Health Care Expenditures in the OECD

Health care expenditures have been growing much more rapidly than income in the U.S. and other developed countries for some time. In the U.S., for example, health care expenditures as a share of GDP have tripled since 1950, from 5% then to 15% today.

Less well understood are the causes of this dramatic increase in particular, how much of the rise is due to an aging population and how much is due to rising expenditures per person at a given age. This distinction is important because benefit levels are determined by government policy, while demographics are largely outside government control. Being aware of the relative contributions of demographics and benefit growth can help policy makers project future expenditures more accurately and make better decisions about how to allocate scarce government resources.

In "Who's Going Broke? Comparing Healthcare Costs in Ten OECD Countries," (NBER Working Paper 11833), Laurence Kotlikoff and Christian Hagist explore this issue. Their analysis uses OECD demographic and health expenditure data along with age healthcare expenditure profiles for each country to measure growth in real healthcare benefit levels between 1970 and 2002 in ten countries: Australia, Austria, Canada, Germany, Japan, Norway, Spain, Sweden, the UK, and the U.S.

Figure 1 illustrates the age healthcare expenditure profile for one particular country, the U.S. The profile rises sharply with age the average expenditure per capita for those age 75 and above is eight to twelve times as large as for the reference group, those ages 50 to 64. In Austria, Germany, Spain, and Sweden, by contrast, the average expenditure per capita is only twice as large for the oldest old as for the reference group. The other countries lie somewhere in between, with relative spending factors for the oldest old ranging from four to eight.

The authors note that the profile for the US is so steep in part because the elderly have virtually universal coverage through Medicare, while many working age adults are uninsured. The steep profile indicates that the U.S. will likely experience very rapid growth in overall health expenditures in coming years, as its population continues to age.

Next, the authors use these profiles, population data, and the growth in total health expenditures to calculate the growth in per capita benefit levels between 1970 and 2002. They find that real per capita benefit levels grew at a 4.1% annual rate on average across the ten countries over the period, a full two percentage points higher than the average growth rate of real per capita GDP.

A simple example illustrates the importance of the growth in real per capita benefits. In 1970, Sweden had the highest level of per capita government health spending, while Norway spent one third less. Over the next thirty years, Norway's benefit level grew at 5% per year, more that double the growth rate experienced in Sweden, and by 2002, Norway's per capita expenditures were nearly 60% higher than Sweden's.

Norway, Spain, and the U.S. experienced the highest growth rates over the period, 4.6% to 5%. Canada and Sweden had the lowest, at 2.3%, while the other countries lay about halfway in between. The authors note that Canada and Sweden's low growth rates are unsurprising, given their use of rationing to limit health care spending. Conversely, the high rate of benefit growth in other countries likely results from costly product innovations, such as the acquisition of CT scanners.

Next, the authors calculate the share of total benefit growth that is attributable to demographics. They find that demographics are relatively unimportant; on average, three quarters of health care expenditure growth can be traced to growth in benefit levels.

Finally, the authors present the budgetary implications of permitting benefit levels to continue to grow at historic rates. Assuming a 3% discount rate, the U.S. is projected to double expenditures as a share of today's GDP in 20 years and triple them in 40 years. Increases in most other countries are smaller, but nonetheless quite significant. The authors note that "letting benefits grow at historic rates even on a relatively short-term basis is extremely expensive. It locks in high benefit levels for years and generations to come."

The authors conclude "growth since 1970 in aggregate healthcare spending by our ten OECD governments reflects first and foremost growth in benefit levels. Although OECD countries are projected to age dramatically, growth in benefit levels, if it continues apace, will remain the major determinant of overall healthcare spending growth." They note that continued very rapid growth in benefit levels is clearly unsustainable and that benefit growth must eventually slow to the growth rate of per capita income. However, the U.S. has not yet made any move in that direction and indeed has recently increased the rate of benefit growth with the addition of a Medicare prescription drug benefit.