How to Recognize a Bent Health Care Cost Curve

As noted in our most recent Health Sector Economic Indicators release, health spending has grown at historically low rates for the past 28 consecutive months (September 2008 through January 2011). In addition, the health spending share of gross domestic product has remained stable since the start of the recovery in June 2009. These would seem to be positive developments, but do they indicate that the health care cost curve has started bending away from its unsustainable rate of increase? How would we recognize success?

The health care cost curve that I will focus on is the health spending share of GDP. Over the past 40 years this share has grown from about 7 percent of GDP to nearly 18 percent. This growing share of GDP is a problem in large part because roughly half of health spending is financed by the government. Thus continued cost growth is at odds with national deficit reduction goals.

Here is the health spending share of GDP curve from January 1990 through January 2011. It is characterized by rapid upturns in and around recessions and stability in between.

Source: Altarum monthly health spending estimates (monthly GDP estimates from Macroeconomic Advisors)

In 1990, the share stood at about 12 percent, but the 1990-1991 recession and its aftermath pushed it upward to about 14 percent. The share stayed at just under 14 percent for eight years until the recession of 2001, where it increased another 2 percentage points to 16 percent. After six years of stability, the recession of 2008 resulted in a jump of about 1.7 percentage points to 17.7 percent, where it has remained for the past 17 months.

There is no doubt that recessions are the reason for the three short-term episodes of upward growth in the health spending share of GDP. Because of widespread insurance coverage, health spending is largely resistant to economic downturns while, by definition, GDP growth declines. This results in a recessionary pattern in which GDP growth drops below health spending growth and the health spending share of GDP increases.

However, this does not mean that recessions are the cause of the long-term upward trend in the health spending share of GDP. Recessions deflect GDP from its long-run potential growth path, but expansions accelerate it back on roughly the same path. This is similar to a passenger plane that is delayed at the gate but then flies faster than usual in order to reach its destination on time. Because of this ability to catch up, long-run GDP growth rates are not overly impacted by occasional recessions. Long-run health spending growth rates are also largely recession resistant. It follows that recessions will have little impact on the long-term growth in the health spending share of GDP. In the absence of recessions, GDP would grow smoothly along its long-term path and the health share of GDP would move steadily upward, without the stair steps, toward the same destination.

With this background, let us define “bending the cost curve” as reducing the long-term growth of the health spending share of GDP to a sustainable rate of change. I’m not ready to take a stand on what this sustainable rate might be. But, for purposes of discussion, let us suppose that the target is no further growth in the share. How will we know when we have begun to meet this target?

The answer might at first seem obvious – just track the share over time and if it shows a few years of stability, declare the curve bent. However, suppose we had applied this criterion in the year 2000. We would have seen eight years of a zero growth in the health spending share and declared success. In 2008, we would have reached a similar conclusion based upon six years of stability. And today, based upon 19 months of stability from June 2009 through January 2011, we might also be tempted toward optimism.

On the other hand, if we had applied this criterion in 1992 or 2002 or 2009, we would reach the opposite conclusion. The problem is that, because of recessions, current trends can be wildly different from long term trends and it is the long-term trend we wish to detect. Fortunately, there are patterns that suggest some rough rules of thumb for identifying when short-term performance is consistent with desired long-term trends.

The chart that follows reproduces the curve shown previously, along with a second curve (the dotted red line) representing how the health spending share of GDP might have looked with successful bending. This hypothetical “bent” curve was constructed by constraining the cumulative health spending growth rate to be equal to that of GDP growth over each complete recession cycle. The first cycle covers January 1990 through June 2000, while the second covers June 2000 through January 2008. The health spending share of GDP in this bent curve is essentially constant at 12 percent except for the temporary disruption caused by the recessions and, according to the definition presented above, is therefore bent.

This curve illustrates the basic shape of a successfully bent curve. It rises in and around recessions and then declines during expansions. Under our example goal of holding the share constant over the long term, the decline during expansions must be just enough to offset the increase that occurs during recessions.

Looking forward, the historically low rate of growth in health spending over the past 28 months is certainly good news. But as the expansion takes hold, we’ll be looking for declines like in the hypothetical red dotted line during past expansions. In the expansion between 2003 and 2008, the dotted line falls by about 0.2 percentage points each year. Over the five years, it results in a 1 percentage point drop in the health spending share of GDP which just offsets the increase associated with the previous recession. Results for the previous expansion (1993 to 2000) are roughly the same. So if we were to observe a steady 0.2 percent annual decline in the health spending share over the next few years, we might be on track for long-term stability.

Of course, the real challenge is not tracking the cost curve, but identifying and implementing the structural changes needed to bend the curve, without sacrificing quality and access. I hope to have something to say about such strategies in future blogs.


Columnist Charles Roehrig is a health economist and director of the Altarum Center for Sustainable Health Spending. His column for the Health Policy Forum considers health economics issues and health spending. Columnists are regular contributors to the Health Policy Forum who pose their own opinions and policy positions in the realm of health care and health policy. As a leading nonprofit health care research and consulting institute dedicated to improving human health, Altarum encourages open discussion and debate about the many challenges in health care today. All postings to the Health Policy Forum (whether from employees or those outside the Institute) represent the views of the individual authors and/or organizations and do not necessarily represent the position, interests, strategy, or opinions of Altarum Institute. Altarum is a nonprofit, nonpartisan organization. No posting should be considered an endorsement by Altarum of individual candidates, political parties, opinions, or policy positions.


You make some excellent points and raise good questions. The first graph is also extremely important, though I think it reveals that we can’t offer a monocausal explanation of growth in health care expenditures as a share of GDP. Citing recessions (which somehow slow the rest of the economy but not health care) works for the latest episode from 2008-2010, but it can’t account for more than a fraction of the jump around the two previous recessions.

For example, health care costs started increasing as a share of GDP well before the 2001 recession, for reasons that I think are pretty well understood (on the commercial side, intense competition among HMOs collapsed and managed care efforts were reduced) and they kept increasing after the recession for reasons that are also well understood (such as increased hospital market power, and legislation that boosted spending for Medicare).

Jonathan – thanks for the insightful comment. Since this blog I have done some additional work that illustrates your point that the growth in the health spending share of GDP around the earlier recessions is only partially explained by reduced GDP growth. Tracking the spending share of “potential” GDP (PGDP) — what GDP would be at full employment — helps isolate the impact of health spending growth from business cycles on the share of GDP. The exhibit in this blog shows that the health spending share of PGDP accounts for a large share of the increase in the health spending share of GDP around the two earlier recessions but not the most recent one – just as you note in your comment. You might also be interested in a more recent blog that isolates the health spending effects from 1965 to present.

Add new comment

Click to Print