Luxembourg spends 5.6% of its economy on healthcare, the smallest slice in the wealthy world. Looks thrifty. Per person, it spends more than Japan, Korea, or Italy.
Both numbers are real. That's the whole trick.
"Share of GDP" is a slice of the entire economy, not money in anyone's pocket. A big enough economy can spend a small slice and still spend a fortune per head. So the measure you pick quietly writes the story before you have read a thing.
Watch it happen. Here is every OECD country in 2023, life expectancy across the bottom, health spending as a share of GDP up the side.
A shapeless cloud. Countries that live past 83 spend anywhere from 5.6% to 11%. If a bigger share bought years, this would slope up. It doesn't. And up in the corner, alone, sits the United States: 16.9% of GDP, nearly double the average, with a life expectancy of 78.4 years. Twenty-nine of the other 37 members outlive it.
Now switch the second axis to dollars per person. The picture reorganizes itself.
Luxembourg leaps from the thriftiest-looking country to one of the heavier spenders. And the US doesn't just stay high, it floats off on its own. About $13,900 a head, more than half again as much as Switzerland, the next biggest spender anywhere. Still at the short end of the lifespan range.
Two charts, two questions. One asks how much of the economy goes to health. The other asks what each person's care costs. Most countries trade places between them. The US is extreme on both, which is the strange part.
The countries that actually live longest, Japan, Korea, Italy, Spain, spend between $4,800 and $6,000 a person. A third to a half of the American bill. They live about five years longer.
So what is the money buying? Here the data goes quiet. Diet, obesity, guns, cars, opioids, whether you can see a doctor without a bill you can't pay, how much a hospital charges for a procedure that costs a third as much one border over. Economists have argued about the mix for twenty years and still disagree.
The number was never the answer. It is the better question: if the money isn't buying the years, where are the years going?
Data: OECD, 2023, via the SDMX API. Life expectancy at birth (total population) from DSD_HEALTH_STAT@DF_LE. Current health expenditure from DSD_SHA@DF_SHA, shown both as a percentage of GDP and as US dollars per person at purchasing power parity. The "OECD average" is the unweighted mean across the 38 members. This is an association between country-level indicators for a single year, not evidence that spending causes the difference in lifespans. Pulled with oecd-mcp.