The case against industrial policy: My long-read QA with Scott Lincicome

James Pethokoukis and Scott Lincicome, "The case against industrial policy: My long-read Q&A with Scott Lincicome," American Enterprise Institute, April 5, 2021, https://www.aei.org/economics/the-case-against-industrial-policy-my-long-read-qa-with-scott-lincicome/

Speech Notes

The case against industrial policy: My long-read Q&A with Scott Lincicome

Pethokoukis: Since we’re going to be discussing industrial policy, tell me what you think of the following definition: Industrial policy is having the government promote — via tax, regulatory, spending, or trade policy — certain sectors, technologies, or even companies for some strategic national goal.

Lincicome: That’s pretty good. I would just narrow it a bit because one of the problems we have in current debates over industrial policy is that industrial policy, even by your definition, can bleed into a lot of things that are very difficult to fit into the traditional definitions of industrial policy. Technology policy bleeds into things like basic research, which then feeds into a lot of the arguments of pro- industrial policy folks who consider things like the internet and iPhone to be industrial policy, or at least the products of industrial policy. I’m actually writing a new paper on some of this now.

I would also make it more explicit that industrial policy is inherently nationalistic. Industrial policy advocates want any benefits from industrial policy to be national benefits, whether that is manufacturing jobs or certain products. Whatever it may be, it inherently must be national and nationalistic. That, again, rules out some things. For example, it’s difficult to say that the Pfizer-BioNTech vaccine is a product of real, traditional industrial policy because that contract quite plainly said that the government would not have control over the supply chain and that it was not demanding an American-made vaccine.

The third thing that needs clarification is that, within all of this, there is a government plan to achieve defined objectives. That’s really another concrete part of industrial policy that, again, tends to get left out in some of these very wishy-washy definitions. What I mean by this is that the government is not merely giving a bunch of money to a bunch of smart people and saying, “Hey, go do stuff.” Instead, the government is saying, “We need to achieve X” — whether X is more semiconductor output or increased manufacturing jobs in high-skilled industries — and then providing incentives, domestic subsidies, or protectionism to achieve that goal.

That strategic element ends up ruling out some of the things that are erroneously called industrial policy. For example, if the government is funding some sort of basic research project and these guys stumble upon a magnificent invention, development, or technology, but the government really had no plan or design for that, that’s not industrial policy. That’s just a bunch of smart people being smart who happened to be on a government grant when it happened. That’s just not what I think is traditionally considered industrial policy.

Pethokoukis: It seems like industrial policy is entering the conversation now in ways I haven’t seen in a long time. Why is it becoming more “popular,” and how is that actually playing out?

I think it’s having a moment right now primarily because of China. About a decade ago, the Chinese government really changed course with “Made in China 2025” and a few other industrial plans. Barry Naughton, a China expert, has a new book out that actually documents this shift. It chronicles how, in 2010, the Chinese government had a major sea change in how they wanted to do economic policy. They went from being more market-oriented and reformist to being much more expressly pro-industrial policy and throwing a lot of money, government subsidies, and government action at specific sectors. Those sectors included things like semiconductors, electric vehicles, AI, and a few other technologies.

That, I think, set off some alarm bells within the US government, particularly going into the mid-2010s when there were also bilateral irritants and problems being created by the Chinese government, particularly in the South China Sea or with the Uighurs in Xinjiang or Hong Kong. So there was geopolitical tension combined with a change in the Chinese government’s economic policy approach, which shifted to just throwing gobs of money at these different technologies.

This set off alarm bells within Congress and the US business community. People were saying “We need to do something, we need to counter the subsidies.” And of course, let’s face it, some very smart and well-connected lobbyists also saw the opportunity for there to be new subsidies for domestic companies and corporations. All that, I think, has created a lot of momentum.

I think another thing creating a lot of momentum is the slow-down in productivity we’ve seen over the last decade, perhaps due to the Great Recession. This noticeable slow-down — mainly in the manufacturing sector and in government funding of R&D, but also elsewhere — is creating a kick-start of sorts. Market skeptics in particular say, “Look, the market is failing, so we need the government to step in to boost R&D spending and manufacturing productivity, and do all these other wonderful things because the market simply isn’t achieving those objectives.” When you combine China with some of those data points, you get a lot of momentum for industrial policy.

Pethokoukis: How do you see this push for industrial policy playing out in the Biden administration? What do they want to do that resembles industrial policy?

I think the biggest areas are in environmental technologies. The Biden administration has been quite clear that they want to throw gobs of money at that sector, whether it is battery technology, electric vehicles, or any other technologies related to climate change. That’s a classic example. And of course, they want it to be here in the US in order to achieve other economic objectives, like boosting manufacturing jobs or restoring “the US manufacturing sector.” I think that’s classic industrial policy: They are picking specific industries and economic objectives in the US and are going to achieve those goals through subsidies, protectionism, Buy American mandates, you name it. All very classic industrial policy levers.

I think things get a bit murkier when they talk about boosting basic government funding for R&D. If we’re talking about just basic research, then, again, just throwing money at a bunch of smart people and saying basically, “Go be smart,” doesn’t strikes me as meeting the traditional definition of industrial policy. But frankly, sometimes I think that’s something a lot more people can get on board with, because there is a stronger case for basic research support as opposed to sectoral industrial policy.

Pethokoukis: Some people think, “We can’t wait for the invisible hand to work in a time of crisis.” So when people are worried about climate change or China, it’s hard to convince them with the argument “Don’t worry, we’re just going to let markets work, and everything will be fine.”

Oh, definitely. But I think there are other arguments that can and should be made. The first one is that the US has a long history of experimenting with industrial policy.

As I wrote in my recent paper, and as I’m writing in a new paper now, the results of American industrial policy in the past have been pretty lousy. In the worst case, we end up with things like the Jones Act. There, we needed a domestic shipbuilding industry on supposed national security grounds, so a law was implemented to effectively force companies to use American ships — made in America, crewed by Americans, so forth and so on — to transit goods from US port to US port.

The result of that has been absolutely catastrophic: a steady decline in the US fleet, an inefficient shipbuilding industry that doesn’t make any good or large ships and certainly doesn’t sell any abroad, and (of course) the mobilization of a very well-connected and efficient lobbying machine to ensure that there no reforms are made. There were also, of course, all sorts of costs for consumers, particularly if you lived in a place like Puerto Rico or Hawaii.

But even when you dig into the supposed successes, you find out they weren’t really successes. For example, there was a study just a couple years ago about pilot funding for green energy. The researchers found that, while these startups did okay, they truly didn’t do any better than other market participants did. There are all sorts of examples where things have gone horribly wrong with US industrial policy — including, quite ironically, in some of the sectors that we’re now obsessed with, like semiconductors. That’s the first thing that I think is important to note: We should have some skepticism if people say that, regardless of the crisis, industrial policy is the answer.

Another reason for skepticism has to do with subsidies. The mere existence of Chinese subsidies does not mean that those subsidies have been effective. You mentioned China’s growth, but China’s growth trajectory is a lot less scary now than its past catch-up growth. China has pretty significant issues in terms of demographics and population, plus its productivity growth is in the toilet. Also, a lot of the subsidy programs have actually been pretty giant failures. Again looking at the semiconductor industry, there have been a lot of bankruptcies in that industry, and it’s still a decade behind companies like Intel and TSMC. Even in sectors where the Chinese government has supposedly done well, like electric vehicles, there’s just tons of waste and graft.

The economic threat of China requires a bit more nuance. It’s not that China isn’t a problem — whether due to geopolitical reasons or other things — but it’s not an unstoppable economic juggernaut that requires abandoning decades of US policy that is more market-oriented, market-friendly, and quite frankly has produced a lot of great results.

The last critical point to note is that there are things that the government can do to improve the competitiveness of the US economy and the outlook for core industries. I list some of these in my paper. You could talk about full expensing, for example, in tax policy. And the most glaring one is immigration — we could and should be rapidly expanding high-skill immigration, as there is a strong connection between innovation, multi-national investment, and levels of high-skill immigration, particularly in our research universities and the rest.

There’s also lot of evidence that US restrictions on high-skill immigration actually end up benefiting other countries like Canada and China. Some of China’s biggest economic impediments have to do with access to human capital. For example, if the US acts as a magnet for the smartest people in the world, they’re not going to be in China. There are also a lot of other policies that we could be implementing to boost US industry in very good ways.

The problem is that none of that is sexy. It doesn’t sound like, “Aha! We’re going to fund AI,” or “We’re going to boost the industries of the future.” Politicians love ribbon-cutting ceremonies, and they love being responsible for economic greatness. But again, I think there are some significant reasons for us to be skeptical that those things work very well.

Pethokoukis: But if there’s a problem hurtling at us, we just can’t wait. Even if we’re going to probably waste a lot of money, some problems are so pressing that we need to act — it’s wartime, and wartime means that you don’t let the market devise new weapons. Instead, you say, “Here’s what we need, and we’re going to fund it. Let’s go.”

Right. I think the danger there is that we may end up in a weaker position as a result, not a stronger one. We may end up diverting resources from more productive investments to less productive investments, undermining core parts of the manufacturing sector, or exacerbating the waste, fraud, and corruption that this instills. I think you risk as much creating a US shipbuilding industry as you do creating some sort of glorious, bleeding-edge, research-intensive industry.

We have evidence of this in the last decade — that efforts to boost certain industries end up harming others. For example, if you look at Trump’s steel tariffs, which were implemented on express national security grounds, the fact is that they ended up hurting a lot of other industries. There’s also the risk of just simply discouraging investment if you inject uncertainty into the markets.

It’s not that you have to do nothing, but we should recognize the downsides of action. Just doing something does not necessarily mean that you’re going to be in a stronger position than if you had a little more faith in markets and did what we would call horizontal policies. In other words, improving the tax environment, immigration, basic research, etc., instead of cherry-picking specific industries because of these perceived threats.

Pethokoukis: The US has such a big, technologically advanced economy with a lot of capital to fund private research. So is there really a huge downside to spending a few hundred billion dollars on applied research or funding some regional tech clusters — just in case? I know I don’t want to wake up in 10 years and have all the leading chipmakers be located in China.

Two things. First, we need to have a pretty sober assessment of where the US really is with respect to semiconductor manufacturing, pharmaceuticals, and a lot of other areas. I think that will show that there’s not nearly the urgent need for government support that a lot of people say there is. Whether it’s in AI or advanced manufacturing, there’s still a lot of very good stuff that’s going on in the US. So that’s the first point: We’re not starting from zero, and we’re not in a very weak position in terms of things like semiconductors. We still produce a lot of them. We still have a national champion in Intel that, while it has had some setbacks, is still pretty darn good, awash in capital, and spending it on R&D, CapEx, and the rest.

The other real concern is that the process of trying to pick these winners will actually crowd out the necessary capital and the necessary process of funding and finding the best and most productive things. That can happen even within the same industry. Let’s take pharmaceuticals for example. The Trump administration was going to spend about $800 million in taxpayer loan money to turn Kodak into a pharmaceuticals company. Because of that implicit government backstop, Kodak’s shares went through the roof — a bunch of investors in the market said, “Aha! This government wants Kodak to be a drug company, so we’re going to all invest in Kodak.” That capital went towards Kodak.

Meanwhile, there were all sorts of other companies in the US that actually made these drug ingredients, what we call API, that weren’t getting access to that money because it was all being tied up in Kodak. And at the same time, there were plenty of other very successful drug companies out there in the US — we could talk about Pfizer, but there’s also Fujifilm down here in North Carolina that’s doing some pretty awesome things with biologics.

The point is this: When the government gets involved, it can distort investment decisions and actually end up inhibiting investment in the best and brightest. So again, we need to be really aware that when you step out of the basic research space and start funding specific industries, technologies, or companies, you might actually end up retarding advancements in that industry.

Batteries are another pretty good example of what I’m talking about in the sense that private capital does a good job of finding the next big thing. About a decade ago, the Obama administration tried to create a battery industry and failed pretty miserably. Today, however, there’s tons of private capital going into the domestic production of batteries, advanced batteries, and battery materials because there is now a clearer and brighter future for that.

Again, do we really need the government directing that capital — directing that process — when we see that it can have a lot of downsides and that the process is still working pretty well?

Pethokoukis: Are those risks acceptable if it means keeping supply chains out of China or making sure that they have little to no role in anything that can be viewed as strategic or essential for national security? It was one thing when it looked like China was not nearly as technologically advanced as it is today or when it looked like it might be slowly going in a more friendly direction, but that’s not the situation today. Is it worth having more government involvement if it means that China will be less involved in our economy?

I certainly think there’s a role for US export controls to the extent that the US finds the need to deny certain technologies to the Chinese government. There are, however, some pretty significant problems. First of all, in a lot of these areas, the US is not the only supplier of those technologies. So at the end of the day, these types of export controls just end up harming American companies and not actually deterring China in any way.

The second problem is that once you decide to provide subsidies or other government support or protection to industries deemed “essential to national security,” a lot of other industries are suddenly going to say that they are essential to national security. We’ve seen this in the COVID space. The textile industry popped up saying, “Aha!ppE is essential to national security, so we need all these new Buy American rules, new protection, and new supply chain mandates to get these essential goods out of China.”

Finally, there is the risk that it moves from a non-China supply chain mandate to an American-only supply chain mandate. That comes, of course, from the usual political incentives and domestic lobbying.

So I think to certain extents, that is fine. Whether it’s US policy encouraging diversity of supply through things like the Transpacific Partnership or building alliances through the national technology and industrial base — which is this Defense Department program to have a bunch of allies collaborate with certain critical technologies — that’s totally fine. However, once you start really trying to force these supply chains to reorient and involve subsidies and protectionism, I think things get pretty dicey pretty quick.

Pethokoukis: I don’t want to over-learn from the lessons of the past. I think the Soviet Union’s failures and Japan’s slide into stagnation are two powerful lessons about the failure of government. So… persuade me that I haven’t overlearned those lessons, and that we haven’t gotten smart enough since the 80s to justify putting more money towards AI or regional tech clusters.

Well, I don’t think that history should foreclose the considering of different policies. Like you note, there’s a categorical difference between Japan and China in terms of the nature of the government, sheer size, and thus economic influence. But my view is that those lessons should give us a lot of skepticism about the new industrial policy fad. That’s even leaving aside that the language and arguments being used today are hilariously similar. I say “hilariously” because I’m reading a book right now called Losing Time: The Industrial Policy Debate by Otis Graham. It’s this pro-industrial policy book from the late ’80s. I kid you not, you could take the language in this book and compare it to the op-eds of pro-industrial policy folks in the current pages of The New York Times or The Washington Post — it is startlingly similar.

Leaving that aside, I think that history does and should give us some skepticism about what can and can’t be achieved via government policy and about the threats we face. Let’s go back to semiconductors. In the ’80s and ’90s, we tried to bolster the semiconductor industry through a combination of subsidies and protectionism. The protectionism was an absolute failure. Not only did we pick the wrong types of chips, but we actually encouraged the offshoring of our domestic computer industry because they couldn’t get a hold of certain semiconductors, so it was a massive net negative.

Yet even on the subsidies side, the creation of this consortium Simatec was, by all accounts, pretty much ineffective at best and was only made effective after it opened itself up to foreign competitors. Simatec was this very nationalistic, classic industrial policy entity that didn’t produce much value beyond what could have been done in the private market. But it did become a bit better after again opening itself up to all sorts of foreign participation — Japanese, South Korean, and the rest.

That should give us some pause about how we’re going to treat semiconductors today, for example. Are we really going to have a wiser government in picking the right products, the right industries, that type of thing? We need to then apply that skepticism to today. While there may be a narrow and targeted approach for US policy to resolve real market failures or counter real threats by the Chinese government and Chinese industry, when you start digging into these things you find a whole lot of misdirection and misinformation, and a lot less reality when it comes to the need for government intervention.

If you look at the semiconductor industry, you see an industry that actually is doing pretty darn well. Even though there is certainly a lot of production offshore these days, the US chip industry is still growing. But you do see some perils of trying to re-nationalize chip production. For example, the major ice storm in Texas actually knocked out a bunch of semiconductor companies there and exacerbated the global chip shortage that US automakers are now facing.

We need to think about the threats we really face and the effectiveness of the policies that are being proposed in actually achieving market-beating outcomes.

Pethokoukis: You’ve been talking about pumping the breaks. To me, it doesn’t seem like we’re going to do that. How far down this road — and in what direction —do you think the US is going to go towards having a lot more government intervention in what people in Washington think are key sectors and technologies?

I think that the train has left the station and is barreling ahead. The goal for those on my side will be narrowing the scope of these measures and trying to get them to, at the very least, target legitimate issues.

For example, maybe we shouldn’t just be giving billions of dollars of subsidies to any company that wants to build some sort of semiconductor manufacturing facility, regardless of whether that’s actually a bleeding-edge technology in the five- or three-nanometer space — i.e. the really, really advanced stuff. We shouldn’t just fund something because it has some nexus to actual national defense — for example, there’s a new facility being built in Upstate New York for that very purpose. I hope that we can work to, at the very least, tailor these industrial policies to actual needs instead of just making them a giant grab bag for anybody who registers as a lobbyist. That, I think, is going to be the victory we see these days.

But at the same time, it still remains critically important to raise the very real questions that industrial policy raises about the ability of policymakers to figure out what the next best thing is and the potential distortions that these interventions can create — whether it’s really high costs or crowding-out private investment in more productive endeavors. That, I think, will remain valuable even if we’re on the losing side.

Pethokoukis: My guest today has been Scott Lincicome. Scott, thanks for coming out on the podcast.

My pleasure, thanks for having me!