“Generals always fight the last war.” — an old World War II saying

Manufacturing jobs were a huge part of America’s post-World War II economic miracle.

In the early 1980’s, 20 million Americans worked in factories, assembling consumer products like cars and appliances.

Well, what happened after that?

There are two narratives here. The shorter story arc is about globalization. American corporations moved all the old manufacturing jobs off-shore to relatively poor countries that still had OK education systems (like China).

This is the story that most people think of when they realize that, as of 2017, your average high school graduate can no longer own a home and raise a family on a single income.

But there’s a second narrative — one that arcs back centuries, to 1794 when Eli Whitney patented the cotton gin. This story’s plot is more complicated, and has quite a few twists that have yet to unfold. It goes something like this: technology keeps making individual workers much, much more productive than they ever were before.

And when one worker — with the help of a robot army — can do what used to require 100 workers… well, you don’t need 100 workers anymore. You just need one.

So here’s the real story of American manufacturing over the past 70 years, told in a single chart:

1uGZ2MW5pWia7UzLt8Pyo9i9hwEqwPkILmhL
Input (human workers) keeps going down. Output keeps going up.

And robots aren’t just boosting productivity in rich countries like the US. They’re starting to affect countries like China, too.

Machines building machines

a5Pf-hjvSviXeO5aNfN24yNPEInzetas-YLE
The robots in a Tesla factory move around so fast that they’re encased in glass to protect human workers from them.

Dongguan — a city near Hong Kong that’s basically the manufacturing capital of the world — recently launched their first automated factory.

The Changying Precision Technology Company manufactures parts for mobile phones. It has 60 robot arms that work on 10 production lines that run 24 hours a day, 7 days a week. Each production line has 3 human workers who monitor the robots.

Before these new robots arrived, the factory needed 650 human workers to be able to operate. Now it just needs 30.

Since this factory laid off 95% of its workers, and handed over the task of manufacturing to the machines, its defect rate has dropped by 400%, and its overall output has nearly tripled.

More production, fewer people.

This is part of a massive automation initiative in China with the twin goals of cutting costs while improving the quality of manufactured goods.

And keep in mind— we’re not talking about the US, where the humans manufacturing workers earn an average of $20 per hour. We’re talking about China, where the average factory worker makes closer to $2 per hour.

This means that if US-based manufacturers took cost cutting as seriously as they are in China, they could afford to spend 10 times as much money automating away each human worker.

In other words, regardless of which country the manufacturing takes place, it’s just a matter of time before most of its physical “building” part of it will be done by robots.

But that physical “building” part of manufacturing has never been all that profitable, anyway.

The text on the back of an iPhone says: “Designed by Apple in California. Assembled in China.” Even though that’s probably an effort to de-emphasize the fact that the phone was “Made in China,” it’s still an important point. Because most of that value of that iPhone was indeed created in California, through the process of designing the hardware and software.

To illustrate this, here’s a quick pop quiz: When Apple sells an iPhone, what percentage of the total proceeds from the sale go toward the Chinese labor it took to build that iPhone?

The answer: less than two percent.

60Myy4D0bl98H0ywpAxMJ5Le78Y10qAQ9wIm

That’s right — for every dollar you spend on an iPhone, about two pennies go to the factories in China that assembled it. The vast majority of that revenue comes right back to the United States.

Whenever some politician talks about “bringing back the good manufacturing jobs” that’s what they’re talking about: trying to bring that 2% back to America, in the form of a few thousand new jobs. Jobs that would only exist with the help of heavy subsidies from the government, and would die off the moment the government quit paying those subsidies.

And this cost doesn’t begin to account for the full side-effects of subsidizing an industry.

Real-life case study: America’s steel industry in 2002

In 2002, American steel was being undercut by cheaper Chinese steel. We tried to keep the American steel industry alive by passing tariffs on foreign steel.

In the end, more than 200,000 people who worked outside of the steel industry lost their jobs due to the macro-economic effects of our artificially inflated steel prices. That means more people lost their non-steel industry jobs than the entire US steel industry employed at the time (187,000 jobs).

In other words, taxpayers spent billions of dollars to destroy jobs, so they could keep even fewer jobs — jobs that were dangerous and low-skill — from inevitably leaving the country. For a few more years, anyway.

The good news to come out of all this misery is that we’ve discovered a better way forward.

“I skate to where the puck is going to be, not where it has been.” — Wayne Gretzky

Instead of trying to bring back the manufacturing jobs that are already being automated away by machines, we have an opportunity to fill the hundreds of thousands of vacant American jobs that aren’t being automated. These are the jobs that involve telling those machines what to do. The programming jobs.

And we can fund this opportunity with all that money that the politicians are wasting on trying to bring back those old manufacturing jobs that would otherwise no longer exist.

Teachers, not tariffs

A vast majority of the value that rolls out of factories is not produced in the factories themselves. It’s produced by people sitting in front of computers in their office — and increasingly, in their own homes.

These are the people who tell the machines what to do. And the real problem our country is facing is that there aren’t enough of these people.

Right now, there are hundreds of thousands of job openings for people who know how to code.

Apple, Facebook, Microsoft, Google, and thousands of other employers spend $5,000 per year in fees just to bring a single foreign engineer over to the US to fill a six-figure salary developer job. These companies aren’t doing this as a cost-savings measure. They’re doing this because there aren’t enough Americans who have the right skills.

So here’s a simple way we can improve our economy and help people get more jobs: every time a politician mentions “bringing back those good old manufacturing jobs” we should tell that politician, “Hey, why don’t we train people to fill the good engineering jobs that never left our country in the first place?”

Here are some concrete steps that we can pressure politicians to take, which will help us get more qualified Americans into these good engineering jobs:

  1. Help experienced teachers to retrain in new skills by paying for them to earn advanced degrees in computer science and software engineering from public universities. Then sponsor summer internships for these teachers to work at software companies, so they can learn even more in the field.
  2. Prioritize technology education in high school, instead of treating it as an afterthought. Require every high school student to take at least one hour of programming each day, just like they take one hour of English each day. What if administrators are unwilling to de-prioritize subjects like French? Well, just make the school day one hour longer. Programming is that important.
  3. Offer preferential financial aid to students who major in either computer science or software engineering.
  4. Leverage our nation’s massive adult education infrastructure — including community colleges — to help busy adults learn to program on nights and weekends.

All the people and all the infrastructure we need to do this are already in place. We just need to adjust the economic incentives, which is something our government has the full power to do if we convince them that it’s a priority.

This will be much cheaper in the long run than trying to subsidize the kind of outdated, low-value-added domestic manufacturing that politicians keep pitching to us.

So the next time a politician tries to use “jobs” as a justification for handing some corporation a bunch of taxpayer money, tell them you know the score. Tell them that China can keep those low-skilled jobs of yesterday. And instead, let’s train more people for the high-skilled jobs of today.

I only write about programming and technology. If you follow me on Twitter I won’t waste your time. ?