Baumol effect

Based on Wikipedia: Baumol effect

The Curious Case of the Expensive Haircut

Here's a puzzle that might keep an economist up at night: a barber in 1900 could cut roughly the same number of heads of hair per day as a barber in 2024. The scissors move at the same speed. The hair grows at the same rate. And yet, adjusted for inflation, that modern barber earns vastly more money than their predecessor from a century ago.

Why?

The answer reveals something profound about how modern economies work—and why your healthcare, education, and concert tickets keep getting more expensive even as televisions and computers get cheaper. It's called the Baumol effect, sometimes referred to as Baumol's cost disease, and understanding it changes how you see everything from government budgets to the future of work itself.

The Violinist Who Started It All

In the 1960s, two American economists named William Baumol and William Bowen were studying an unlikely subject: the performing arts. They made an observation that seems almost comically obvious once you hear it, but carries enormous implications.

A string quartet performing a Schubert piece in 1820 required four musicians playing for roughly forty minutes. A string quartet performing that same piece today? Still four musicians. Still forty minutes. The "output per man-hour," as economists like to say, hasn't budged in two centuries.

But here's the thing: those musicians today earn far more than their 1820 counterparts, even after adjusting for inflation. The Schubert hasn't gotten any faster. The violins haven't become more efficient. So where does the extra money come from?

The Competition for Workers

The answer lies in a simple economic force: competition for labor.

Imagine you're a talented, hardworking person in the modern economy. You could become a classical musician, or you could become a software engineer. Over the past few decades, software engineers have become dramatically more productive. The tools they use—faster computers, better programming languages, cloud infrastructure—mean a single developer today can accomplish what might have taken a team of fifty in 1980.

This productivity explosion means tech companies can afford to pay software engineers handsomely. And here's where the barber and the violinist come in: if orchestras and barbershops paid 1900-level wages, all their workers would quit and learn to code.

So wages rise across the entire economy, even in sectors where productivity hasn't improved at all. The barber gets a raise not because haircuts have gotten more efficient, but because the barber has alternatives.

This is the Baumol effect in a nutshell: wages in low-productivity-growth sectors rise to match wages in high-productivity-growth sectors, because workers can move between sectors. The consequence? The things that can't be automated or made more efficient keep getting relatively more expensive.

The French Connection

Before Baumol and Bowen published their famous analysis, a French economist named Jean Fourastié had noticed something similar. Writing in 1949, he observed that entire professions had barely changed their working methods in half a century:

The men's barber does not cut more clients' hair in 1948 than in 1900; entire professions have not changed their working methods from 1900 to 1930.

Fourastié predicted this would reshape entire economies. As manufacturing became more efficient, requiring fewer workers to produce more goods, people would shift into service jobs—teaching, healthcare, personal services, entertainment. He foresaw what we now call the post-industrial economy, or as he put it, a "tertiary civilization."

He was right. In the United States, manufacturing once employed a substantial portion of the workforce. Today, the vast majority work in services. This transition wasn't driven by declining demand for manufactured goods—we consume more physical products than ever. It was driven by productivity. Factories simply need fewer workers to produce more stuff.

The Great Divergence of Prices

If you want to see the Baumol effect in action, look at price trends over the past several decades. The pattern is striking and consistent.

Things that can be manufactured with ever-improving technology get cheaper: televisions, computers, clothing, furniture. A television that cost a month's salary in 1980 costs a few days' wages today, and it's incomparably better.

Things that require human time and can't be easily automated get more expensive: college tuition, healthcare, childcare, legal services, live entertainment. These haven't become dramatically more efficient, but the workers who provide them need to be paid competitive wages.

In 2008, the economist William Nordhaus analyzed American data from 1948 to 2001 and found an almost perfect correlation. Industries with low productivity growth saw their relative prices rise, almost percentage point for percentage point matching their productivity disadvantage. The hypothesis of a "cost-price disease" was, as Nordhaus put it, "strongly supported by the historical data."

Why This Isn't Actually a Disease

The term "cost disease" is somewhat misleading. It sounds like something has gone wrong, like an economy suffering from some kind of infection that needs to be cured. But Baumol himself pushed back against this framing.

Think about it this way: the reason service prices rise is because productivity is growing in other sectors. That productivity growth makes society richer overall. Yes, healthcare consumes a larger share of national income than it did fifty years ago. But total national income has grown so much that even after paying for more expensive healthcare, we have more money left over for everything else.

Baumol offered a simple thought experiment. Imagine an economy with $100 billion in national income, spending $20 billion on healthcare—leaving $80 billion for other things. Now fast-forward fifty years. Thanks to productivity growth, national income has doubled to $200 billion. Even if healthcare costs explode to $120 billion—a 500% increase—there's still $80 billion left for other purchases. Exactly the same as before.

And in reality, if healthcare only grew to, say, 40% of a doubled national income, you'd have $120 billion for other things. More than before. Society gets richer even as certain services take up a larger share of the pie.

The Baumol effect isn't a disease. It's the inevitable consequence of uneven technological progress, and it accompanies genuine improvements in living standards.

The Inequality Problem

There's a catch, though. While the math works out fine for society in aggregate, it doesn't work out equally for everyone.

The gains from productivity growth aren't distributed evenly. If your wages rise more slowly than the cost of services affected by the Baumol effect, those services become less affordable to you—even as the economy grows overall.

This is exactly what has happened over recent decades. Healthcare, education, and housing have become more expensive faster than many workers' wages have grown. A person whose income tracks average productivity growth can still afford these things. But for workers whose wages have stagnated, the Baumol effect means being increasingly priced out of essential services.

Baumol himself acknowledged this: the cost increases "disproportionately affect the poor." The problem isn't that society can't afford these services—it's that the benefits of productivity growth flow mainly to some people while the costs are borne broadly.

The Government Budget Squeeze

Perhaps nowhere is the Baumol effect more visible than in government finances. Think about what governments actually spend money on: police officers, teachers, nurses, soldiers, judges, social workers. Almost all of it is labor-intensive services.

These are precisely the sectors most affected by the Baumol effect. A police officer today doesn't patrol ten times more territory than one in 1960. A teacher doesn't educate ten times more students per hour. But both need to be paid wages competitive with the private sector, which has enjoyed those productivity gains.

The result? Government services get more expensive over time, even if the government is doing exactly the same things. Citizens feel like they're paying more in taxes for the same services—and in a sense, they are. The Baumol effect makes this almost inevitable.

This creates persistent political pressure to cut government services or find savings, but the underlying economic force doesn't go away. You can cut the number of teachers, but you can't make teaching dramatically more productive. The cost disease is baked into what government does.

The Great Labor Shift

One of the Baumol effect's most dramatic predictions has come true before our eyes: the wholesale reallocation of workers from high-productivity sectors to low-productivity sectors.

The logic is straightforward. If manufacturing productivity doubles, you need only half as many workers to produce the same output. Those workers go somewhere—and they go into services, where productivity growth is slower.

The numbers are remarkable. In 1930, about 21.5% of the American workforce was employed in agriculture, producing about 7.7% of the gross domestic product. By 2000, only 1.9% of workers were in agriculture, producing just 0.7% of output. Farming became so efficient that the vast majority of agricultural workers were freed up to do other things.

Manufacturing followed a similar pattern, though less extreme. The United States has transformed from a manufacturing economy into a service economy—not because we make fewer things, but because we need fewer people to make them.

The Growth Slowdown

There's a troubling implication lurking here. If an ever-larger share of workers moves into low-productivity-growth sectors, what happens to overall economic growth?

Baumol's original analysis suggested it should slow down. This follows logically: if you have more workers in slow-growing sectors and fewer in fast-growing sectors, the aggregate growth rate falls. Since long-term economic growth is driven largely by productivity growth, a shift toward stagnant sectors implies slower economic progress.

Some economists have pushed back on this pessimistic conclusion. The economist Nicholas Oulton argued in 2001 that things might not be so simple. Many services—accounting, consulting, software development, logistics—are actually inputs that help manufacturing become more productive. A growing service sector that supports manufacturing might boost productivity growth rather than drag on it.

The debate continues. But it's worth noting that productivity growth in developed economies has indeed slowed since the high-growth decades of the mid-twentieth century. How much of this reflects the Baumol effect versus other factors—exhaustion of easy innovations, measurement problems, policy changes—remains contested.

The Performing Arts Paradox

Remember those musicians who started this whole analysis? Their story has an interesting twist that challenges the original Baumol and Bowen conclusion.

In 1984, the economists Gambling and Andrews pointed out something that might seem obvious in retrospect: performance halls have gotten bigger. A violinist playing to a hall of 500 people in 1820 is, in a sense, less productive than one playing to 3,000 people in a modern concert hall. Output per performance has increased, even if output per hour of playing hasn't.

But that's just the beginning. In 1995, the economist Greenfield noted that recorded music, radio, and eventually streaming meant a single performance could reach millions of listeners. In productivity terms, this is an almost incomprehensible increase. The Vienna Philharmonic's recording of a Beethoven symphony might be heard by more people in a single year than every live performance of that symphony in the nineteenth century combined.

This complicates the simple Baumol story. Some services that seem stagnant have actually found ways to scale dramatically through technology. The question becomes: which services can be amplified this way, and which really are stuck at inherently human scale?

What Can and Can't Be Scaled

Consider the difference between a recorded concert and a haircut. You can stream a concert to a million people at almost zero marginal cost. You cannot stream a haircut.

This distinction matters enormously. Some services that seemed subject to the Baumol effect have found ways to scale: entertainment through recording and broadcasting, education through online courses, certain types of medical diagnosis through telemedicine and artificial intelligence.

But other services are inherently local and personal. Cutting hair. Caring for the elderly. Performing surgery. Teaching a small child to read. These require physical presence and human attention. No amount of technological progress eliminates the need for a human being to be there, spending time.

The future of the Baumol effect may depend on how many services can be transformed from the second category to the first. If artificial intelligence can diagnose medical conditions accurately, fewer doctors are needed per patient. If robots can assist with elderly care, fewer human caregivers are needed per person. Each such transformation would reduce the Baumol effect in that sector.

But for services that remain stubbornly human-scale, the effect will persist. As long as people want experiences that require another human's time and attention, those experiences will keep getting relatively more expensive.

The Opposite of the Baumol Effect

It's worth understanding what the opposite of the Baumol effect looks like. In sectors with high productivity growth—manufacturing, agriculture, some forms of information technology—we see the reverse pattern.

Prices fall relative to wages. Output grows faster than employment. Fewer workers produce more goods. These sectors shrink as a share of employment while growing or maintaining their share of output.

This is the engine of modern prosperity. When it takes fewer people to grow food and make things, those people can do other work. Society gets richer because we can have both the manufactured goods and the services.

The Baumol effect is the shadow side of this progress—not a failure, but an inevitable consequence. You can't have productivity growth in some sectors without relative price increases in others. The alternative would be wage stagnation across the entire economy, which would mean no improvement in living standards at all.

Why This Matters for Everything

Understanding the Baumol effect changes how you interpret many policy debates.

When people complain that healthcare costs are out of control, the Baumol effect reminds us that some cost increase is normal and expected—not a sign of waste or corruption, but of uneven technological progress. This doesn't mean there's no waste in healthcare, but it does mean that even a perfectly efficient healthcare system would get more expensive relative to other goods.

When people ask why college is so expensive, the Baumol effect is part of the answer. Teaching remains labor-intensive. Professors haven't become dramatically more productive in the classroom. But they need to be paid wages that compete with sectors that have become more productive.

When people wonder why governments seem unable to provide services without ever-increasing budgets, the Baumol effect explains a large part of it. Government is disproportionately in the business of labor-intensive services. Those services will cost more over time, no matter how efficiently the government operates.

And when people worry about economic growth slowing in rich countries, the Baumol effect offers one explanation: as economies mature and shift toward services, the opportunities for dramatic productivity gains may narrow.

The Limits of the Theory

The Baumol effect is a powerful lens, but like any economic theory, it has limits and critics.

Some argue that the shift to services is driven more by changing demand than by productivity differences. As people get richer, they want more experiences, more personal attention, more custom services—not just more stuff. This is related to what economists call Engel's law: the share of income spent on necessities like food declines as income rises, while the share spent on other things increases.

Others point to globalization. Manufacturing employment in rich countries has declined partly because manufacturing jobs moved to countries with lower labor costs, not just because factories became more efficient.

Still others emphasize changing patterns of work. The rise of women in the workforce, the growth of specialized business services, the outsourcing of tasks that were once done inside firms—all these have contributed to the growth of service employment in ways that aren't fully explained by productivity differences.

The Baumol effect is real and important, but it's not the only force shaping modern economies. It's one thread in a complex tapestry.

Looking Forward

What happens next depends on questions we can't yet answer. Will artificial intelligence transform healthcare, education, and other services the way mechanization transformed manufacturing? Will new technologies find ways to scale personal services that currently require one-on-one human attention? Or will we find ourselves in an economy where an ever-larger share of income goes to human-scale services that resist automation?

The optimistic view is that technology will eventually reach even the stagnant sectors. Artificial intelligence might diagnose diseases better than doctors. Robot tutors might personalize education at scale. Virtual reality might create social experiences that substitute for some in-person services.

The pessimistic view is that the most valuable human services are inherently resistant to scaling. People don't just want their health managed—they want a doctor who listens to them. People don't just want information transmitted—they want a teacher who knows their child. The irreducibly human elements of these services might be precisely what makes them valuable.

Most likely, the future will be somewhere in between. Some aspects of today's stagnant services will be transformed by technology, and their costs will fall. Others will remain human-scale, and those will keep getting more expensive.

Either way, the Baumol effect gives us a framework for understanding what's happening—and a reminder that rising costs in some sectors aren't necessarily a sign that something has gone wrong. Sometimes, they're just the price of progress.