Will the Job Outlook be Great Again? (Part 2: Energy Extraction)

In my previous blog, I discussed the likely impact of the Trump administration on manufacturing jobs and concluded that he is unlikely to deliver on his promise of a renaissance in low-skill manufacturing jobs. Although in his campaign Trump gave the greatest emphasis to manufacturing, he also assured voters that he would boost employment in other industrial sectors. This blog is about another sector.

Energy Extraction. The government has considerable powers to regulate the energy-extractive industries: coal mining and petroleum. The majority of lawmakers in Donald Trump’s party have been saying for some time that global warming either does not exist or is not caused by human activities, and in doing so they have positioned themselves as champions of this industry sector. Although at times Trump has made vague concessions that there may be human causes, most of the time he hews to the more extreme version of the party line. He even tweeted that human causation is a myth concocted by the Chinese, although later he denied having said so.

 For both Trump and the Republican Congressional majority, two main policy recommendations result from this denial of the scientific evidence: (1) increased oil drilling and coal mining (“Drill, baby, drill!”); and (2) reduction or elimination of limits placed on carbon-releasing industrial activity (“Burn, baby, burn!”). Another result is encouragement of oil-pipeline construction, which has frequently been blocked by environmentalists.

Among the existing laws, regulations, and treaties that bind the United States to reducing carbon output, some will easier than others Trump and Congress to reverse. Trump has said he will “cancel” the Paris agreement, in which nations have promised to cut their output of greenhouse gases, and he can back out of this agreement by one of several ways. On the other hand, it will be difficult for him to roll back the regulations that make up the EPA’s Clean Power Plan and the fuel-efficiency standards for cars and trucks. To do so, he will need to propose alternative regulations and will face litigation from environmentalist groups that could delay or even block any changes. Any changes Congress wants to make will face possible filibustering from Democrats. If Trump and the Republican Congress succeed at rolling back regulations, it is possible that states will take a more active role. For example, California already has more stringent standards for vehicle emissions than the nation as a whole, and other states can adopt the California standards.

Market forces are already promoting the transition from dirty energy sources to cleaner sources. As cheaper natural gas has become available, power plants have been switching from coal to gas. The Energy Department found this transition the chief reason why carbon dioxide emissions in the first half of 2016 reached the lowest levels since 1991. During that period, in which the weather was relatively mild, consumption of coal fell by 18%, while consumption of natural gas fell by only 1%. Energy from renewable sources increased by 9% during that same period. Meanwhile, the cost of alternative-energy sources has continued to fall, and energy-saving consumer goods, including hybrid and electric cars, have become more affordable. If you haven’t shopped for LED light bulbs lately, you may be surprised at how cheap they have become.

Even if the government managed to walk away from the anti-carbon policies of the past few years, it is questionable whether market forces would allow the petroleum-extraction industry to expand and create jobs on a large scale. Employment in this sector was at 538,000 in October 2014 but is now at about 175,000 workers. Anti-carbon policies had very little to do with this slump. Instead, the main culprit was the glut of capacity that was created by America’s widespread adoption of fracking technologies, together with the development of new oilfields in other countries, such as Brazil. Opening up new areas (such as national parks and sensitive offshore ecosystems) for drilling will not produce a cornucopia of jobs as long as the glut of cheap oil and gas continues to make it uneconomical to set up new drilling rigs and pipelines.

Coal-mining jobs face a similar hurdle. Employment in 2015 reached the lowest levels since the Energy Information Administration began collecting data in 1978, and the number of mines decreased in all three major coal-producing regions.  The EIA reports (PDF) that between 2014 and 2015, “the average total number of employees at underground mines and surface mines  declined  by  13.6%  and  9.3%,  respectively.” More significantly, “the   average production per employee-hour increased by 5.4% to 6.3 short tons per employee hour.” Like the manufacturing sector, coal mining has increased its use of automation and other technologies for extracting coal more easily.  Coal is now extracted mainly by excavation from the surface (including mountaintop removal), using colossal machines operated by relatively few workers. In 2015, all of the top 10 coal-producing operations were surface extractions. The industry is vastly different from the labor-intensive days when Loretta Lynn’s father went to work with a pick and shovel down in a tunnel .

So the policies advocated by Trump (most of the time) and by the Republicans in Congress are unlikely to create a boom in employment in the energy-extractive industries. That’s not to mention the damage that these policies will do to the alternative-energy industries, to the extent that solar and wind power are still being subsidized.

There is one more wild card that may influence the energy-extraction policies of the new administration: Trump’s relationship with Russia. Oil and natural gas are Russia’s major exports, accounting for 68% of total export revenues in 2013. For reasons that are not entirely clear, Trump has made good relations with Russia a high priority. So although removing the barriers to petroleum output in the United States might create a limited number of jobs here, it would depress worldwide prices for petroleum and thus inflict a blow to Russia’s economy. Perhaps Congress will be less eager than President Trump to prop up Vladimir Putin.

Of course, unemployment in the oil patch and especially in the coal belt is no trivial matter that either party can afford to ignore. But the answer is not to turn back the clock on energy policy as if the energy markets and extractive technologies have not changed since the 1970s energy crises. Wayne Gretzky used to say that you skate to where the puck is going, not to where it’s been. As was true for manufacturing, the answer is to retrain displaced workers and train young people for the changed economy. But, as I pointed out in the previous blog, Trump and the Republicans are not advocating policies that will expand the availability of low-cost, high-skill training.

Will the Job Outlook be Great Again? (Part 1)

As a new presidential administration approaches, you may be wondering about the impact that the change in leadership will have on job prospects in the United States. I don’t claim to have a crystal ball, but I believe I can glean useful insights from what economists are saying and from past employment trends. And, in fact, I have a record of forecasting trends brought in by a previous sea change.

Eight years ago, when a major recession was looming and I was working for JIST Publishing, my editor—Susan Pines—assigned me to write a book that eventually was called 150 Best Recession-Proof Jobs. The book came out just as the full force of the Great Recession came crashing down on the U.S. economy. Thanks to this good timing, the book was so newsworthy that I was being interviewed on television approximately once a week for a month and more. (A lot of credit goes to JIST’s crackerjack publicist at the time, Selena Dehne.)

I selected the occupations that I featured as “recession-proof” by mathematically comparing the past ups and downs in the gross domestic product with the ups and downs in the workforce size of each occupation. Thus I was able to identify occupations that were least impacted by past downturns. The main limitation of this approach is that each recession is different from previous recessions. For example, the Great Recession was caused by a sudden drop in the value of real estate after a long bubble of overbuilding, so certain industries related to real estate—especially construction—suffered more than they did in previous downturns. Nevertheless, over the following several years my predictions were more often confirmed than disconfirmed.

The pending change in administrations presents a similar problem for those who would forecast employment trends. It might seem sensible to look at the employment effects of past transitions when a Democrat was succeeded by a Republican. However, each such transition, like each recession, is different in many ways. The year 2017 can’t be expected to repeat 2001 precisely. For example, while the last part of Barack Obama’s administration has seen a long streak of growth, it is not comparable to the technology bubble that ended Bill Clinton’s second term. And Donald J. Trump is not as conventional a Republican as George W. Bush.

In fact, it’s difficult to tell what kind of Republican—indeed, what kind of politician—Trump is. His policy pronouncements tend to lack specifics and frequently change depending on his audience. Some of the policy goals he has stated at various times run counter to the stated goals of Congressional leaders of his own party, raising the question of who will prevail if he tries to bend Congress to his will.

Because of these uncertainties, I am basing my forecasts on Trump’s general goals, plus an analysis of possible policies that he might or might not employ in pursuit of these goals. And I am organizing my forecasts by industries. In this blog, Part 1, I discuss only manufacturing.

Manufacturing Jobs. One of Trump’s most consistent messages has been his desire to bring manufacturing jobs back to the United States. What’s often overlooked in his rhetoric is that manufacturing output in the United States is now at nearly an all-time high. The kind of manufacturing that is now done here uses highly automated processes and employs a comparatively small workforce of highly skilled technicians. (Since recovering from the Great Recession by 2015, the number of employees has leveled off at about 12.3 million.) Most low-skill manufacturing tasks are performed either here by robots or overseas by low-paid workers, such as the ones who are making the garments branded with Trump’s name and his daughter’s.

That leaves the next president with these policy choices:

·         He might impose high tariffs that make it uneconomical for offshore manufacturers (whether American-owned or otherwise) to export to the U.S.  market—a policy that Trump has explicitly endorsed at times. But the policy could backfire. To the extent that imposing or raising tariffs is feasible under existing trade agreements, it would encourage other nations to retaliate with their own tariffs, reducing exports of our own manufactured goods and increasing the costs of the supply chain—imported parts and raw materials that U.S. manufacturers use. Also, once American manufacturers are protected from foreign competition and start hiring low-skill American workers, they will not be able to pay the rock-bottom wages that third-world workers earn. So, although many jobs will open in manufacturing, consumers will find American-made manufactured goods costing a lot more than the cheap foreign-made goods that now fill the shelves at Walmart. In summary, this policy would hurt high-skill manufacturing workers and create an economy where the low-skill manufacturing workers might have no more purchasing power than is now possible from work in service industries. And, finally, almost all Republican lawmakers, as well as many Democrats, are committed to tariff-free trade (although Republican voters have become more hostile to it), so this policy stands little chance of getting through Congress.

·          

·         He might mandate that American-made products and components be used when federal funds are expended. He hinted at this policy at the third presidential debate with Hillary Clinton when he said that he had used Chinese steel in one of his building projects because Congress had done nothing to stop him. It seems unlikely, however, that Congress would go along with this policy. Congressional leadership has not changed since House Republicans defeated an amendment to impose just such a mandate on infrastructure projects for the nation’s waterways.

·          

·         He might offer tax breaks to American manufacturers so that their operations here are more profitable. Trump actually tried this approach in October when he persuaded Indiana to give tax breaks to the Carrier division of United Technologies as a way of preventing jobs from being shifted to Mexico. However, Carrier has indicated that it intends to use the resulting savings to increase the company’s use of automation—thus subsequently cutting more low-skill manufacturing jobs. Cutting labor costs produces savings that are more dependable than easily-reversible tax breaks. So tax cuts seem unlikely to be an effective solution.

·          

·         He might reduce federal regulations on manufacturers that add to the costs of doing business in the United States. (I am not saying I favor this policy, but it is one that is sometimes proposed.) For example, it has been argued that the main reason Carrier wanted to relocate jobs to Mexico was regulations, not wages, and Trump is in agreement with Republican leadership in Congress on the need to reduce regulations on industry. It is not clear that Trump can get enough cooperation from Congress to make sufficient reductions in regulations to bring a true renaissance of manufacturing. And for some regulations (e.g., on the formaldehyde levels in plywood), the way to reduce competition by offshore manufacturers is to apply the same regulations to imports that are imposed on goods manufactured here, rather than rolling back regulations.

·          

·         He might accept the reality that low-skill manufacturing jobs are gone for good and instead focus on preparing (or retraining) workers for high-skill jobs in the industry. One way to accomplish this is to make community college as free of charge as high school. This is what President Obama proposed and that became the America’s College Promise Act of 2015, but Congress sent the bill to die in committee. I have been unable to find any statement from Trump himself about this proposal, but Trump’s campaign co-chair Sam Clovis, in an article in Inside Higher Education, stated that the campaign rejected the call for free community college.

·          

·         A related policy to encourage manufacturing would be to find a way to increase the number of manufacturing engineers working here. According to Walter Isaacson’s biography of Steve Jobs, the Apple  CEO once told President Obama that Apple’s suppliers in China are able to employ 700,000 factory workers because they have ”30,000 engineers on-site to support those workers. ‘You can’t find that many in America to hire.’” The skill level that Jobs was referring to seems to have been closer to engineering technicians than to what we normally think of as engineers. So, again, increased funding of community colleges and technology schools would be a way to achieve this goal, but that was not a priority of the Trump campaign. Alternatively (or additionally), a reformed immigration system might welcome engineers and engineering technicians trained abroad or coming from abroad for training here. Trump has both welcomed and rejected immigration of high-skill workers at various times. Congressional leadership has generally favored it despite the reluctance of many Republicans and therefore has sometimes had to find underhanded ways to encourage it. Therefore, it is difficult to predict whether or not this policy will be pursued in the coming administration.

·          

·         Yet another way to create a high-skilled manufacturing workforce is to encourage unions to partner with manufacturers on forming apprenticeship programs. Although Trump starred in a TV series called “The Apprentice,” I have not been able to find any statements from his campaign about his attitude toward actual apprenticeships, with the meaningless exception of a tweet he issued in October: “Did Hillary just say she wants more Apprenticeships? I created The Apprentice!” Judging by Trump’s past record as an employer, his attacks on an Indiana union leader, and his appointment of a foe of unions for Labor Secretary, he seems no friend of unions, and Congress certainly has not encouraged them, so this last outcome seems unlikely.

·          

In conclusion, then, I do not expect the Trump administration to deliver on its promise of a renaissance in low-skill manufacturing jobs. I expect employment in these jobs to remain at its current low level—or worsen if there is an economic downturn.

Will a Drone and a Robot Take Your Job?

In the past, when talking about the use of robots to replace human workers, I have often given the example of ground transportation at the airport. To get from Terminal A to Terminal B at many airports, you take a robot-controlled trolley. No human judgments are needed to navigate the rails, make the stops, and open and close the doors. However, to get from the airport to your hotel, you take a shuttle driven by a human, because a robot cannot make the many judgments that are required to navigate through traffic out on the streets.

This example used to be a way I would indicate that some types of jobs may never be replaced by robots. But recently I am using this example to illustrate how robots may soon be extending their reach. Google has been experimenting with robot-driven cars for several years and has already logged hundreds of thousands of accident-free miles. The self-driven cars use GPS to understand their route and can consult a database of information to learn about speed limits and other considerations that we human drivers learn from signage. They avoid accidents with other cars or careless pedestrians by means of the same radar technology that is now being offered as an accessory in human-driven cars. Google’s technology is still experimental, but in a few years we may see it being used in airport shuttles, probably beginning with trips that involve the fewest variables, such as to and from the airport’s rent-a-car lot. I suppose the robot shuttle vans will also need to provide some mechanism that lifts heavy suitcases in and out of rear storage. And you won’t need to tip the robots.

When will these robot drivers take over? First, jurisdictions will need to change traffic laws that do not presently allow driverless vehicles on the roads, and you can expect some pushback from the Teamsters Union and other representatives of the people who earn their living by driving. Secondly, the cost of the technology will need to come down to the point where companies that deploy fleets of cars and trucks will save money by switching to robots. Besides saving on wages and benefits, fleet owners may realize savings if robot drivers prove to be safer than human drivers, as preliminary data indicates. It may take many years before all of these stars align, so human drivers can probably expect at least a decade’s reprieve.

The outlook changes, however, when you look at occupations with a shortage of human workers. There are lots of people who are qualified to drive airport vans. As far as I can tell, most states do not require a special license for the drivers, although employers look for a clean driving record. A modest level of fitness is necessary to handle passengers’ luggage, and the driver must speak English well enough to understand passengers’ destinations. But millions of Americans have these qualifications, so it is not hard to find workers to fill these jobs.

Long-distance truck driving requires a higher level of skill, and there currently is a shortage of qualified drivers. However, the higher skill requirements, which are reflected in the special licensure needed for this work, also mean that robots will probably take longer to make inroads into this occupation.

Japan furnishes a fine example of how a shortage of human workers can accelerate the adoption of robots. You may have already read about how Japan is using robots to perform certain routine health-care tasks, such as moving a patient from a bed to a wheelchair. Japan’s aging population means there is a growing number of elderly patients and a diminishing number of health-care workers with the physical strength needed to do the work. This provides the opening for robots.

Japan also has a shortage of workers who can drive heavy construction vehicles, probably also largely because of the physical demands of the work. The Komatsu company is planning to fill this employment gap by using self-driven bulldozers and excavators. Unlike long-haul trucks or even airport shuttles, construction vehicles function in a closed location and don’t have to deal with traffic or random pedestrians.

One thing that is particularly intriguing about Komatsu’s plan is that it also involves another new technology: drones. At a construction site, drones made by the San Francisco company Skycatch will survey the terrain from above, and the mapping data the drones gather on the actual lay of the land will be compared to a computerized map of how the site is meant to be shaped. The self-driving construction vehicles will then move earth as needed to achieve the desired result; their work will be periodically monitored by the drones.

Note that this arrangement displaces not only heavy-vehicle operators, but also surveyors. The Komatsu manager overseeing this project notes that the old way of surveying a site typically required a week’s work by two people, whereas the drones can acquire the data in only an hour or two.

Understand that this kind of construction will require some highly skilled human operators to program the machines, monitor their progress, and sometimes jump in to take control of a machine as needed. So consider this an example of how yet one more industry, construction, is seeing a trend toward eliminating many low-skill jobs and creating a smaller number of high-skill jobs. I have often said that construction jobs can’t be offshored, but the other trend eroding jobs—automation—is about to take its toll.

UPDATE.: Drones strike again: An Israeli company is marketing a self-piloted drone that reads water meters remotely. Also, a Dutch student has prototyped a drone that delivers a defibrillator to a heart-attack victim much faster than an ambulance could. Such drones presumably could also deliver other medical supplies needed in an emergency, plus a webcam to allow on-the-spot diagnosis that would enable helpful bystanders to be coached and thus provide better-informed first aid. Such drones certainly would not replace the need for EMTs, but they might mean that fewer EMTs would be needed to cover a geographic area because proximity would no longer be quite as important.

When Robots Create a Need for Human Workers

Yesterday a Facebook friend of mine wrote about an “eerie and creepy” experience. An advertisement had just appeared in the margin of the Facebook page she was reading, showing a tee-shirt with the lettering, “Just a California Girl in a New Jersey World.” It is no coincidence that my friend is a native of California and now lives in Flemington, New Jersey. Facebook obviously sold this user-profile information to advertisers, one of whom found a way to develop a highly personalized product.

But my friend did not buy the tee-shirt, and if enough other targets of this pitch find their personalized tee-shirt unappealing, the manufacturer will lose money on the ad campaign (plus the costs of tooling up to produce the tee-shirts; the manufacturer surely doesn’t have an inventory of shirts for every possible two-state match-up). What this tells me is that although Facebook’s data-gathering (a robotic function) makes it possible to create highly personalized products, the actual creation of products that people will buy remains a human function. In the brave new world of Big Data, human creativity is still needed to make a sale.

Something very similar became clear in this month’s congressional elections. As The New York Times pointed out,

Modern political campaigns home in on their key voters with drone-like precision, down to the smallest niche — like Prius-driving single women in Northern Virginia who care about energy issues. They compile hundreds of pieces of data on individuals, from party registration to pet ownership to favorite TV shows. And they can reach people through Facebook, Pandora, Twitter, YouTube or cable television.

The only problem: They do not have enough messages for them all.

The Big Data era of politics has left some campaigns drowning in their own sophisticated advances. They simply cannot produce enough new, effective messages to keep up with the surgical targeting that the data and analytics now allow.

…Or, as Joe Rospars, the founder of the Democratic digital agency and technology firm Blue State Digital, put it, “The science is ahead of the art.” An analytics team can help a campaign make “a much more targeted buy,” he explained, but that alone will not offer a particularly efficient return on investment if the ad is still “just a white guy in a suit.”

As a result, the people who design the advertisements for electoral campaigns end up trying to tease out certain large slices of voters with something in common, such as “soccer moms” or “angry white males.” The campaigns do not have enough creative people to craft the highly personalized ads that should be possible given data-analysis tools.

On a recent broadcast of NPR’s “On the Media” (sorry, I can’t remember which date), I found another example of the limitations of technology. You may remember that YouTube moved quickly to take down the videos of the recent beheadings in Syria. You may not know that YouTube takes down many other videos because of pornographic or sadistic content, and so do most photo-posting sites, such as Flickr and Photobucket. Computer technology makes it easy to post photos and videos, and to make them searchable by keywords, but the explosion of content that has appeared on the Web requires human eyes to decide which photos and videos violate site policies.

One indication of how subtle the human decision-making must be is the fact that photo sites prefer to offshore this work to the Philippines rather than to India, where the workers are lower-paid. The reason is that Filipino workers have a better understanding of American culture and therefore can decide whether (for example) a shot of someone in a bikini is too revealing, whereas an Indian worker might reject every bikini shot.

The takeaway is that technology sometimes creates a need for more human workers, and not just those employed in creating, manufacturing, or repairing the technology. It sometimes creates opportunities for work that requires great creativity or subtle judgments. Robots may be driving cars, but they are not yet metaphorically in the driver’s seat.

Complacency About an Inadequate Recovery

We have officially emerged from the economic downturn that began at the end of 2007, but for many workers, it really doesn’t feel that way. Yes, employment has finally returned to prerecession levels, but many of the jobs that have come back are inferior to the jobs that were lost.

A report by the National Employment Law Project (PDF) finds that “lower-wage industries accounted for 22 percent of job losses during the recession but 44 percent of employment growth over the past four years. Today, lower-wage industries employ 1.85 million more workers than at the start of the recession.” The report finds the opposite is true for mid-wage and higher-wage industries, with greater losses and smaller gains.

Yet it’s surprising how often I encounter cavalier attitudes toward the economic plight of working Americans.

Sunday’s Review section of The New York Times featured an op ed piece called “Fear Not the Coming of the Robots,” by the investment advisor Steven Rattner. The article features a nifty graph that you really must view to appreciate. It shows one column with bars representing occupations that used to employ large numbers of workers but have been decimated by automation. These occupations include word processors, telephone operators, computer operators, proofreaders, travel agents, and switchboard operators, among others. For each occupation, the graph shows not only the number of jobs lost but also the median wage, which tends to be in the thirties. Bars in a second column represent occupations that have gained workers, including computer systems managers, physical therapists, financial analysts, registered nurses, financial managers, and accountants. The wages for these occupations are much higher, with the lowest (for accountants) at $63,550.

Rattner concedes that technology has changed the nature of work, making advanced training increasingly necessary, and resulting in a sharp rise in income inequality. But he argues that the main culprit is globalization, not technology, “particularly the ability of companies to substitute far less expensive and increasingly skilled labor in developing countries.”

So what is Rattner’s proposed solution? “To address these very real challenges, we should be embracing technology, not fearing it. That means educating and training Americans to perform the more skilled jobs that cannot yet be performed by workers in developing countries.” I’m all for that, too, but where is the funding for this education and training going to come from? Not from a Congress that is incapable of passing legislation that invests in human capital. And how much more can young people and displaced workers mortgage off their futures for tuition loans?

There’s also the problem of those people who lack the ability to learn advanced skills even if somehow a training program were available. Rattner breezily comments, “Of course, not every worker can be retrained, and so we must help those who aren’t suitable for the new jobs through more robust social welfare programs.” Do you see any indication that Congress is about to beef up support for jobless people? I see the exact opposite. Even the idea of creating jobs through a program such as the Civilian Conservation Corps is a nonstarter in the current political climate.

So, yes, I do fear the one-two punch of robots and global competition. Rattner is like the captain of a ship who tells the passengers, “Don’t worry that the ship is sinking, because you can go merrily sailing in the lifeboats”—but the lifeboats aren’t there.

Recent Trends in Work Conditions

Money isn't everything. When choosing a career, most people also consider the work conditions. Will the job be mainly indoors or outdoors? Will it involve a lot of time pressure or high-impact decisions that add stress? It's nice to be able to choose work conditions that agree with you, but sometimes the jobs that are available do not offer exactly what you want, so you have to make compromises if you want to earn a paycheck.

As the economy changes over time, there are changes in the salience of various work conditions that people deal with on the job. I decided to identify recent trends in work conditions by examining recent data about occupations and about the changing levels of employment in occupations. Which work conditions are becoming more prevalent and which are receding?

Here is the procedure that I used to track several work conditions:

• I used workforce estimates for the years 2006 through 2012 from the Occupational Employment Statistics survey of the BLS. Note that this covers most but not all workers; self-employed workers are not included.
• For each occupation, I multiplied the workforce size at each year by the numerical ratings for various work conditions in the O*NET database (where they are called Work Context elements).
• For each year and for each work condition, I summed the products for all occupations and summed the workforce sizes of all occupations.
• For each year and for each work condition, I divided the sum of products by the sum of workforces to get an overall quotient that indicated the level at which that work condition was significant for the nation’s workforce during that year.
• Finally, I graphed the changing levels for each work condition.

Note that each chart below uses a different vertical scale, and that this scale is selected to emphasize the vertical movement of the curve. If I had plotted all the curves on a single chart, the vertical movement of some of the curves would not have been as evident. I'm more interested in the direction of movement than in the absolute value of the movement.

The first two charts show the trends for outdoors and indoors work. (Actually, in O*NET, the specific work conditions I used were "Outdoors, Exposed to Weather" and "Indoors, Environmentally Controlled.") As you might expect, the two curves are almost mirror images, and the general trend is away from outdoors work (although there has been a leveling-off in the most recent years). One factor contributing to that trend is the increasing mechanization of agriculture. Another factor is the decline in construction work that followed the collapse of the housing bubble.

You may wonder why the two preceding curves are not exact mirror images--why indoors work falls off very slightly during the recovery even though outdoors work does not pick up by a comparable amount. I found that the slight dip in indoors was offset by slight increases in work done in interior environments without climate controls (such as the jobs done by some warehouse workers and mechanics) and in enclosed vehicles (such as the jobs done by transportation workers and many sales workers). These mini-trends reflect the job growth that happened following the Great Recession, in which office jobs did not return as fast as some skilled blue-collar jobs.

You can see more evidence of this mini-trend in the following chart, which shows the trends in sedentary jobs.This chart shows a curve almost identical to the curve for indoors work.

The trend in what O*NET calls "Degree of Automation" may surprise you by its downward slope. Isn't automation constantly increasing? Understand that the curve represents not the amount of automation being used but rather the number of workers whose jobs involve a lot of automation. Most employees in manufacturing have worked with automation for a long time, so when automation reaches the point that it eliminates a manufacturing job, that's one less automation-related job. An additional drain on automation-related jobs was occurring in the years 2006 through 2009, when many highly automated jobs were being shifted to plants in foreign countries.

On the other hand, you'll note that the slope has pretty much leveled off in the most recent years. I interpret this to mean that automation is not a threat as much as it is the new normal. It's getting harder and harder to find new ways to automate jobs, and the great bulk of automation-using jobs that can be shipped overseas already have been offshored. Nowadays we're actually seeing an increase in advanced manufacturing jobs.

If automation is the new normal, what does this mean for you as a worker? It means that you need a high level of skills, either for doing the things that automation can't do (making sophisticated judgments or using people skills) or for doing the technological work of creating and programming automated equipment. And this is true no matter whether your work is indoors or outdoors, standing or sitting.

Do Robots Dream of Electronic Paychecks?

This week, I’m going to continue my year-end tradition of looking back at the tweets I’ve issued on Twitter in order to follow up on some and comment on the significance of others. (If you’re not already following me, my handle is @LaurenceShatkin.) But this week, I’m going to focus on a single trend that I’ve tweeted about numerous times: the growing use of robots.

In fact, if it were up to me to choose Time’s Person of the Year, I’d choose the robot. For example, this year I learned about a new trend in warehouses: using robots to pull merchandise off the shelves and bring it to the loading dock. Human forklift operators need not apply for these jobs. Robots are also threatening some jobs in the green career field of solar panel installation. Part of the reason this trend is accelerating is that technology is creating a generation of robots that are more deft and therefore more capable of doing work that until now has been done by hand. (“Fine motor control” takes on a different meaning when it refers to actual motors.) For example, a Philips Electronics factory in Holland is using robots to do the same delicate work—assembling electric shavers—that humans are doing in a Philips factory located in the Chinese city of Zhuhai. One of the biggest manufacturers in China, Foxconn (which assembles iPhones, among many other products), has announced plans to add up to one million industrial robots to its assembly lines inside of three years.

But despite these improvements, robots still can’t think as well as human workers do, and it seems unlikely that they will soon be able to make complex decisions or recognize nonroutine visual patterns as well as the human brain can. Job erosion in the United States has been slowed partly by this limitation on robots, combined with a limitation on foreign workers: much work in America needs to be done by on-site workers. But perhaps a perfect storm is coming that will overcome both of these obstacles: what one writer calls “the Avatar economy.”

If you’ve seen that movie, you saw how one of the characters remotely controlled a physical body, his avatar, to accomplish work that his disabled body could not do. In a way, we already have workers doing this when they control aerial drones for surveillance and for making kill shots in foreign countries. Why can’t that be done in reverse—by foreign workers—for civilian work tasks here? For example, some delicate surgery is now performed by robotic hands under the control of a physician who watches the progress of the operation through stereoscopic video cameras. I’ve seen this being done from across a room. Couldn’t this be done from across an ocean? In surgery, the major remaining barrier may be the few seconds of delay in response time, which could be critical in some situations. But for less life-and-death work tasks, it may become less expensive to hire a foreign worker to control an on-site robot.

It’s important to understand that not all robots are engaged in physical manipulation of objects. Robots are also making inroads into some American jobs that are concerned mainly with manipulating data. For example, “e-discovery” software is replacing lawyers and legal assistants by rapidly reading through reams of documents to find material that is relevant to a lawsuit or other trial. One California firm analyzed 1.5 million documents for the bargain-basement cost of less than $100,000. The software is sophisticated enough to go beyond keyword recognition and actually identify key concepts. You saw something similar happening if you watched the computer that bested the human opponent on “Jeopardy!” Less clever computers are working with numbers to make decisions that used to require human loan officers and tax accountants.

Machine translation has reached the point where you can easily ask Facebook to translate a posting made in a foreign language. One indication of where this is going was a demonstration this year by Microsoft researchers in which a computer translated someone’s spoken English into Mandarin with only a very short delay and, most remarkable of all, spoke the translation in a voice that sounded like that of the original speaker.

Fears about the Mayan apocalypse now seem rather quaint, and I’m not going to predict an imminent robotic apocalypse for American workers. But if you’re planning for a career with any longevity, you need to think about developing skills that robots will find difficult to master.

The Hollowing-Out of the Workforce

One of the most important trends affecting your job prospects is what is often called the “hollowing-out” of the workforce: High-skill and low-skill occupations continue to grow, but middle-skill occupations are shrinking. This means that if you want to live a middle-class lifestyle (or better), it is increasingly important for you to acquire and maintain a high level of skill. An average level of skill won’t bring decent rewards in the emerging labor market.

Here is a graph I created showing the change in workforce size for occupations at three levels of skill at three points in time, 2000, 2007, and 2011. At each point, you can see the change from the previous year in the percentage of the workforce.

What are the reasons for this trend? Blame automation and offshoring. Many medium-skill occupations consist of routine physical and decision-making tasks that can easily be programmed into a computer and/or robot. I can remember when tall buildings still employed elevator operators. Since then, we’ve seen automation replace human workers who used to route phone calls, take phone messages, drive airport trolleys from one concourse to the next, make travel reservations, format business letters, ring up retail sales, and weld automobile bodies, to name just a few work tasks.

Offshoring has eliminated other jobs consisting of routine physical and decision-making tasks or jobs where the physical proximity of the worker is not needed. The most obvious example is manufacturing. The next time you’re in a big-box store, try to find a manufactured item that was made in the United States. When you’ve had problems using your computer, you may also have encountered a foreign help-desk worker. The highly labor-intensive work of making animations from hand drawings--as opposed to making them from computer images, Pixar-style--is being farmed out to foreign shores. (The next time you watch “The Simpsons,” note the names in the ending credits.) Airplanes can easily be flown to foreign hangars for routine mechanical inspection and repair.

So what jobs resist automation and offshoring? Some involve physical effort that must be done on-site and that involve unpredictable physical settings and nonroutine tasks--unlike an automobile assembly line. Most construction jobs fit this description. So do automobile service jobs. With regard to the latter, note the bifurcation of the workforce into low-skill, low-wage workers, such as those who change your oil at Jiffy Lube, and high-skill, high-wage workers who do the more difficult diagnostic and repair tasks--and who require initial training that takes two to four years, followed by continual training updates as the technology changes. You can observe a similar bifurcation of wages and training requirements in health-care jobs, with home health aides and orderlies at the low end and technicians and professionals at the high end. It’s no accident that job prospects are not increasing as rapidly for licensed practical nurses and for registered nurses with only a two-year degree compared to registered nurses with a four-year degree. Even in our fastest-growing industry, health care, the middle-skill jobs are being hollowed out.

Another category of jobs that resist automation and offshoring are those that require high-level decision-making or people skills. Although the concourse-to-concourse airport trolley is now robotized, the airport van that threads its way through city traffic is still driven by a human (so far). Red-light cameras can detect one form of careless driving without human intervention, but we still need police officers for many other complex decisions about what is legal or illegal behavior. (They also are needed when lawbreakers must be physically arrested.) An automated questionnaire can elicit structured responses from an interviewee, but for employment interviews, businesses still want to use human workers, who can make judgments about which questions do and do not need follow-up and about the interviewee’s body language.

Work that is highly collaborative or creative often resists automation and offshoring. For example, routine computer programming can easily be offshored, but development of a new smartphone app will probably be done by American workers. Locations such as the Silicon Valley allow programmers, computer graphic artists, marketers, electronic engineers, and venture capitalists to come together rapidly for a collaborative project and to shoot ideas back and forth.

Jobs that require a knowledge of American language and culture also resist automation and offshoring. Many companies refuse to offshore their help desks to foreign countries (or at worst locate them in Canada) because customers complain about the difficulty of conversing with foreign workers. A story on the radio program “This American Life” recently revealed that some highly routine newspaper articles--covering real estate transactions or city council meetings, for example--are being written by foreign workers and printed under fake, American-sounding bylines. Nevertheless, most of the news stories you read have been created by American workers. Automation has made it easier to aggregate links to news stories, but when you follow the links, the actual writing has still been done by humans.

So some jobs are relatively secure despite these trends. Robots are not yet capable of taking over everyone’s job, and many American workers cannot be replaced by some low-paid drudge on a foreign shore. What the secure jobs tend to have in common is either a high or a low level of skill. If you are not content to earn the low wages that come from a low-skill job, you’re going to need to acquire and maintain a high level of skill.