The US government is spending billions of dollars to build out state-of-the-art domestic semiconductor manufacturing capacity. But spending money is no guarantee of success. In fact, there are already worries that the CHIPS Act passed by the Biden administration isn't succeeding, due to various roadblocks, speedbumps and unforced errors. So what are the odds that it will pay off? And what should we be watching for as evidence of its efficacy? On this episode of the podcast, we spoke with Dan Wang, technology analyst at Gavekal Dragonomics and Adam Ozimek, chief economist at the Economic Innovation Group. This episode was recorded live at Decades, Adam's bowling alley in Lancaster, Pennsylvania, during the #EconTwitterIRL event in April. This transcript has been lightly edited for clarity.
Key insights from the pod:
Can industrial policy in the US even work? — 2:51
What is the test to show that CHIPS is working? — 5:35
Why we need to focus on specific goals— 8:22
Is the CHIPS Act solving a real problem? — 10:33
How does national security factor into this? — 11:38
Is job creation an important aspect of policy? — 17:36
Where China is winning and losing — 21:39
Why China's problem is the opposite to the US's — 30:38
How can the US get better at building infrastructure? — 33:15
Why the US can't just replicate China's strategy — 38:21
Will lack of focus derail the CHIPS Act? — 40:37
When will we have a verdict? — 43:29
---
Tracy Alloway: (00:21)
Hello and welcome to another episode of the Odd Lots podcast. I'm Tracy Alloway.
Joe Weisenthal: (00:26)
And I'm Joe Weisenthal.
Tracy: (00:27)
So Joe, we can't seem to get away from semiconductors.
Joe: (00:31)
No, we can't. And, you know, we recently talked to two of the people basically running the CHIPS Act in a recent episode. But I think if we're gonna ask the question “how is the CHIPs Act going?”, we can't just talk to the people running it.
Tracy: (00:45)
No. We need some outside perspective, and I'm very happy to say that in this very special episode of Odd Lots, which we are recording live at the Decades bowling alley, we are going to be speaking to two of the perfect guests. We are going to be speaking to Dan Wang, visiting scholar at Yale Law School's China Center and tech analyst over at Gavekal Dragonomics, perhaps the only man who's ever read a full copy of the Chinese Communist Party's internal policy journal from cover to cover...
Joe: (01:18)
What's it called again?
Dan Wang: (01:20)
Seeking Truth.
Joe: (01:20)
Seeking Truth, yeah.
Tracy: (01:22)
… And special guest — a last minute addition just to keep our producers and our audio people on their toes — we're also going to have Adam Ozimek. He is of course, Chief Economist at Economic Innovation Group, and also our host for this wonderful event. So thank you both so much for coming on Odd Lots. Without further ado, Dan, you recently moved back to the US and now you're in Lancaster at a bowling alley. Isn't it great that we're all here?
Dan: (01:49)
No one is more surprised than me, Tracy, that we've all gotten over here. If I could heighten the surrealness of the moment just a little bit, I want to point out that my parents are also here in the back, from their home in suburban Philadelphia. They are shy. I hope we can involve them in the bowling.
Joe: (02:13)
Adam, are you surprised to be an economist who owns a Bowling alley and to issue your own tokens?
Adam Ozimek: (02:19)
I'm a little used to it at this point. But I enjoy being an economist who issues my tokens.
Joe: (02:25)
Okay. That's fair enough.
Tracy: (02:26)
So maybe just to set the scene, I mean, Joe and I started talking about semiconductors years ago now because of supply chain shortages, which we've since gone on to explore in various ways. But in the years that have passed since, semiconductors have become a huge policy point, why is this of interest in particular to someone like Adam? Why does this interest you?
Adam: (02:51)
Well, it's part of a broader discussion around industrial policy. And I think that if we're going to be trying to pick specific industries that we want to grow I sort of put myself in the camp of, “I think this is possible, but we have to do it very smartly.” It's not easy. And so my interest in the discussion is trying to make sure we're doing this the right way to maximize the odds of success.
A lot of times, you know, you sort of find two camps. There's the critics of industrial policy who just say, “It can't work, it's impossible. There's public choice problems. It's doomed.” And then you sort of have another side that's just like, “As long as we're spending the money, we are already mission accomplished.”
They might not say that, but that's kind of how they act sometimes. Sorry guys. But I think we need to be critics who think it can succeed and trying to make it succeed by being critical and ensuring that it's done smartly.
Joe: (03:48)
So one of the things, the first time we ever did a Covid-related episode, it was actually with Dan, and I think it was before we knew quite how much it was going to change the world. And it was a very specific, like “Well, what's Covid going to do to essentially electronics manufacturing in China?”
And that was the extent to which we were thinking. I think it was like February 2020. Obviously, then the world changed. We don't need to go over all of those things. But let's just start with this simple point — when it comes from your perspective, can it succeed? Not like the specifics of the CHIPS Act yet or whatever, but do you accept the premise that the attempt to turn around domestic manufacturing of high-end semiconductors in the US is conceivably possible?
Dan: (04:37)
I think it is possible. I think it is conceivably possible. And I think that we are off to a good start with an emphasis on “start.” Now, the Financial Times recently tabulated that there's something like $200 billion of new investment in clean tech as well as in semiconductors after the passage of the IRA as well as the CHIPS Act. And $200 billion is quite a lot. We're seeing these flurries of investments, in semiconductors from the likes of Samsung and TSMC as well as Intel. And we're seeing also a lot of battery makers. We're seeing a lot of solar, photovoltaic makers announcing these very large facilities in the US. So we are off to a start.
I think what I really wonder about is, you know, what exactly are the criteria for success? Is spending money really just sufficient or as you've had the CHIPS Act administrators come on to tell you. You know, there has to be a little bit more. And the test that I would propose for, you know, everyone to think about, are probably to focus on these two things.
The first is, you know, what exactly is the economic criterion for success? If we are going to have a thriving chip industry, if we're going to have a thriving clean tech industry, then, you know, the economics have to work out in some big way. And I think the challenge for a lot of this is that, when I spend a lot of time looking at the clean tech supply chain in China, it is going to be pretty difficult, I think, for the United States to become a lower cost producer than China on a lot of these important manufactured products. But is there some criteria on here that could be met in a bit of a lower way? Are we going to see some self-sustaining dynamics that is not being primarily driven by foreign investment? Are we going to be able to see, you know, the cost curve perhaps with some help from tariffs, be able to get us to some competitive products.
And then the second test I would propose that, you know, we really think about is some threshold of a national security test. Now this is really up for the national security folks in DC to debate what this should really mean. But what I would look at is the structure of these investments. Because a lot of what we see right now in terms of the actual investments in chips, the actual investments in clean tech, that's overwhelmingly going to the downstream industries.
In both of these sectors, we're seeing a lot of investments in the battery cells. We're seeing a lot of the investments in solar modules, but really these are only the assemblies of all of the components that you need. There hasn't been quite so much investment in the critical minerals for batteries in the US. There hasn't been quite so much investment in the solar cells, the solar polysilicone. This is much more difficult stuff.
So if the US is spending a lot of this money, you know, $200 billion, potentially far more than that, mostly to import products mostly made in Asia in China, and then assembling a lot of these things here, well, it doesn't look to me like they are really going to be pushing past that threshold. So for ultimate measures of success, I would look at the economics and I would look at the structure of these investments.
Tracy: (07:41)
Adam, I kind of want to ask you the same question, which is, how are we judging the success of this program? Because it does seem, as Dan just laid out, there are all these different targets. So on the one hand, the CHIPS Act is sort of the poster child for a more activist industrial policy. I know that's a negative term in some economic circles, I've heard people call it industrial strategy instead. So maybe I'll use that.
People trying to fix supply chain shortages, create new jobs, [take into account] geostrategic interests as Dan just mentioned, the list of potential goals here is quite long. So how should we be judging the success of this type of program?
Adam: (08:22)
Well, I think we need to focus. I think that's really important and we need to focus on the goals that actually make sense versus the goals that not only do they not make sense, but they're going to be a distraction for achieving our primary goals. So if you think about, you know, national security interest, I think we have a genuine economic interest in making sure that specific mature-node chips that we rely on for our military, that we have a safe and steady supply with them, doesn't necessarily mean they need to be here.
It means they need to be some of them here and some of them on allies that we can trust. That to me seems like a completely reasonable national security concern that I don't see why even libertarian, neoclassical style economists would reject. I think having the most advanced production here, there's an argument for that too, at least having some of it for the purpose of making it less geographically concentrated.
A lot of the production, what is it, 90% of advanced chips is made in Taiwan? That's a hugely risky situation. So I think we think of that as like insurance. So we don't want all the advanced production concentrated in one place. We want it to be insured. This is an area where we should be relying on allies and we shouldn't see it just as a goal of producing it here, but a goal of making the supply of that more resilient, less risky.
I see those as reasonable policy goals. I see the idea that this is going to be something that creates good jobs, skilled or or low skilled as being a distraction. I think that the idea that we're going to, you know, have more resiliency in a huge variety of chips is not really realistic. I mean, there's so many chips made, so many countries and you just, you're not going to know which ones we need to insure against.
If you look at like the auto industry, the reason they had a chip shortage wasn't because we couldn't make enough chips for them. It's because they canceled all their chip orders early in the pandemic. And then by the time they realized they had made a mistake and they went to chipmakers and said, “actually, we do want those chips.
Tracy: (10:32)
Classic bull whip effect.
Adam: (10:33)
Right. And so how are we going to solve that? And I don't really hear any clear answers about that. And even when I do hear an answer, it's sort of not fully thought out, in my mind. So we're going to create mature node production here? What's going to happen next time something like this happens, right? Are the automakers going to cancel their orders again? If not, then we don't have a problem. And if so, then what are we going to do? Use the Defense Production Act to seize it because it's here? It seems like a solution that's a little disconnected from the problem.
Joe: (11:05)
Well, this actually gets to another thing. Dan, I mean, let's say we all were to sort of accept the national security case for domestic advanced semiconductor manufacturing. Is there another problem that we need that actually needs to be solved beyond that? I mean, can we just sort of have maybe more diversified globally or something like that? Like ok yes, it seems pretty crucial, the military part. What about just going back to the status quo outside of that, would that be fine?
Dan: (11:38)
Well, the wonderful thing, Joe, about national security in the US is that no one can ever define it . You know, the president is very reluctant to define it. The federal judges are always very deferential about how US president defines what national security is. Maybe we should just come up with our own definition.
You know, I think that I take an expansive view of US national security. I think there should be quite a lot more manufacturing jobs. You know, we're speaking in Pennsylvania. This is a site of quite a lot of manufacturing. I value manufacturing for its own sake. I think that I love heavy industry. The heavier the better.
And so, you know, it's really great that we're thinking about these sort of things. If I were able to think about what is a good vision of all of these things, I would ideally like to see, you know, that even if automakers make some big mistakes in the beginning of the pandemic, cancel all their orders, that was no doubt a mistake. That somehow the US manufacturing system is able to be somewhat robust, to be able to be self-correcting enough to rescue a lot of these manufacturers from their mistakes.
Because what I was quite surprised by in the early days of the pandemic was that, you know, in the earlier days of March, 2020 when the US had a pretty big deficiency then of masks and swabs, which were simple products by any measure, there were a lot of manufacturers that were not able to quickly retool and make these pretty simple products that manufacturing employment even up until I think something like the last three months did not, you know, surpass the peak in March, 2020. That it is pretty surprising that both the manufacturers and the workforce were not really able to respond in a pretty robust way and be supple enough to, you know, make a lot of these types of products.
And so I would start there by thinking about, you know, from a workplace issue, less from an optimization, less from an efficiency issue, are we able to build some sort of, you know, workforce economic system that res responds robustly to emergencies?
Tracy: (14:01)
So just on the national security point, I take Dan's point that this is sort of a moving and very flexible goal. But I did see a headline float by this week basically saying that Germany is in talks to limit the export of certain chemicals that are important to chip production. I guess my question is what is the ultimate goal here?
Is it just to cut off chip production within China, or is it to limit chip production in other countries that might be interested in developing it? Is the future goal that we just have, you know, a strategic industry of semiconductors that is concentrated in a few select western countries? I'm being very careful not to use the word western “bloc,” but that kind of feels like we're heading there.
Dan: (14:49)
Yeah, it's a pretty important question, Tracy. And so, you know, what I wonder about with something like the IRA, with something like the clean tech bill is what exactly is the ultimate objective of the United States government? Is it to decarbonize as quickly as possible? Or is it to build out a clean tech supply chain in the US as quickly as possible? Because to some extent, these two are somewhat different intentions.
You know, China does have all the solar as well as quite a lot of the battery technologies. It would be much simpler to decarbonize if you just imported all of that. But that is totally not acceptable in the present day. Now with respect to chips, that is a little bit of a different story. But again the targets here are not terribly clear. You have the National Security Advisor, Jake Sullivan, say that it can't be enough for the United States to be just two generations ahead of China on semiconductors. They really should be seeking some sort of an absolute advantage.
This is someone else on the National Security Council Tarun Chhabra said we no longer want a comparative advantage against China's technology capabilities. The US really needs to have an absolute advantage. Now, if you're removing comparative, if you're in the realm of the absolute, then it becomes quite a lot more difficult to figure out exactly the degree to which the US really needs to be ahead.
But I think in general what it seems like here is that the US really wants, you know, you know, quite a lot of the most advanced stuff, everything else as well. And it would be ideal if China did not get that much further out of here.
Joe: (16:26)
So I want to go back to something where I sense there is some disagreement and get some debate going. And that is essentially about employment as a good thing per se. And I could see sort of [Adam’s] argument is “Look, if the goal is we want to have advanced chips here, then the goal should be we want to have advanced chips here. And the fact that it creates jobs, jobs are good, but that is a secondary goal, and we wouldn't want the job creation aspect to be a distraction from the advanced chips.”
To [Dan’s] point, however, there is this sort of, it seems like a slightly different take, which is that “No, we want to have an economy in which a lot of people are capable of working in complex, heavy industry. And you don't get there if you don't start with employing more people in complex heavy industry in the training.”
And I'm thinking about a recent conversation we had with Henry Williams and David Ochs, economic complexity is a really good thing and is part of what drives wealth. But maybe we could explore this a little bit. Maybe start [with] what are your concerns about the degree to which some of these secondary goals may undermine the main goal? And maybe hear both of your perspective on whether jobs per se is a good part of the ambition?
Adam: (17:36)
Yeah, I mean I think Dan acknowledges that we already are seeing the tradeoff between domestic employment and the main goal when it comes to clean energy, right? I mean, we could be importing more clean energy. We'd be getting, you know, cleaner, greener faster, and we're not — to preserve jobs.
And the reality is the wage premium that a low skilled person earns in manufacturing has declined massively over time. It's just not there anymore. These aren't great jobs anymore. It's just not the case. We don't have an economic interest in taking someone from services to manufacturing anymore.
I think a lot of what we're seeing is a relic of 10 to 20 years of a bad economy, right? I think we had a long, long, Great Recession, and before that we had the China shock. So it's been a long time since we've seen a legitimately good labor market, and I think people lost faith in that, they lost faith in basic macroeconomic policy's ability to generate decent wage growth for people with less than a college degree. And I think that's just misplaced. I think the fundamental mistake there is macro policy. I don't think it's industrial policy. I don't think the solution to a 10-year recession is getting more people in manufacturing. And I think if we do that, we don't need to do that.
Joe: (18:57)
Dan, do you want to defend the idea of job creation as part of industrial policy?
Dan: (19:03)
Sure. Well, you know, I think it's impossible to disagree with Adam that the wage premium here is not good in terms of a lot of manufacturing. And I think it is pretty difficult to entice these workers who are working in semiconductors to go from Silicon Valley bean bags and then go into these fab rooms where, you know, if you ever go into one of them, they have a very strange light. It is yellowish, purplish light. You're sitting there, you know, looking at, you know, these transistors for eight hours a day.
It is pretty strange work to try to do something like this. Whereas if you can, you know, get into those bean bags or if you are able to just work as a home healthcare worker where your wages can be just as high as working in these grueling factory jobs, I think that is pretty difficult to entice a lot of people to take a look at the situation and say, “well, let's do quite a lot more manufacturing.”
You know, I'm sitting at a law school now. I'm hardly turning wrenches all all day. And so I want to be the first to put up my hands and say you know, this is quite challenging for me to envision myself doing something like that as well. But I think, you know, in general, what I would love to do is, you know, imbue manufacturing with a special sense of dignity.
Let's imbue complexity with a special sense of economic dignity. And we should say that, culturally in general, that we should, you know, exalt manufacturing, that we should exalt economic complexity, that we should exalt the types of technology that involve a lot more of this creation, because I think that is good for both the economy as well as for national security.
Tracy: (20:48)
So we managed to mention the bull whip effect and beanbags. So we just need to mention ‘mint the coin’, and then we'll hit like all the traditional talking points...
Dan: (20:59)
We already talked about his token operation.
Tracy: (21:00)
That's true, yeah. So just on this point Dan, maybe you can explain. So one thing I've always wondered is, you know, we talk about US-China rivalry when it comes to semiconductors, but China is far behind a lot of other countries when comes to semiconductor technology, but it's doing pretty well, as you already mentioned, with clean tech.
What are the differences between those two industries as they relate to China's economy and social fabric? Because I think that'll also give us some insight into whether or not this type of traditional ‘industrial policy’ will actually work in the US.
Dan: (21:39)
So you know, maybe think a little bit about where China is on most technology. And so if I can give a very quick snapshot of where China is on most of its technology endeavors, roping up clean tech as well as semiconductors. What I would say is that at this point I think China is pretty competitive with the US in manufactured products outside of two big areas. The first is semiconductors, where China has built pretty much the basics of competition in all of semiconductor production. But it is leading in absolutely nothing. That it is at best 10 years behind the market leaders here. China is also really weak in something like aviation. So in China's answer to Airbus and Boeing, it is really behind.
But outside of these two big areas, again, manufactured technologies only, I would say that China has broadly caught up to the west-at-large on most manufactured products. It is leading the US and Europe in most clean tech, by which I mean solar, wind, batteries, hydrogen electrolyzers. It is leading in all sorts of these boring industrial products, things like hydraulic pumps, which would never grace the headlines of, you know, fine papers like Bloomberg, but...
Tracy: (23:04)
The Hydraulic Pumps Act isn't coming anytime soon?
Dan: (23:07)
But we definitely need something like that. But China is leading in all of these broad manufactured products. Where I think it is weak relative to the US is that I think that any scientific area, any technology sector that involves the complex integration of different scientific disciplines, China tends to be pretty weak.
So in semiconductors, that involves the integration of chemistry, electrical engineering, you know, computer science aviation involves the integration of material science aerodynamics many other things. This is something where the US is still really, really strong at you know, building these advanced manufacturing products.
Where the US tends to be weak is where the science is pretty mature, but all of the execution risk lies with manufacturing. So anything where the manufacturing operation is pretty complex. If I'm thinking about something like putting together a battery cell. It involves about a dozen different steps. Everything from cell filling to final sealing, all of these demand, you know, perfect handoff between each one of these steps. And this is where the Chinese are really, really strong. They learn from building iPhones, building complex electronics. They're just really good at building products, with high intricacy at high volume. And so this is where I would love to see the US become a little bit better at this more volume production.
Joe: (24:28)
Adam, I want to go back to something you said, which I think is really key, which is that perhaps in some the diagnoses of the US, we forget that we had really many years of slack labor markets, poor demand, maybe 10+ years of weak economic growth, even going back to well before the 2008-2009 economic crisis.
One thing that strikes me now that I've been trying to think about is, okay, we're in this tightening cycle with the Fed, etc., but there's still a lot of money coming, high multiplier money, it seems like building all these factories, batteries, chips, etc. To what extent, looking from a macro lens, do you see some of this investment allowing the US to maintain robust job growth, strong economic activity, strong wages, etc., and sort of helping us avoid going right back into the old slump?
Adam: (25:24)
Yeah, I mean, I think ongoing capital investment great. It's nice to see that the interest rates haven't fully shut that down. And I think that that's sort of the risk that we're playing with right now. I think all eyes should be on growing GDP. How do we increase GDP? How do we do more? Because the reality is we have had, we've had tight labor markets over the last two years, but a substantial part of that has been reduced labor supply. And also just the speed at which things moved, right?
So I think that looking forward the next five, 10 years, I think when we think about how many people are working, we can do better. The labor market can do better. And so we shouldn't look at the amount of people working now and say we've hit the limit. That's the capacity. It's over. It's really a lot of things happening that, you know, short-term NAIRU is a lot realer than it ever has been.
And so I think we should think about capital investment growth and higher employment all as that's the next five to 10 years. That's what we should be aiming for.
Tracy: (26:24)
So just on this note though, is there a risk that we're front loading CapEx or pulling forward too much investment at a time when the economy is arguably running hot? And at the risk of not having “Oh, God, I'm about to veer into MMT territory...”
Joe: (26:44)
The water's warm...
Tracy: (26:47)
When the economy is slowing, maybe we won't have that fiscal policy or investment lever to pull?
Adam: (26:55)
I think the economy's confusing right now, and the resilience of spending, the resilience of durable good spending. And I wouldn't be surprised if there are a lot of factory owners and a lot of manufacturers across the country who don't know yet what the permanent trajectory looks like. And so given that there's uncertainty it would be surprising if there weren't people building factories that it's going to turn out that they don't need. You just hope that the economy has enough strength to sort of reabsorb that stuff as companies do realize some of that was malinvestment — to put my Austrian hat on...
Joe: (27:32)
Dan, going back to actually some of the national security questions, is there a level at which the US can feel comfortable? And so in terms of like, okay, “we have hit enough.” Is there any way of even knowing or being able to quantify the right level of factories that are up and running? Yes, probably the majority of manufacturing is gonna be in Taiwan or China or elsewhere in Asia, but it's like “oh, we're good here… we actually did this.” Do you have any sort of analytical way of like figuring out what that number is?
Dan: (28:05)
I think that is pretty difficult to say. If I had to come up with my own test, it is that, you know, the United States regularly runs into problems of oversupply of a lot of these things, rather than undersupply and having, basically these spiraling cost curves chasing after fewer and fewer goods. You know, again, it is hard to say what the US national government declares to be national security.
But you know, what I would say is, hopefully that the folks at #EconTwitterIRL can agree on is bring down the cost curve for infrastructure that, you know, it's still kind of mind blowing that these subway extension lines in New York City cost about 10 times more than in European countries, that it becomes super difficult to build these solar farms, to build these sorts of transmission grids.
I never knew what an obscene four letter word NEPA was until folks at IFP told me. So NEPA, CEQA, its close cousin, just really block all sorts of development in the US. And if you don't have, you know, some sort of a robust system to save automakers from their mistakes, although these automakers make, you know, probably too many mistakes, to really try to save them from, but you know, unless you are able to have, you know, robustness bring down these cost curves, have regular problems of oversupply rather than persistent undersupply where, you know, prices just keep rising to chase after these goods where it's not at all clear where a hundred million dollars in the New York public school system goes, it's not very clear where, you know, tens of billions, hundreds of billions of dollars of infrastructure investment goes, then I think we can be a little bit more at ease in thinking about having a robust economic system.
Tracy: (29:57)
And just on this point, there is a perception out there that when it comes to China, it's a command economy. And, you know, Xi Jinping can wake up one day and say, “we're going to make semiconductors a strategically important industry.” But actually the way that happens, it usually isn't a bunch of money thrown at that problem. It's sort of a wink and a nod to the banks and the credit providers like, “Hey, you should lend these guys some more money” at cheap rates. I guess my question is, what kind of policy response would we expect to see from China on this front, and would they start to tweak their own way, their own version of industrial policy in response to what the West is doing?
Dan: (30:38)
Yeah, I tend to think China is mostly pretty focused on building all of this capacity without thinking too much about these other competitive forces, and you know, that creates a lot of these problems. It creates mutual issues of oversupply, but I think the China is mostly thinking about its own internal system rather than trying to figure out how to make the rest of the world more happy.
So when the Chinese decide that, you know, semiconductors are a critical technology, or that solar is a critical technology, as you say, the central government would do something like name solar a strategic emerging industry, as the state council did in 2010, it would trigger a vast cascade of subsidies which leads to business creation. It encourages the local governments to give free land to favorite industries, have the banks lend at very cheap rates. Sometimes these, you know, direct grants.
But you know, in general, what the Chinese tend to do is have oversupply. They have no ability to stop themselves from building. And I think, you know, if I had to choose, I would, rather choose that set of economic problems that trigger that has made it so that China builds basically as much renewable infrastructure every year as the rest of the world combined, that these Chinese local governments cannot stop themselves from building more favored industries in something like solar or wind farms.
That what they really try to do, is they can't stop themselves from building. Thy build even when there is no interconnection, when these farms aren't really connected to the grid, they build you know, hydroelectric dams that displace millions of people they build even when the central government tells them to stop building because it just wants this sort of capacity and employment out there. And that is just what the US is up against. And I wonder how these systems can be more adaptive.
Joe: (32:38)
So it really does feel like we have the opposite problem where people are always writing some posts about how terrible we are at building things. And then you have Dan talk about they literally can't stop the building even when they say “stop the building.” Is it really as simple as like, “oh, there's some laws from the Carter era that we have to get rid of,” and then suddenly we could build subway and rail as fast as they do in Europe and Asia? What do we get to do to get the US building again? How do you like rank the priorities or get good at that?
Adam: (33:15)
I mean, I think there are important laws to change, but I think what we have to understand is that a lot of the blockers, like why are those laws there, right? We can't just erase them and say they should be gone to be gone. A lot of the blockers to lower cost are there because someone wants them to be there. And because we don't have enough, you know, bipartisan cross ideological agreement that these things should go, you know what I mean? All we have to do is get enough people to agree that Buy America is a problem, and then we can get rid of Buy American provisions. But is the law the problem? Yes, to an extent, but the problem is for whatever reason, we can't overcome the forces that push for it.
Joe: (34:00)
You know, but there's also, I mean, Buy America, I presume, is not why it took forever with the Second Avenue subway, which I don't know if they even exists, or maybe it has a few stops and why there are large swaths of the country where they can't put electrical wires and stuff. It seems like there's more things.
Adam: (34:18)
Sure there's a lot of problems. And behind most of those problems, there's some party pushing for them. And I would argue that what we need is more ideological alignment among the people who wish to push back against those things. You know, like NEPA is a huge issue, right? But it exists for a reason, it's defended for a reason.
There are people who sue using NEPA because, you know, they want those outcomes. And so we need to have more agreement. Here are the set of problems and sort of work in a bipartisan way to push behind them. And you have to be realistic about people are on the other side of the table of the everything bagel problem, and that's who we're up against if we're gonna get rid of it.
Tracy: (35:19)
Just going back to the beginning of this conversation, when we were talking about the goals of the CHIPS Act and the IRA as well, it does feel like there is a tension between a recognition that these are strategically important things for the US — whether they're semiconductors or clean energy tech — and then the ability to actually scale these and sell them in a way that is appealing to American people and corporations. Is that tension insurmountable? I'd like to hear from both of you on this point.
Dan: (35:54)
I think it is probably going to be a pretty serious tension. And I think that the US government has to work through all these tensions. Some of these things will be resolved through the executive process.
So I think about something like, you know, the Treasury Department will give guidance on which electric vehicles are actually subject to the full $7,500 in subsidies from the IRA. And so it has to thread these needles. If it is, you know, a little bit too tight, then you know, the EV industry uptake will be pretty slow because most of these batteries are used with minerals processed in China.
On the other hand, if it is too strict, then no one will buy electric vehicles. If it is not strict enough, then none of these automakers feel that they have to get rid of their China dependence on all of these batteries. And so these are basically a lot of these tensions that you know are going to be pretty difficult to solve.
Tracy: (36:57)
Adam?
Adam: (36:58)
I think there's a couple of important facts that are that matter here. One is that we have a huge cost gap, especially when it comes to advanced semiconductor production. You know, TSMC says it's 40% to 50%, so it's too expensive to start. We used to make it here and we don't anymore. And that's for a reason. And the reason is because of costs and, you know, it made more sense to make it in other countries.
So you have to be laser focused on those costs because of the third reason, which is this is a globally competitive industry. It's ruthless, and you have to be focused on costs, and you have to be focused on prices. And policy should reflect that. And if policy doesn't reflect that, I think that's a problem.
Dan: (37:39)
If I can amplify that point a little bit, I think that what is really strange when it comes to something like clean tech is that I think the US is in a very strange position where it is trying to engage in technological catch up with a lower wage competitor, which is China and that China makes solar photovoltaic panels that are both cheaper as well as more efficient than the US. And often the same goes for batteries.
And so, you know, usually one of these things is not true, that you know, you are either technologically ahead or you are cheaper, but the US really has a hard time here, which is why I want to echo Adam's point that you really have to focus on bringing down this cost curve.
The other slightly novel thing is that, you know China developed technologically by embracing a lot of foreign investment into China. These involved, you know, Apple through Foxconn building enormous electronics factories in China. Tesla has a fully-owned factory in Shanghai that, You know Intel, Dell, whatever, all of these American technology giants invested a lot in China. And Beijing was extremely welcoming in their investment and has not retaliated against them throughout the many years of President Trump's trade war.
By contrast, the US has become pretty hostile towards attracting investment from, you know, what is the technology leader in a lot of these spaces. China, when it comes to solar and batteries especially. And so it is pretty hostile towards Chinese battery makers from setting up in Virginia including in Michigan.
And so I think, you know, the US is making this bet that it is going to be able to be okay in something like batteries by, you know, mostly working with allies in Japan as well as South Korea. And perhaps that is all very fine, but I would also encourage it not to be, you know, boxing with one hand tied behind its back to be a little bit more like China, and then try to get as much investment as possible, solve and mitigate these national security problems where they exist, but don't really, you know, reject foreign investment from the technology leader.
Joe: (39:46)
So going back to chips specifically mentioned at the top, Tracy and I recently interviewed two of the top people running the CHIPS Act. And it was totally lucky timing because like a week before that episode, our colleagues on the Bloomberg Opinion side who were like, you know, separate wall between us, they totally savaged the state of the CHIPS Act. And you know a lot of these accusations this is becoming like a progressive Christmas tree of like environmental concerns child childcare, etc.
How concerned are you about some of these other elements? Or how overblown [are they] or where do you stand on some of the other priorities that some are saying are seeping into the CHIPS Act?
Adam: (40:37)
Yeah, I mean, I think a hundred percent that there are other priorities that are just — they're just going to raise costs. Like this is what they do. There's no world in which forcing project labor agreements onto the construction of these fabs doesn't raise costs. And you have to acknowledge that that's the problem. You gotta be realistic about that. So I think that they are absolutely in tension.
I want to be a two-handed economist here, and there's a lot of stuff that they are doing and saying that I think is good. There's a recognition from the administration that we are going to need to nest semiconductor production within globalization, and there's a realism there and a desire to work with allies on this that I think is missing from a lot of the other discussion around this.
So we're not going to make it all and globalization is not the enemy here. Globalization is a key factor in why Moore's law has continued over time. If we didn't have globalization, Moore's law would've been dead a long time ago. And so you cannot run away from globalization when it comes to semiconductor production. So I do want to say that I do think the administration has been good on that. It's the other things, it's the Christmas tree stuff that I think is a problem.
Dan: (41:46)
On the other hand everything bagels are delicious. I don’t order plain bagels. And Christmas trees are beautiful. I agree, Adam, that it is really difficult to try to make everything here at home. But you know, I wonder to what extent that these childcare provisions, that these other labor provisions are really going to be mostly inconveniences that these companies are going to be able to write off or whether they actually are crippling.
It may be a little bit too early to tell, but I agree with Adam. Certainly it is pretty difficult, which is why I go back to the test I laid out at the beginning. But we are seeing as already huge amounts of investments. Is it going to be invested in most of the right things? And, you know, is there going to be some way to, you know, is the US going to be able to say that these economics are going to be pretty self-sustaining over the longer run?
Tracy: (42:44)
So I just have one last question, sort of a classic interview question, which is in 10 years time, or I don't even know if 10 years is long enough because we're talking about technologies with very, very long lead times. So I guess I have two questions. One, what timeframe do you think is reasonable to judge the success to declare a verdict?
And then two, what are you looking out for? Is there something specific where you would wake up in 20 or 30 or 50 years’ time — hopefully we're all still around — and say this has either been a massive success or a desperate failure.
Dan: (43:22)
I look forward to getting back into this bowling alley in 10 years, one decade from now.
Tracy: (43:29)
Decades! Perfect.
Dan: (43:29)
Decades, yes. Or perhaps many decades. You know 10 years is not a very long time, as you say, Tracy, technologically and politically, that also tends to be not too long of a time for the Chinese that is only two turns of the five-year plan. And so I think they judge things on a slightly longer time scale. Let's give it at least three five-year plan cycles.
What I slightly worry about is that, you know, there is quite a lot of investment in things like solar and batteries and also in semiconductors in the US today. What I worry about is that, you know, somehow TSMC facility in Arizona turns a little bit more like a showcase factory that doesn't end up producing too much.
And I think the example here is something like Apple's Mac Pro factory in Texas, which doesn't have very high volume after all this investment, a lot of of firms will fail, I think that is certainly going to be the case even two, three years from now. You know, we take a look back, you know, if, you know, one of the parties points out a lot of these failures and says, you know, this is why industrial policy cannot work.
We're getting a lot more Solyndras. They're still talking about these Solyndra failure of about you know, 15 years ago now. That politically the appetite disappears to continue investing in these sort of things and that is something I am worried about. And so therefore, I would say that, you know, we have to think about these things on the longer timeframe.
Joe: (45:04)
Adam?
Adam: (45:05)
I would say it depends what we're talking about specifically. So if we're talking about advanced semiconductor production, it'd be nice to have a little less geographic concentration there, both onshore and friendshored.
When it comes to mature node chips, I'd love to see more direct accounting for here's why we need it, and here's how we've improved resiliency when it comes to solar. What I'd love to see is the cost curve come down. I think that's everything. That's everything. And I don't care about making it here.
I don't care where we make it. I care that the price comes down and I care that people install it. And I think that we should think about the technological frontier as being something really important here. And CHIPS, to be fair, is spending billions of dollars on R&D too. So I'd love to see that some of that money helped move the technological frontier forward.
Joe: (45:54)
I think Alec Stapp is here somewhere. He's always posting the solar chart, and it always looks very encouraging that that line is going down. So that does seem to be happening. One last thing, and then I want to open [it] up and then I think we can go to Q&A. But Dan, you know, I'm really struck by the point that you make about, it's kind of rare to be trying to catch up to a country that both is cheaper and more technologically advanced. Like usually maybe there's one the other, and you exploit the other.
Should we be doing things like export discipline and subsidizing the companies that could produce here and sell well in the global market? Should we be thinking about that type of thing for the US? Or should we be thinking about, you know, more buyer of last resort things. Like that seems to have been a part of the sort of when we have had successful military driven semiconductor policies like back back in the Cold War?
Dan: (46:50)
I would love for US firms to have export discipline, which is the strategy for our firms in Asia to sell to the rich American market, and then you know, that gets them to raise the quality of their products. The challenge for the US is that there is no bigger and more sophisticated markets, so it's hard to figure out who is able to discipline the American firm.
And I think to add to the point of, you know, how, you know, the US is in a bit of a tougher position. Again, I would think, you know, slightly politically the challenges here, to go back to the example of CHIPS, whether the Chips Act is working, I think about something like TSMC in Taiwan. And so TSMC is by miles and miles, everyone's favorite employer in Taiwan. Engineers in Taiwan would say, “I would sell my liver to go work for TSMC”.
You know, it's much more difficult to think that people would say that of any chip firm over here. You know, Taiwan is periodically in a drought. It is an island where they give the fresh water to TSMC and make the citizens drink the treated water in a drought. They park a lot of these firetrucks outside of the facilities of TSMC in order to make sure that TSMC has fresh water to produce the chips powering our iPhones. And so, you know, that is a different sort of thing where politically, is it really easy to imagine that TSMC'S facilities in Arizona, a fairly dry state, is able to park a lot of firetrucks outside of TSMC'S facilities in a drought?
Joe: (48:31)
We're going to be doing an episode soon with an Arizona alfalfa farmer, so he will be making the case that we should not.
Tracy: (48:40)
But thank you, Dan, for leaving with us with that wonderful vision of late stage capitalism.
Joe: (48:45)
And thank you to Adam Ozimek for joining us for this conversation. We welcomed questions from the audience, and if you'd like to hear those, we invite you to be on the lookout for live recordings where you can be part of our audience in the future.
Tracy: (49:19)
Joe, I found that conversation fascinating. I loved it.
Joe: (49:21)
It was so great. I appreciated both perspectives. I like Dan's case for we should do big heavy stuff here, right? I guess I have some sympathy for it, but no, that was great. I really enjoyed it.
Tracy: (49:37)
One thing that I think is pretty special about the discussion is it has to be a first for a semiconductor policy discussion in a bowling alley?
Joe: (49:45)
And [for] the listeners who are not here, we're on this stage, we're looking out in bowling alley, we're looking at this beautiful light fixture in this old building. I had the exact same thought. No one is having this conversation about first world industrial policy in this context before. So if nothing else, [it’s] quite an accomplishment. It makes me really bullish on America
Tracy: (50:09)
We can draw the line between bowling pins and pin diodes. How about that?
You can follow AdamOzimek on Twitter at @modeledbehavior and Dan Wang on Twitter at @danwwang.