Technology and e-commerce companies have a reputation for being drivers of creative destruction, sometimes at great cost to local communities. Economic nostalgia tells us to lament those jobs and fear the changes that come with technological progress. But it's worth remembering that tech companies are also a major source of high-wage job growth in the US economy. On this episode of Faster, Please! — The Podcast, I'm joined by Michael Mandel to consider the role of tech companies in the American economy.
Michael is vice president and chief economist at the Progressive Policy Institute. He's also the author of "Investment Heroes 2022: Fighting Inflation with Capital Investment," co-authored with Jordan Shapiro.
In This Episode:
Tech sector job growth (1:23)
How technology affects the labor market (6:08)
Job-replacing tech vs. job-creating tech (10:46)
Encouraging the digitization of US manufacturing (15:00)
America’s tech firms: investment heroes (18:34)
Below is an edited transcript of our conversation.
Tech sector job growth
James Pethokoukis: Last year you said, “We’ve seen in recent years [that] the tech/broadband/e-commerce sector has been the main source of job growth in the economy.” Do you think this is a widely understood fact either among the public or among policymakers here in Washington?
Michael Mandel: That's such an excellent question. No, it's not a widely understood fact. I've just calculated the latest numbers, and if you look at full-time equivalents, all of the job growth since the pandemic started has been in what I call now the “tech/e-commerce” sector. And the rest of the economy and job growth has been much, much weaker.
Is this purely a pandemic-era phenomenon, or do you expect it to continue to happen?
It was happening before the pandemic. It is going to continue after the pandemic, too. I think what we've learned in the past is that whichever sectors grow during a recession tend to lead the next recovery as well. The fact that we've had all this growth in the tech/broadbrand/e-commerce sector during the pandemic suggests that's going to be the job leader going forward as well. The Bureau of Labor Statistics has just released its occupational projections for the next 10 years. I haven't had a chance to look through them yet. I suspect that they will understate the future job impact of the tech/broadband/e-commerce sector as they have in the past.
Is that an accurate forecast that they put out?
It is about as accurate as just extending long-term trends. In terms of looking forward [at] telecom-related jobs or app-economy jobs or computer-related jobs, it has consistently under projected. They actually make no real claims. They don't say it's a forecast. They say it's a projection. Probably, if you ask them privately, they would tell you they really don't want to do it. But it's really widely read.
The part of that sector I think people might be surprised by is e-commerce. I'm guessing that a lot of people view e-commerce as a jobs killer: It's replacing all the people who work at in-person stores with kiosks. Is that your perception? That is wrong, though.
That is my perception, and that is wrong. The way that I think about e-commerce is it doesn't pull jobs out of brick-and-mortar retail. It actually pulls hours out of the household sector. So what happened is that people used to put an enormous amount of hours into driving to stores, parking, walking around, and standing on line, and so forth. And if you look at the data that comes out of the Bureau of Labor Statistics on the American Time Use Survey, you see a really sharp drop in the number of hours that people spend shopping for goods. It's gone down by about 20 percent over the last 15 years. And it dropped about 10 percent just over the course of the pandemic. All of these hours, which is an enormous number of unpaid household hours, are being moved into the paid market sector. Instead of you going into a store and picking out the stuff yourself, somebody else is doing this using robots in an e-commerce fulfillment center. And instead of you driving to the store by yourself and spending all that time parking, somebody else is putting the stuff in a big truck and delivering it to you, using more capital, doing it more efficiently. There's been a very sharp drop in the number of hours that households are spending on shopping, which (A) creates a lot of jobs in the market sector, (B) really distorts the productivity numbers, and (C) leads us to misunderstand the sources of growth in the economy: what the effect of productivity is, what the effect of technology is.
I know you and I have talked about this in the past, for many years we used to wonder, when was technology going to start generating jobs for the ordinary person? And that's what e-commerce has done: generate tech-enabled jobs in e-commerce fulfillment centers, in the entire supply chain, that pay better than the old retailing jobs, that pay a lot better than the non-paid jobs in the household sector where people used to spend this. You're creating a lot of income that wasn't in the economy before.
How technology affects the labor market
People don’t think about those warehouse fulfillment center jobs. If they do, they probably think they pay worse than they actually do. And they probably underestimate how many there are and figure it's just all robots or something.
I think we've managed to break the “all robot” canard, because what we've seen here is that the ability to put robots into the fulfillment centers has lowered the cost of doing e-commerce so much that it's actually made it open to all consumers for everything, basically. There are no restrictions on it. You're producing enough surplus that you can actually do returns correctly. There's a big economic surplus being generated by the automation of the fulfillment centers that enables us to hire a lot more workers.
That's a classic case of technology affecting the labor market, right?
If you look back historically, this is very much the same sort of thing that happened with Henry Ford and the assembly line, which is that you think that when you have an assembly line, your adding productivity that’s going to reduce the number of jobs. But it lowered the cost of cars so much that all of the sudden the ordinary person was able to buy them. That created a lot more demand for workers. And if you think about, why were people buying cars as opposed to just using horses? It's a time thing. The thing that's in most scarcity for households is time, because they can't create more of it. Anything that saves people time, they're going to be willing to pay a lot for. In that one case, this was the automobile creating jobs. In this case, it's less shopping time creating jobs in e-commerce fulfillment and delivery.
If you've never been in one of those fulfillment centers, there was a wonderful movie Nomadland that starred Frances McDorman, and she would work during the busy season at an Amazon fulfillment center. It did not seem like a miserable job, but it seemed like a busy job.
It's a busy job. I think about these as the equivalent of manufacturing jobs for the technological age.
They're mixed physical-cognitive jobs, just the way that assembly line jobs were mixed. They actually required some skill, and at the same time they required manual labor. They pay about the same as entry-level manufacturing jobs. In many areas of the country they are in fact becoming the substitute entry-level job that manufacturing once was. If you look at the data for occupational health, they're kind of where they should be. They're physical jobs, you can't deny that. Which actually kind of gives a lot of people problems because they think, “Well, what is an ideal job? Is an ideal job an office job?” It turns out for a lot of people, it’s not. It’s something that involves some measure of physical labor, too. Let me give you a number here: Since July 2019, the tech/broadband/e-commerce sector has produced about 1.3 million jobs out of a total of 2.2 million for the economy as a whole. And that's pretty amazing. That's more than a majority and much more than healthcare and social assistance, which should be your next question: What's going to happen to healthcare jobs with automation?
What's going to happen to healthcare jobs?
If you think about the shift to telehealth during the pandemic, people are realizing that there are less expensive ways of doing what they were doing before. Better ways of communication. One of the biggest phenomena I think we're going to see going forward is that the long healthcare job boom may be over. We may actually end up with a surplus of healthcare workers. Rather than retraining manufacturing workers to go into healthcare, we may be retraining healthcare workers to go into technology.
Job-replacing tech vs. job-creating tech
On that issue, I know there's been some research by Daron Acemoğlu about how technology is affecting the modern job market. Are we producing the kind of innovation and automation that replaces jobs? Are we producing the kind that creates new things for people to do? Are we creating the kind that helps people do their jobs better?
I think there's some concern that we've produced too much of the job replacing rather than the job creating/enabling.
We're going to have both types. We haven't actually had any of the job replacing yet—at least not in the measure that people were worried about. Remember, we were worried about all the losses of jobs for truck drivers from autonomous trucks. Instead, we have shortages of truck drivers.
I was told there'd be riots.
If you look back historically, you see that some technologies generate jobs and some technologies replace jobs. I think what we've seen, which we hadn't seen before in the e-commerce sphere, is we know this is a case where we've created new jobs. If you actually add together e-commerce jobs and the brick-and-mortar retail jobs, what you see is that there has been a 650,000 job increase since the beginning of the pandemic in the combined retail/e-commerce sector. There's a net job increase from technology here. And there's a wage increase because the e-commerce jobs pay about 30 percent more than brick-and-mortar retail.
And they are more diverse, which is really interesting. Diverse racially and diverse ethnically. People usually think of retailing being poor people of color. But in fact, you look see that there's a lot of discrimination in brick-and-mortar retail. I think, in the end, the retail sector broadly extended, including e-commerce, is going to be a net job gainer from technology. The real interesting question is going to be, what's happening to manufacturing? I'm watching this very closely. Of course, we've lost a lot of manufacturing jobs.
We still make a lot of things, though.
We make a lot of things, but the non-high tech manufacturing capacity peaked in 2000 and has been coming down since then. The actual size, however you want to measure the manufacturing sector in the US, has actually been shrinking. What we need to be able to do is adopt more advanced manufacturing techniques, more automation, more digitization: drive down the cost of making goods, drive it down in a way that starts increasing the ability of people to buy them, increasing the capacity, and increasing the jobs associated with them. This kind of goes to your question about, is it job replacing or job creating? What happened was that people got scared. We’ve been replacing manufacturing jobs with technology up to now, but there's nothing that says that has to keep going that way.
I actually think that's really a crucial question for the US economy going forward: Are we going to actually invest in manufacturing digitization, not just on large scale but on a small scale as well, on entrepreneurs? One of the things that I watch really closely is this new census data on business formations. You don't see the business formation growth. You see a lot of business formation growth across the economy, but not manufacturing yet. That's going to be a crucial turning point for the economy.
Encouraging the digitization of US manufacturing
How do we make that happen, that will have manufacturing here in the US using the latest technology, robotics or what have you?
One of the things we have to realize is that our small businesses are still cash poor and credit poor. And they also don't have access to the latest technology. The way that I think about this is if you go back to the auto industry and auto dealer franchises, which created a lot of wealth on the local level. We have to think about manufacturing franchises on the local level where the technology is prepackaged, where people start small businesses and do a lot of creation and production on the local level, in a lot of different places. There may be some signs that could be happening; there may be some signs that it isn't. But this would be one of the big turning points for the US economy in terms of moving towards a really strong, sustainable future.
You're not just talking about big companies with big factories, you're talking about far smaller companies able to use the latest technology: an off-the-shelf robot or something who could do things. Is the technology almost there? Is there a role for government? Do we just need the technology to keep progressing? What's the key?
We need the technology to keep progressing, but it's almost there. The real question is financing for small entrepreneurs and exposure to the technology.
They're just unaware that this is out there?
They're just unaware. They have to be able to experiment with it. Right now the small manufacturers are scared.
Who is a small manufacturer? What do small manufacturers manufacture in this country?
There are far more small job shops out there than you might think, and there’s potential for far more than you might think to grow up. For example, suppose you had an old appliance that was missing a part. In theory, there's no problem producing that part with 3D manufacturing or some other technology, if you had the plans for it. And if you had that set up. What you have right now is manufacturing networks so you could contact a manufacturing network company and give them your plan and they would find a job shop around the country that could do that. You could set up a manufacturing operation tomorrow. A lot of this is not difficult. There are some areas that are more difficult. The art of automating a lot of apparel manufacture is still not all there. It's getting there, too. You've got apparel manufacturing, small tools, small objects. You should be able to have the ability to make customized furniture much cheaper than you do. You go through the different lists of things. It becomes harder the more complicated that things get, but things that are really simple should be able to be manufactured with these new technologies in ways that are less costly and more customized.
America’s tech firms: investment heroes
Every year, you folks at PPI put out an Investment Heroes report showing who are the companies really investing. A lot of very well-known tech companies. Not just tech companies on that list, but there are a lot of tech companies. If technology companies are creating a lot of jobs, if they're investing a lot, why do they seem to be so wildly unpopular here in Washington?
Let's actually say some more of the good things they do as well. They also pay their workers well. They did not participate in the inflationary surge. Inflation in the digital sector was accelerated a little bit, but much less than the rest of the economy, which is what you would expect if you had high productivity growth. I think that when push comes to shove, people just don’t like “big.” Big worries them. If you compare these companies to the big manufacturing companies in the past, if you compare them size-wise to the global economy, they're about the same size, relatively speaking. They're not out of scale. But what happens is that there's a regulatory push. And that's only natural.
When you're regulating, you want to avoid throwing out the baby with the bath water. As you know, at PPI we believe in light-touch regulation. We think that regulation is an important part of a market economy, but you really want to make sure you don't go overboard with it. In this case, I think about regulating large tech companies as the big bear theory: If you're sleeping in a bed with a big bear and it rolls over, it's going to crush you, whether it wants to or not. So you have to distract the big bear with a stick every once in a while to keep it alert and say, “No, don't roll over.”
What is true if you look historically at the way productivity growth spreads, productivity growth doesn't spread from technology moving from big companies to small companies, or from highly productive companies to less productive companies. It comes because the highly productive companies expand their share of the market. They are good at doing productivity and they expand. That's kind of where we are in this process. The highly productive companies look around and they see other areas of the economy that they think they know how to fix. They see a market opportunity and historically that's what usually happens. What we're seeing now from my perspective, as long as it doesn't go overboard… which some of the bills in Congress did. Some of the bills in Congress made no sense whatsoever. You want to have a kind of push and pull, which is that these companies are highly productive, great for workers, great for consumers, great for suppliers. And so you want to see them expand and you want to see them take cognizance of some of the side effects of what they're doing.
Thanks for listening to part one of my two-part interview with Michael Mandel. Next time, we'll discuss US productivity growth, industrial policy, and more.