Friday, May 10, 2013

Editor’s note: Eugene Steuerle is Institute Fellow and Richard B. Fisher Chair at the Urban Institute in Washington, DC. He serves on the Altarum Center for Sustainable Health Spending National Advisory Committee. Please join us on July 30, 2013 at the Pew Charitable Trusts in Washington, DC for our “Symposium on Sustainable U.S. Health Spending: Public Sector Imperative, Private Sector Urgency.”  Dr. Steuerle will be one of the presenters.

In the past few years, health costs have begun to grow more slowly. A debate has arisen as to whether these effects are more permanent in nature.

The simple answer is that the jury is out. Both health cost growth and the economy contain cycles that do not always run in parallel, but even a  few years do not a cycle make. My own preferred way of looking at these numbers is to see what share of growth in per capita income is absorbed by additional health costs. By this criterion, the share has been building over time and, in no small part due to the Great Recession, absorbed almost all of our overall modest growth in per capita income in the first decade of this century. Some cyclical fallback from that extraordinary pressure is only natural.

A more complex answer entails understanding the economic forces at play. Remember Superman comic books, when the question regarding an irresistible force meeting an immovable object was raised? The current design of insurance, including the large portion covered by both government and employers, encourages us and our doctors to bargain over what everyone else will pay. At a zero price at the margin for services that have value, individual demand and business incentives for supply are each potentially infinite. That’s the original sin that must be absolved before any long-term solution to health care costs can be found, a finding highly, though not perfectly, correlated with the oft-stated notion that fee-for-service medicine is not sustainable. This original sin creates an irresistible force for higher costs as long as it remains an important component of most health insurance design.

But that irresistible force often runs into countervailing, if not yet immovable, objects. In private industry, health cost growth now builds upon such a large base that it threatens cash wages, sometimes even to the point of decrease.  This pressure need not be universal among all firms to have an impact. The private sector fights off these costs in no small part by insuring fewer people. State governments also fight against health costs through such methods as discouraging the building of nursing homes whose clientele often receive substantial state Medicaid dollars. As health care absorbs ever larger portions of the economy, many of the structural cost-reducing changes highlighted in recent Health Affairs articles by Ryu et. al., Cutler and Sahni, and KFF (e.g., more cost sharing in employer-provided insurance plans) are the types of responses one would expect.  Most, if not all, of these, to be clear, are not primary causes for health care slowdown; rather, they are causes only in the sense of being current mechanisms through which these economic and budgetary pressures play out.

The federal government often takes up the slack created by private and state agents, both directly through legislation like the Affordable Care Act, and indirectly through expansion of numbers of eligible participants under existing programs. When it does so, costs again go up, and these have been a major source of past growth in health spending.  

There’s also an inside-the-beltway issue that affects how to interpret both future and past health cost growth. Most federal agencies are required to project as if federal laws are never changed, even when those policies are unsustainable. Yet, projections are based upon past experience, which involves ever-changing policy. In technical terms, projections treat future federal policy as independent of economic and budgetary pressures, when in fact the data underlying those projections include those very types of policy changes.     

To make matters yet more complicated, there is a considerable amount of flexibility in the implementation of existing policy. My sense is that both regulators and the Congressional overseers are more liberal in the prices and services they accept when times are good. Given current massive deficits, one might project that Congress would be more willing to accept spending-reducing proposals of advisory and regulatory groups like MedPAC and the future Independent Payment Advisory Board.

Here are the raw data on annual percentage change in national health expenditures per capita.   

Graph of annual percentage change in national health expenditures per capita

Looks a bit like a downward trend, right?  But economic growth has also been weaker over time, and much of the long-term increase in health costs as a percentage of gross domestic product (GDP) derives from its continued growth even when incomes fall throughout the economy.

Here, then is a calculation of health cost growth over and above GDP growth. This is a traditional measure often cited for health cost calculations, largely because sustainability is typically assessed by noting that health cost growth in excess of GDP growth cannot go on forever.

Graph of percentage change in national health expenditures in excess of GDP growth from 1965 to 2012

Note that the extraordinary growth in recessions has typically been followed by a period in which it dips down below zero. Since health costs continue upward even when incomes go down, the upturns should not be surprising. But the cyclical downturns have also occurred in every case. We might relate this to higher-than-normal economic growth, but it might also be due to a rebound pressure from the non-health part of the economy, e.g., wage earners trying to catch up on their cash income lost in the previous period. Growth relative to potential GDP, as shown by Charles Roehrig, is one way to deal with this effect, though it leaves open some questions on the extent to which economic changes have a delayed or lagged effect on health costs as discussed in his recent blog. The most recent period, of course, knows no postwar parallel either in depth of the recession or slowness of recovery.

There is, I believe, a better way to conduct empirical research on this question than to look at growth rates or growth rates relative to GDP.  For many years now I have suggested that a superior variable on which to focus is the share of per capita income growth that gets absorbed by growth in health care costs. This measure, I believe, better accounts for the reaction within the non-health part of the economy, and, in theory, it could approach some constancy over time. One has greater difficulty imagining health costs growing at some constant rate over and above the economic growth rate over any extended period of time because it must simultaneously mean than all non-health costs will grow at a non-constant, ever-declining, eventually negative rate.   In effect, it tends to ignore the impact on the non-health part of the economy both when health’s share rises, and when economic growth rates decline over an extended period of time. Further complicating matters, economic growth includes growth in real health care services, so the two are not independent.   

Below I present the historical numbers for each decade up until 2010, as well as the projection of the health care (CMS) actuaries for 2010 to 2020, converted to shares of per capita income growth. The 2000 to 2010 period ends, of course, in the midst of a slow upturn, so the health care juggernaut essentially absorbed almost all per capita income growth of that period (In 2010 dollars, growth in per capita GDP was $2,613 while growth in per capita health costs was $2,275).

Graph of health spending as a share of total and per capita GDP growth

Suppose we are on a path where there is a longer-term slowdown in economic growth. Then it is possible simultaneously for both health costs to absorb a very high portion of income growth, yet at the same time fall historically relative to its absolute excess over GDP growth. Note that any health cost growth at all can absorb all of economic growth if the latter approaches zero.

Having said all that, how to interpret the current period on the economic side is wide open. Will it continue the slower growth of the very recent past, or might it eventually entail some above-average growth as the population ages and we more fully recover from the worst downturn since the Great Depression?  Looking at even longer 20-year periods, the irresistible force seemed to be growing more, not less, powerful in the 1990 to 2010 period compared to 1970 to 1990. But much of that played out through growth in government programs, whereas today Medicare, Medicaid, and (now) exchange rate subsidies are recognized almost universally as unsustainable, even if scheduled.

I’m hesitant to draw any very strong conclusion about what might happen while waiting for the significant government reform that is almost inevitable. Yet I find it hard to believe that health care still isn’t on track over the next 20 years to absorb at least a third of per capita economic growth, as it did in the 1990 to 2010 period. The share would likely be larger if the economy remains on a low growth path, though, as noted, such a low growth scenario could simultaneously entail a lower rate of growth in the variable on which I would place less focus—health cost growth over and above the rate of growth of GDP.

Graph of health spending as a share of total and per capita GDP growth for 1970 to 1990 and 1990 to 2010

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.



Subject Matter Experts