Friday, May 20, 2011

US Manufacturing Employment

Paul Krugman has an editorial this morning expressing some optimism about US manufacturing and manufacturing employment in particular:
By the middle years of the last decade, I used to joke that Americans made a living by selling each other houses, which they paid for with money borrowed from China. Manufacturing, once America’s greatest strength, seemed to be in terminal decline.

But that may be changing. Manufacturing is one of the bright spots of a generally disappointing recovery, and there are signs — preliminary, but hopeful, nonetheless — that a sustained comeback may be under way.
The story so far: In the 1990s, U.S. manufacturing employment was more or less steady. After 2000, however, it entered a steep decline. The 2001 recession hit industry hard, while the bubble-fueled expansion of the decade’s middle years — an expansion marked by a huge rise in the trade deficit — left manufacturing behind. By December 2007, there were 3.5 million fewer U.S. manufacturing workers than there had been in 2000; millions more jobs disappeared in the slump that followed.

Only a handful of these lost jobs have come back, so far. But, as I said, there are indications of a turnaround.

Crucially, the manufacturing trade deficit seems to be coming down. At this point, it’s only about half as large as a share of G.D.P. as it was at the peak of the housing bubble, and further improvements are in the pipeline. The Boston Consulting Group, which is now predicting a U.S. “manufacturing renaissance,” points to major U.S. firms like Caterpillar that once shifted production abroad but are now moving it back. At the same time, companies from other countries, especially European firms, are moving production to America.
As context for this discussion, I looked at the BLS data for manufacturing employment, and normalized it with the civilian non-institutional population over age 20.  This is the result (along with recession bars in pale blue/grey):

You can see that this fraction has been declining since about 1970.  It's worst in recessions.  It used to recover somewhat between recessions, but since the 1980s it's tended to be flat or declining even during recoveries.  Krugman is quite right that it was worse in the 2000s than in the 1990s.  It's also true that its performance after the great recession is slightly better.  However, it's only flat - it's not actually recovering.

I don't personally see much medium term cause for optimism here.  The reason the 2000s were worse than the 1990s was the huge explosion of manufacturing in China.  Here's the BLS data for Chinese manufacturing wages as a percentage of US manufacturing wages:

In 2008, it was up to 4.2 cents on the dollar.  It is growing rapidly - 20-30% a year - but starting from such a small base, it's still going to be decades before it equalizes.  As long as there is such a massive gap, US manufacturing is going to continue to erode more years than not, and it will take a far greater collapse in the dollar to stop that trend.  Admittedly, that cannot be ruled out in coming decades.

If you were just to do the simplest thing and extrapolate the trend of the past 40 years, that suggests US manufacturing employment would reach zero some time in the 2030s:

Of course, many things could intervene to prevent that: some kind of revolution in China, a collapse of the US dollar, a shift to protectionism in the US: these are just a few examples.  But, absent a really big shift like that, I doubt this trend is going to turn around in the near future on account of moderate changes in the value of the dollar.


James said...

I believe that (i) the good news is that economic imbalances are self correcting in the long run, (ii) the bad news is that this imbalance is going to be corrected via a currency crisis.

I hailed a cab in downtown Buenos Aires about a year ago and my cab driver was about 72 years old. I am sure he did not intend to be driving a cab at 72.

The symbiosis between the desire of the average American to live beyond his means and the desire of the Chinese government to pursue a mercantilist trade strategy cannot go on forever.

[Of course it is not just Americans who like to live beyond their means - the Greeks have an even greater propensity to do so and lots of western countries are in the same league as the US in this regard.]

Anonymous said...

Don't rising fuel costs act as a tariff on imported goods? At what oil price would shipping costs equalize the wage gap?

Draft said...

Stuart - why not the obvious cause of a trend reversal: peak oil? Seems like shipping costs have no direction to go but up, and that's not something that can be substituted for. (i.e. I don't expect to see solar powered cargo ships anytime soon.)

Eventually we'll have to source our goods locally.

(Though in writing this I was reminded that I'm still puzzled on where you stand - sometimes I think you expect peak oil to be a major issue and sometimes it seems you think it won't be that major.)

Anonymous said...

Another possibility is that people will become easier to employ through relaxed regulations governing wages, hours, and safety. USA: the whiter Third World nation.

Anonymous said...

@Mythodrome... Doesn't that work both ways? Any exports that we ship to China will incur those same increased transportation costs.

Now, this hurts the Chinese more at a gross level because they export more to us than we do to them. But if both sides have increasing costs, does that significantly benefit either side? Increasing transportation costs are basically tariffs on distance not on imports. If you are an American company located in LA, with supply lines coming from all over the country, and deliveries of goods all over the country, then increased transportation costs are going to hurt you as well.

What is interesting though is looking at the percentage of final cost of goods made up by transportation vs. labor. If the difference is large in one way or the other, then even moderate moves in one will not materially affect where manufacturing goes. However, if they are similar in scale, then significant jumps in transportation costs could offset the advantage of cheap labor enough to change manufacturing siting decisions. I suspect that this ratio is very product/technology dependent, so rising transportation costs would tend to move some (but how many?) manufacturers back to the states.

I just don't know whether the effect would be very big absent really high oil prices. And, at really high oil prices, is there an economy that requires all of the existing manufacturing let alone additional capacity returning from abroad?

James said...

There are some solutions to rising transportation costs driven by rising fuel prices - electrified rail and fission powered container ships for example. Interestingly, China seems to be working like mad to develop their strengths in both those areas. (Electrified rail and fission power plants at least.)

KLR said...

The other day I found this article which has a graph depicting fuel costs as % of freight revenue: Shipping Intelligence Network 2005 - Features : Ballooning bunkers blister bank balances. This was something I'd been curious about for years but never found any data for. Jeff Rubin alluded to this phenomenon in his book, that with crude >$100/bbl fuel costs would be so onerous that manufacturers would give up on globalization and relocate closer to home, never mind that cheaper labor was to be had elsewhere.

Stuart Staniford said...

KLR - that's certainly a very interesting piece of data, but it's not exactly what we need to know. What we really need to know is what is the share of shipping fuel in the overall cost of manufactured goods, versus the share of labor. I suspect the share of fuel is still much smaller, but would like better data.

KLR said...

I like how shippers negated the cost of fuel as a % of the whole by simply jacking rates up. Yeah, that works, sort of! Here's another paper:

Ocean shipping, which constitutes 99 percent of world trade by weight and a
majority of world trade by value, also experienced a technological revolution in the
form of container shipping, but dramatic price declines are not in evidence.
Instead, prices for ocean shipping exhibit little change from 1952–1970, substantial
increases from 1970 through the mid-1980s, followed by a steady 20-year decline.
That is not to say that the container revolution is unimportant; after all, estimates
in this paper show that increasing the share of trade that is containerized lowers
shipping costs from 3 to 13 percent. However, these savings were trumped in the
1970s by sharp increases in fuel and port costs. Indeed, ocean freight costs in recent
years have again begun to increase with the cost of crude, and port congestion has
become an especially severe problem in those countries with rapidly growing trade
volumes (Bajpai, Carruthers, and Hummels, 2003).

Transportation costs and international trade in the second era of globalization. Published 2007.

In the news, from 2 days ago: Dry bulk, container shipping rates firm up

Bunker fuel constitutes up to 65% of the total operational cost. Shipping freight rates, which plunged during the global economic crisis in 2008, have been improving since while grappling issues of demand and overcapacity.

There's a surprising dearth of relevant articles on 'bunker fuel shipping cost' in Google Scholar; the 1st paper quoted here seems to be about it. Maybe Hamilton or Rubin have delved into this in further depth. You're right of course that as long as shipping costs can be born by consumers via price markup it will still be profitable to build in developing nations and ship to OECD etc. The news item mentions world's #1 shipper AP Moller-Maersk Group stating that they expect growth in containers regardless of high fuel costs, after all.

Stuart Staniford said...

Mythodrome, Draft, etc: I'll try to write further, with data, on this, but very roughly my working hypothesis is that given peak oil taking the form of a longish bumpy plateau, the solution the global system may find to continued operation is to displace US and, to a lesser extent other OECD, workers driving big cars with Asian workers on scooters, electric hybrid bikes, etc. This will save the necessary fuel (as the western workers will be home unemployed playing video games instead of working and driving around) while allowing the system to avoid making fundamental change (I think large systems always resist making big sudden changes if there's any possible way to avoid it).

Michael Cain said...

Don't forget the impacts of automation on manufacturing employment. Over the course of the 20th century, employment in agriculture declined from 40% of the workforce to 2%. In the early years the decline looked linear, but eventually tailed off.

From Henry Ford's original assembly line to the 1960s or so, the hours of human labor required to assemble a car went from 100 to about 20. More recently, computer control has made it possible to automate an wide range of tasks, eg, precision machining of complex forms. If transportation costs bring manufacturing back to the US, it is not clear how many jobs will be generated.

Nick G said...

I'm baffled that Krugman is being so sloppy - it really misleads his readers to not cover the key factor: manufacturing labor productivity.

The US manufactures 50% more now than it did in 1978. People are misled by the fact that US manufacturing employment has dropped substantially in that period. But, that was caused by sharply rising manufacturing labor productivity, rather than by a decline in absolute levels of manufacturing output. See nice charts at and data at, including , Historic Timeseries - SIC (1958-2001), "Shipments"

What happened in the 90's was that US domestic manufacturing stopped growing, not that it declined. Flat production meant that rapidly growing labor productivity reduced manufacturing employment equally quickly.

It would be very hard for US manufacturing to grow faster than labor productivity, which tends to grow 3-5% per year. So, the best we can hope for is flat employment levels. That, of course, would be a relief for US workers in manufacturing.

Harish Kumar said...

I am just wondering guys, given the advances in computer technology and robotics(chips and software), is it possible to make a car without any human interference at all? Given the level of technological advances in computer vision (which we see everyday on TV as cruise missiles find their target unerringly), I am surprised that we still have human workers in many manufacturing facilities. What has prevented our workforce from being completely replaced by robots? Given todays technology, robots should be able to build more robots, in fact given a problem to tbe solved, a robot should be able to design another robot (number of manipulative units (arms), level of processing power(brains), etc.)

At least given a blueprint, one robot should be able to manufacture another.

I am pretty amazed to see so many human workers in factories and mines.

Has there been a deliberate policy to keep humans working?