Welcome to part two of my conversation with Michael Mandel, vice president and chief economist at the Progressive Policy Institute. In the last episode of Faster, Please! — The Podcast, we considered the capital investments and job-creating power of America's major tech companies. In this episode, we discuss the Biden administration's CHIPS and Science Act, industrial policy, and whether we should expect an uptick in US productivity growth.
In This Episode:
Innovation and industrial policy (1:14)
Looking at productivity numbers (4:14)
How technology affects jobs (7:19)
The future of productivity (12:38)
Investing in bioscience and materials science (15:00)
Policy for societal resilience (19:29)
Below is an edited transcript of our conversation.
Innovation and industrial policy
What do you make of this recent CHIPS and Science Act and perhaps a move in the United States toward what some people call industrial policy—a phrase that can mean a lot of things. I think in this case it means subsidizing sectors that government thinks are important, especially in competition with some other countries.
I have to say, quite honestly, that I took my eye off the semiconductor industry for a couple of years because I assumed we were in good shape. And then when I looked over, I said, “Wait a second, something happened here. All of a sudden, we're not in good shape anymore.” I support investment in this sector. I don't consider this to be classic industrial policy at this point. I just consider this to be doing what we've done in the past. We did this with memory chips: There was government intervention with SEMATECH. You sort of say, “Here's a sector, we need to fix it. Let's just go ahead and spend some money here.” We haven't gotten to the point of being strategic yet. This was not really a strategic investment. We're just saying, “Let's throw money at this problem.” We know that at the margin, throwing money at this problem is going to get us further along than we need to be.
Do I think that more of this is needed? The country you didn’t mention, of course, was China. I do think China's innovation policy is a really interesting question because we haven't had an experience with authoritarian countries that were successfully innovative. For a lot of reasons, because it seems that capitalism works better to produce good innovation. If it turns out that authoritarian innovation works, many countries around the world will want to imitate that model because it's much more comfortable for governments to run innovation from the top. The only reason why they allow innovation to bubble up from the bottom is because doing it the other way doesn't work.
What I would expect to see in the US is a combination of the two, a lot that is bubble up from the bottom. We will be faced with technological and social and environmental challenges that we can't imagine. And we have to have invested the money in the new technologies before we get there. We don't know what the problems are going to be. We don't know what the technologies are going to be. We discovered this in the pandemic, where it turned out that mRNA technology, which was sitting on the shelf for 20 years, was a solution to a problem that we didn't even know it was a solution to. But if we hadn't been investing in it so it wasn’t there, it wouldn't have been available as quickly as it was.
Looking at productivity numbers
Statistically, we had this productivity boom during the pandemic, at least in 2020, 2021. And people read about a lot of technologies happening: maybe AI spreading, mRNA, CRISPR, rockets. The first half of this year, statistically, was not so good with productivity. These numbers tend to jump around a lot. What's the reality going forward?
As you know, productivity numbers, especially total factor productivity numbers, are useless over any period less than 10 years. We mentioned earlier the shift of hours from the household sector to the market sector as part of e-commerce. Remember: Hours in the household sector are not measured as part of the productivity basis. If you actually include them, it significantly adds to the productivity growth in this period. Because what's happened is, if we take the total amount of hours being put into consumer distribution, which is both the market hours and the non-market hours, market hours has gone up, which is what shows up in the official productivity numbers. If you look at retailing, you don't actually see very much productivity gain because, in fact, the hours have gone up a lot. But they're not counting the fallen hours in the household sector. What has happened is when you count the fall of hours in the household sector, productivity growth—I haven't done these calculations recently, but it goes up a lot: quarter percentage point a year, half a percentage point a year. It’s actually a significant increase.
In the sector or economy-wide?
Economy-wide. Because it's a lot of hours. The degree to which telehealth, for example, removes the necessity of people to drive to the doctor's office, if we are not including those hours in our calculation of productivity growth, we're missing the big effect. And you can go through the economy like that: places where there were movements outside of the hours in the market, in the household sector, just not being counted. That even leaves out increases in output in the info sector that's not being measured. I need to go back to something else that you said, which was the productivity boom that we saw in the past: My belief is that a lot of that was mismeasured, too. But over-measured.
When? What period?
I'm talking about the early 2000s. There was the apparent boom from 2000-2007, increase in productivity.
How technology affects jobs
We had the ‘90s boom, and then we had the pop of the internet stock bubble. But, statistically, we still saw a lot of productivity growth after that.
My belief, looking at the numbers, is that a lot of that is mismeasurement of a shift in purchasing from US manufacturing firms to, say, overseas manufacturing firms, which were being picked up as a productivity gain rather than a price drop. Now we're getting to really abstruse stuff, but it doesn't really matter.
That doesn't make me feel good, because I like seeing years of high productivity growth and we haven’t seen as many as I would like since 1973.
I understand, but that actually explains why it is that people are so pissed.
Because in the ‘90s we had high productivity growth driving high wage growth.
That’s right. And then you did not have high wage growth after that … Retailing was in some sense a bellwether industry. Originally McKinsey was writing reports about retailing being a high-productivity industry. And then they realized it was a low-productivity industry. And in fact, real wages stayed low for many years and did not start increasing until Amazon and the other e-commerce companies came in and started taking away the really low-wage jobs which were moving out of retail, into e-commerce and fulfillment as much higher-wage jobs. What I look for is wage growth. If I'm not seeing real wage growth, I assume that I'm not seeing productivity gains, because I'm seeing real wage growth in the areas that I think real productivity gains are happening—whether or not they're being measured or not.
Do I think it's going to spread to the rest of the economy? I do. We know what it looks like. The question is, are we ready for this? Are we ready for telehealth? Let’s just stick with telehealth for a second. You could eliminate big chunks of healthcare workers and costs on the consumer side by shifting as much as you could to telehealth. That becomes a byproduct of the money that's invested in broadband and 5G. And then the question is, are you measuring this correctly? And are you doing what you need to do to make this work? And the case of telehealth, of course, is a licensing problem: being able to get healthcare connections in a state that's different than yours isn't always the easiest thing.
Some people who listen to this will say, “That economist is being flippant about job loss. This is another job-killing technology.”
I think what you want to think about is that we have not seen any evidence of job killing at all. Let’s go to the autonomous vehicle and the truck drivers. Your autonomous truck is going to have to be kept in really good repair. It’s going to have to be kept highly tuned, because it's out there by itself. If you want to do this, you want to run it all night. You’re either going to have somebody sleeping in the cab or you actually have to have something that is kept in as good repair as the average airplane is. Which is really a lot. And so you're talking about having a very large repair force, and you're shifting truck drivers from a dangerous job to a less dangerous job that is better paid.
You also might need more road maintenance people. If you imagine a future where you’ll have cars driving 80 miles an hour, six inches apart from each other, you better not have too many potholes.
That's exactly right.
Some people don't want to switch jobs, though.
I think that's important for us to respect. But I also think people like their lifestyles and they don't want to necessarily switch from a job that is partly physical to a job that's all nonphysical. What the e-commerce example tells us is that we can actually produce a lot of jobs that are of varying types, that are technologically enabled. What the telehealth tells us is that we have a lot of telehealth maintenance people that we didn't have before that are very practical. I think that if we stay on the track that we are, I'm not scared of [inflation]. I find it really weird when people say, “We produced too many jobs this month in the job report, because we’re scared of inflation.” You shouldn't be scared of inflation. You should be scared of low productivity. Jobs are good. Productivity is good.
The future of productivity
I like both. When we look back on this decade from 2030 or maybe 2035, will we say, “That was a high-productivity gain where we sort of stepped up,” or we still be having this conversation of “What do we need to do to boost productivity growth?”
I'm going to take a step back here. I think we're going to discover that a lot more people are being kept out of the labor force by long COVID than we think right now, and that we're going to be running into labor shortages. And as we run into labor shortages, there is going to be incentive for companies to invest in technology in a way that they didn't do before. We are going to start seeing real growth and productivity as investments in technology spread from the digital sector and in a few other sectors into the rest of the economy. And we'll circle back around to healthcare. What we want from healthcare…
Look at those capital investment numbers.
If you look at capital investment over the last 10 years, it's been running at about half the rate as it was in the previous 10 years. Not just in the US, but in Europe—not in China, though. That's really what the big distinction is. China did not have the capital investment slowdown that the developed world had. We need investment in technology. We need a willingness to change. We need investment, not just in information technology, but in the biosciences. And that we need a regulatory structure that is flexible enough to adjust to this.
What’s your best guess? I've brought this up several times in this podcast: Erik Brynjolfsson and Robert Gordon, the economists, have a public bet about productivity growth.
In the end, I've got to go with Erik. Erik has been excessively sanguine up to this point. I think the numbers have, up to this point, leaned in favor of overestimating productivity growth. But I do think that coming out of this pandemic the combination of information technology and biosciences and whatever more investment we do in materials sciences, is going to be extremely important.
Investing in bioscience and materials science
I understand how IT might affect productivity growth. How would the biosciences? Because we would be healthier and work longer?
That is one thing. Another thing has to do with agriculture. And related to that, energy.
I know [CRISPR pioneer] Jennifer Doudna has an agricultural startup.
The agricultural stuff is really important at this point, because if we're moving into a period of changing climate and we're moving into a period where food and water supplies are really important, then anything we can do to increase the productivity of the agriculture sector and also its ability to adjust quickly is just really important. The fact is that we collectively as a global economy have survived the worst pandemic in 100 years, basically without touching the growth rate of the economy. I mean, it touched it, whatever it did, but mainly we kept going. And the reason why we kept going is that we had invested so much in biosciences, especially in the US. We had the technology on the shelf that we needed fully operational. We could say, “Well, it doesn't do exactly what we wanted to do.” That's not important. It was there, it was ready to the degree that people were willing to roll it out.
I think what we're going to find is that we're going to have a lot of other challenges that come up for which having a strong biosciences capacity is absolutely essential. Information technology is great, but it doesn't cover the full range of innovations. The place where we're missing is materials sciences. Other countries have spent more on materials sciences than we have. If you go back to your question about industrial policy, I would say that the main thing that we have to do is actually with semiconductors because semiconductors is basically about materials sciences, is more investment in materials sciences.
The old nanotechnology initiative, despite that people had thought this was going to create tiny machines that built things, it was basically materials sciences.
It was basically materials science. Once again, that's something that we spent some money on, then we stopped spending money on it. It’s still lurking out there as a possibility we may have. There may be stuff on the shelf right now that we can reach out for when we need it. The glass on smartphones was originally a Corning glass that they had made.
They didn’t know what they could do with it.
It was not good. They had designed it to be shatter-proof auto windows. And it was just bad for that. But it was in their drawer. And the thing about Corning, of course, is that they had such continuity in their research capabilities they actually remembered it. I'm on the plus side of this. I think that I'm of the school of, the future happens slowly then all at once.
It's also how we go bankrupt.
Let me tell you a little bit about my theory about innovation, both positive and negative black swans. We have very little ability to predict technology. We have very little ability to predict what the problems we're going to face are. What we do have is the ability that when we have something bad happen, can we ameliorate the negative consequences? And when we come up with a positive, good surprise, can we take advantage of that? We had a big negative happen with the pandemic and we managed to deal with it. The question is, can we take advantage of new technologies to push things forward, or are they going to languish on the shelf? And that's really the answer to your question: chop off the bottoms of the down rungs, boost the top rungs and the overall growth is higher.
Policy for societal resilience
What you've also described there is kind of a societal resilience, the ability to do that. Since I work at a think tank, you work at a think tank, what is the five-point policy plan there?
Let's actually just go back to manufacturing, because that's the one that I've thought about the most. In the broader sense, in terms of regulating technology, don't destroy the goose that is laying the golden egg. You can regulate it and you should regulate it. If you see the things that are wrong, if you have definitive things that you think are wrong and you can say, “Don't do that. We can punish you.” And then you can sort of judge for yourself whether or not people have followed that or not. If companies are doing well, encourage them to expand. Encourage them to expand because that's the best way to make sure that the higher productivity is in more places of the economy rather than fewer. In terms of manufacturing, which is so crucial, make sure that the technology is available at a local level for anybody to use so that they have a chance to experiment with it. The problem is we don't have enough experimentation going on.
How does government do that?
On the state level, you can imagine setting up centers that anybody could come into and use the latest—not a consumer model 3D printer, but the latest production model one, or have access to the latest-model robot, not an older one—and be able to say, “What could you do with this that is different?” Because you want to be able to throw smart people at the technology. One of the great things about information technology, the personal computer, is that it was available to everybody.
What you've described almost reminds me of a World's Fair, where technology can be presented to people and they can interact with it.
We haven't had a World's Fair in a long time, have we?
We've covered this topic in the newsletter. What you're describing is maybe kind of World's Fair, but for small business.
You can imagine that, with spinoffs for it. I'm not talking about industrial policy in the classic sense. There are a lot of technologies that are out there that don't have enough people working with them, that don't have enough financing available at the entrepreneurial level that we want to be able to make sure that they have available to them because then we'll have the creativity that we need to move to the next stage. But having said that, I'm feeling more positive going forward. In the next 10 years, I'm not going to put a number on productivity growth because I'm really getting more and more doubtful of our ability to measure it…
If you did, remember: a number and a date, but not both. That's the classic stock market strategist.
As you know, Jim, I’ve been at this a long time. What I usually forecast is big ups and downs, with the ups being bigger than the downs. How could that be wrong?
🚀 Faster, Please! — The Podcast #10