The good news: Automation is creating factory work for humans, not just destroying it. And working in today’s auto factory is safer and more engaging than ever. The bad news: The U.S. can’t find the right humans for all this new work.

BRINK talked with Ron Harbour, who co-authored and developed (with his late father Jim Harbour) The Harbour Report, the authoritative guide to auto manufacturing worldwide since it debuted in 1981. Harbour still produces The Harbour Report as an analyst with Oliver Wyman.

BRINK: You’ve been watching the automotive industry for decades. Tell me about the history of automation at these plants. What is getting automated and what isn’t?

Ron Harbour: The first significant blitz was in the 1980s. Robots were more likely to replace humans when there was a safety or ergonomic reason, or if they could do a task more consistently than a person. So first it happened in the body welding shops, because hoisting big welding guns is very tiring and not very safe. Then it happened in the paint shops. If you have a bad day, you can’t apply paint as consistently as a robot. But about two-thirds of the jobs in an auto plant are along the assembly line. And here there are a lot of tasks that can’t be automated. They likely won’t be for some time.

When it comes to what functions to automate, we’ve seen differing philosophies. The manufacturers we saw as most competitive—the Japanese, the Koreans—needed a business case for automation. Whereas automakers in Western Europe and even the United States went more aggressively towards automating things, even when that didn’t pay off. We’ve seen examples of companies installing automation that required more people, with higher skills, than were required before. They did it just to display their technological prowess. I’ve always been dismayed by that.

Ironically, the most [heavily] automated factories in the Harbour Report are not in the top quartile [in the productivity ranking]. Many are in the bottom.

BRINK: According to the Bureau of Labor Statistics, U.S. manufacturing output since 1990 has increased by 72 percent even as employment has fallen by 31 percent. What’s going on here? Is this solely due to automation?

Harbour: Certainly automation has contributed to that. But there have also been a lot of advancements in “design for manufacturability.” That’s when you design a product so it is easier to assemble with less labor. There have also been a lot of pure productivity improvements—more efficient delivery of materials, less wasted motion on the job. Production concepts like lean manufacturing have gone well beyond the auto industry and helped manufacturers operate much more efficiently.

I’m also wondering if that figure from the Bureau of Labor Statistics is adjusted for sourcing. Over the last 10 to 15 years you’re seeing U.S. manufacturers doing final assembly here and calling it “Made in the U.S.A.” but the parts are bought from overseas, where they’re being built with a tenth of the time and labor. Some of it is smart strategy at a lower cost. But sometimes it just makes it seem like there’s more labor productivity.

BRINK: So we’re talking about work done overseas. That’s another factor in U.S. manufacturing job losses?

Harbour: Well if someone says it’s all due to automation, that’s a simplistic answer. The degree of job loss due to automation [compared to offshoring] depends on the sector and what they make. … If the objective is to have as many jobs as possible in the U.S., then you have to create an environment in which companies want to do [labor-intensive] work here: lower regulations, lower taxes, whatever it takes.

BRINK: You’ve said that the number of labor hours expended per vehicle has actually increased a bit in the last ten to fifteen years, despite all the automation. How can that be?

Harbour: Much of the additional productivity has been offset by the complexity of the vehicles. The number of models offered by most automakers has simply exploded. So a plant will be expected to make four or five models and not just one. And within those models there are a lot of new features: side airbags, Wi-Fi, navigation and other components. This has happened in other segments of manufacturing, where the automation has been partly offset by the increasing complexity and customization of the product.

BRINK: So the number of labor hours per vehicle has increased. But the people who do this labor don’t have the same kinds of jobs, do they?

Harbour: The mix has changed. There are more experts in programming, controls and electronics. They’re really hard to find and train. Generally, work is done in small teams of five to seven people, and there’s a lot of cross-training in different roles, which helps with safety and ergonomics and makes the line more flexible.

“We don’t just have an unemployment problem. We have a skills problem. Maybe not everyone belongs in college.”

The manufacturing environment takes a lot more skill than it did before, in the skilled trades as well as in regular production. As part of that team you need to have problem-solving skills, and you’re continuously working on productivity and quality improvements.

BRINK: Why is it so hard to find and train the right people for these factories?

Harbour: Many of these jobs are highly skilled: To be an electrician or in controls, you need at least a four-year apprenticeship program. You can be an electrician in an automotive plant and make very good money. But the attitude in this country is that if you went into a trade, it’s because you failed to get into college. People tend to look down on you and your job.

Compare that to a country like Germany, where they identify your skills and aptitude in your teens and coach you in one direction or the other. If you ultimately become an electrician, that job is viewed with a great deal of respect. There is a popular belief that everyone should go to college, but maybe not everyone belongs in college. Some people may fit more into a trade, and trades can offer great opportunities.

We don’t just have an unemployment problem in this country. We have a skills problem. Many plants we work with have a hard time finding the right skills for the jobs they have.

BRINK: You’re talking about a cultural or policy failing. What about the training and education infrastructure that we already have?

Harbour: You’re seeing more and more collaboration between industry and academia. These are partnerships between companies and local universities and colleges to train not only new workers, but also existing workers on a continuing basis. Manufacturers and local schools are designing and developing programs cooperatively.

But the skills problem starts long before this stage. It’s 2017 and I still go into plants where they’re putting on a problem-solving course and there are people who can’t even read. They have to go to remedial reading classes first. They can only do third or fourth-grade math. This is a failing of the educational system.

If you build something in the city of Detroit, 50 percent of your workforce must be a resident of the city. But contractors can’t find tradespeople in the city. In a city of 700,000 people, there are 83 licensed electricians. The contractors and the city set up an educational institution to build up the skills of residents, and it’s been an almost total failure. People would only come on a sporadic basis, or they lacked basic reading or math skills.

BRINK: Let’s go back to the shop floor. You said that in cars and other manufactured items, you’re seeing increased complexity and customization. What does that mean for the manufacturing industry and for the workforce?

Harbour: Not long ago, you would drive home in your new car and find that six of your neighbors had the same vehicle. Now I can give you a car that you won’t see again in a year. Customization gives consumers the feeling that they’re getting more for their money. Flexibility has become a differentiator for automakers, because some are far better at it than others. It has to do with how you engineer a plant and also your people. People who make model changes have to be very flexible and team oriented.

In today’s factories, it’s increasingly easy to develop or test new procedures or components. We’ve seen plants that have 3-D printers line-side. If a team identifies a problem, they can go make a tool to fix it in a matter of minutes. Once the workers learn this technology, it makes their job a whole lot more interesting than drilling screws into a car. If you properly balance technology and people, you can make the work environment a lot more fun and interesting.

I’ve seen this any number of times. You take a guy on the line with a lot of seniority, who’s done something the same way for many years. You put him in charge of a team of five, give him the authority to run his team almost like a small business. You train the team in problem solving; give them the tools to fix their own issues. Then they invent a solution to a really bad quality problem. They eliminate a defect. That’s something they can brag about to their family. Their sense of pride in their work changes dramatically.

The biggest transformation in factories in my 35 years hasn’t been automation—it’s been in the way we work. People have been more involved in the process. They’ve been given more authority and accountability to run the floor themselves. The supervisors become coaches and trainers and not firefighters. And that’s been fun to see.