The world’s manufacturing juggernaut

From the Boston Globe of all places:

Americans make more “stuff’’ than any other nation on earth, and by a wide margin. According to the United Nations’ comprehensive database of international economic data, America’s manufacturing output in 2009 (expressed in constant 2005 dollars) was $2.15 trillion. That surpassed China’s output of $1.48 trillion by nearly 46 percent. China’s industries may be booming, but the United States still accounted for 20 percent of the world’s manufacturing output in 2009 — only a hair below its 1990 share of 21 percent.

“The decline, demise, and death of America’s manufacturing sector has been greatly exaggerated,’’ says economist Mark Perry, a visiting scholar at the American Enterprise Institute in Washington. “America still makes a ton of stuff, and we make more of it now than ever before in history.’’ In fact, Americans manufactured more goods in 2009 than the Japanese, Germans, British, and Italians — combined.

American manufacturing output hits a new high almost every year. US industries are powerhouses of production: Measured in constant dollars, America’s manufacturing output today is more than double what it was in the early 1970s.

HT:  Greg Mankiw

PS.  The four countries mentioned in paragraph two have a combined population greater than the US.



38 Responses to “The world’s manufacturing juggernaut”

  1. Gravatar of Luis H Arroyo Luis H Arroyo
    6. February 2011 at 09:49

    Scott, is that perhaps a case against Cowen´s book and his Great Stagnation?

  2. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. February 2011 at 10:04

    Not at all a surprise to me but then I follow such numbers. Americans loudly complain about all the Chinese manufactured consumer goods at “the Walmart” but they are unaware of all the capital goods America exports (and which are then used to manufacture the consumer goods that we get back in return). (Yes indeed, comparative advantage and the benefits of trade.)

    P.S. Also, a trip to the mall in Europe or Asia will reveal a surprising amount of American manufactured consumer goods.

  3. Gravatar of Hillary Hillary
    6. February 2011 at 10:14

    (I bought The Great Stagnation but haven’t had time to read it, so my understanding of the argument is coming from blog posts about the book, including Cowen’s.)

    Luis – not necessarily. A lot of that manufacturing is probably industrial equipment. Has there been revolutionary improvement or innovation in that sector since the 60s? I’d say yes in electronics, medical technology, automation systems, and offshore drilling. But not in most industrial equipment.

    My last job was for a company that designs & builds very large equipment. Their designs hadn’t fundamentally changed since the forties (and there’s equipment they built in the forties and fifties still in service). They’d upgraded the controllers on the electric-controlled systems, but that was about it. The big change in the last fifteen years was color screens instead of black & white.

    A combine is still a combine and an engine is still an engine. We’ve made incremental improvements, but we’re still doing things the same way.

  4. Gravatar of Scott Sumner Scott Sumner
    6. February 2011 at 10:25

    Luis, Not really, although I see why people might make that assumption.

    Mark, Yes, I knew this too. But I hear so much about the deindustrialization of the US that I thought it was worth mentioning.

    Hillary, That sounds right to me.

  5. Gravatar of david david
    6. February 2011 at 10:44

    It gets worse; when Those People do hear about the capital goods exports, they then switch complaints to that about technology theft.

  6. Gravatar of Mike Sandifer Mike Sandifer
    6. February 2011 at 11:11

    I’ve made this point many times to people with disbelieving eyes. But, it doesn’t take long for them to agree, usually because they haven’t looked at any data at all.

    It was Robert Reich who first pointed my attention to data like this, as he’s long focused on automation as the principle factor in the loss of American manufacturing jobs, and indeed jobs in many other countries in the world, China included.

  7. Gravatar of Luis H Arroyo Luis H Arroyo
    6. February 2011 at 11:56

    I´m reading Cowen´s Book. I don´t agree totally -for the moment- but he have got good points. for example, the (lack) of marketable valuation of public services as a source of decadence.
    But, If US is stagnant, how is Europe! I hope he is not right.

  8. Gravatar of nanute nanute
    6. February 2011 at 12:56

    We may still be a leader in manufacturing. The problem is that workers haven’t shared in the gains of productivity and profits from manufacturing. And, we’ve got this insatiable appetite to import more cheap shit from China than we export. Then again, we’re a consumer economy, what would you expect? We can’t make a pair of pants for a dollar.

  9. Gravatar of Lorenzo from Oz Lorenzo from Oz
    6. February 2011 at 13:32

    Nanute: The problem is that workers haven’t shared in the gains of productivity and profits from manufacturing. Oh really? That is not what this suggests. (I love a good graph.)

  10. Gravatar of nanute nanute
    6. February 2011 at 14:03

    Lorenzo: I should have been more precise: I wasn’t referring to US vs. the rest of the world. I was thinking more in terms of the growing disparity in income among Americans:
    I love a good graph, too.

  11. Gravatar of johnleemk johnleemk
    6. February 2011 at 14:47

    “And, we’ve got this insatiable appetite to import more cheap shit from China than we export. Then again, we’re a consumer economy, what would you expect? We can’t make a pair of pants for a dollar.”

    I really don’t see why this is a big deal. Honestly, I’m perplexed as to why this is a problem. Importing more than you export is an accounting impossibility in any event; as Milton Friedman noted, what are the Chinese going to do with their dollars — eat them?

  12. Gravatar of Mike Sandifer Mike Sandifer
    6. February 2011 at 15:16


    I point out that if inflation-adjusted wages have fallen slightly over the past 30 years or more, but the prices of many goods and services have fallen more rapidly, then real income has increased some ways even for the lowest paid workers. Among exceptions are health insurance and college education, but the costs of food, clothing, household items,consumer electronics, cars, etc. are lower.

    I agree that too many of the benefits may have flowed upward, but it’s a definite advantage to have people elsewhere willing to work for so little to make items more affordable for us.

  13. Gravatar of W. Peden W. Peden
    6. February 2011 at 15:27

    Mike Sandifier,

    There’s also the usual qualifier with “real” statistics, which is that these are generally based on indexes of prices that overstate inflation e.g. CPI. Charts like the one to which nanuntes linked also hide the gains of immigrants and women.

    That doesn’t address the question of inequality, but it does mean that the stagnation of income is largely a product of naive (at best) statistics.

    As for inequality, there is the fascinating assumption (that Krugman in particular likes to make) that every worker has had increasing productivity in the last 30 years i.e. we should take productivity gains as an aggregate, divide them equally, and compare them against gains in income. Now, it may well be that productivity gains HAVE been very evenly distributed, but that’s (a) not shown in the standard statistics and (b) still irrelevant.

    I would be interested to see the productivity gains of the lowest quintile compared against the gains in real income using the GDP deflator and comparing the income of immigrant Americans against Mexican/Puerto Rican/etc. income circa 1980.

  14. Gravatar of Goldilocksisableachblonde Goldilocksisableachblonde
    6. February 2011 at 15:36

    I’m not sure about the U.N. data , but BEA data on manufacturing output is suspect because of the significant , and misleading , impact of the offshoring of various inputs. They’ve been trying to get a handle on it for a few years now , as I understand it.

    Here’s a Business week summary from 2007 :

    “BusinessWeek has learned of a gaping flaw in the way statistics treat offshoring, with serious economic and political implications. Top government statisticians now acknowledge that the problem exists, and say it could prove to be significant.

    The short explanation is that the growth of domestic manufacturing has been substantially overstated in recent years. That means productivity gains and overall economic growth have been overstated as well.”

    “By BusinessWeek’s admittedly rough estimate, offshoring may have created about $66 billion in phantom GDP gains since 2003 (page 31). That would lower real GDP today by about half of 1%, which is substantial but not huge. But put another way, $66 billion would wipe out as much as 40% of the gains in manufacturing output over the same period.”
    Whatever the state of our manufacturing sector , Enron-style output accounting our producing sector is a very bad sign.
    And , naturally , the same problem exists in the service sector.

    The good news is that similar accounting problems can occur with domestic outsourcing , so that some things , like gov’t worker productivity , may actually be much BETTER than we think. ( Morgan’s dream realized ! )

  15. Gravatar of StatsGuy StatsGuy
    6. February 2011 at 18:31

    I feel dumber every day. I looked at the UN statistics, and I don’t know what they mean. Seriously, I don’t. I cannot recall (and am having difficulty finding the answer) whether ISIC D is net of manufacturing imports, or gross. Grr. I should know this.

    If it’s net, then surely ssumner has an excellent point.

    If it’s gross, well, gosh. It’s difficult to know exactly what this says, since it depends on the relative balance of imports to exports – many US plants “assemble” final products, but import significant components.

    Also, I think the data is measured in final market value price, but I can’t even remember if that’s wholesale or retail. (Yes, more dead brain cells.) If its final consumer value, then much of the manufacturing cost is dwarfed by sales, distribution, and marketing expenses, and I suspect that these expenses are relatively less for components.

    Dell computer, for instance, manufacturers most of its computers in Austin Texas, Tennessee, Florida, and North Carolina.

    Of course, most of the component parts (except processors) are made overseas. Since Dell maintains decent gross margins, the cost of the final machine dwarfs the cost of components. Also, because wages are so much higher in the US, the price attributed to “assembly” is much higher than it would be if assembly was conducted in Thailand – so the relative “value” is somewhat dependent on the relative wage rate.

    So it rather matters whether ISIC C counts the value added in the US, or the total output price, and even then I would be concerned that measurement based on output price for consumer goods (vs. input components) dramatically overstates the amount of production activity actually conducted in the US.

    That’s not to say the US doesn’t maintain considerable high end production – Boeing, Intel, Caterpillar… Paper and fabric mills (very capital intensive), also oil refining (counts in ISIC D, I think) for all the oil we import… Still, I rather strongly suspect ssumner’s argument above is more than a little overstated.

    But I could be totally wrong, because I don’t know what the data means.

  16. Gravatar of StatsGuy StatsGuy
    6. February 2011 at 18:40

    Now I’m even more confused. I figured, I’d get some clarity by looking at merchandise trade balance numbers, bilateral and total. So, US manufacturing output dwarfed China, but somehow the US imported more merchandise from china than we exported. Over 200 billion dollars more in 2009.

    My neurons, all four of them, are shutting down now.

  17. Gravatar of Carl Lumma Carl Lumma
    6. February 2011 at 21:00

    From 1960 to 2009, the U.S. population grew by a factor of 1.7, workforce by a factor of 2.4, and the number of people working in manufacturing by a factor of 0.8. Manufacturing share of GDP fell from about 25% to 10%, placing us below China, Korea, Germany Japan, Italy, and Brazil. Then you have the current account…

    But we added a ton of high-flying information-era jobs right? Well, not exactly. IT employment grew by a factor of 1.6. Leisure and hospitality: 3.8. Education and health: 6.5.

  18. Gravatar of Bryan Willman Bryan Willman
    6. February 2011 at 21:28

    The article cited is of course not news.

    What is also not news, but people seem to miss, is that this pattern has happened before in agriculture – where society when from something like 90% employed in agriculture to about 3% so employed, in more or less a century. You don’t normally hear people complaining about “loss of farm jobs” any more. Prices? Yields? Regulatory, environmental, labor, capital, issues? Sure.

    And of course some pool of people who could have had a job as farm hands in 1910 (say) are effectively unemployable now.

    By the way, at least some industrial equipment has changed a great deal in the last 4 decades – CNC machine tools being one example. The changes in productivity have been huge. Oh, and while Haas is the largest volume CNC machine maker in the world, and a few other very high end players remain in the US, that industry is now mostly based in Europe and Asia.

    But the real long term issue (well, I hope longer term than my lifetime) is what do we do when say 30% of the population can make everything we need? How do society, employment, and the patterns of commerce work in such a cirumstance?

  19. Gravatar of Matt R. Matt R.
    7. February 2011 at 07:40

    The articles numbers and the UN source don’t match up.
    Check the original data source.

    The comparison really isn’t a clean one. The author used the line titled “Mining, Manufacturing, and Utilities” to get his numbers. Unless mining and utility output is zero we don’t have clean manufacturing numbers.

    The UN does provide manufacturing specific data for the US as $1.68 trillion. The similar entry for China is just a repeat of the “Mining, Manufacturing, and Utilities” number.

    The manufacturing sector in the US appears larger, but really have no idea by how much. At least, not from this UN data.

  20. Gravatar of Chris T Chris T
    7. February 2011 at 14:06

    How is this not a strike against Cowen’s argument? American manufacturing has grown unabated (except during the recent unpleasantness) even while the number of people employed has gone off a cliff since 2000:


    Either workers were really lazy until recently or industry has really taken to automation in the last decade. I somehow doubt the former.

  21. Gravatar of dtoh dtoh
    7. February 2011 at 15:11

    Bryan is spot on with his analogy to agriculture. Manufacturing jobs in the U.S. are going away…..manufacturing is not.

  22. Gravatar of Scott Sumner Scott Sumner
    7. February 2011 at 15:24

    David, Good point.

    Mike, Yes, I’ve also made this argument many times. People just don’t look at the data.

    Luis, I agree.

    nanute, You said;

    “And, we’ve got this insatiable appetite to import more cheap shit from China than we export.”

    Bi-lateral trade deficits are meaningless. We shouldn’t even keep data on the bilateral relationships, they just confuse people. Many of our “imports” from China are disguised imports from other countries.

    Johnleemk, I agree.

    W. Peden, Those are good points.


    Nice article. It doesn’t have much effect on the argument that manufacturing in the US is 20% world manufacturing. But it does suggest that RGDP and inflation numbers shouldn’t be trusted, which is a point I strongly argue in this blog.

    Statsguy, It would be more accurate to say it’s Jeff Jacoby’s argument, not mine. But this is all pretty much common knowledge among economists, just not the general public.

    Never look at bilateral trade patterns, there are literally meaningless. They don’t even measure what they claim to measure. And if they did, they’d still be meaningless.

    Carl, You are confusing manufacturing output as a share of GDP, far more than 10%, with manufacturing employment as a share of employment, less than 10%. We produce lots more stuff, with fewer workers (due to automation.)

    Bryan, Yes, it’s widely known among economists, but not the public. Even in the 1980s (rust-belt period) we saw US industrial output rise much faster than in Europe. You said:

    “But the real long term issue (well, I hope longer term than my lifetime) is what do we do when say 30% of the population can make everything we need? How do society, employment, and the patterns of commerce work in such a circumstance?”

    Easy, then we switch over from producing what we need, to producing what we want. Since our wants are unlimited, we always have jobs.

    Matt R. Good point, but since industrial output is mostly manufacturing, and since mining and utilities haven’t grown that fast, the qualitative conclusions would probably be the same either way.

    Chris T. I don’t trust productivity growth rate numbers in manufacturing. Too much fiddling around with quality improvements. I think Tyler is saying that a larger and larger share of our economy is in health, education, and government, where productivity growth is weak. I’m sure that Tyler agrees that manufacturing productivity is still rising.

  23. Gravatar of Carl Lumma Carl Lumma
    7. February 2011 at 16:06

    Hi Scott,

    I’m not


  24. Gravatar of Chris T Chris T
    7. February 2011 at 16:23

    That wasn’t productivity growth data; it was employment data. The point still stands, American manufacturing has maintained a relatively stable share of global output even while shedding over 4 million jobs. This after a period of fairly constant level of employment for 20 years.

  25. Gravatar of James Davies James Davies
    7. February 2011 at 17:57

    Is the relative size of the US and China’s manufacturing sectors quoted in that Globe article based on exchange rate or purchasing power parity?

    This of course matters, as it’s the same mistake that papers made last year when China overtook Japan as the second largest economy in the world based on official exchange rates. But of course everyone knows the yuan is undervalued, and likewise, everyone knew China’s economy had overtaken Japan’s in size years earlier when based on purchasing power parity.

  26. Gravatar of “Hope is in the air” « Historinhas “Hope is in the air” « Historinhas
    7. February 2011 at 18:54

    [...] on an up note: Despite all that, as Scott Sumner notes, the US can still be described as “The world´s manufacturing [...]

  27. Gravatar of Chris T Chris T
    7. February 2011 at 19:09

    I think Tyler is saying that a larger and larger share of our economy is in health, education, and government, where productivity growth is weak.

    True, but we’ve also forced a lot of workers that would have been employed in manufacturing in prior decades into other fields. If the same changes affecting manufacturing are affecting other fields (a lot of white collar jobs no longer exist), we’re going to end up with a lot of disruption. Wouldn’t having a lot of displaced workers depress wage growth, particularly when the jobs were middle income? Would it not also lead to much higher demand for education?

    It would seem to me a better hypothesis is that some industries are innovating very rapidly and people are getting displaced into jobs that are less susceptible to technological change. Since these jobs currently have a large pool of people to draw from, there is little pressure on wages.

  28. Gravatar of Goldilocksisableachblonde Goldilocksisableachblonde
    7. February 2011 at 19:21

    “It doesn’t have much effect on the argument that manufacturing in the US is 20% world manufacturing.”

    I hope you’ll elaborate on this. If US manufacturers represented 20% of global manufacturing , but 100% of that took place in the offshore factories of those US companies , it wouldn’t be too meaningful to talk about US manufacturing output and productivity , would it ? The US-located facilities wouldn’t be manufacturers , they’d be wholesalers.

    What these studies show is that a big part of value-added attributed to U.S. manufacturing facilities is actually attributable to foreign factories , due to import accounting distortions.

    Here’s a good summary paper by the same group mentioned in the Business Week piece – Houseman of Upjohn and Kurz , Lengermann , and Mandel of the Fed :

    A few clips :

    “…We estimate that average annual MFP growth in manufacturing was overstated by 0.1–0.2 percentage points and real value-added growth by 0.2–0.5 percentage points from 1997 to 2007. And, although the bias from offshoring represents a relatively small share of real value-added growth in the computer industry, it may have accounted for one-fifth to one-half of the growth in real value added in the rest of manufacturing. Moreover, our work only examines biases to manufacturing statistics from the offshoring of material inputs. Additional biases may have arisen from the substitution of imported for domestically produced capital equipment and from the offshoring of services.

    These biases have implications not only for the industry statistics, but also for the analyses based on them. Because the growth of these imports will be understated in real terms, offshoring will, at least to some degree, manifest itself as mismeasured productivity gains….”

    “…Excluding computers, real value-added growth for manufacturing is biased by 0.2–0.4 percentage point, which accounts for 21–49 percent of the growth. The annual growth rate of real value added for manufacturing excluding computers falls under a half percent per year in some of our
    adjusted estimates, while that for nondurable goods turns negative for all of our adjusted estimates…”


    Keep in mind that they only examined the effects related to mispricing of materials imputs ( the largest input overall ) , and did not examine capital and services inputs. Since capital inputs typically includes computers and machinery , I’d have to think the numbers for computer manufacturing ( which come out looking better in this study than does manufacturing ex-computers ) would be suspect , as well.

  29. Gravatar of Mark A. Sadowski Mark A. Sadowski
    7. February 2011 at 22:00

    Here’s an interesting article by Kash noting that for the first time in a coon’s age (a very long time) US manufacturing is leading the recovery:

    “Something interesting is happening with this recovery. Actually, it’s something that was also an interesting feature of the Great Recession of 2007-09: the manufacturing sector of the US economy has been performing surprisingly well. In fact, it wouldn’t be too much of a stretch to say that manufacturing has been one of the important drivers behind the economic recovery (such as we’ve had) over the past two years. Normally manufacturing output and job growth lag overall economic recoveries in the US. This time manufacturing is leading it.”

  30. Gravatar of marcus nunes marcus nunes
    8. February 2011 at 08:35

    I did a post critical of Kash´s “musings”

  31. Gravatar of Scott Sumner Scott Sumner
    10. February 2011 at 06:14

    Carl, 13% is not 10%. In any case we produce more than the Europeans and Japanese (per capita). We just do it with a smaller population because we are more productive. What’s the problem?

    Chris, I agree.

    James, I don’t know. But PPP is a much smaller problem for manufacturing (which deals in traded goods) than overall GDP.

    Chris, I don’t know enough about productivity trends to comment on your hypothesis. It sounds possible.

    goldilocksisasbleachedblonde, The article you cite says RGDP may be overstated by as much as 1/2%. I don’t see how that could materially affect the 20% of GDP estimate.

    BTW, another problem is determining what exactly is meant my “manufacturing.” If GM outsources cleaning services or accounting services, then employment in manufacturing drops, even though nothing changes in reality. Do workers at McDonalds “manufacture” a Big Mac out of components? Deep philosophical questions . . .

    Mark and Marcus, I agree with Marcus, I don’t see much difference from the earlier recessions.

  32. Gravatar of Chris T Chris T
    10. February 2011 at 09:37

    MSNBC had an article on this very topic earlier this week. It relies somewhat on anecdotal evidence, but it is suggestive:

    The NY Times had an article on how factories are having problems finding workers with the skills they need:

  33. Gravatar of Scott Sumner Scott Sumner
    11. February 2011 at 19:10

    Chris, I think that’s right, we are doing much better in terms of output than employment in manufacturing.

  34. Gravatar of Carl Lumma Carl Lumma
    12. February 2011 at 15:10

    @Scott I don’t have a problem with U.S. manufacturing per se, though I don’t know much about it. I do have a problem with the idea that mechanization implies less manufacturing involvement (share of employment). That may be the case, but efficiency improvement can also cause an increase in demand/involvement. Has manufacturing involvement decreased worldwide?

    I’m very optimistic about the U.S. (especially the next 10 years) but I’m critical of certain trends since 1970:
    * we monetized much of what women used to do for free
    * created mostly jobs in leisure/hospitality/education/health
    * got a bunch of inflation in positional goods (ref E Warren)
    * saw a decline in primary education quality and a produced a raft of “communications majors”
    * suffered a decline in many facets of domestic and civic life (ref R Putnam)

    I think some of these may be related.

  35. Gravatar of ssumner ssumner
    13. February 2011 at 08:29

    Carl, Yes, manufacturing jobs are being replaced by automation everywhere–even China.

    The monetization of women’s job has been wonderful news. It’s boosted living standards.

    It makes sense to use our improved productivity to have more leisure, health care, etc. Do you really want two dishwashers? Three stoves? Four washing machines? Five dryers? What else would we do with all the labor? I suggested building bigger houses and more houses. But people didn’t think that worked out well (I can’t imagine why.)

    I agree about education and civic life. I’m a big believer in decentralization, which I think improves civic involvement. Instead the federal government is gradually taking over education, so why should the average person get involved, if they have no say in things?

  36. Gravatar of Carl Lumma Carl Lumma
    13. February 2011 at 11:21

    @Scott Huh, I would guess more people in China are working in manufacturing than 30, 20, or 10 years ago. :P I agree about decentralization.

  37. Gravatar of Carl Lumma Carl Lumma
    13. February 2011 at 11:21

    Hm, sorry about the emoticon.

  38. Gravatar of ssumner ssumner
    15. February 2011 at 19:19

    Carl, A few years ago I read an article claiming that the number working in manufacturing had fallen sharply in China, due to improved efficieny. Recall thy used to have like 50,000 workers in SOE steel mills that needed maybe 2000. It’s pretty easy to expand manufacturing output sharply while cutting employment. I have seen the most recent figures.

Leave a Reply