|
TRENDSHealth Care Spending Growth: How Different Is The United States From The Rest Of The OECD?
This paper compares the long-term (1970–2002) rates of real growth in health spending per capita in the United States and a group of high-income countries in the Organization for Economic Cooperation and Development (OECD). Real health spending growth is decomposed into population aging, overall economic growth, and excess growth. Although rates of aging and overall economic growth were similar, annual excess growth was much higher in the United States (2.0 percent) versus the OECD countries studied (1.1 percent). That difference, which is of an economically important magnitude, suggests that country-specific institutional factors might contribute to long-term health spending trends.
HEALTH ECONOMISTS in the United States frequently assert that although health care spending per capita might be higher in the United States than in other high-income countries, the long-term rates of spending growth have been similar.1 In fact, the United States has had an unusually high rate of long-term health spending growth, as this paper demonstrates. That fact suggests that institutional features that differ between the United States and other high-income countries might have affected long-term rates of health spending growth.
This paper compares long-term U.S. health spending trends with trends in countries in the Organization for Economic Cooperation and Development (OECD, hereafter assumed to exclude U.S. data) during 1970–2002 (the most recent year for which OECD data are available). The OECD figures consist of all members of the OECD for which the necessary data were available for 1970–2002 (Exhibit 1
Other researchers have compared the rates of health spending growth in the United States with those of other OECD countries.2 This paper incorporates several methodological advances on this prior research. First, rates of growth in health spending per capita are decomposed into population aging, economic growth, and excess growth. The rate of excess growth is used by the Medicare Trustees and other U.S. analysts as a key parameter in projections of long-term growth in health spending; it provides a useful yardstick for international comparisons.3 Second, prior research has compared the rate of growth in the United States with the OECD median.4 The median growth rate gives equal weight to small countries, such as Luxembourg and Iceland, and large countries, such as Japan and Germany. This paper instead compares the United States with an aggregate of the OECD countries, which is generated by summing the populations, GDPs, and health spending. This OECD aggregate weights more-populous countries more heavily and therefore provides a more meaningful comparison than the median. Third, this paper uses more recent data (through 2002) than prior analyses have used, and it illustrates trends in recent years.
Methods. For each country and year, I measured real (inflation-adjusted) health spending, GDP, and population by age group. For countries other than the United States, I first converted GDP and health spending to U.S. dollars using purchasing power parity (PPP) exchange rates.5 All GDP and health spending figures were then inflated to 2004 dollars using the U.S. GDP deflator. Data source. The key data source is the 2004 OECD Health Data. All OECD countries were included in the analysis if data on health spending and demographics were available for the period 1970–2002. The countries that were excluded because of missing data are Czech Republic, Hungary, Italy, Korea, Mexico, New Zealand, Poland, Slovak Republic, and Turkey. Although it is unfortunate that the data for those countries were not available, those countries are mostly lower-income countries (compared with other OECD countries), so the comparability of the United States and OECD is strengthened by their exclusion. Analyses. I measured the rate of growth in health spending per capita using a logarithmic growth rate concept. For example, the annual rate of growth over the period from year X to year Y equals the natural logarithm of spending per capita in Y divided by spending per capita in X, divided by the number of intervening years (Y minus X). Logarithmic growth rates were used rather than simple annual growth rates because they have desirable additive properties.6 Population aging. The rate of population aging refers to the rate at which health spending per capita would grow if health spending per capita within each age group remained constant but the share of the population in each age group changed. Population aging reflects the combination of the age-spending profile (older people, on average, have higher health spending per capita than younger people) and the fact that older people are making up larger and larger shares of the populations in OECD countries. Countries such as Japan have high rates of population aging because the elderly are growing very rapidly as a share of the population. The first step in calculating the rate of population aging was to calculate a measure of health spending per capita within each of the following age groups: ages 0–14, 15–49, 50–64, 65–69, 70–74, 75–79, and 80 and older (groupings chosen because of data availability). Spending per capita within each age group was calculated using the Medical Expenditure Panel Survey (MEPS) from 1996 through 2000. I used the MEPS spending measures, which are based on the U.S. population, because data are not available on health spending by age group for each OECD country. Out of necessity, I assumed that the patterns of relative health spending by age group in the United States reasonably approximate those for all OECD countries. This assumption does not appear to be unreasonable, based on published data on relative spending among people over versus younger than age sixty-five among a subset of OECD countries.7 The next step was to calculate a hypothetical level of health spending per capita for each country and year by combining the actual demographics with the age-specific spending per capita from MEPS. Changes over time in those hypothetical demographics-based health spending measures were then converted in rates of population aging using the logarithmic growth method described above. Excess growth. The rate of excess growth equals the rate of growth in health spending per capita minus the rate of real growth in GDP per capita and the rate of population aging. If excess growth were 0 percent, then health spending per capita within each age group would grow at the same rate as GDP per capita.
In 1970, the five countries with the highest health spending per capita were Canada, Sweden, the United States, Switzerland, and Denmark (Exhibit 1  ). By 2002, health spending per capita in the United States had grown to $5,475, far higher than in the next-highest country, Switzerland ($3,581). By that simple measure, growth in U.S. health spending per capita appears to have been unusually high.
During 1970–2002, the annual rate of growth in health spending per capita was higher in the United States than in the OECD (Exhibit 2
Has excess growth in the United States always been high, relative to the OECD? To examine this question, I measured the rate of excess growth for each country for two periods: 1970–1985 and 1985–2002. (The year 1985 was chosen because it is near the middle of the time series and because it represents a point of divergence, as will be shown.) During the earlier period, the rate of excess growth in the United States was slightly higher than in the OECD (2.0 percent versus 1.7 percent), but the United States was near the middle of the pack (Exhibit 3  ). In contrast, during the later period, the rate of excess growth in the United States was the highest among all OECD countries and was much higher than the overall rate for the OECD (2.0 percent versus 0.6 percent, Exhibit 4). What stands out in Exhibits 3 and 4 is that in the United States, excess growth was essentially unchanged between the two periods, whereas excess growth in other OECD countries dropped substantially.
The final question is whether the observed differences in excess growth rates between the United States and the OECD are of an economically important magnitude. To address this question, Exhibit 5 illustrates the cumulative effect of the difference between the rates of excess growth in the United States and the OECD. Two versions of the health share are shown for the United States: the actual health share and the hypothetical health share that would have occurred if the rate of excess U.S. growth in each year matched the rate of excess growth in the OECD. By construction, the actual and hypothetical health shares in Exhibit 5 begin at the same point in 1970. During the 1970s, excess growth in the OECD was actually higher than in the United States (note that the hypothetical health share is greater than the actual health share). But beginning in the 1980s, the rates of excess growth in the United States and OECD diverged sharply. By 2002, the cumulative difference in excess growth rates accounted for a 3.6-percentage-point difference between the actual and hypothetical U.S. health shares (14.6 percent versus 11.0 percent). In 2002, that difference—3.6 percent of U.S. GDP—totaled $436 billion. To put that figure in perspective, it is slightly larger than combined federal spending for Medicare and Medicaid in that year.
In the United States, real health spending per capita has grown rapidly over the past three and a half decades, at an annual rate of 4.3 percent. Four factors might explain this growth: (1) population aging, (2) general economic growth, (3) expansion in the technological capabilities of medicine, and (4) other factors, including expansions in insurance coverage and changes in health care financing and delivery systems. The analyses presented in this paper can help quantify the roles of each of these factors.
The first factor, population aging, explains only a small part of the growth in real U.S. health spending per capita (0.3 percentage points of the 4.3 percent annual growth; see Exhibit 2 Technological change. The third factor, technological change, is commonly cited as the primary driver of growth in health spending as a share of GDP.11 The technological-change argument is that as a result of expansions in the capabilities of the medical sector, new types of treatment are provided and broader segments of the population receive treatment; those changes result in a greater share of the economy’s resources being devoted to the medical sector. David Cutler argues that the growth in health spending as a share of GDP, because it is primarily driven by technological change, should be viewed not as a problem to be fixed but rather as a positive outgrowth of welfare-improving advances.12 Medical technology and research findings flow fairly freely among high-income countries, so it makes sense to think of the expansion in medical capabilities as a phenomenon shared among the United States and other high-income OECD countries. To the extent that we find similarities between the United States and OECD in rates of excess growth, that would tend to support the technology argument. Over the long term, there has been a positive rate of excess growth in health spending in both the United States and OECD, and rates of excess growth were similar over the period 1970–1985. But over the period 1985–2002, the rates of excess growth were quite different between the United States (2.0 percent) and the OECD (0.6 percent). That large disparity strongly suggests that institutional factors have played an important role in determining the rate of health spending growth. The question is what those factors are and, more specifically, how the United States differs from the OECD. Later role for insurance. One way in which the United States stands out is as a relative latecomer in the expansion of insurance as a source of health care financing. In 1970, among those OECD countries for which data are available, government and private insurers already covered a large share of health spending (84 percent), and that share has remained fairly constant since then. In contrast, the share of health spending covered by insurance in the United States was relatively low in 1970 (65 percent) but rose fairly steadily thereafter, reaching 77 percent in 1985 and 86 percent in 2002. That gradual increase likely spurred growth in health spending somewhat. The timing of the increase in the U.S. insured share does not appear, however, to explain the divergence in excess growth rates in recent years. Differences in health care financing. The United States also differs from other countries of the OECD in how health insurance is financed. Elizabeth Docteur and Howard Oxley have identified three broad approaches to financing health care: (1) public-integrated—government acts as both insurer and provider of services; (2) public-contract—government or a centralized social insurer purchases services from private providers; and (3) private insurance/provider—private insurers purchase services from private providers.13 Nearly all high-income OECD countries other than the United States use either the public-integrated model or the public-contract model. The United States stands out in its reliance on the private insurance/provider model. Private insurers financed 41 percent of health care spending in the United States in 2002, compared with a weighted average of 12 percent among the OECD. As Docteur and Oxley point out, both the public-integrated and public-contract models give a central authority (either the national government or the social insurance administrators) a great deal of leverage over medical care providers, and that leverage can be used to constrain health spending. One clear example of the use of that leverage is the setting of hospital budgets in countries that use the public-integrated model, such as the United Kingdom. Henry Aaron and William Schwartz, in their classic work from the mid-1980s, observed that the financial constraints imposed on hospitals by the British National Health Service led to the provision of a much different level and mix of services compared with the United States.14 Cost containment in the OECD. In the 1980s and 1990s, the policy-making focus in the OECD shifted from ensuring access toward constraining growth in health care spending. Those countries using public-integrated and public-contract financing models might have had more success in constraining spending than the United States has had. The United States, outside of Medicare, does not use a centralized authority to set health spending budgets or negotiate prices with providers, and this could contribute to a relative lack of spending constraint. Perhaps because of their relative success in constraining costs, the health care reforms currently under way in the OECD are no longer generally focused on cost constraint and are instead focused on quality of care, output and efficiency in the production of health care services, and responsiveness to patients’ needs. It remains to be seen whether the U.S.-OECD divergence in excess spending growth will continue. If OECD countries conclude that they have "overdone" their cost containment in the past, their rates of excess growth could rise and resemble that of the United States. On the other hand, the structural differences between the health financing arrangements in the United States and the other OECD countries might contribute to a continued divergence in spending trends.
Chapin White (chapin_white{at}post.harvard.edu) is an analyst at the Congressional Budget Office in Washington, D.C. The views expressed in this paper are those of the author and should not be interpreted as those of the Congressional Budget Office.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Study Data And Methods
Study Results
