The Two-Speed Economy: Services vs. Goods

Since 1980, the CPI for services has risen 481%. The CPI for commodities has risen 177%. That 304-percentage-point gap is the defining economic reality of the past half-century. It explains why Americans feel squeezed even when inflation reports look tame. It explains why a flat-screen television costs a fraction of what it did twenty years ago while a hospital visit costs multiples more. It explains, in a single pair of lines on a chart, the central tension of the modern American economy.

Physical goods — manufactured in factories and shipped in containers — have benefited from decades of automation, globalization, and relentless supply-chain optimization. A television factory in 2026 produces a hundred times what a 1980 factory did with fewer workers. A container ship carries ten thousand units where a freighter once carried hundreds. The result is deflation so dramatic that some goods now cost less in nominal dollars than they did forty years ago.

Services are a different story. Medical care, education, housing, insurance, childcare, legal services — these are labor-intensive activities where productivity gains are structurally limited. A doctor’s visit still requires a doctor. A semester of college still requires professors. A house still requires carpenters. Wages in these sectors must keep pace with the broader economy, but output per worker cannot rise the way it does on an assembly line. The economist William Baumol identified this dynamic in 1966 and called it the “cost disease.” Sixty years later, the BLS data proves him right.

The chart below tells the story in three lines. The blue line is CPI for all items — the number that makes the evening news. The red line is CPI for services. The teal line is CPI for commodities. All three are indexed to 1980 = 100.

In the early 1980s, the three lines tracked closely. Services and commodities both rose through the Volcker recession and the recovery that followed. But starting around 1985, the lines began to separate. Services pulled ahead. Commodities lagged. And the gap widened every single year for the next four decades.

By 2026, services have risen to an index value of 582 — nearly six times the 1980 level. Commodities sit at 277 — less than three times. All items land at 419, pulled upward by services and dragged down by goods. The services line did not merely outpace commodities. It lapped them. The annualized rate for services was 3.8% per year; for commodities, 2.2%. That 1.6-percentage-point annual gap sounds modest. Compounded over forty-six years, it produced a chasm.

The two-speed pattern becomes even more vivid when we plot individual categories on a single chart. Every line below starts at 100 in its base year and traces the cumulative price path through January 2026. The result is a fan shape — narrow at the left edge, wide at the right — that reveals the full spectrum of American inflation.

At the top of the fan: college tuition, which has risen 1,338% since 1980. A dollar of tuition in 1980 buys about seven cents of tuition today. Tobacco sits nearly as high at 1,326%. Motor vehicle insurance has risen 687%. Medical care, 726%. Shelter, 456%.

At the bottom of the fan: televisions, which have fallen 99% since 1996. A TV that cost $500 in 1996 costs the equivalent of $7.40 today. Toys have fallen 66%. Computers have fallen 78%. Even apparel — a physical good that still requires human labor — has risen only 51% in 46 years, less than 1% per year.

The fan chart is Baumol’s Cost Disease made visual. Every category above the All Items line is dominated by human labor. Every category below it is dominated by manufacturing and technology. The pattern admits almost no exceptions.

In 1966, the economist William Baumol published a paper with William Bowen titled “Performing Arts: The Economic Dilemma.” Their observation was deceptively simple: a string quartet in 1966 required the same number of musicians and the same amount of time as a string quartet in 1866. Productivity in live performance had not improved at all. Yet the musicians had to be paid wages competitive with the broader economy, where productivity had improved. The result was that the cost of a live performance, relative to manufactured goods, had to rise over time.

Baumol generalized this into what became known as the “cost disease”: in sectors where automation cannot easily replace human labor, productivity grows slowly, but wages must keep pace with the rest of the economy — or workers leave. The consequence is that labor-intensive services become progressively more expensive relative to goods.

The BLS data from 1980 to 2026 is a forty-six-year empirical validation of Baumol’s theory. Consider the extremes:

A teacher in 2026 is not five times more productive than a teacher in 1980. A classroom still holds thirty students. A lecture still takes fifty minutes. But that teacher must be paid a salary competitive with an economy where manufacturing output per worker has quintupled. The result: tuition rises 1,338%.

A television factory in 2026 produces screens that are larger, sharper, and more energy-efficient than anything imaginable in 1980 — and it does so with a fraction of the workforce. Robots handle assembly. Software handles quality control. Global supply chains deliver components from wherever they are cheapest. The result: TV prices fall 99%.

Between these extremes lies the entire American economy, sorted by its exposure to Baumol’s dynamic. Medical care: labor-intensive, rising 726%. Shelter: land-constrained and labor-intensive to build, rising 456%. New vehicles: increasingly automated to produce, rising only 110%. Apparel: offshored to low-wage countries, rising only 51%.

The chart and table below rank every major CPI category covered in this series by total percentage change from 1980 (or earliest available year) to January 2026. The range is staggering: from college tuition at +1,338% to televisions at −99%. That is a spread of over 2,400 percentage points between the fastest-rising and fastest-falling categories in the American economy.

The color coding tells the story at a glance. Categories in red have outpaced overall inflation by a wide margin — these are the services that squeeze household budgets. Categories in green have deflated in absolute terms — these are the manufactured goods that global trade and technology have made cheaper. The dividing line is clear: if it can be manufactured at scale and shipped in a container, it got cheaper. If it requires a human to show up in person, it got more expensive.