Odd Lots - How to Make the U.S. Semiconductor Industry Boom Again

Episode Date: May 3, 2021

This year, everyone's become aware of the hollowing out of U.S. semiconductor manufacturing capacity. Whether it's the rise of TSMC, the stumbles at Intel, or the inability of car companies to acquire much-needed chips, semiconductors are becoming a major political issue. But how can you actually turn things around? What would the right policy mix look like? On this episode, we speak with Alex Williams, a research analyst at Employ America, and Hassan Khan, a tech procurement expert with a PhD in semiconductor policy, to discuss how the new administration can make American chip manufacturing roar back to life.See omnystudio.com/listener for privacy information.

Thanks for listening to Odd Lots. Follow the show on Amazon Music for more future episodes or just ask, Alexa, play the Odd Lots podcast on Amazon Music. Today's show is brought to you by Vanguard. To all the financial advisors listening, let's talk bonds for a minute. Capturing value and fixed income is not easy. Bond markets are massive, murky, and let's be real. Lots of firms throw a couple flashy funds your way and call it a day. But on Vanguard, at Vanguard, institutional quality isn't a tagline. It's a big line. It's a very big. It's a firm. It's a few. commitment to your clients. We're talking top-grade products across the board of over 80 bond funds, actively managed by a 200-person global squad of sector specialists, analysts, and traders. These folks live and breathe fixed income. So if you're looking to give your clients consistent results year in and year out, go see the record for yourself at vanguard.com slash audio.

That's vanguard.com slash audio. All investing is subject to risk vanguard marketing corporation distributor. Hello and welcome to another episode of the Oddlots podcast. I'm Joe Wisenthal. And I'm Tracy Allaway. Tracy, we're doing another semiconductors episode today. There's no escape. Once you're in, you can't get out, right? We're going to be doing semiconductors episodes for the rest of our career, I'm sure. But, you know, I think there's actually like two reasons, maybe three or maybe a bunch, why we keep going back to this well. I mean, for one, there's just a lot of news. Like so many things have happened, whether it's sort of new ventures from, say, Intel or new spending plans from Taiwan Semi or new legislation being proposed. So there's all kinds of new news.

But I also think it's just a really good, it's a good topic for us. Yeah, I think it's an interesting mix of technology and sort of business competition and also government policy. And we've spoken a lot about this, but government policy plays such a big role in sort of incubating semiconductor industry. And I think we spoke about this probably the most with Taiwan and TSM. But even in the U.S., you have that history of government support nurturing the early stages of the semiconductor industry as well. Yeah, that's exactly right. Like, it touches on all these things, technology, competition, markets, stock market.

But as you say, you know, there is a long history in China now currently, Taiwan, Japan, the U.S., of governments playing a role. And I think there's like this sort of broader macro thing, which we talk about all the time of sort of more activist role for government for democratically elect our elected leaders, like thinking, taking a more active role in shaping the economy and managing the economy. And so like because of that and because of the history of chips, semiconductors fit right into our wheelhouse. Yeah, it's sort of a nice petri dish, I guess, for a greater fiscal role for governments. Yeah, exactly, exactly right. And every time we do semiconductor episodes, people listen to them and they want more. So we are going to deliver. So here we are again.

Yeah, so here we are again. So, you know, like I've forgotten all the ones we've done. We've done the fall of intel, the rise of the... of Taiwan semi, China's own ambitions, the stumbles of the U.S. I don't know if we've like really done one yet on what it would take for the U.S. to really have a sustained recovery in its semiconductor manufacturing prowess. And of course, we know that that's important because we have this acute semiconductor shortage, particularly affecting autos, but also there is this sort of long term.

Maybe it's a defense. maybe it's a geopolitical concern beyond this acute moment, that it is problematic that such an important piece of basically everything that gets built is almost entirely important. I'm pretty sure we've touched on it in various episodes, but you're right. We haven't done a whole episode dedicated to U.S. policy in the same way that we've done it for Taiwan or China. Yeah, exactly right. So today, I'm very excited because we're going to finally talk about some of the history of semiconductor policy in the United States and also perhaps what policies could revive the domestic industry. All right, let's do it. All right. So I'm very excited to speak to

our guests. It's sort of actually the perfect combo because we actually have two guests today. One of them is an economist and one of them is a technologist. Alex Williams and Hussein Khan wrote a piece for Employ America back in March titled A Brief History of Semiconductors, How the U.S. cut costs and lost the leading edge. And Alex is an economist and research analyst at Employ America. Husson, he's a PhD from Carnegie Mellon on semiconductor policy specifically, and he now works in tech in both operations and procurement. So between the two of them, kind of the perfect way to talk about the macro and the micro at once.

Alexson, thank you so much for joining us. Really excited to be here. Yeah, glad to be here. So what is it from your perspective? I mean, as I said, you guys wrote this piece on Employer America, which normally is sort of very focused on

bigger picture, macro stuff, labor market stuff in general. Why, from your perspectives, is semiconductors as sort of a thing that fits into that rubric? So I think the big overriding theme behind this series that we've been doing on semiconductors,

we have two pieces out so far. We have a third one that should be ideally coming out right around when this episode goes public and then a fourth one after that. But the idea is that- You're just like us. Once you start the semiconductor series, you can't stop. It's so true. We originally were going to write one piece and then it was so big and then it kept going. Same with us.

Yeah. There's something about this industry that every time you look a little bit closer, there's an entire other world that opens up. All right. Sorry, keep going. I didn't mean to interrupt you. Oh, no worries. So really what we wanted to talk about was semiconductors as a lens for thinking about sort of the intersection of, you know, industrial policy, trade policy, stagnant demand over the period since the 2000 crash and dot-coms and things like that. And the idea was that we are a labor market focused institution, but industrial policy, you know, of its nature has a strong, you know, labor market focus.

and that the idea is that you're directing the economy to make best use of its resources, even if sort of short-run profit considerations or efficiency considerations, don't necessarily support that. So it seems, you know, a little strange. We write about the Fed a lot. We write about labor markets a lot. But the idea is that we wanted to bite off something a little bit bigger to say, okay, cool.

We do care a lot about labor markets, but part of that is upskilling, and part of that is focusing economic policy around specific industrial. objectives. And when you do that, you naturally, you can't help but intervene in the labor market to do so on the one hand. And on the other hand, you know, there's been real sort of bipartisan support for sort of doing something about semiconductors. Like there's really, regardless of what political party you look to, there's, you know, energy behind doing something about this situation. And we wanted to kind of spell out, you know, from a perspective that sees. these labor market outcomes as critical and not just accidental. We wanted to spell out what that

kind of policy intervention should look like. If I can add one thing on just the industrial policy lens, it's really interesting because the semiconductor industry has been at the forefront of a lot of these debates in the past. And if you revisit the industrial policy debates of the 1980s, when they were talking about saving steel and auto, they weren't focused on the semiconductor industry as a place that needed intervention. And so I do think given this moment in time and the focus on the industry is a good chance for us to reflect on, you know, we had a chance to do IP in the past and we chose not to because we thought, hey, the booming industries will save us.

And we're kind of back at that same debate with an industry that that was ignored in those previous debates. So it is a good chance for us to kind of revisit the history too and say, what can we learn from the past debates that we had? So Alex mentioned this idea of industrial policy being tied to labor markets and also allowing an industry to sort of look past short-term inefficiencies for the sake of like longer-term competitive gains. Can you talk a bit about how that relates to semiconductors specifically? Like, why is that important for the semiconductor business model?

One of the lessons that gets lost when we talk about offshore. of production is that it's not just moving a facility offshore. You actually move that expertise offshore. So a factory is not just a box with a bunch of machines in it that are outputting widgets. It is that CAPX, but it's also the people who come together to produce out of that factory. And so one of the major consequences of sort of the last 30 years of the like offshoring revolution in manufacturing from the United States perspective, is that we also lost a lot of latent expertise within the economy. So actually, my advisor has worked with other researchers and students

that showed that when plants shut down, a lot of the folks who are in the scientific research department leave the industry. They don't patent in the industry anymore. So now you have scientific expertise exiting the industry entirely. And similarly, think of like blue-collar workers who, you know, they might be doing servicing of machinery, or they might be floor operators, they have to go find work elsewhere.

And they have a lot of latent knowledge about processes and how to actually build things that gets lost when they go work in a different industry altogether. The upshot is that that human capital is not just like jobs, but they're actually the people who figure out solutions on how to make, how to push technology forward, right? And not just from a academic perspective of pushing technology forward with invention, but actually with, you know, tweaking on the shop floor and translating research and development ideas into production

capabilities. And the role that policy can play in this is, you know, after say the 2000, you know, dot com bubble bursting, you had a situation where, you know, in a lot of sort of economic models, you think, oh, there's a crash. It's going to force out the less competitive companies so that the more competitive ones can restructure themselves and win the day. But the issue is, is that you had substantial labor market attrition. So this sort of, you know, sort of squeezing effect like an accordion where people get pushed in and out of the labor market creates just an incredible excess waste of capacity when under a policy-supported softer budget constraint regime, you could actually preserve and inculcate that, you know,

that human capital, as it were. One of the things that you guys talk about in this piece, sorry, and as you mentioned, there's two pieces by now, but I'm looking at. at the one that you wrote in March. You draw this distinction between industrial policy and I hadn't heard the term before, but it's very useful science policy. And I think, you know, would people think, oh, what's it going to take to revive U.S. industry or revive tech expertise like, oh, what about R&D and spending? And you sort of put that in this category of science policy. But can you talk about the sort of the distinction between these two things and the some of the approach. that have been taken in the past because this isn't the first time that there has been anxiety

about chips elsewhere sort of pulling away from us. So I'm going to say one quick thing before leaving it to Hassan to talk about the history of this split. But I think that the core thing to remember here is there's this idea that technology is just some overlay that sits on top of existing productive processes, whether you're thinking of like, you know, sort of total factor productivity or other ideas of technology, that it just sort of is this fairy dust sprinkled over existing production? The issue is, is that for technology to really be incorporated in the capital stock, it has to feed out through and become incorporated in the actual physical products used in production. And so if you don't have an industry that is booming that is getting

turnover in CAPEX, where older technology production equipment is being replaced with, newer technology production equipment, the things that you know how to do with the frontier don't necessarily translate into production outcomes. But I will leave the history of, you know, this situation to Hussin. What Alex said is very apt. I think where the policy mistake happened in the past was this framing of America needs to maintain its lead. In order to maintain our lead, we're going to fund R&D. The production and competition and aspect of technology will be left up to firms in the market. And so there was a broad policy consensus to say we'll fund R&D, whether that's through the military or through institutions

like the NSF or National Labs. The problem becomes that there is this translation that has to happen between a researcher who is, you know, doing cutting edge research and what is commercializable. So just as one small anecdote, as part of my work on alternatives to Silicon Seamus in graduate school, the industry was evaluating proposals by different academic researchers. And a researcher got up in front of an audience of industry researchers and was very confident that I solved your problem, I have your next transistor, and went through and presented all sorts of data on its ability to operate and showed that it was better in some cases. And the first question was, what temperature does your transistor operate at?

And very confidently, the researcher said, oh, 77 Kelvin, it's cooled by liquid nitrogen. So just very, spend two seconds thinking about it, that's not something you or I could put in our laptop or cell phone. And that's sort of, you know, but that's kind of the, where the disconnect is. We fund a lot of researchers who have leeway to go do all sorts of crazy things. And they do invent very interesting ideas. But at the end of the day, to take that from the realm of science to the realm of technology requires a lot of translation and attention to, you know, different types of variables. Like, can I make a million of these? Can I put them in people's homes?

and something like a transistor at, you know, this liquid nitrogen cooled fails that test for a technologist, but maybe not a scientist. Today's show is brought to you by Vanguard. To all the financial advisors listening, let's talk bonds for a minute. Capturing value and fixed income is not easy. Bond markets are massive, murky, and let's be real. Lots of firms throw a couple flashy funds your way and call it a day. But not Vanguard. At Vanguard, institutional quality isn't a tagline.

It's a commitment to your clients. We're talking top grade products across the board of over 80 bond funds, actively managed by a 200-person global squad of sector specialists, analysts, and traders. These folks live and breathe fixed income. So if you're looking to give your clients consistent results year in and year out, go see the record for yourself at vanguard.com slash audio. That's vanguard.com slash audio, all investing in subject to risk vanguard marketing corporation distributor.

Can we get into like the actual historical aspects of some of this. So just in terms of the industrial versus science policy split, you talk a lot about it in relation to what happened to the semiconductor industry in the 1990s. Could you describe that period and exactly what happened and how it sort of set the U.S. up for being less competitive in the 2000s? To come to where we were, the decisions that were made in the 90s, it helps to take a step back. So at the founding of the semiconductor industry, you have this breakthrough at Bell Labs, and the military is extremely interested in this. And it plays a very heavy-handed role in shaping the diffusion of transistor technology across basically the entire economy. It requires Bell Labs to

openly license it, and it requires second-sourcing contracts for any firms that are manufacturing transistors. The end result is that because of the military implications of this technology, we had this de facto industrial policy saying we need to diffuse semiconductor expertise throughout the economy. So that kind of lays the foundation for the U.S. having a durable lead for a couple decades. Beginning in the late 1970s, Japanese firms begin to catch up, beginning with their targeting of DRAM markets, and they take the process lead from U.S. firms. And this is sort of a major watershed moment in the industry because the industry does actually ask for help, But its first ask for help is on trade grounds, and it basically gets, it asks for like anti-dumping measures against Japanese suppliers.

There's all sorts of debates that happen. The big electronics manufacturers, think firms like HP or IBM, kind of sit out these debates because they actually like having cheap chips from Japan. And there isn't a major push towards more heavy-handed government involvement. The industry does get some wins in the 1980s where the government allows them to do collaborative. research, but even there, it's pretty hands-off. The government basically says all we're going to do is fund advanced lithography and help you coordinate some of your investments across firms, but we're not going to take an active hand in shaping where the industry needs to go. We're going to look to the industry to kind of define where we need to go. And that kind of gets

institutionalized in a few different institutions, the first being the SRC, which is a semiconductor Research Corporation, which funds university research. Then Semitech, which is the most talked about industrial consortia, which was jointly funded between U.S. semiconductor firms and the DoD. And then the lastly is the national technology roadmap for semiconductors, which is basically the industry saying, here's where we think semiconductors are going to be in 15 years, and here are the technical capabilities we'll need. And hey, government, it'd be great if you help to coordinate some of the early stage research

that we're going to need, you know, 15 years from now for future products. So how did that work out? The good side is that in the early 1990s, you actually have a resurgence for U.S. Semiconductor firms. They begin to recapture market share. And it is kind of like a, everyone celebrates, hey, this policy regime that we, we established seems to be working. But the U.S. takes its foot off the pedal and actually begins to wind down a lot of these

institutions. One, they no longer have a national focus. They begin internationalize. So Semitech begins to introduce international members beginning in the late 90s. The technology roadmap begins to introduce international members also in the late 90s. So it's no longer a national technology roadmap. It's an international technology roadmap. The end result is that a lot of entrepreneurs outside of the U.S.

kind of have a blueprint for what it's going to take to compete with U.S. firms. And they begin to catch up in key industry processes. So you see it in things like DRAM, where Korean firms have, have majorly caught up to even Japanese firms, and most Japanese firms have left the market. And you see it now with the rise of foundries, where foundry firms were able to catch up to integrated device manufacturers and overtake, most recently with TSM and Samsung, overtaking Intel in terms of their process capabilities. And then the third sort of trend that kind of makes this more exacerbated is you have a major

technology revolution in the late 2000s where traditional manufacturing processes get more expensive. And U.S. firms, which were funding heavily in R&D and were previously integrated device manufacturers, so they were vertically integrated, they begin to pull out most of their manufacturing and rely more and more on foundries. And so that helps those foundries also increase their overall KAPX spend because they now have more customers, kind of, they have a stronger customer base to drive their R&D in KAPX and help them further push the frontier of manufacturing. So ultimately, it worked out well in the near term in that a lot of these companies became more profitable because they were able to coordinate their research and development

more tightly with one another and they were able to sort of hone in on a particular vision of industry organization. But the problem is, is that by distributing who is working on what in that manner, you don't build out redundancies in the system. If every, individual firm is competing on every individual product line or, if not product line, approach to production. So like sort of production methodology or what have you, you just from an evolutionary perspective, have more people working on the same thing. So more sort of opportunities for kinds of mutations in the method of production that may or may not be worthwhile. If you cut out that entire, you know, base, once what you're working on starts to, you know,

not quite work as well. You don't have a rolodex of strange ideas or mutations or slight differences that you can point to on the one hand. And on the other hand, you've self-consciously shrunk the market for semi-cap for producing machinery to produce semiconductors for domestic firms. And so these two together lead to this kind of self-reinforcing situation where you're running really quickly, but the moment you trip, you're on the ground. Just one more historical question before we move on to modern-day efforts to boost U.S. semiconductor expertise and capacity. Can you talk a little bit more about Japan in the 1980s?

Because this was such a big deal in terms of competition with the U.S., and they kind of, on the one hand, they went down a similar path in that the government was promoting their own domestic industry. But on the other hand, they didn't have nearly as much involvement from the military because Japan's military, it kind of exists, but it's nowhere near the size or scale of the U.S. Did you describe that contrast? So the major institution in Japan that drove industrial policy for semiconductors and also computing is Miti, which is, and I always forget the exact name, but it's the Ministry for International Trade. They actually played a very heavy-handed role in reshaping both the semiconductor and computing

industries during the 1980s. They set targets for what their internal firms needed to be capable of at a technology level. So they said, by 1984, we need to have XYZ process and product capabilities. And they would actually reshape firms by forcing collaborations between competitors on specific technology approaches. The end result was that it allowed them to leapfrog American firms in product categories like DRAM, where they took a more conservative path to reaching sort of the targets that had been laid out by Mitty and capturing huge portions of market share. The apex of fears over Japanese domination of the industry came with the announcement

of Japan's fifth-generation supercomputing project, which I believe was in 1984, sort of based on the heels of their success with their VLSI program, where they basically said, now that we've caught up to the frontier in semiconductors and computing, we're going to define the next generation of computing with advances in semiconductors and computing technology. However, that program mostly was a failure. It petered out.

And sadly, going into the 1990s from Japan's perspective, the economy more broadly stagnated. And, you know, today, Japanese computer and semiconductor firms are nowhere near the leadership they were in the 80s and 90s. So they did take a very different tact to sort of play catch up to U.S. firms, but they were also unable to sort of build on that catch-up phase to establish long-term dominance the way the U.S. industry was able to do in the decades prior. If I can just offer a short summary of that in a certain way, so much of industrial policy

for ketchup is effective, especially in East Asia, because the rest of the world presents a target. And the goal of policy is to hit that target. When you're at the technological frontier or when you're attempting to gain the technological frontier, you need to not only be good at hitting targets, but you need to become good at specifying or imagining targets that no one sees yet, which is a different necessity and sort of speaks to the fact that there is a role for industrial policy, but the fact that there is a role for industrial policy doesn't mean that there's no role for science policy, which sort of funds people thinking about where targets might be or where targets ought to be.

That's super, that's super interesting. So you still have to have that R&D, you have to have the crazies that think about can we have a chip that's at negative 77 Kelvin and, okay, maybe that's never going to be in a phone, but there have to be people out there and investment in figuring out what that true, like, pure science tech frontier is. You know, I know we're supposed to, we got to get to like the sort of present tense, like what's to be done. But, you know, one of the questions I'm still, like, curious about it that I feel like we need to talk about is we have part of the consternation right now is the sort of model that you talked about where we have design in the U.S. and a lot of companies that are still extremely good at it, but manufacturing overseas, the separation of

design and production, the foundries, and that's what has left us sort of like, I guess, geopolitically vulnerable, perhaps. How much is the role of Wall Street, specifically in the history, a contributor to companies opting for this business model? So I think that is a good question. And I think where you, again, I don't have like a deep study on it, but I do think you saw a change in the industry when you saw generational change. and its leadership. So in the late 80s and early 90s, many of the leaders of U.S. semiconductor firms like Intel or Fairchild or AMD were still their founders. And there was actually, when you talk about this split between design and manufacturing, a motto from integrated,

vertically integrated manufacturers that, quote, real men have fabs. It was meant as an insult towards fabulous firms who, you know, were relying on foundries to build. their chips. And there was a generational change in that leadership that also sort of saw, I think the financialized approach of, hey, you know, we can shed some of our assets and reduce our overall costs and increase our margins if we rely on foundries. So I think it was a mix of both Wall Street's influence, but also when that first generation of leaders left the industry, new managers came in and took a different approach to technology development. An influence in terms of mood and approach rather than profit pressures, you know, necessarily

are a priori. So maybe just to set the scene for the present day, when you look at the U.S. Semiconductor Industries, competitiveness, its place in the world, what do you see as lacking and where are the biggest areas for potential improvement? And then we can talk a little bit about how we might get there. So I think if you were to look at the U.S. Semiconductor industry today, there are certainly pockets of excellence. We have world-leading firms in electronic design automation like Synopsis and Cadence. We have world-leading firms in equipment manufacturing like applied materials in KLA-10

core. And we have world-leading fabulous design firms like Qualcomm and Nvidia and increasingly a new crop of tech giants like Google, Apple, and Facebook who have realized they can get into chip design for their specific workloads. Where the U.S. industry has gotten weakest is it's specifically in its semiconductor manufacturing ecosystem. Although we do have quite a few fabs still in the U.S., both at the leading edge and at the trailing edge, we don't have sort of a thick ecosystem of equipment suppliers and manufacturers on the ground working together to push the edge on what's possible. And that's one place where, for example, a weakness is we don't have any EUV lithography machines in the U.S. producing commercially outside of Intel, right?

We have just one company in the U.S. who is using that technology, and we don't have that learning environment between multiple manufacturers and suppliers who are visiting multiple sites who are kind of learning best-in-class techniques from different firms to continue to push the envelope. And that kind of hinders our ability to innovate because you don't get innovation across the entire stack. You kind of focus your innovation on the places where we do lead. So you don't kind of bring that full package together where you're,

where you're developing new technologies in the fab and helping leverage those findings with suppliers and equipment manufacturers and design firms. And this is exactly the role of industrial policy and speaks to the sort of target metaphor that, you know, I used earlier, which is that if you have these cutting edge firms at the absolute peak of your industry while you are at the technological frontier,

it's clear that you have people who know how to set targets. the issue becomes translating that into a industrial organization, you know, sort of a, it's like an ecosystem is a very good word for it, where, you know, sort of there are a variety of firms engaged in and, you know, who in principle can help hit that target that is being specified by those sort of absolute front of the line firms. So the idea is that an introduction of industrial policy into the existing industry, like, will sort of allow it to do what it does well even better by virtue of adding

support and supply lines. Eating well shouldn't be complicated, but somehow it turns into recipes, prep, cleanup, and half your Sunday gone. Factors solves all that. These are fresh, ready-to-eat meals designed by dieticians, delivered to your door, and ready in just minutes. No prep, no cleanup, no excuses. And it's not just about convenience. You're getting real food, talents, nutrition, and zero artificial stuff. Meals that help you stay on track for all of your goals, without the grind of doing it all yourself. Grilled chicken, roasted veggies, steak plates, postables.

They taste like something you get in a restaurant, but they come out of your microwave in two minutes flat. If time, cost, or effort have been holding you back from eating better, Factor just took those off the table. Right now, get 11 meals, free shipping, and free sides for life. Hurry, this offer won't last long. Go to FactorMeals.com. and use code fit.

That's 11 meals, free shipping, and free size for life, but only with the code fit at factorneals. Cia. Factor, Canada's number one ready-to-eat meal delivery service. Visit Don Valley North Lexus for interest rate reductions of up to 3% for lease and finance rates, resulting in lease rates as low as 0.9%. Plus delivery credits of up to $1,500, only until April 30th. Visit Don Valley Northlexus.com. So I realized there's one more question.

I want to get to, and it's related to this, before we get to this sort of pure question of like, all right, what is the best policy it looks like? And, Hussein, you sort of hinted at it when you're talking about Japan's rise and fall. And it's something Alex that your colleague, Skanda, Amranaeth has talked a little bit about. But what is also the role of just a sort of sluggish economy in all this? The lack of capital deepening, the lack of capital investment, of course, from the great financial crisis to, you know, the virus, the pandemic, economic growth was weak. Tech really never really, tech spending never really recovered in the same way after the dot-com bubble. What is the role that just like our acceptance of poor growth contributed to the sort of hollowing out of tech know-how?

So our first piece in the series, actually, which is called supplying demand, the chip shortage in macro context, delves into this question, you know, pretty intensively. And the simplest answer is a basic, you know, sort of Keynesian answer. If you don't have strong demand, manufacturers are going to see that and say, ah, okay, our present capacity in whatever technological quality it's in is sufficient to meet what we think is going to happen. So there's no need for us to invest in new capacity here because at that point we'll just flood the market and we'll begin to eat into our own margins. So if you can durability predict that there's going to be low demand for a long period of time, you can cut CAPEX and not have it negatively impact you because there's no demand for anyone in that market space. But the problem is, is that if you have, say, technology advancing at a constant rate and like I said before for that technology to enter into the market space, but the problem is, is that if you have, say, technology advancing at a constant rate and like I said before for that technology to enter into the market space, the production process, it has to come through, it has to become embodied in new capital goods for production. If you don't have a macroeconomic environment that has enough, you know, effective demand to persuade producers to invest in new capital goods and to add capacity, you're not going

to see that technology be taken up because you're not going to see the CAPEX happen. An environment of low effective demand is going to redound not only on poor employment, which, you know, creates a less skilled labor force, you know, and even without even speaking to layoffs caused by, you know, sort of market slowdowns, but it's just going to not have technological advances that do happen be incorporated into the economy as a whole as quickly as it could be. And so you'll see a productivity slowdown as well.

So does that suggest that one of the fixes for the market right now could be setting some sort of demand floor, like just ensuring that the government, is there to buy up excess capacity of semiconductors? So I think that that's a necessary part of the fiscal spending part of an industrial policy plan. However, it's very difficult to specify beforehand where that should be done. As I'm sure, Husson can speak to, semiconductors are an astoundingly differentiated market where, you know, There is very little that you could think of as a simple baseline in the same way that

price supports for commodities like milk. You can define a basic, this is milk. Everywhere you go, this is what milk is. So before we can do that, really what needs to be done is the U.S. government would need to build out a data gathering apparatus to get a full picture of what supply chains look like now. because with the level of differentiation, it's difficult to say what that would look at before that data is gathered, though I'll let us speak on that if he wants to. And actually, this is a place where the Biden administration sort of started on the right foot. It was one of the first executive orders that they announced was a supply chain review of semiconductors. And sort of as Alex hinted at earlier, I think it's really important because there is major bipartisan consensus on doing.

something, but to cut through the noise and get to the, what is the something we should do, I do think the government has to take time to take stock of where are our biggest vulnerabilities in the supply chain today? We've had a lot of noise related to the shortage, especially with, say, automakers, but I don't think there's anyone who's saying we need to completely resure the automobile I see supply chain. I think instead, you know, The conversation we should be having at an industrial policy level is what are the internal capabilities that we need to maintain in order to maintain a resilient semiconductor ecosystem in the United States? And how can we best support that? So there may be areas of that ecosystem

that are best supported with purchase guarantees, you know, for specific markets to make them competitive. But I do think other areas would have other support mechanisms as well. Right. So with autos, although it's a problem right now, my understanding is that the actual technology that's missing is they're cheap, their low end. It's not like necessarily the most urgent thing ultimately. It's just sort of a weird thing going on right now. So in your view, like, where should we be looking? Like, where do you think we're going to find the supply chain vulnerabilities and what are some of the perhaps best policies that we could be putting forth in terms of like legislation and dollars that. that might go towards addressing it or maybe building a system that can resiliently address issues as they come up? I think I'll talk a little bit about the specific deficiencies that I see, and then Alex can maybe talk more about some of the proposals that we've seen and where those are most promising. So I think where the U.S. has a deficiency is we've weakened sort of our overall ecosystem for manufacturing innovation.

And I think I would give to give a good counter example of what that might look like is you've had Dan Wang on this podcast to talk previously about how magical of a place Shenzhen is for its electronics ecosystem where you have engineers from across the world kind of coming together and tinkering to build new things. And if you go back to Silicon Valley in the 1980s, that's what you had. You had engineers from different firms coming together, meeting up at local bars and kind of learning from each other and learning from local suppliers on what best in class cases. capabilities were. Because today our domestic ecosystem has been so thinned out, you really only have a few company towns where you don't get that same sharing of information. That's one weakness. I think the other weakness that we have is we have reliance on individual suppliers for key links in the supply chain. So the one that's gotten talked a lot about is TSM at the leading edge where you can't manufacture a trip at five or seven nanometers without relying on TSM.

But we're also seeing it in trailing edge nodes where a lot of materials and components are single-sourced, sometimes out of Taiwan or sometimes out of other countries like China. And those create vulnerabilities for our entire industrial ecosystem because, hey, guess what, that $2 integrated circuit that's nothing to write home about, you can only get from one plant. And so I think that leads you to the first policy point that Alex introduced, which is we need to do a review of our supply chains to understand. where those gaps really are, and then think about remedies. And I would say in terms of taking seriously that, the first step towards coming up with a truly efficacious industrial policy program, in terms of proposals on the table that do this, the Endless Frontiers Act doesn't really address this too much. A lot of its funding is still focused through academia, even though it does engage with the Manufacturing USA program. The Chips

Act does provide for a survey of manufacturers, you know, about national security concerns, particularly across the supply chain. And it does the, you know, sort of in my opinion, good move of anchoring that in the Department of Commerce rather than the, you know, the NCSC, the national counterintelligence and security council or something like that, has a similar, you know, national security in the semiconductor supply chain, a project that it's working on. But really, the the ideal is the sort of spec infrastructure plan that the Biden administration put forward, which includes $50 billion for the establishment of an office of the Department of Commerce to monitor domestic industrial capacity just across the board, which, you know, insofar as

the problems in semiconductors are, you know, significant and mealurable through industrial policy, but there's also a critical sense that this industrial policy toolkit is something that we are going to have to get better at as we sort of confront the absolute necessity of adapting the economy to climate change in ways other than just reducing fossil fuel usage. In order to adapt to a significantly changing landscape, having active monitoring of domestic industrial capacity and factors impacting it from a variety of sources is going to be something that we're going to have to get much better at, and semiconductors provide through their complexity and through their product differentiation, just an incredible sandbox for testing out methods for doing this.

I have a devil's advocate question, which is, what do you say to people who are sort of markets fundamentalists, let's call them, who would say that, well, the U.S. has squandered its edge in semiconductors. Intel has made a ton of missteps in its own business. It doesn't deserve to have help from the government in any way because it's just not going to be efficient or it's not going to be competitive with Taiwan or China and other manufacturers. What's the response there? I agree that any industrial policy should not be viewed as a giveaway to national firms. And I think any administration looking to implement industrial policies should be also looking to make sure that they're either getting concessions from firms that they're providing help to or,

using it to seed not just existing firms, but also helping, you know, new firms take advantage of those. But the other comment I'd make is we should be kind of blunt that there is no free market operation in the global semiconductor industry. When Taiwan is subsidizing the production of fads, China has openly stated its ambitions for the, for its domestic semiconductor industry. And Korea and Japan, of course, have a long history of supporting their domestic industries as well. And there's also talk now of Europe funding leading edge fabs. So the question kind of comes, given that national governments across the world are going to be subsidizing and protecting their industries, what do we need to do in order to make

sure that we have a sufficiently capable and resilient ecosystem to service our economy? Because as Alex kind of hinted at, semiconductors are so pervasive. They touch every part of a modern 21st century economy and to find ourselves, at the mercy of far-flung suppliers, because that's what the market requires, I think is a risk that a government doesn't actually want to take, regardless of its adherence to free market principles. And worse than that, it's bad economics. The idea that, you know, free market principles would say that you should, you know, pursue comparative advantage and let these other countries with their cheaper labor do these lower value ad processes and, you know, sort of damn the torpedoes to the domestic

economy, you know, relies on this notion that, you know, we are specializing in high value ad and they're specializing in low value ad. And then we sell our high value add and buy their low value ad. And then it's better for everyone. This is sort of the, you know, happy family of the Ricardian models, you know, from, you know, a thousand years ago. The problem is, is that in an actual dynamic economy, everyone, firms and nations alike, are trying actively to move up the value chain. So if you sort of say, hey, sure, you can have all of this, you know, sort of low value ad work, like, have fun. We're going to be, you know, just doing the ultra high tech, you know, we're just going to design the chips. And you'll never catch us up at that. And then,

you know, five years, 10 years, 15, 20 years later, these firms that were doing what used to be low value add work have stacked up enough process improvements that all of a sudden, hey, if you want to produce something at five nanometers or seven nanometers, you have to go through them because at the time we decided, oh, you know, this is sort of Schlepp work and, you know, sort of Ricardian comparative advantage will take care of us forever. Markets don't work that way. Was anyone ever like warning about this back then where there people are like, oh, maybe it's not a good idea that all these companies for all these products are just sort of assuming, you know, just sort of thinking of Taiwan is their factory and, you know, let's just do the high-tech stuff here.

Chalmers Johnson famously. I think Hussein can speak to more. I mean, there was an enormous freak out in the 1980s when Japan became ascendant. And that's where the DOD really did get involved. But I don't think anyone in the 80s and 90s, sort of as we were crafting the path that we ended up now, foresaw a future where, you know, the U.S. economy could be ground to a halt by foreign ship supplier availability. Right. I think it's kind of gotten to a point where well beyond what the architects of our current policy regime sort of imagined. Do you, what do you, like, how optimistic do either of you feel based on, you know,

we have this sort of emerging bipartisan consensus? It might be like one of the only bipartisan issues that there is. Like it really does seem like both Republicans and Democrats are kind of on similar pages on this. And there's willingness to spend money and identification of some of the issues. So how optimistic are you that the sort of like long decay of our production know-how can start to be reversed? I just speaking from a policy perspective, I think it really all comes down to the correct choice. and I'll let Husson speak a little bit more to what that correct choice of policy program is. I was going to say two things on how optimistic I am.

One is, I think in the short term, I'm very optimistic that we will pass policy. Whether it's the right policy, depends on making sure we get policymakers to listen on what approaches we have. I think in the long term on whether or not this episode helps the United States adopt a forward-looking, industrial policy strategy that can be leveraged for, say, climate change is still a much bigger question. And that's the level that I would like to see us get to for success. I think what Alex has done and has seen in proposals that are out there today is that the right pieces are in proposals, whether that's the Endless Frontiers Act, the Biden infrastructure bill or work from some senators like Senator Coons, the pieces are there. I think.

think to be really successful in the short term, it's about pulling out the right specific proposals and merging them together into the right approach with a focus on supply chain analysis, setting clear goals for what the U.S. semiconductor ecosystem should be capable of, and then building up that supply chain monitoring muscle to sort of look forward and beyond the semiconductor industry to where else we'll want to manage our industrial. capabilities to tackle, you know, future challenges with climate change. Anything else, Alex and Husson, any sort of last things we didn't ask about or key ideas you want to get across? I guess just to look out for sort of the next two pieces in this

series, which the next one will be a sort of thorough account of what the industrial policy toolkit ought to be for for governments looking to implement this. And the one after that being sort of a longer theory-based, you know, sort of argument against this, this Ricardian comparative advantage approach to trade policy, which the semiconductor industry illustrates particularly well, but which holds, yeah, across sectors. Can't wait to read them. All right, Hudson and Alex, thank you both so much for coming on. Thanks so much for having us. This was wonderful. Appreciate it. Thanks, guys. That was great. I found that to be a very helpful conversation to sort of like frame where we are right now. I really like I like this idea. I mean,

we've talked about this, you know, Dan Wong and a few others. You know, obviously this idea of like learning by doing the importance of actual like getting your hands dirty know-how. But I thought like, you know, as Alex put it, that, you know, technology isn't just something that's like, is handed down to producers and then they like upgrade the thing. Like production is like what technology is like, what technology is. It's just like a very like sort of like useful way to think about the challenge. Yeah. And I also like this idea that, you know, people tend to think about industrial policy and government involvement in industry as a sort of like a really boring, very controlling, like in stifle stifling of innovation, I think. But the way Alex and Hassan kind of described it,

it can be a backstop for getting really creative innovation in different products. And I think a lot of people don't necessarily think of it that way. Totally right. And I just think, you know, I really appreciate the sort of like more granular discussion about the different periods of our different policy regimes and thinking like, oh, we can do this just by funding a lot of R&D. And of course, I don't think anyone's going to deny that R&D is extremely important. but then also the demand side element, but you can't just get the demand because there's no such

thing as just a chip or saying, you know, can't say, oh, we're going to buy all the chips, having to identify them. It's a difficult problem. And I think, you know, that point that Hussein made at the end, like, what matters is getting the policy right. Like, there isn't going to be a sort of a nice, a nice clean, a one-shot solution. It's got to be a sort of a multi-prong focus. Yeah, absolutely. I'm sure. I have a. have this terrible feeling that in like two decades, we're going to be having a conversation about this moment in time talking about whether or not the U.S.'s attempts to revive its semiconductor industry were successful or not. It's going to be a never-ending semi-series.

We'll definitely talk about it in two decades, but maybe we'll be looking back at how it was how the U.S. turned it around. Yeah, that would be pretty excited. Okay. Shall we leave it there? Yeah, let's leave it there. This has been another episode of the All Thoughts podcast. I'm Tracy Allaway. You can follow me on Twitter at Tracy Allaway. And I'm Joe Wisenthal. You can follow me on Twitter at The Stallwork. Follow our guests on Twitter. Alex Williams. He's at Tragic Bios. Follow Hussein Khan on Twitter. He's at Hussein Khan. And check out all of their writing. And, you know, there's ones coming up at the Employ America Medium page or at the Employ America website. Lots of a super. deep stuff there. Be sure to follow our producer Laura Carlson. She's at Laura M. Carlson.

Follow the Bloomberg head of podcast, Francesca Levy, at Francesca Today. And check out all of our podcasts at Bloomberg onto the handle at podcasts. Thanks for listening. You can get the news whenever you want it with Bloomberg News Now. I'm Amy Morris. And I'm Karen Moscow here to tell you about our new on-demand news report delivered right to your podcast feed. Bloomberg News Now is a short five-minute audio report on the day's top stories. Episodes are published throughout the day with the latest information and data to keep you informed.

Yes, there are other products like this from a variety of news organizations, but they usually rerun their radio newscasts throughout the day. That's not what we do. We create customized episodes that can only be heard on Bloomberg News Now. And we don't wait an hour to publish breaking news. When news breaks, we'll have an episode up in your podcast. podcast feed within minutes. So you're always getting the latest stories and developments. Get the reporting and the context from Bloomberg's 3,000 journalists and analysts. We're all over

the world. Listen to the latest from Bloomberg News Now on Apple, Spotify, or anywhere you listen. What separates good leaders from transformational ones? I'm Jessica Chen, and in season two of Leading By Example, we'll sit down with executives like Grace Chen of Bertie Gray, to find out. It's important to understand where you spike, but also really acknowledge where you don't and find people who can fill those gaps. Listen to leading by example, executives making an impact on the IHeart radio app, Apple Podcast, or wherever you get your podcasts.