is willing to provide in tax revenues and the share of revenues going to non-health spending.

In this blog, I apply that model to assess the sustainability of a health spending growth that equals GDP growth between now and 2035. This rate of growth, commonly referred to as “GDP+0”, would keep the health spending share at its current level of 18 percent ad infinitum. 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 18 percent. For example, the Centers for Medicare and Medicaid Services project that the share will rise to nearly 20 percent by 2020. A 2007 paper by Hall and Jones states that "optimal health spending is predicted to rise to more than 30 percent of GDP by the year 2050 in most of our simulations. ...".(1)  

Clearly GDP+0 is below what most observers would consider the upper limit to sustainable growth.  We might therefore expect this rate to be easily sustainable. Let’s see.

I begin by identifying the share of GDP required in federal tax revenues to sustain health spending growth of GDP+0 under different assumptions about non-health spending requirements and compare it to the historical average. Then I identify the underlying insurance “cost trend” associated with this rate of growth in spending and compare it to the likely growth in median incomes.  I conclude by briefly commenting on the implications for sustainability.

Federal Tax Revenues Required to Support Spending Growth of GDP+0 

The chart below, taken from my previous blog, shows sustainable growth rates associated with tax revenues between 18 and 22 percent of GDP, combined with assumptions about what the government must set aside for spending on non-health items such as social security and national defense. The blue line corresponds to 13.1 percent of GDP allocated to such non-health spending and is the lower of two estimates given by the Congressional Budget Office (CBO). They characterize the discretionary component of this non-health spending as the smallest in over 50 years, so it is low indeed. The red line corresponds to only 10.7 percent of GDP set aside for non-health items and is taken from a CBO analysis of the Ryan budget path.(2) This is obviously much lower than the already low CBO estimate.  The yellow area in the chart represents my estimate of the most plausible combinations of revenues and non-health spending set asides, though I must say the yellow area that is closer to the blue line than the red line seems more feasible. The horizontal dotted line represents my projected average annual growth rate in GDP of 4.8 percent.(3)  

The chart shows that if GDP+0 is to be sustainable through 2035, and tax revenues are to be kept at their long run average level of 18 percent of GDP, nonhealth spending must be reduced to less than the  exceptionally low amounts specified in the Ryan path.(4) If non-health spending is kept to the low estimates from CBO, tax revenues must grow to 20.6 percent of GDP—well above the historical average of 18 percent. The takeaway is that for growth of GDP+0 to be sustainable, it will almost certainly be necessary for tax revenues as a share of GDP to increase to 20 percent or more.(5)

 
_{Source: Altarum Center for Sustainable Health Spending}

GDP+0 and the Per Beneficiary Cost Trend

Suppose that there is a single underlying rate of increase in the cost of delivering health care that affects all insured groups equally.(6) Specifically, let’s assume an equivalent rate of growth in spending per beneficiary for Medicare beneficiaries, Medicaid beneficiaries, and the privately insured.(7) I’ll refer to this as the “cost trend”. The overall rate of increase in national health spending is primarily the result of this trend.  But it is inflated by population growth, population aging, and increased spending associated with expanded coverage (under the ACA or its replacement).(8) I take the impact of population growth, aging and expanded coverage as given.  Thus, achieving a target rate of increase in national health spending must be accomplished through adjustments to the cost trend.

What trend is needed in order to hold health spending to GDP+0? The answer is 3.4 percent, as shown in the stacked bar chart below.(9) Population growth adds 0.8 percent to the annual growth rate and population aging adds 0.5 percent.(10) The costs of expanded coverage under ACA (or a comparable coverage plan) add about 0.1 percent to average annual growth through 2035. In summary, if national health expenditures are to grow at GDP+0  between now and 2035, the insurance cost trend must be held to 3.4 percent – about 0.6 percentage points below the growth rate in per capita GDP.

_{Source: Altarum Center for Sustainable Health Spending}
   
Cost Trend and Median Income  

Holding the cost trend below the growth in per capita GDP would seem to bode well for employers offering insurance. However, for each of the past three decades, median income growth has lagged per capita GDP growth by a percentage point or more.(11)  If this continues through 2035, median income would be expected to grow at about 3 percent per year – 0.4 percentage points below the 3.4 percent cost trend. Thus, even under a GDP+0 scenario, health spending as a share of income would continue to increase for median income workers.  Between now and 2035, this would amount to an increase of about 13 percent.

Is GDP+0 Sustainable?

So is GDP+0 sustainable? My answer is yes, but not nearly as easily as might have been expected. Since this health spending growth rate keeps the health spending share of GDP constant, year after year, one might expect this to be a simple matter of maintaining the status quo. Instead we find that sustainability requires a significant increase in tax revenues as a share of GDP while simultaneously reducing non-health federal spending (apart from social security) to historically low shares of GDP. In addition, employers paying median incomes are likely to find health spending rising faster than wages for their workforce.  So why doesn’t status quo in the health spending share of GDP translate into status quo in terms of government spending and health spending shares of median incomes? The two main factors are the aging of the population and the (assumed) failure of median incomes to keep pace with per capita GDP.

Of course this begs the question as to whether GDP+0 is attainable in the first place. Our latest Health Sector Economic Indicators^(SM) brief shows health spending has been growing at about 4 percent for the past 12 months – well below the 4.8 percent long-term rate of growth projected for GDP.(12) There is much discussion about whether this low growth will persist or is simply the result of poor economic conditions. But, at least for the time being, health care costs are growing below the long run GDP+0 rate.

Please note: The second chart and accompanying text have been corrected for an error in the original posting.  The corrected cost trend is 3.4 percent (versus 3.6 in the original post) and the correct population aging impact is 0.5 percent (versus 0.3 percent in the original post). (Changed on July 16, 2012).

References

1) Hall, R., & Jones, C. (2007). The value of life and the rise in health spending. Quarterly Journal of Economics, 122 (1), 39-72.

2) The period of analysis is 2011 through 2035.  The difference between revenues and non-health spending represents what is available for federal health spending in 2035 and this determines the growth rate in health spending that the government can afford (see the previous two blogs for underlying assumptions).

3) This is based upon annual growth of 4.5 percent in potential GDP (PGDP) – 2.3 percent real, 2.2 percent inflation – and 0.3 percent for GDP to catch up to PGDP from its current depressed level.

4) For GDP+0 to be sustainable, 7.5 percent of GDP must be available for federal health spending in 2035. If revenues are 18 percent of GDP, non-health spending must be kept to 10.5 percent of GDP – which is 0.2 percentage points below the Ryan path. [One should not try to compare my non-health percentages to what might be computed from CBO data since I made adjustments for consistency with the national health accounts.]

5) This is not surprising in the context of health care and related social services, given that the nation has never before experienced an aging wave like that of the baby boomers.

6) This assumption is built into the model I use for estimating sustainable growth rates.

7) I refer here to individuals covered by Medicaid but not also covered by Medicare. The assumption of an equivalent insurance cost trend is consistent with a view expressed by many (e.g., Alice Rivlin, Len Nichols, and Arnold Milstein) that more efficient models of delivery and lower price growth must eventually permeate the entire health care system.

8) Our definition of sustainable spending includes the provision that the nation is meeting its access and quality goals. Therefore, my estimates all assume expanded coverage, and at the same cost as projected under the ACA (I estimate this as a 3.6 percent addition to health spending in the year of implementation). As shown in the chart, this adds only about 0.1 percent to the average annual growth rate in health spending over the 24 year period.

9) This is based upon my projected 4.8 percent growth rate for GDP which depends, most critically, on my assumption of 2.2 percent annual economy-wide inflation (as measured by the GDP deflator).

10) The Medicare population grows by 74 percent from 2011 through 2035 while the non-Medicare population grows by only 9 percent.

11) Since 1980, per capita GDP has grown 1.1 percentage points faster than median income.

12) Economy-wide inflation has been at about 2 percent, consistent with what is assumed in the long term GDP growth projection.