In an op-ed in the Wall Street Journal econ prof Martin Eichenbaum argues against the idea that we need to do more manufacturing:
According to the Bureau of Labor Statistics, manufacturing’s share of nonfarm employment declined from roughly 32% in 1947 to approximately 8% at the end of 2023. Yes, there was a slight increase in the rate of decline around 2001, when China entered the World Trade Organization. But you have to stare pretty hard to see the effect. In any event, that effect is trivial compared to the long, slow, inexorable decline in the importance of manufacturing as a source of U.S. jobs.
I found that puzzling. It’s not hard at all:
Here’s his conclusion:
I started my career teaching at Carnegie Mellon University in Pittsburgh. There was real suffering when manufacturing declined in that area. Retraining a 45-year-old steel worker to become a biotech engineer is hard. Proponents of free markets sometimes forget that there are real political consequences of ignoring the human cost of change.
The right answer to the challenge of change involves at least three initiatives. First, the government must materially help the people who are affected and help them find jobs in which they can take pride. Second, parents and educators must provide children with the skills they need to thrive in a constantly changing world. Third, politicians must remove unnecessary regulations and other barriers to growth in emerging sectors of the economy.
The wrong answer is to pine for a mythical golden age that never existed. Pittsburgh is now a thriving center of education, research and health services. It didn’t get there by trying to bring back the 1950s.
He fails to define “thriving”. Pittsburgh’s population in 1950 was 677,000 people. Today it’s 303,000—less than half of that. Most of that decline had been accomplished by 1990, long before trade with China was a factor but it dropped by 1% between the 2010 and 2020 censuses. Pittsburgh’s real per capita income has declined slightly from 1950 to the present (from $44,000 in 1950 to $42,000 now in today’s dollars). That does not sound like thriving to me. It sounds like “declining but not collapsing”.
I suspect he’s making a false analogy. Let’s consider a highly simplified version of what’s happened to agriculture, looking at wheat production only:
Do I need to document that we are producing wheat with a lot fewer farmers than we used to? But:
I think those tell different stories. I wonder what Dr. Eichenbaum thinks?
Rather clearly for the last 25 years we have increasingly been transitioning to a service economy. It’s easy to measure the productivity of wheat farmers or steelworkers. You just divide the amount of steel by the number of workers (or thereabouts). Measuring the productivity of workers in a service economy is considerably more difficult. It’s almost entirely political.
Just to take one controversial example, how do you measure the productivity of physicians? Milton Friedman thought that the productivity of physicians was, essentially, flat. Physicians strenuously object to using any outcome-based measurement which is handy for them. The metric being used is something called “relative value units” (RVUs). There’s a discussion of RVUs here. Basically, you count “procedures” and then adjust that based on the relative values of the procedures. That’s basically circular.
My view is that measuring productivity based on procedures is not only circular but actually perverse. Our rate of iatrogenic morbidity and mortality (i.e. physician-caused) is too high. Just how high is controversial but it’s something between 45,000 and 450,000 deaths per year. IMO one of the factors in that is the predisposition to treat which is encouraged by using procedures performed as a metric for physician performance.
A similar story can be told for every service. How do you measure the productivity of an economics professor? Without such metrics things like “gross domestic product” (GDP) have no validity and formulating policies based on whether GDP is increasing or decreasing is fraught. Trying to measure in dollars leads you to the parable of the three merchants.
Yeah. The Neocons thought it playing games with statistics was cute, and then, they realized that smart bombs and artillery shells do not come from the munitions fairy. I guess nobody learned anything from the COVID shortages.
Rebuilding using materials based upon statistics from some dumb-ass is not easy or quick. You really do not want Chinese drywall in your house.
Post-WW2, the US was the only country with an intact manufacturing base. In order to strengthen US allies, the US helped to rebuild their manufacturing bases, and Bretton Woods was to facilitate trade between allies. (Dollarization was an outcome, not a nefarious plan.)
So, US manufactured goods were being exported to countries where the US was helping to rebuild their manufacturing base, and naturally, those jobs would eventually be exported. Additionally, automation should continuously eliminate jobs, and if the number of products produced in 1947 were still being produced, the amount of manufacturing jobs would be miniscule.
Apparently, none of the dumb-asses have entered a Walmart in the past 50 years. Almost nothing in there, including the store itself. existed 50 years ago, and an American worker at an automated manufacturing plant could produce it.
NOTE: China is not an ally. Israel is not an ally.
I think I’ve quotes this Keynes argument before:
“But I am not persuaded that the economic advantages of the international division of labor today are at all comparable with what they were. I must not be understood to carry my argument beyond a certain point. A considerable degree of international specialization is necessary in a rational world in all cases where it is dictated by wide differences of climate, natural resources, native aptitudes, level of culture and density of population. But over an increasingly wide range of industrial products, and perhaps of agricultural products also, I have become doubtful whether the economic loss of national self-sufficiency is great enough to outweigh the other advantages of gradually bringing the product and the consumer within the ambit of the same national, economic, and financial organization. Experience accumulates to prove that most modern processes of mass production can be performed in most countries and climates with almost equal efficiency. Moreover, with greater wealth, both primary and manufactured products play a smaller relative part in the national economy compared with houses, personal services, and local amenities, which are not equally available for international exchange; with the result that a moderate increase in the real cost of primary and manufactured products consequent on greater national self-sufficiency may cease to be of serious consequence when weighed in the balance against advantages of a different kind. National self-sufficiency, in short, though it costs something, may be becoming a luxury which we can afford, if we happen to want it.”
I take his argument here as suggesting that productivity in manufacturing not only drives down costs, but it makes policies that encourage internal production more affordable. He doesn’t think all countries can afford it and it’s possible that the U.S. could not afford it if high energy costs are too high relative to countries that are not pursuing high energy cost policies. But I would tariff on the basis of removing comparative environmental regulatory advantage.
Add safety, labor, etc. regulatory advantage and that’s my view. Although in fairness China’s environmental regulations are more stringent than ours. They just don’t enforce them.
The bottom line is that manufacturing produces a high wage high tax revenue economy, and services produces a low wage low tax revenue economy. Which do you think wins. Make your children learn Russian or Mandarin.
PS. Re iatrogenesis:
https://www.theburningplatform.com/2024/10/04/the-covid-vaccine-all-risk-no-benefit/
I guess the bit about doctors is supposed to be bait for me. There are a number of ways to measure productivity. I have offered some of them before. For example cataract surgery. It used to take several hours, was done as an inpatient and required a week or so of bedrest in hospital after surgery and you didnt really return to function for weeks. Now, you are in and out fo the outpatient center in 1-1/2 hours and are functional the next day. Besides which the surgery was limited to fairly healthy patients in the old days and there is almost no health limiting factor now.
Anyway, we know how fix the manufacturing issue, we just have to subsidize it. We did that for semiconductors and the TSMC plant in Arizona, as I read recently, is already turning out chips.
As an aside, there has been a lot fo literature looking at the premium in pay for manufacturing with many claiming it is gone and the rest saying it has declined. If it exists it’s probably about a 10% premium, down for about 17% in the 70s-80s. This should be pretty obvious given that a lot of manufacturing jobs have moved to states where they dont have to have unions so they can pay lower wages. Along with this it’s notable that many areas of US manufacturing have little adoption of robotics and automation, as noted in the dockworkers strike.
If we want more manufacturing we need more robotics/automation to increase productivity. It’s not clear it will really back many more jobs and its not clear that they will pay much, if any more.
Steve