Arnold Kling  

Cost Disease and Class War

Trade Controversies... Comment of the Week, 2003-07-2...

Steve ("econopundit") Antler says that Baumol's Cost Disease is the basis for a new form of class warfare, pitting service producers against goods producers.

Much of what's normally called "technical progress" is actually the "iron law" of service pricing in action. Postwar advances that replaced household servants with home appliances evidenced this law. We can say the same for the desktop computer revolution. Machines replace people as people get more expensive...

The Democratic Party is quite close to a grand alliance of service-providing re-distributors. The Republicans, on the other hand, seem fast becoming the party of goods-providing producers.

Beyond the tendency for government workers to prefer political parties that favor big government, I am not convinced that economic interests will determine voting behavior in the way that Antler describes. However, if the Republicans are to become the party of goods industries, then they will be the minority party. Just as agriculture has declined to a small share of employment and GDP, the same thing is happening with the goods industries.

For Discussion. Will technological progress ever lead to rapid productivity growth in service industries, such as health and education?

Comments and Sharing

COMMENTS (16 to date)
Eric Krieg writes:


I don't mean to bust anyone's bubble, but SERVICES increased their productivity faster than manufacturing during the 1990's. Think Walmart and warehouses, and the revolution that information technology has brought to retailing and the movement of goods.

Now, there is definitely a problem with productivity in health care and government services, but it is a management problem, not a fundamental problem inherant to services themselves.

I'll give you an example. According to my local councilman, there hasn't been an increase in productivity in snowplowing in 25 years. BTW, this man is a manager at a huge multinational auto parts manufacturer, so I THINK he knows a thing or two about productivity.

To which I ask, why can't there be productivity increases in snowplowing? First, you use graphical information system (GIS) technology to map the most efficient routes. Then you use infrared cameras mounted on the trucks to tell you how cold the ground is, and thus precisely how much salt is needed to keep roads clear. Maybe you can embed sensors in the roads to do something similar. GPS on the trucks would allow managers to know where the trucks are at all times, to be routed to where they are most needed in the most efficient manner.

Maybe one day those trucks will drive themselves.

According to my councilman, one big problem is that there simply aren't the software tools for local governments that the big corporations have. One example is enterprise resource planning software. SAP isn't available for government.

But does anyone really think that this will be the situation forever? One day there WILL be ERP for government, and productivity will increase.

Bernard Yomtov writes:

Is it really true that technological advance hasn't improved productivity in health care? If so, maybe we're using the wrong measure.

There are lots of people alive today because of medical technology (broadly defined), and lots of people who are healthier than they would otherwise be.

Shouldn't that count as productivity?

Eric Krieg writes:

>>Shouldn't that count as productivity?

Bernard Yomtov writes:

"Productivity usually makes things cheaper. "

OK. What does a coronary bypass operaton cost? How much would it have cost 50 years ago?

This sort of measurement problem is hardly restricted to health care, of course. It applies to any new product. It's easy to compae the cost of a car 50 years ago to one today, but don't forget that the CD player, anti-lock brakes, etc. were not part of the 1953 package.

Eric Krieg writes:

>>It's easy to compae the cost of a car 50 years ago to one today, but don't forget that the CD player, anti-lock brakes, etc. were not part of the 1953 package.

dsquared writes:

The point here is surely that we can't manufacture hours in the day (Schopenhauer remarked how wonderful it would be if as you bought a book, you could also buy a couple of hours in which to read it). Services take time, and time is a fixed input to a lot of them.

Eric Krieg writes:

Why is time a constraint to services, but not to manufacturing?

Even the haircut example is flawed. Maybe the actual process of cutting hair resists change, but that isn't the big time waster. Usually, I have to wait whan I go to the barber. He doesn't take appointments. And I assume that there are other times when no one is there.

But if he used a web site to take appointments, he could cut the time that many of his customers waste waiting, and he could shift some of those people to unproductive times.

Matt Young writes:

I never understood the difference in a good and service.

Dell Inc will assemble your computer parts after you order them, is that a good or service? Why is Dell different than Walmart?

I can obtain a typed out legal brief on-line, without a lawyer. Isn't that a product? If I pay a lawyer $500 to do the same thing, why do we call that a service?

The distinction is artificial.

If a service is that human component of a result, then improving the productivity of the human machine requires evolution or genetic engineering.

Matt Young

Eric Krieg writes:

>>If a service is that human component of a result, then improving the productivity of the human machine requires evolution or genetic engineering.

Scott writes:

“Try starting a business filling out legal forms for people and see how long before you get hit by a lawsuit based on your violation of the licensure laws.”

You need to move to Arizona. The legislature got mad at the lawyers a few years back and the law prohibiting the practice of law without a license went away. Lots of people will help you fill out the forms around here.

Scott writes:

Productivity in education seems to be stuck. However, at the college level, it’s getting better at minimizing student time. Using internet classes and compressed one class at a time formats, private colleges are getting people undergraduate degrees in a couple of years. Graduate degrees take 18 months – all in the evening while you work full time.

If we could get the community colleges to work with the high schools, I’m sure we could graduate a bunch of high school seniors with AA degrees.

On the other hand, a 4 year degree seems to take 5-6 years to complete at most state universities.

Eric Krieg writes:

This paper might have some relevance. What do you guys think of the claim that manufacturing has a larger "multiplier" effect than services do?

Matt Young writes:

I still don't get this services/goods distinction. Services seem to be activities that have more human capital and manufacturing more machine capital.

The more machine capital an activity has, the more prone it is to productivity improvement; and the greater the multiplier. We have the following dilemma; humans machines have barely improved their inherent functionality in 200,000 years; but machines have come a long way. Compare the "tractor driver" of 2,000 years ago and the "tractor" of 2,000 years ago. Which has improved the most, and which has had the greater multiplier effect.

Eric Krieg writes:

I don't think that's what they mean by multiplier. I think they mean "subconstractors". When my company, which builds oil refineries, gets a job, we subcontract out things. We might buy a pump. The pump vendor goes out and buys material for the pump, and then has it machined, and buys other parts, and pays to have his payroll outsources, etc. etc. etc.

Maybe services don't have as many opportunities for outsourcing, so there is less of a multiplier?


Mark Bahner writes:

"Maybe one day those trucks will drive themselves."

I predicted, at Brad DeLong's website, that by 2020, the first production car without a steering wheel (i.e., no human driver) will come off the line.

I also predicted that, by 2030, no production cars will be built with steering wheels.

Finally, I predicted that the grandchildren of anyone under (I think it was) 20 years old would never learn how to drive.

I'll stick to those predictions. Until 2015, anyway. ;-)

So, there is absolutely NO doubt in my mind that, someday, snowplows will operate without drivers.

Mark Bahner writes:

"Will technological progress ever lead to rapid productivity growth in service industries, such as health and education?"

I do a wide range of environmental research. Mostly air pollution, but also solid waste and water.

In many tasks (e.g. general literature research) the Google-ized Internet makes me...maybe 10 times more efficient than I was a decade ago. That is, in 1 day, I can get as much information as would have taken me 10 working days, 10+ years ago.

However, such tasks are probably only 10-30 percent of my total work. For the rest of my job, there probably hasn't been anywhere near that productivity improvement.

Comments for this entry have been closed
Return to top