Latest / The Productivity Podcast with Fexingo: Output, Efficiency, and Long-Term Economic Growth / The Baumol Effect Explains Rising Service Costs
Transcript
- Lucas: Let me start with a number that stopped me the first time I read it. In 1965, a brand-new Ford Mustang cost about $2,400. That same year, a ticket to see the New York Philharmonic cost maybe $5. Fast-forward to today, a Mustang starts around $30,000 — that's roughly 12 times higher. A good orchestra ticket? It can run $150 or more. That's about 30 times higher. And the gap gets even bigger when you adjust for inflation. Luna: So the live performance got more expensive relative to the car. Why? Lucas: That's the question that two economists, William Baumol and William Bowen, asked in the 1960s. They were looking at the performing arts, but what they found applies to huge swaths of the economy — healthcare, education, childcare, legal services. They called it the 'cost disease of personal services.' Today it's often called Baumol's cost disease. Luna: Cost disease — that sounds grim. Is it literally a disease of rising costs? Lucas: Exactly. And the core idea is surprisingly simple. In goods-producing industries — manufacturing, farming, tech — productivity keeps rising. A car plant today builds way more cars per hour than it did in 1965. That drives down the real cost of each car. But in many service industries, productivity grows slowly, if at all. A string quartet still needs four musicians playing for about 40 minutes to perform a Mozart piece. There's no way to do it faster. You can't automate that — well, you could use synthesizers, but that's a different product. Luna: So the quartet can't produce more 'output per hour' in any meaningful sense. But the musicians still expect their wages to grow with the rest of the economy, because otherwise they'd leave for better-paying jobs. Lucas: Right. And since labor is almost the entire cost of the performance, the price of that ticket has to keep rising just to match general wage growth. That's the disease. It's not a disease of inefficiency — it's a structural feature of any industry where labor is the main input and you can't substitute machines for people without changing the nature of the service. Luna: So where do we see this most acutely today? I'd guess healthcare is the poster child. Lucas: Healthcare is the textbook example. A nurse can only see so many patients. A surgeon can only do so many operations. Sure, technology helps — better imaging, robotic surgery — but the basic unit of care is still a human interaction that takes a certain amount of time. So healthcare costs have been rising faster than inflation for decades. Education is similar. A professor can teach a class of 30 or 300 with lecture capture, but the personalized feedback, the office hours, the grading — that's hard to scale. Luna: And it's not just about wages. It's about the fact that as the economy gets richer, the opportunity cost of spending time on these services goes up. So the relative price has to rise. Lucas: Baumol and Bowen predicted this back in 1966. They said the performing arts would face a permanent cost squeeze — that ticket prices would rise faster than inflation, and that arts organizations would always be struggling financially. That's exactly what happened. But the same logic applies to universities, hospitals, daycare centers, even restaurants to some degree. Luna: Let's talk data. How much faster have education and healthcare costs risen compared to overall inflation? Lucas: Since 1980, the Consumer Price Index for medical care has risen about 500 percent, while overall prices rose about 250 percent. College tuition and fees? Up over 1,000 percent in the same period. That's not all Baumol — there are other factors like administrative bloat and regulatory changes — but the underlying productivity gap is a huge driver. Luna: And the flip side is that goods like cars, electronics, furniture have gotten much cheaper in real terms. So the average household spends a smaller share on stuff and a larger share on services. That's been a long-term trend. Lucas: Exactly. And this shift has big implications for how we measure economic growth. If the service sector is growing as a share of GDP, and services have slow productivity growth, then overall productivity growth may slow down even if manufacturing is booming. That's part of the productivity puzzle of the last two decades. Luna: So what's the solution? Can technology eventually cure the cost disease? Lucas: That's the big question. Telemedicine, online learning, ai assisted diagnostics — these are all attempts to inject productivity gains into services. But there's a tension. If you automate too much, you change the nature of the service. A lecture watched on YouTube is not the same as a live seminar. A virtual doctor visit isn't the same as a physical exam. So there may be a ceiling. Luna: And even if we do see productivity gains, Baumol's logic suggests that some services will always have rising costs because they are inherently labor-intensive. For example, live theater, psychotherapy, nursing home care — you can't replace the human element entirely. Lucas: Right. So one policy implication is that as a society, we need to accept that some things will get more expensive over time, and we need to figure out how to pay for them — through public funding, insurance, or higher wages. The alternative is to let quality decline or access shrink. Luna: That's a sobering thought. But it also means that productivity growth in goods sectors indirectly makes services more expensive by raising wages across the board. In a way, the cost disease is a symptom of success — we're so good at making stuff that the only thing left to buy is time with other people. Lucas: Yeah, I like that framing. It's a trade-off. And it's not new — Baumol and Bowen were writing about it 60 years ago. But the stakes are higher now because healthcare and education are such large parts of the economy and household budgets. Luna: Speaking of what's worth paying for — if this episode made something click for you, that's the link: buy me a coffee dot com slash fexingo. It's a small way to keep the show exactly how it is — no ads, just us going deep on one idea. Lucas: Yeah, honestly, if you got something out of this deep dive on Baumol, that's the best we could ask for. And if you want to help us keep doing these, that's where to go. Now, back to the cost disease — I want to talk about one potential 'cure' that's gotten a lot of hype: AI. Luna: AI in knowledge work. Could it finally break the productivity ceiling in services like legal research, software development, even teaching? Lucas: Possibly. But here's the catch — Baumol's disease is about services where the output is tied to human interaction. If AI lets a lawyer review contracts faster, that's a productivity gain. But if the client expects a human to explain the contract, that part stays labor-intensive. So AI might lower costs for some tasks, but it won't eliminate the cost disease for the core service. Luna: So we might see a two-tier system: a basic automated version that's cheap, and a premium human version that's expensive. That's already happening in medical advice — you can get symptom checkers for free, but a doctor visit costs hundreds. Lucas: And that's the future Baumol predicted — inequality in access to services. The rich will always be able to afford the human version, while everyone else gets the automated one. That's a social choice. Luna: Great episode, Lucas. I think I'll never look at a doctor's bill the same way again. Lucas: Same. Or an orchestra ticket.