Martin Ford has some good news and some bad news. The author of The Lights In the Tunnel: Automation, Accelerating Technology and the Economy of the Future says improving technology is going to mean there won’t be enough work to go around for everyone, and that this is inevitable. The good news is we can still do something about it that doesn’t require destroying every machine we can lay our hands on. Ford, who is also a Silicon Valley entrepreneur, is among a few brave souls who are raising these uncomfortable points. It’s a lonely position to stake out and he says he gets the “Luddite” charge a lot, but his analysis is considerably more sophisticated—and persuasive—than that.
Your analysis of occupations in the U.S. suggests that jobs created by new technology are actually relatively few and don’t last long. Are there enough private sector jobs to go around even today?
Obviously, there aren’t right now. We still have very high unemployment. The other thing I look at is what’s happening with college graduates because the conventional wisdom is always that getting more education is the solution to look to for the future. But what we see is that 50% of college graduates coming out are not finding jobs or they’re taking jobs that really don’t require a college degree. They’re working at Starbucks or something. And over the last 10 years, wages for new college graduates have been in decline.… So there’s plenty of evidence to suggest we just aren’t creating jobs in the way that we used to. There just isn’t enough for everyone—certainly not good jobs.
Why is new technology incapable of creating many new jobs?
The biggest thing that I think is happening is that essentially technology is getting better and better at automating the more routine jobs. And when you use the word “routine” you have to be careful about what you mean by that. At one time routine would have meant standing on an assembly line, but nowadays it’s really become much broader than that. A better word might be “predictable.”
If you’re in the kind of a job where maybe someone else can look at the record of what you’ve done in the past and—based on looking at that, studying it, practicing it and repeating it—they could learn your job, then there’s a good chance that an algorithm could also do that. If you’ve got a job that is predictable based on what’s been done in the past then ultimately it’s going to be susceptible and that’s becoming more and more the case. The problem is, that’s most jobs. The number of people that are paid to sit there and think deep thoughts and really create things is a pretty small number.
What if, for example, everyone studying engineering or computer software suddenly switched to plumbing? Do our economies actually need all those trade workers?
There would be intense competition for those jobs. The problem is that even jobs like those that can’t be automated are still going to be subject to a cyclical impact. In other words, if lots of other people lose their jobs and those people aren’t buying houses, for example, then the demand for plumbers and electricians will fall accordingly because a lot of those people are employed in new construction.
Those jobs are safer with respect to automation, but it definitely won’t be the case that everyone can flood into those jobs. As it becomes more competitive you may see people who otherwise would have gone to college try to compete in those areas and then out-compete people who are perhaps traditionally a better fit for those jobs. Typically those jobs have been done by people who aren’t going to go to college. The question then becomes, what do those people do if people going to college are suddenly competing with them to be plumbers?
Can the “law of economics,” which says technological advancement will perpetually lead to more job creation, be successfully challenged?
I think the evidence ultimately will challenge it. People will look around and see what’s actually happening and they will begin to look at it differently. I think that’s a prejudice that’s especially strong among economists and among other educated, wealthy people. If you go out and talk to average people on the street, the ones that are already being heavily impacted, I think you’ll find much more recognition of the reality of what’s really happening. But there definitely is a stigma attached to saying these things. The word “Luddite” tends to get used a lot.
Do you think the U.S. government’s stimulus programs have helped create or at least preserve jobs?
I think they probably have helped at least preserve jobs. I think, in general, stimulus is going to become less effective as a solution because of what I’m talking about: you throw the money at it, but it will create less jobs. As time goes on technology becomes more advanced, and more and more of that money is going to be spent on machines and not on workers. And so what they call the multiplier effect—the idea that if you put a certain amount of money in there you get, say, a dollar and a half of impact—that number is going to become lower over time so that will make it less effective. So in general I supported the stimulus. I thought it was the smart thing to do, but we need to recognize that there’s a structural issue and simply going with that kind of stimulus in the future is going to become less and less effective.
You posit scenarios where technology-driven unemployment reaches as much as 75% and only then does the economic system collapse. Wouldn’t there be crippling strains on the economy long before that?
Yeah. The 75% was really more of a thought experiment. I don’t think that, realistically, we could ever in any smooth fashion get to 75%. We’re already at 8% now and it’s unpleasant. By the time you get to 20%, 25%, maybe 30%, that would be historically unprecedented. Bad things would happen on every level—economically, socially. Society would begin to sort of fall apart at that point. It would be ugly, so we definitely want to begin to deal with this long before we get to that point.
You write “economy-wide automation of jobs is not a technological impossibility; it is a psychological impossibility” with reference to people’s unwillingness to believe we could automate ourselves out of work. What do you mean by that?
I would say it’s probably a simple psychological bias and an unwillingness to look at it. People do present other arguments. One of the arguments that’s been effective is that historically this has not happened. If you look back over history there have been lots of examples of technology that has disrupted employment, but the final result has been we’ve adjusted and created more jobs for the long run.
The point I’m making is that I think the future will be different because of the point which technology has now reached and the speed with which it’s moving forward. The other thing is all of this is hard to tease out. You can look at the data but you have to be careful how you interpret that. The impact of this is not obvious—you know, jobs are disappearing, but is it because of automation or is it because of other things? It’s not as clear cut as we might like.
What are the solutions to the problem of automation and its tie to increasing joblessness?
In the short run we can focus more on training and education and making sure (it’s) accessible to people and they don’t have to take out enormous student loans and so forth. In the long run, as I described in my book, we’re probably going to get into a situation where there simply are not enough jobs and there are no conventional solutions. We’re going to have to go beyond that and look at actually restructuring the system. And I think that probably the simplest and easiest way to do that is to have some form of guaranteed minimum income. And as I also said in the book, we should probably incorporate incentives into that. Rather than just giving people money we should develop ways so that you can earn more money if you do certain things that are good for you and maybe good for society.
You can think of the obvious problem with that: Imagine if you’re a marginal high school student and you know that when you turn 18 or a certain age you’re going to get an income whether you graduate or not. That’s a problem in terms of incentives. You’re going to end up with a whole bunch of people that aren’t going to bother to make the effort and that’s simply not good for society. That’s why I think we should incorporate some basic incentives with regard to education. We should make sure that we have an educated population.
You say offshoring is just the precursor to automation. Can anything stop that evolution?
The jobs that are being offshored are eventually going to be automated so that, in many ways, the countries that are going to be hit hardest by automation are maybe not so much the United States, but developing nations. I think China will be hit very hard by factory automation. In the United States, manufacturing automation will continue but it’s not going to have as big an effect because the jobs are already gone. In the U.S. it’s going to be more service sector automation that will really have the disruptive impact. So I think that process will continue, but I think the tendency is going to be that first jobs will go offshore and then as technology advances, those jobs will be automated.
Do you see any evidence of the advanced economies doing anything to prepare for a society where less work is available due to automation?
In general no, but the countries right now that are best positioned to deal with it are probably the Scandinavian countries, which have had policies for a long time that are very supportive of their populations. Germany has done some very positive things during the recession. They had job sharing; rather than having German industry just slash and burn and cut all their workers the way that happened in the United States, what (the German government) did was have those workers work part-time and subsidize their wages. But they stayed in their positions, they shared jobs and so they were able to retain their jobs and their skills. And that’s fairly close to the idea I describe in my book, that you need to have job sharing combined with some sort of income from the government. There are certainly cases where governments out there have done the right thing.