That’s the question asked in a paper by Sheila Smith, Joe Newhouse and Mark Freeland that just came out in the journal Health Affairs. The paper reports that real per capita health care expenditures increased by a factor of nine between 1960 and 2007, implying an annual growth rate of 4.8%. (This is slightly higher than I reported in a past post, but if Joe Newhouse says it, then I believe it.) The paper looks at data from a variety of countries over a number of years to try to tease out what, exactly, is driving this growth. [Note, this is less a paper about why the US is different than other countries than a paper about drivers in all countries.] They posit a number of possible drivers.
Technology: Technology improves over time, and new technology is expensive. Inasmuch as health economists agree on anything, it is that new technology is a major driver of cost growth.
Rising Medical Prices: If the prices of medical care increase faster than the economy as a whole, this could also be a driver of cost growth.
Insurance Coverage: Insured people tend to consume more health care than the uninsured. So, growth in the proportion of people covered by health insurance could increase health expenditure. In addition, increases in the generosity of health benefits might also lead to spending growth. The paper estimates that the share of health expenditures paid out-of-pocket by consumers dropped from 55% in 1960 to 14% in 2007. People tend to consume more when they pay 14 cents on the dollar than when they pay 55 cents on the dollar.
Rising Incomes: If health is a normal good, and if health care contributes to health (or if richer people just like more health care), then rising real incomes could also contribute to higher real health expenditures.
Demographic Factors: The population is getting older, and older people use more health care.
So, how much does each factor contribute to the increase in health care costs? Technology is, indeed, a major driver of growth, accounting for between 29 and 43 percent. Rising incomes are similarly important (29 to 48 percent). Price increases contributed about 5 – 19 percent of the growth, and increases in insurance coverage was about 10%. Demographics didn’t seem very important in the historical data, although the paper notes that this may change as the baby boom generation ages.
What does all this mean? Well, I don’t think anybody is going to argue that we should reduce incomes as a method of controlling medical costs. So, that’s out. Insurance coverage and increasing medical prices account for relatively modest amounts of growth. This leaves growth due to technological innovation as the most likely target of efforts to slow down the pace of cost increase. However, it is unclear that this kind of growth is a bad thing (raise your hand if you want to return to 1960s medical technology!), and it is difficult to see how we could take steps to rein in technological growth without evoking cries of “Rationing, Socialism, DEATH PANELS!”
Maybe we should just focus on speeding up the growth of the rest of the economy relative to health care rather than the other way around. That might be easier.