Your job is next

A while back I wrote an article about why all the jobs have disappeared from the coal industry, and why they aren’t going to come back. It’s possible that some people read it and thought “Well, that’s sad, but those unemployed people should have retrained”. That’s somewhat true, but only in the short term. The unfortunate reality is that in the long term, what happened to coal is going to happen to most other major industries.

Blame NAFTA? #

It’s popular to blame the loss of US jobs on free trade deals. NPR’s Planet Money recently ran a series of podcasts about NAFTA, looking at its effects over the last 20 years. I recommend the entire series.

The short summary is that from an economic point of view, NAFTA was pretty good. Fresh food is more affordable and available all year round. Overall trade has more than tripled. Yes, the US auto manufacturing industry suffered, but on the other hand the US services industry increased exports to Mexico from $27 billion to over $82 billion, and agricultural exports rose as well.

Overall, 800,000 new jobs were created. The problem, though, is that at the same time 800,000 new jobs were created, 1.5 million jobs were eliminated through trade, for a net effect of minus 700,000 jobs or thereabouts.

So the corporations got richer, faster — the rate of increase of corporate profits ramped up sharply in the mid 90s.

source: tradingeconomics.com

However, that didn’t translate into middle class jobs or middle class wealth. And now, in the 21st Century, we’re facing a new job killer that will make NAFTA look like the Public Works Administration.

Tariffs to the rescue? #

A popular answer to the problem of job loss from trade is to impose hefty tariffs. Trump has threatened 45% tariffs against China, 35% against Germany, and so on.

To see how a tariff against China might work, take a look at what happened when the US tire manufacturing industry started to see competition from China. You can read the full details in a report, but I’ll give a quick summary.

Starting in late 2009, Chinese tires were subjected to a 35% tariff. This was reduced to 30% after 1 year, 25% after 2 years. Imports of car tires fell from 13 million per annum to 5 million per annum. This, in turn, saved around 1,200 US jobs.

So it worked, right? Well, yes, in pure job-saving terms it worked. The problem is that the price increase on tires cost US consumers an estimated $1.1 billion in 2011. Spending $900,000 per job to save tire manufacturing jobs is probably not such a good idea.

It was worse than that, though. China retaliated by imposing tariffs on some US goods, which lowered retail sales and caused a loss of jobs. Once you factor that in, it’s estimated that the tariff decreased the number of US jobs by around 2,500.

The tire tariffs did, of course, boost the corporate profits of Goodyear, Bridgestone and other tire manufacturers with US factories. It just didn’t result in overall gains for workers.

Many problems have simple and obvious steps to take that will make the situation worse. When it comes to unemployment, tariffs are one of those simple and obvious steps.

What about the jobs saved at Carrier? #

Back on December 1st 2016, Trump made an appearance at Carrier in Indianapolis. They’re a big manufacturer of heating and air conditioning equipment; my home AC system is one of theirs. Trump was there to take credit for saving 1,000 jobs which were due to move to Mexico. He had threatened tariffs, but fortunately he hadn’t tried to impose them. Instead, juicy tax breaks had enticed Carrier to play along. As well as saving jobs, the deal was to include the company investing millions of dollars in US manufacturing.

A few right-wingers decried the deal as crony capitalism, and a few troublesome individuals pointed out that 1,000 jobs is an inconsequential number to the US economy, but mostly everyone expressed polite approval that at least some jobs were being saved.

Just a week later, though, the estimated number of jobs saved had dropped to 800, and Carrier quietly announced that those millions of dollars they would be investing would be spent on automation.

HVAC isn’t a dying industry the way coal is. Nevertheless, persuading Carrier to keep its business in the USA isn’t going to keep the jobs around. It’ll just delay the inevitable and prevent the automated factory from being built in Mexico.

Make iPhones in America! #

Another discussion point during the 2016 election campaign was whether Apple might move its iPhone manufacturing to the USA, another Trump demand.

If Trump wants to try to get some of those disappearing jobs back in the USA, he’ll have to compete in a bidding war with China — which already gave Foxconn $1.5 billion in perks to build “iPhone City” where the world’s iPhones are made.

Even if Trump decides a bribe in the billions of dollars is worth paying, there’s a snag. Earlier in 2016 Foxconn gleefully announced that they had replaced 60,000 of their 110,000 factory workers with robots. They plan to continue with the automation, in a three phase project, until the factories are fully automated.

What might that look like? Well, another mobile phone manufacturer, Changying Precision Technology Company, used to employ 650 people at its factory. Now they only have 60 employees. Replacing 90% of their staff with robotsincreased production by 250%, and defects dropped by 80%. The company says it will eventually need just 20 staff.

So, will Trump offer Apple billions of dollars of perks to bring a few dozen long-term jobs to the USA? He’s dumb, but I don’t think he’s that dumb.

Manufacturing is dead, what’s next? #

So, the manufacturing jobs are all going away eventually. That’s bad, but we’ll find something else for those people to do, right?

Don’t count on it. There are other entire industries vanishing right now.

The first one you’ll probably notice is truck driving. Currently, truck driver is the most common job in 29 US states, largely because it has been impossible to offshore to China or automate.

That era is now at an end. Uber successfully delivered a truckload of beer 120 miles via robot truck. After California banned Uber’s pilot self-driving car program, Uber shipped the self-driving cars to Arizona — by self-driving truck.

You might be skeptical about self-driving cars, but Google’s experience is that the difficult areas for self-driving technology are cities. Long haul freeway trucking is a much easier problem.

Currently there are 3.5 million truck drivers in the US, and the trucking industry employs 8.7 million people. While the industry expects double-digit growth and currently has a shortage of drivers, ultimately that will only accelerate the deployment of robot trucks.

Even given the high demand, current predictions are for robots to replace 1.7 million truckers in the next decade.

Let them eat burgers! #

Many of the people indirectly kept working by the trucking industry are the workers at the nation’s truck stops and fast food restaurants. Right now, there are 3.5 million people employed in fast food restaurants.

Unfortunately, as the demands for a $15 minimum wage have escalated, McDonalds has responded by starting to roll out 14,000 self-service kiosks. I first saw these in McDonalds in London a few years ago, and they were in New York soon afterwards. A single restaurant can have six constantly open ordering kiosks, with no wage overheads.

There have been companies building burger cooking machines since the early 2010s. There are machines which cook and serve pizza, and machines which cook and serve french fries, so it shouldn’t surprise you that McDonalds has also opened some completely robotic restaurants as a test. The owner of Carl’s Jr and Hardees is following suit. Wendy’s is adding self-ordering kiosks in 16% of locations by the end of 2017; ROI is less than two years.

Computers taking the orders, cooking the food, ordering supplies, cleaning the floors — all you need is a human per store to deal with any robot breakdowns. The only thing which will slow this trend is the initial capital cost of the robots. Expect a million or so fast food jobs to disappear over the next decade or two.

The farmbot to waiterbot system #

When the robotic restaurants order produce, chances are it will have been grown on a farm using robots too. Remember the self-driving tractor and combine harvester in “Interstellar”? Those are already here, controlled via GPS. Right now, you still need someone in the tractor to watch for collisions, but once the sensor technology from the self-driving cars is incorporated into new tractor designs, that will change.

If Trump’s great wall ever gets built, the crop-picking jobs often filled by undocumented migrant workers might have disappeared before it is completed. Robots can already harvest fruit, lettuce, and many other crops. Again, the limiting factor is the cost of the equipment. Also, farms have already become somewhat industrialized, so the total number of regular jobs is already under a million. Still, another half a million lost will hurt the rural economy.

The packaging and order fulfillment can be automated, too, whether it’s produce or any manufactured goods. Look at Amazon, which has over 45,000 robots working in its shipping centers. Multiple companies are working on deliveries by autonomous drone, as well as the previously mentioned self-driving trucks.

Building a world of unemployment #

Another major industry in many states is construction. Texas has around 800,000 undocumented workers in the construction trade. You might think that deportation and border security would therefore be a boost to employment, but again there are signs of trouble in the long term.

China successfully 3D-printed an apartment building in 2015. Dubai 3D-printed an office. There’s a house in Russia that was 3D printed on site. Another possibility is using robots to construct buildings using structural insulated panels. I’ve seen buildings being constructed with SIPs, have you?

White collar jobs aren’t safe #

So far I’ve talked about “blue collar” jobs. You might think that because your career involves sitting at a desk and using your brain, you don’t need to worry about robots and computers putting you out of a job. If so, think again.

Let’s start with managers. The world’s largest hedge fund is already working on replacing managers; they expect to automate hiring, firing, and other day-to-day management tasks.

Goldman Sachs has replaced nearly 600 traders with software. Most fund managers are no better than an index fund anyway; for situations where you need to make speculative investments, computers can do most of the work.

A Japanese insurance company is replacing its payments department with AI; savings will start almost immediately. An Israeli insurance company is replacing the brokers. H&R block is offering AI tax advice.

Prescription for mass unemployment #

Medicine is a high cost industry, so automation will probably start to hit any medical area it can quite soon. Obviously the insurers will want to cut costs by having claims processed by software, but a lot of routine diagnostic medicine can probably be automated too. IBM has an AI radiologist’s assistant, and an oncology program which has diagnosed cancers human doctors have missed.

Japan is working on robotic nurses to ease the shortage of help for an aging population. 30% of medical laboratories in the west have started automating. An autonomous robot surgeon is being tested on pigs, and beating human surgeons. While you’re being operated on, perhaps your anesthesiologist will be a robot. Mail-order pharmacies are already automated, your corner drug store will be next.

Think fast, but you won’t be fast enough #

So, manufacturing’s going, management and administration will be next, bankers and doctors will follow… what about disciplines requiring creativity?

Bad news, companies are already using AI techniques to partially automate graphic design. Web design is being automated — sure, the results may not be as good as those from an inspired human, but most companies don’t need anything that good. A lot of web developers already lost their jobs to cheap developers in India churning out templates and customizing them; now those tasks can be fully automated.

AP has started using AI to write stories instead of journalists. So far it’s only corporate earnings reports and sports stories which have been automated, but as the software gets better, it’s probably only a matter of time before political reporting and product advertorial can be churned out by computers.

Meanwhile in the realm of inspired engineering, NASA used AI to design a space antenna which was successfully deployed and used.

A good legal strategy can require human creativity, right? Well, maybe not so much. A simple AI chatbot has helped overturn 160,000 parking tickets by offering legal advice. There’s a lot of straightforward work done by lawyers which could be automated. For instance, if I had the time I’d like to work on a chatbot to help people with immigration processes…

How bad is it going to be? #

All of the above are just examples. Furthermore, in many cases the robots and computers won’t completely eliminate the humans; they’ll just help the remaining humans to do work that used to take many more people. A McKinsey study says that sure, only 5% of jobs can be 100% automated — but at least 30% of the work can be automated in at least 60% of jobs.

So the big picture is that experts are estimating that around half of US jobs can be automated. They’ve evaluated hundreds of jobs against nine criteria which determine how easy they will be to automate; you can search for your job to see how likely it is that you’ll be replaced in the next couple of decades, or you can check out a table of jobs likely to be safer.

If you want to move towards a job that will still be around in 20 years, aim for one that requires negotiation, coming up with new ideas, being empathetic, dealing with unpredictable situations and messy environments. A plumber fixing people’s toilets and sinks? Probably a safe bet.

Any job which just requires that you do the same stuff over and over again in an organized space, well, that’s probably going to be something a machine can do, whether it’s bookkeeping, stacking shelves, or flipping burgers.

You know what’s probably a solid long term career? Law enforcement. Because in a decade or three when half the population is out of work and the 1% are still raking in more and more money, there will be riots in the streets…