The Economist surveys recent research and concludes, “An ageing society might not cost too much.”
Spending on health care depends on how much it costs to provide and how much treatment is needed. Generally, the former has increased over time. Economists attribute this to Baumol’s cost disease, diagnosed by William Baumol, an American economist. Technologically intensive sectors (think of manufacturing or software development) get more productive. Labour-intensive service sectors do not. A doctor’s consultation takes roughly as long as it did a century ago. Yet physicians’ pay must match increases in faster-growing skilled industries, or all the doctors will go to work in tech. (Baumol pointed out that a string quartet still took 45 minutes to perform a Schubert work, but was paid vastly more than when the composer was alive.)
[…]
Two new papers from the National Bureau of Economic Research, a repository of economic thought, suggest that the vision of health care gobbling up tax revenues may not come to pass. In the first, David Cutler and Lev Klarnet, both of Harvard University, observe that in 2024 America spent $1trn less on health care than official forecasts from 2010. The second is by Liran Einav, of Stanford University, and Amy Finkelstein, of the Massachusetts Institute of Technology (MIT). Using a survey of users of Medicare, America’s system of health insurance for the old, they find that, whereas expected lifetime spending on Social Security (ie, pensions) rose by 14% between 1993 and 2017, the expected cost of Medicare increased by only 6%. That is partly because elderly Americans have been collecting their pensions for longer but spending less time seriously ill.
The papers point to improvements on both sides of the equation. Start with the cost of providing health care in hospitals. Messrs Cutler and Klarnet point out that, between 2000 and 2010, it consistently rose by around 2.3 percentage points above inflation—but that later something changed. Between 2011 and 2024 hospital costs rose by only 0.5 percentage points more than overall prices. Technology has played a role. Technological improvements in health care come in two types: some are adopted because economies get richer and can afford to treat new things; others make it cheaper to provide the same outcomes. For much of the past century, the first type dominated the second. Now, it seems the second type has taken over. Baumol’s cost disease turned out to be curable.
On top of that, less health care is needed than was once expected. Health-care systems may have become more efficient at targeting treatments, and insurers better at saying no. The old are also getting healthier. Mr Einav and Ms Finkelstein point out that since 1993 Americans’ life expectancy at age 66 has risen by 2.4 years, which are (on average) entirely healthy. They can expect an extra three years of healthy life. The amount of time spent in the severest mental and physical distress has declined by 0.6 years. That means less spent on extremely pricey residential care or home help for the infirm.
Improvements are not limited to America. As this newspaper pointed out three years ago, health-care spending has slowed (as a share of GDP) across the OECD, a club of mostly rich countries. Another new paper, by Sheila Diane Smith and Joseph Newhouse in the American Journal of Health Economics, makes the same point—and predicts that spending per person will grow by around 2% over the next decade in the OECD (excluding America), compared with 4% before 2009. Spending on health care as a share of GDP will rise by only a tenth of a percentage point each year.
One hopes these projections are right. I remain skeptical that an ever-increasing share of our resources can go to elderly retirees (whose ranks I hope to join in the not-too-distant future) while the share of the population that’s young and productive continues to shrink. That’s doubly true if AI and other technologies deprive them of the ability to make a living.
