Scott Sumner  

Fracking boom won't create many manufacturing jobs

A British Perspective on Ameri... A Few Dangers of Heroin Prohib...

The Wall Street Journal has an article entitled; "Gas Boom Rejuvenates Manufacturing." There certainly are some manufacturing sectors that will be helped by the energy boom. However people shouldn't expect too much from this development.

In a recent post I pointed out that population growth in Texas had slowed since the onset of the fracking boom. One problem is that as more engineers and skilled workers are drawn into fracking, fewer are available for other industrial jobs. Thus fewer manufacturing firms will move to Texas. The WSJ article expresses a similar concern:

In an about-face, the U.S. is drawing foreign manufacturing investments, Mr. Witte said. Inexpensive gas is luring Canada's Methanex Corp. MX.T -2.69% to pack up its one-million-ton-a-year methanol plant in Chile and move it to Louisiana at a cost of $550 million. But all that building could cause construction costs to balloon as companies compete for a limited supply of labor and materials, particularly in Gulf Coast states, according to IHS.
The lesson here is that the rules of classical economics (opportunity cost) still apply in a depressed economy, just not as tightly. It is possible (though not certain) that fracking will lead to more manufacturing jobs. But any gross estimates of the gain in energy-intensive manufacturing sector jobs will wildly overstate the net impact on all manufacturing. The overall impact is likely to be so small that it doesn't show up in the aggregate data. As our comparative advantage shifts, we'll have more jobs in some industrial sectors and less in others.

Machines are much more easily adapted to producing things (agriculture, mining, manufacturing) than to producing services. Thus the relentless move toward a lower and lower share of all jobs being in agriculture and industry will continue. Not just in the US, but in Germany, China, and elsewhere.

Long term I don't see machines replacing service jobs (in net terms.) Rather I see machines adding more and more consumption to the workplace. As when people surf the internet at work.

Comments and Sharing

COMMENTS (10 to date)
Tom m writes:

In the short term, if tend to agree, but with more higher paying engineer jobs the shortage should be corrected in the mid-long term. Unless you believe fracking is a short term boon and won't be long lived?

Andrew_FL writes:

You're right, it doesn't really show up in the aggregate data. Manufacturing as a fraction of overall employment is lower than it was at the nadir of the recession. And growth in the absolute level of manufacturing employment was slower in 2013 than in 2012 or 2011.

It's true that both are rising, but it's to be expected that most sectors see absolute gains in a recovery-however little actual recovery there is. But their relative gains will vary a great deal. Some may see very small gains or no gains at all.

I've never understood the fetishization of manufacturing employment.

Rajat writes:

Scott, I don't quite follow the penultimate sentence. I realise people use the internet for personal purposes at work, but I don't see more generally how the rise of machines and technology will increase private consumption at work. Do you mean machines will increase the blur between work and home? The times of the day and week when professional service workers work are becoming more and more diverse thanks to technology. Or do you mean that machines will make the proportion of 'busy times' (ie when service workers such as retail and hospitality workers must be fully engaged in their work) less frequent, thereby enabling them to consume privately for more of the time? For example, do you mean that machines do most of the work, and humans just hang around like baby-sitters and act only when something goes wrong with the machine?

Scott Sumner writes:

Tom, Focus on the concept of comparative advantage. As our advantage in some sectors increases, we will lose comparative advantage in other industrial sectors.

Rajat. Both.

Andrew_FL writes:

Having thought about it for a while, I don't think I agree with the statement that machines won't replace net service jobs. But only because I suspect in this case I would use long term to mean centuries and I suspect you are thinking decades.

When it comes to technological development, there's the long view, and then there is the loooooong view.

Michael Byrnes writes:

@ Andrew:

I think Scott is thinking from a macro perspective, not a micro perspective.

It seems a virtual certainty that, on a micro level, there are a great many service jobs that will be replaced by machines.

But does this mean that less work will be done by people because it can be done by machines? Or does it mean that people will find other types of work since machiens will have freed them from the need to do the jobs that were done before?

Keynes, in "Economic Possibilities for Our Grandchildren" postulated that the most people would work only a couple of days per week, thanks to improvements in productivity. Obviously he was wrong - the productivity gains simply allowed more and different types of work to be done. My "service job" didn't exist in Keynes' time - and I'm sure that is true for many of us.

Although, Scott's remark about surfing the internet at work sort of suggests that Keynes was maybe a bit more correct than in generally appreciated. :)

Mark Bahner writes:
Having thought about it for a while, I don't think I agree with the statement that machines won't replace net service jobs. But only because I suspect in this case I would use long term to mean centuries and I suspect you are thinking decades.

I found the 15 most common jobs in the U.S. To me, it appeared that at least 12 of them were vulnerable to computers/robots in the next 10-30 years.

We're all doomed

RPLong writes:

I think the robot worker thing is overblown, for a lot of reasons. But one of the most important reasons is the fact that electricity costs a lot of money. There is an upper limit on the kinds of jobs robots can replace even setting aside engineering constraints. It may be physically possible to run the universe with robot labor, but as our energy use grows, so too grows the price tag of automation. Worth keeping in mind.

Regarding Prof. Sumner's post, I have no real objections, but I think it's worth adding that there has been significant tech-sector growth in Texas. I think Sumner is correct about the fracking/manufacturing, but the piece being left out is the extent to which the fracking boom has fed a peripheral tech sector boom in TX. I'm not sure how that all ultimately plays out, but it's a factor worth remembering.

Mark Bahner writes:


Sorry to be a bit (or more) off-topic, but RPLong writes, "I think the robot worker thing is overblown, for a lot of reasons. But one of the most important reasons is the fact that electricity costs a lot of money."

Here are the top 4 job categories in the U.S., according to the link I provided:

1. Retail Salespeople, 4.2m

2. Cashiers, 3.4m

3 Office Clerks, 2.8m

4 Fast food prep and service, 2.7m

None of those are very active jobs, in the sense of running around or using great strength. Let's say one could run the computer brain at 1 kilowatt, and the body at another kilowatt. That's only 2 kilowatts. At 12 cents a kilowatt-hour (about the nationwide average), that's only 24 cents an hour for electricity. We can already see cashiers being at least partially replaced, with stores that have self-checkout. To me, it's very easy to view some near-term version of Siri taking orders at a fast-food restaurant...with Roombas cleaning the floors and tables. Retail salespeople...some sort of rolling thing with a pleasant voice and voice recognition. A lot of jobs could be taken by robots/computers with very little electrical consumption.

Vangel writes:

I think that the much bigger problem with fracking is the destruction of capital that goes with the shale boom. If anyone bothers to look it is easy to see in the 10-Ks that shale gas and oil are not economic outside of a few small core areas. While there are some wonderful, highly profitable wells in those areas the average well is uneconomic.

The best production for shale wells comes in the first few days of production and the best year for production is the first. The way the companies tell the story this decline rate levels off and the wells produce for more than two decades. This means that high levels of drilling activity will increase production rates AND will be able to keep the average well production growing for at least five years.

But the production data tells a different story. The decline rates for a typical well is much steeper than the estimates used to create the depreciation schedules and the aggregate production level is not growing fast enough to prevent the average production rate from falling. If we look to the ND data, we see that in January 2012 there were 3390 producing wells that had an average rate of 142 barrels per day. The latest data is for January 2014. It shows that the number of wells has more than doubled to 6926 but the average production rate is now down to 126 barrels per day.

While I may not be as good at math as all those analysts who keep telling me that shale is the saviour of the US economy and one of the strongest supporters of the USD I know more than enough to figure out that there is no model that can explain the data if the production assumptions are valid. For the record boys and girls we have seen this story before. In the 1990s many of the equipment manufacturers were busy building new facilities to make products in a rapidly changing environment that made those facilities essentially worthless within two or three years of completion. While honest accounting would have required that the value of those facilities be written off fully the accountants used the slack offered in the rules to create depreciation schedules that assumed value where there was none. Exactly the same thing is going on in the shale space right now. Some of the majors have already written off their acquisitions while some of the smaller players have sold off pieces of themselves to close funding gaps that are mentioned in most of the conference calls time after time without enquiry from the analysts.

Eventually this story will end just as the previous ones did. But when it does I hope that we do not get the story about how we were lied to and how none of the problems were disclosed. Everything we need is in front of us and tells us what reality looks like. The fact that we refuse to see it tells us more about us than about the people behind this scam.

Comments for this entry have been closed
Return to top