Conversable Economist

There are often substantial controversies over whether a merger should be allowed to happen, but then relatively little follow-up after the event. Of course, if a merger was blocked, then it’s hard to know whether it would have led to good or bad outcomes. But when a controversial merger is allowed, it’s fairly straightforward to see if the negative predictions actually happened.

Brian Albrecht, Dirk Auer, Eric Fruits, and  Geoffrey A. Manne take on this task in “Doomsday Mergers: A Retrospective Study of False Alarms” (International Center for Law and Economics, March 22, 2023). Here’s their summary:

Amazon-Whole Foods

The first merger we look at is Amazon’s purchase of Whole Foods [in 2017]. Critics at the time claimed the deal would reinforce Amazon’s dominance of online retail and enable it to crush competitors in physical retail. As now-FTC Chair Lina Khan put it: “Buying Whole Foods will enable Amazon to leverage and amplify the extraordinary power it enjoys in online markets and delivery, making an even greater share of commerce part of its fief.”

These claims turned out to be a bust. … [S]several large retailers have grown faster than Amazon; Whole Foods’ market share has barely budged; and several new players have entered the online retail space. Moreover, the Amazon-Whole Foods deal appears to have delivered lower grocery prices and increased convenience to consumers.

Beer-Industry Consolidation

… ABI’s acquisition of SABMiller in 2016, … critics claimed would increase the price of beer and decimate the burgeoning craft-beer segment. Instead, the concentration of the beer industry decreased after the mergers, prices did not increase on average, and the craft-beer segment thrived. …

Bayer-Monsanto

… Bayer’s acquisition of Monsanto [in 2018] … was met with stern rebukes from policymakers and academics. Critics argued that the merger would raise the price of key seeds, such as corn, soy, and cotton. Perhaps more fundamentally, the deal’s opponents argued it would further concentrate the agri-food industry, forcing farmers to deal with only a handful of seed providers. … Fast forward to today, and these fears appear overblown. Seed prices have remained roughly constant … and there is little evidence that the life of farmers and rural communities has been significantly affected by the merger. …

Google-Fitbit

[Before] Google’s acquisition of Fitbit [in 2019]… The deal’s opponents claimed the merger would reinforce Google’s position in the ad industry and prevent new entry; harm user privacy by enabling Google to integrate Fitbit health data into its other ad services (or sell this data to health insurers); and crush burgeoning rivals in the wearable-device industry. … [A]vailable evidence suggests the exact opposite has occurred: Google’s share of the online-advertising industry has declined, as has Fitbit’s position in the wearable-devices segment. Likewise, Google does not use data from Fitbit in its advertising platform; not even in the United States, where it remains free to do so. Meanwhile, the merger enabled Google’s entry into the smartwatch market as an upstart competitor against the market leader, Apple. …

Facebook-Instagram

Facebook’s acquisition of Instagram provides a different perspective. At the time, basically no one worried about it from an antitrust perspective and many pundits lambasted the purchase as a poor business decision. It is only in retrospect that people have started to see it as the merger that got away and evidence of the problems with allegedly weak enforcement. … Even in retrospect, however, it is far from obvious that the acquisition was anticompetitive. Immediately upon purchase, Facebook was able to bring Instagram’s photo-editing features to a much larger audience, generating value for users. Only later did Instagram turn into the social-media giant that we know today. The recent rise of TikTok casts further doubt on claims regarding the supposed market dominance of a combined Facebook and Instagram. A merger that benefited consumers without generating impenetrable market dominance hardly seems like overwhelming proof of the failures of enforcement.

Ticketmaster-Live Nation

…[P]eople have complained about Ticketmaster being a monopolist ever since it came to prominence. Yet, there was little outrage at the merger with Live Nation [in 2010]. … Ticketmaster’s market share appears to have fallen following the merger with Live Nation. There is thus little sense that the deal harmed consumers. So why the disconnect between longstanding frustration and antitrust enforcement? The agencies have seen the beneficial effects of mergers in a difficult multi-sided market between fans, venues, and artists. After investigation, the agencies found that the merger was primarily a vertical one between a ticketing website (Ticketmaster) and a concert promoter (Live Nation), which could be pro-competitive for the overall multi-sided market. The DOJ placed behavioral remedies in place and allowed the merger.

The article goes through these examples in some detail. I don’t mean to endorse all of their interpretations of events and outcomes, but they do show pretty clearly that dire predictions about ill effects of mergers need to be taken with a few spoonfuls of salt. I would only add that from the perspective of investors and shareholders, the promised benefits of mergers–those “synergies” that are conjured up every time a few consultants and investment bankers get in a conference room together–need to be taken with few spoonfuls of salt, too.

I am by no means an expert on things related to Taylor Swift or Bruce Springsteen, but even I am aware that ticket sales for their recent concert tours have been controversial. From an economic viewpoint, there are really two choices for pricing tickets for the show of a megastar who can sell out large arenas. Tyler Smith offers a short interview with Eric Budish on this topic in “Market design and live events” (March 20, 2023, AEA Research blog).

The tickets to high-profile concerts may seem disconcertingly expensive. But from an economics point of view, if the arenas are selling out at their official price and there is an active resale market at a substantially higher price, then the tickets are “underpriced”–in the specific sense that others are willing to pay more than their face value. In the old pre-internet days, this willingness to pay more than face value of tickets could show up in other ways as well, like starting the line hours or days early, so that getting tickets involved paying with time as well as money.

Budish argues that there has been an equilibrium shift. In the old days, while there was some resale of tickets for big shows, it was relatively limited. Thus, the performance groups were willing to sell ticket at a face value that was probably somewhat lower than they could have gotten, but the performers wanted a sold-out show and a long-term relationship with fans (and perhaps to encourage sales of recordings and merchandise), so they were willing to take a lower ticket price. But when ticket sales went online, the situation changed. Budish says:

I think there was an old equilibrium, and it’s old in the sense that it describes the ticket market probably for the whole 20th century. In this pre-internet equilibrium, tickets were often underpriced and often sold out quite quickly. Many tickets were purchased by speculative resellers, but the large majority of tickets were purchased by ordinary fans. The best numbers come from a great research paper by Phil Leslie and Alan Sorensen. They estimate that in the pre-internet era of the ticket market, it may have been about 5 percent of tickets that went through the secondary market. The internet really transformed the tickets market starting in the first decade of the 21st century. The internet changed the technology of both the primary market and the secondary market. 

The internet technology made it a lot easier to amass large numbers of tickets for events all across the country, really all across the world. There were economies of scale in the amassing of large quantities of tickets, whereas in the good old days ticket resellers had to stand in line. But there is just a lack of scale economies to that. Whereas if you can use an algorithm to scoop up underpriced tickets in the primary market, you can do it at substantial scale. The resale of tickets also changed with the internet because now most of the secondary market goes through internet platforms like StubHub and SeatGeek and TickPick. The secondary market with the rise of the internet is also a lot easier to use.

I think that there was this kind of uncomfortable tension in the pre-internet era when it was a little weird that artists were often pricing their tickets significantly below the market-clearing price. But you can kind of make up stories for why it might be in an artist’s long-run interest to do that. Most of the value from the underpricing is at least accruing to the intended recipients of the underpricing, i.e., to the fans. But the internet kind of broke all that with the rise of algorithms—bots as they are sometimes described—in the ticket market. And in a very low friction, globalized resale market, a lot more of the rents are just accruing to brokers. Ticketmaster estimated that it’s 20 percent. They’ve described to me in a meeting that in the right tail, in some events, it might be as much as 90 percent. 

Basically, if the resale market was not too extensive, performers were willing to charge less for tickets than the maximum they could get. Fans who saw the show benefited. But when performers recognized that they were selling tickets to brokers, who were then marking them up substantially and reselling them, the benefit of lower-priced ticket sales was going to the brokers and not the fans. Ticket prices for many prominent shows skyrocketed.

There are two ways to approach this situation. On one side, accept that brokers and the ticket resale market are going to be active. As a result, many of those attending any particular show will have purchased above-face-value tickets. In this situation, performers will raise ticket prices substantially so that a larger share of the higher prices come to them, rather than the brokers. Some performers will have an online auction of their own to sell tickets, as Ticketmaster is essentially doing for some events already.

The other approach is to try to choke off the role of ticket brokers and the resale market in some way. As a simple example, selling tickets only in blocks of four or eight is a step in this direction. But these kinds of rules just lead to an arms race between the programmers, where ticket brokers have an incentive to write programs which can purchase MANY blocks for four or eight tickets for resale.

What seems like the ultimate way to turn off brokers is that when you buy a ticket, your name is attached to it. Think about airline tickets, for example. Your name is on the ticket, and broker can’t just buy large blocks of plane tickets for popular flights and then just resell them on the internet. In practice, it might work this way: You buy tickets in advance. But when you show up at the event, you need to show them the credit card that was used to buy a ticket and a photo ID. The ticket-taker has a hand-held machine that scans your credit card, and then prints out a slip of paper with your seats on it.

One can imagine various creative ways to get around this arrangement, but it comes pretty close to turning off the resale market. The only way to exchange your ticket is to sell it back to the original seller, who is the only one that can link a ticket to a personal credit card. Budish says:

In an ideal world, I think there’s essentially two economically logical ways to sell tickets. One is to sell tickets at a market-clearing price. The auctions were an effort to do that. The other way to do it is to set a below market-clearing price, but turn off the rent-seeking. The way to do that is essentially to put names on tickets, which makes tickets nontransferable or harder to transfer—in the same way that if I buy an airline ticket and my plans change, I can’t resell that ticket on eBay or StubHub. One of the policy goals ought to be if artists want to sell their tickets at a market-clearing price, great. That’s sort of standard Econ 101. And that’s already quite feasible with the current market structure. I think the other thing that policy might have a role in facilitating is enabling artists to have a sincere option to set a below-market price and turn off the resale market if they want. There have been some efforts. In the United States the euphemism is “paperless ticketing.”

In other words, there is a technical fix for performers who say that they would like to sell tickets at lower prices and shut out the ticket brokers. You can even combine this system with a lottery approach: that is, everyone who wants tickets registers at the website before a certain time, specifies what tickets they are willing to buy, and random chance determines who gets seats. Of course, the real problem is that we all want both lower-priced tickets and the ability to get the seats we want–and for high-profile events, that’s not possible for everyone.

For politicians, the choice between handing out energy subsidies imposing additional taxes is pretty clear.

With subsidies, companies and firms and research institutions get checks and tax breaks. There may be ribbon-cutting publicity opportunities at the start, and even if only one project out of every five or ten is actually successful, that’s still plenty of places to go and hold a self-congratulatory political event a few years down the road. Meanwhile, the costs of these subsidies–the opportunity cost of other priorities that don’t get funded, along with higher budget deficits or taxes in the future–are largely invisible. The tradeoffs that happen in the background when, say, the subsidies for one form of non-carbon energy like wind or solar just lead to less use of another form of non-carbon energy like hydropower, or where supporting technology from a politically favored firm crowds out better technology from an unfavored firm, won’t be noticed either.

With carbon taxes, on the other hand, consumers/voters everywhere will see the effect on gasoline and home energy prices. The costs are up-front. Meanwhile, the altered incentives that lead to innovation and growth of efforts to produce cleaner energy and to conserve on fossil fuels are largely invisible. The effects happen flexibly all over the economy, but the opportunities for ribbon-cutting ceremonies and political favoritism are scarce. The benefits of additional revenue for other priorities like (choose your priority here) a permanent child tax credit, saving Social Security, rebuilding defense stocks after what has been sent to help Ukraine, cutting other taxes in an offsetting way, or reducing the budget deficit are likely to seem indirect and disconnected from the revenue gains.

But in this case, at least, the politically easy choice also imposes high costs. John Bistline.
Neil Mehrotra, and Catherine Wolfram discuss the “Economic Implications of the Climate Provisions of the Inflation Reduction Act” (Brookings Papers on Economic Activity, Spring 2023, conference version of the paper), a law which is actually much more about clean energy subsidies than inflation-fighting. Their essay goes through the law in some detail: detailed provisions, costs, distributional and macroeconomic effects. (Content warning: It’s a research paper, so some of the chunks that describe mathematical modelling choices won’t be easy reading for the uninitiated.) Here, I focus on just one section of the paper, where they compare the subsidy-driven approach to reducing carbon emissions to a carbon tax approach that would have similar reductions in emissions over the same timeframe.

The authors point out various differences between how a subsidies for non-carbon energy and a carbon tax affect the goal of reducing carbon emissions.

For example, carbon taxes do a better job of encouraging conservation than clean energy subidies. “Relative to a carbon tax, subsidies encourage electricity consumption and discourage conservation. If household and industrial demand for electricity is sensitive to price, a carbon tax would have a relatively large effect on electricity consumed and
hence emissions. By contrast, a subsidy policy –by encouraging electricity consumption — would partially undo the switch from fossil to clean energy by raising overall electricity consumption …”

The higher price of carbon emissions from a carbon tax focuses specifically on fossil fuel emission. However, “Under IRA, clean energy that displaces zero-carbon energy such as hydropower is subsidized at the same rate as clean energy that displaces the dirtiest resources.”

A carbon tax is also automatically scaleable–that is, those who emit more carbon bear a higher cost. However, the subsidies of the IRA are based on production, not use: “Other provisions of IRA subsidize the energy-using or energy-producing asset, irrespective of
how much it is operated. The investment tax credit for zero-carbon electricity subsidizes the
construction of the facility rather than its operation. … Similarly, the electric vehicle tax credits subsidize vehicle purchases without regard to how much they are driven. Electric vehicles that are used as second cars and driven less will offset fewer emissions than vehicles that replace a household’s only car.”

Fixed tax credits and production subsidies are relatively inflexible as technology and market conditions change: “Carbon pricing enables households and businesses to select their preferred approaches to lower emissions, which can help to reduce costs and account for other welfare-relevant considerations that vary across individuals and firms.”

Our model abstracts from many dimensions of difference between subsidies and carbon pricing. One important difference is that pricing carbon, depending on how it is implemented, could generate revenue for the government. These revenues could be used to offset other distortionary taxes (Barron et al., 2018; Goulder, 1995), address equity concerns (Goulder et al., 2019), or be directed toward other policy objectives. A subsidy-based approach costs the government the subsidy amounts and imposes the marginal cost of raising government funds on the economy.

The IRA includes various components like requiring firms producing non-carbon energy to have certain percentages of “domestic content,” meaning that they are required to buy certain inputs from American suppliers even if the price is higher than it would otherwise be. There are also rules about labor practices and wages in order to quality for the subsidies, again, even if it makes production of the non-carbon energy more expensive.

When the authors run some of these differences through their model of energy and the economy. They find that the cost of reducing carbon emissions through subsidies approach of the Inflation Reduction Act costs $83 per metric ton of reduced carbon emissions, while getting the same reduction in carbon emissions through a carbon tax would cost about $12-15 per metric ton of reduced carbon emissions. This large gap is probably an underestimate, because the model they are using focuses on differences in incentives for energy conservation and the efficiency of investment in new technology, but doesn’t include factors like the additional revenue raised by a carbon tax or the costs of added-on rules like “domestic content.”

I’m fully aware that for the US political system, a carbon-tax is a heavy lift. But for those who dislike the idea, consider that if the alternative is paintballing government subsidies for various aspects of non-carbon energy–with all their inflexible rules and administrative requirements–then the carbon tax begins to look pretty good.

When economists or policymakers talk about industrial policy, Korea usually enters the conversation. There’s no question that Korea has had remarkable economic success in recent decades. There’s no question that various government policies have contributed to that success. But there is continuing controversy over what broader conclusions to draw from this experience. Shahid Yusuf offers useful background by laying out what actually happened Korea’s development experience in “Could Innovation and Productivity
Drive Growth in African Countries? Lessons from Korea” (Center for Global Development, Working Paper 635, March 2023).

As a starting point, here’s a graph from the World Development Indicators database maintained by the World Bank. It shows per capita GDP for Japan and South Korea in inflation-adjusted dollars. In 1990, Japan’s per capita GDP was triple that of South Korea. By 2021, Japan’s lead was down to just 7%. I try to adhere to the creed of never making predictions, especially about the future. But it is not at all inconceivable that a few years down the road, Korea will surpass Japan in per capita GDP.

What policy lessons can be learned from Korea’s remarkable growth trajectory? Yusuf hits the key themes here, but I’ll quote some of his comments in an order of my own.

As a starting point, remember Korea’s situation after World War II and the Korean War that followed from 1950-53. As Yusuf notes, Korea had an egalitarian redistribution of land after World War II: “The distribution of land formerly owned by the Japanese to Korean farmers in 1949 may have contributed to the egalitarian distribution of income and political stability that buttressed Korea’s later development. This redistribution was undertaken by the US. Military Government.” Moreover, South Korea had an external enemy–North Korea–to unite against: “One should not overlook, the threat South Korea faced from its neighbor to the North, a threat that drove the government to accelerate industrialization. Industrial diversification and deepening enabled Korea to meet more of its defense requirements and neutralize the pressure from its hostile neighbor.”

In the 1960s, Korea’s government and business leaders determined to begin a drive to industrialization:

“The unwavering commitment of the political leadership and the business elite, starting with President Park Chung Hee in the mid 1960s and sustained by his successors, to a relatively inclusive, export-led industrial strategy entailing systematic diversification into more complex manufactures, is arguably the most frequently retailed. The strategy itself was choreographed and implemented by Korea’s economic bureaucracy headed by the Economic Planning Board (EPB) in consultation with the leading business groups. The Five-Year Economic Development and Science and Technology Comprehensive Plans spelled out the government’s vision and objectives. Presidential focus on industrial and export outcomes with reference to assigned targets and the attention given to cross sectoral coordination by the EPB mandarins, minimized failures that can short circuit linkage effects and stymie industrial change (Rodrik 1996).”

It’s useful to emphasize several themes about this broad policy approach.

1) There is a built-in emphasis on export-led development. This is a useful policy guideline, because a government can do a lot of things to favor its companies in their domestic sales, but expanding market share in international markets suggests genuine growth of competitiveness. Korean companies that couldn’t meet their targets in international markets found that their government subsidies were reduced or eliminated.

2) Korea’s development strategy was very broad-based, going way beyond offfering subsidies and cheap credit to certain industries. For example, World Bank data says that back in 1970, about 15% of the over-25 population in Korea had completed “upper secondary” education (basically the equivalent of high school. By 1990 the proportion was about half, and now it’s about three-quarters. “Education especially in STEM disciplines and the development of industrial skills was also a priority from the very outset. Thousands of Korean students went abroad to study and links with foreign universities also facilitated a sharing of information on curricula and teaching modalities. Industrial diversification could not have succeeded had the supply of human capital and workforce skills fallen short .”

This graphic shows a range of policies: interventionist support for manufacturing, but also development of human capital (including heavy use of vocational training, science and technology education, and overseas training and education), strong support for research and development; and policies to bring technology from around the world and diffuse it into Korean firms.

3) Korea followed a step-by-step approach to development, with each step building on the previous one.

Korea’s manufacturing in the 1960s revolved around light, labor intensive activities such as
garments, footwear, toys, food products, and light consumer electricals. The kind that were the norm in other low income economies. But starting in the early 1970s, Korea initiated a structural break and launched its heavy and chemical industry promotion plan (HCIPP) so as to diversify into more complex and technology intensive products. It constructed a state-owned iron and steel complex at Pohang financed in part by Japanese grants, a machinery production complex at Changwon, a petrochemical complex at Wulsan, an electronics complex at Kumi, and a major shipbuilding yard at Ulsan.

The development of individual industries followed a step-by-step approach. For example, Korea’s shipbuilding industry started with domestically produced steel, cheap local labor, and subsidized government loans. But then:

[L]eading shipyards such as Hyundai Heavy Industries sought foreign assistance on such
areas as ship designs, operating instructions, the design of dockyards, and production processes. They hired European engineers and technicians to assist with the running of the shipyard and training of the workforce and adopted quality control measures modeled on the best practices of leading competitors. This plus the recruitment of newly minted engineers from Korean universities—with some having received advanced training overseas—aided in the rapid upgrading of the workforce and allowed the shipyards to largely dispense with foreign assistance by the late 1980s. A more capable, tech savvy workforce plus learning by doing delivered substantial gains in productivity as well as in quality (Kim and Seo 2009). These have continued into the present day with Korean firms among the frontrunners in the production of smartphones, autos, consumer durables, and engineering equipment.

Design capabilities took longer to acquire—close to fifteen years. For some years ship designs were acquired from overseas and technology licensed to build new types of higher value ships such as LNG carriers. This dependence gradually tapered once in-house research and testing facilities had matured. During the latter half of the 1980s, Korean shipbuilders were responsible for advances in protective coatings, welding techniques and in core technologies related to ship propulsion, engine performance, and hull design to minimize pressure and friction drag.

4) Korea’s government invested heavily in infrastructure: seaports, airports, roads, rail. But as a more recent example, World Bank data shows that the share of Koreans using the internet went from 7% in 1998 to 73% just six years later in 2004–and the share has been above 90% and rising since 2016. A US-based techie friend of mind said to me: “We have the fastest home internet access available where we live, which means it’s almost as good as the crappy internet access in Seoul.”

5) Continual technological advance was seen as essential. Korea spends more of its GDP on R&D than just about any other country.

Technology was seen as essential to the success of industrialization and export mpetitiveness. MOST (Ministry of Science and Technology) and KIST (Korea Institute of Science and Technology) were established in order to promote technology transfer and absorption by Korea’s nascent manufacturing sector. Incremental institutional additions continued through the 1970s and the 1980s with the creation of the Korea Advanced Institute of Science (now KAIST, the leading S&T university), as well as a flock of specialized government research institutes (GRIs), many located in the Daedeok Science Town, which later morphed into the Daedeok Science Valley, housing public and private research entities employing thousands of highly trained professionals.

6) Korea had some advantages from its global neighborhood: “East Asian neighborhood effects that conferred reputational advantages and attracted the attention of foreign buyers and investors …”

How does one sum up this policy approach? It’s true that Korea’s government subsidized certain industries. As Yusuf writes: “”The state created a controlled environment in which competition among Korean producers was encouraged but the market was protected by tariff barriers, and by restrictions on the entry and exit of firms. Moreover, Korean exporters were assisted by export subsidies, tax benefits and subsidized financing.”

But the Korean government also did a good job of choosing the industries to be subsidized, in a way that didn’t just pump money into existing production or doomed sectors. It emphasized broad education and improving the skills of its workforce. It emphasized modern infrastructure. It emphasized moving up the technological frontier, by importing expertise when needed, but also by taking many steps to develop its own technology. Also, Korea embraced the continual evolution of technological expertise–and the continual disruption that it brings.

South Korea’s economy faces ongoing challenges moving forward, like other economies around the world. For South Korea, a big shift is that much of its economic growth has been based on large industrial conglomerates (“chaebol”) doing various kinds of increasingly sophisticated manufacturing. But for the world economy as a whole, with robotics and automation on the rise, economic growth is increasingly in the services side of production–innovation, design, and support services like finance and legal–not in the making of physical objects. Thus, Korea’s current government plans emphasize continued technological skills and sophistication, but also application of that technology in small- and medium-size enterprises, as well as in the services sector.

There’s an ongoing argument about whether the specific subsidies to industry were more important than the more general improvements in human capital, infrastructure, and support for high levels of investment and technological growth. I won’t try to untangle that knot. But it seems clear that the industrial subsidies by themselves, without the array of broad supporting policies, would not have had the same success.

About a decade ago, a concern emerged that China was dominant in global production of what are called “rare earth elements”–including notably yttrium, neodymium, europium, terbium, and dysprosium. (Giving the atomic number for each of these is left as an exercise for the reader.) These materials are, at least with current technology, essential for manufacturing certain clean energy technologies, like components of electric vehicles and wind turbines, as well as some materials used in defense-related technologies, like permanent magnets and certain coatings for jet engines.

In a post back in 2015, I pointed out that China had been making threatening sounds about refusing to export these materials, but also that market had been reacting to that threat by figuring out how to conserve on the demand side and how to find alternative sources and recycle on the supply side. A group of researchers from the Rand Corporation–Fabian Villalobos, Jonathan L. Brosmer, Richard Silberglitt, Justin M. Lee, and Aimee E. Curtright– provide an updated view of the situation in “Time for Resilient Critical Material Supply Chain Policies” (2022).

Here’s a figure offering a timeline of events. Focus your attention on the graph at the bottom, which refers to production of REOs, or “rare earth oxides,” which is the amount produced after the ore has been refined. From about 1998 up through 2013, more than 90% of the rare earths being produced originated from mines in China (black line). This was the period when China was threatening to limit or cut off supplies to the rest of the world. But starting around 2016, global mining of rare earths rose substantially (red line), while mining in China rose much less (blue line). China’s market share of rare earth elements remains high, but has now fallen to about 55%.

The Rand researchers emphasize that China’s market share remains large, but they also make the point that China’s share of processing these rare earth elements is even larger–a legacy of the years when China was almost the only global producer. They write:

Current production and processing activities demonstrate that China’s window for
disruption of these critical material supply chains may be narrowing, but not eliminated, and remains a strategic geopolitical tool for its leadership. …

As shown in Figure 2, China’s share of extraction peaked in the late 2000s, when it provided upward of 95 percent of global REE [rare earth element] output; it now accounts for 55 percent. However, China accounts for about 80 to 90 percent of processing and separates nearly all heavy REEs; because of this dominance in the market, the processing sector
is where the risk of disruption is greatest.28 is the same sector that China leveraged when Japan was threatened with REO restrictions in 2010. In fact, nearly all rare earth ore mined in the United States (43,000 tons in 2021) is sent to China for separation and purification. This dominance resulted from China’s ambitious mining programs, an inability to enforce national policies that led to illegal mines, and a disregard for pollutants and by-products that create
environmental waste.

The Rand researcher go into some detail about risks of disrupted supplies, but in broad terms, the policy steps here are fairly straightforward. (You will be stunned to know that the Rand researchers believe a “whole-of-government approach” may be needed here. Personally, I’m not holding my breath for the announcement of some policy problem that is said not to require a “whole-of-government” approach.) For the short-run, having stockpiles of rare earth elements that could last a few months may be useful.

For the longer run, the answers involve allowing expanded mining for these rare earth elements, as well as finding ways to substitute other more common minerals for some of their uses and to recycle already-mined supplies. It also requires building processing plants for these materials. Some of this, like searching for substitute materials, the private sector will eagerly do on its own in response to supply shortages and higher prices. But for other parts–like opening new mines and processing facilities–governments are going to have to strike a balance between environmental concerns and actually allowing a reasonable number of such projects to go ahead at a reasonable time and cost. That part of the challenge may be more political than economic.

It’s been said that colleges and universities have an edifice complex: they like big buildings. Major donors like have having their names on buildings–or at least chiseled into the lobby. Faculty (setting aside the precarious position of adjunct faculty) feel that they need separate offices–even if they often work from home, or the laboratory or library, or even if they go off for a few weeks every year for seminars and lectures at other institutions. The conventional wisdom is that potential students, making their choices, judge by the evidence of their eyes whether the buildings of an institution seem stable and substantial.

My wife and I sometimes note that when you are a child, houses seem like nearly permanent objects, like the Egyptian pyramids. Then when you become a homeowner, it comes as a shock to discover that your house is instead a bunch of systems–roofing, windows, heating/air conditioning, water, electrical, large appliances, carpets, siding, painting–that are continually wearing out and breaking down. Across the country, many colleges and universities have been tempted to take a parallel approach to campus buildings: build them as if they were permanent and unchanging, and then be astonished at the concept of maintenance.

Scott Carlson provides a readable overview of the issues in “The Backlog that Could Threaten Higher Ed’s Viability: A Big Bill for Deferred Maintenance is Coming Due” (Chronicle of Higher Education, March 31, 2023). Carlson writes:

Meanwhile, bricks, steel, concrete, and mortar follow the laws of entropy. As a rule, buildings have two critical stages in their lifetimes: At 25 years, a building needs significant updates and renovations; at 50, a major overhaul of its structure and systems. In recent decades, colleges went through two peaks of construction, one in the early 1970s and another in the late ‘90s and early 2000s. Do the math: Two building life cycles will come due in the 2020s.

Lander Medlin, the president and chief executive of APPA, an association of higher-education facilities managers, points out that the construction costs of a building represent only about 25 percent of the total lifetime expenses. Recurring annual costs, like utilities, everyday maintenance, and operations, represent another 35 to 40 percent. The rest is periodic costs in the lifecycle of essential building systems: replacing the roof after 50 years, updating the heating and cooling system after 20, the plumbing and wiring, the building’s skin, and more.

Carlson cites a number of examples: for example, the major state university branch in my own metro area, the University of Minnesota-Twin Cities campus, “has a 10-year renewal need of $4.2 billion, with more than seven-million-gross square feet of space in poor or critical condition.” It is not an exceptional case.

Moreover, when money is tight, postponing some deferred maintenance often feels easier than the other options. But the number of college and university students has been declining for a decade now, and the combination of US demographic trends combined with pandemic and political factors that have made the US a less attractive choice for international students, the decline seems likely to fall. For a lot of what faculty members do, including the rise in online classes, work-from-home is a very plausible option.

Thus, the supply of space at institutions of higher education is to some extent already determined by the decisions of the past, and some major categories of demand for the use of that space have been in decline. At some places, it’s not just a matter of deferred maintenance for existing space, but of an overload of space. If you walk around a college campus during the standard workweek, you will often see a substantial office and classroom spaces not being used. At a lot of places, this underuse is especially apparent first thing in the morning, as well as Monday morning and Friday afternoon, when students and faculty would prefer not to be in class.

These issues have been around for awhile. Ronald G. Ehrenberg, an economist who ended up doing a stint as an administrator at Cornell, described some of the issues in a 1999 article: “Adam Smith Goes to College: An Economist Becomes an Academic Administrator” (Journal of Economic Perspectives, Winter 1999). With regard to academic buildings, he wrote:

[O]ur trustees have long been aware that new buildings add to the operating and maintenance cost of the university. A rough estimate is that if the building was expected to have a total project cost of a given amount, it would take an equivalent endowment to provide the funds for utilities, custodial services, and routine and planned maintenance over the useful life of the building. This estimate derives from these costs averaging roughly 4 percent of the project costs and 4 percent is what Cornell ‘‘targets’’ as the annual payout, after investment expenses, on its endowment funds. …

Cornell’s trustees have long required that a plan for meeting operating and maintenance costs be present before construction of a building can begin. Realistically, however, once a major donor has committed to funding half the cost of a building and it has been publicly announced that the building will be named after that donor, the idea that construction on the building would be held up because an endowment for maintenance had not been raised is a non sequitur. Furthermore, our ability to raise the additional construction costs, let alone the endowment for operations and maintenance, was somewhat uncertain and based upon forecasts of our development staff. So, while the university hoped that funds to endow a
maintenance fund for the building will be raised, we instead planned to pay for the needed operating and maintenance funds that will not come out of indirect cost recoveries from our annual operating budgets.

Inevitably then, this new building will compete for funds with faculty positions, graduate student support and faculty salaries. The very same faculty members who vehemently argued that the institution needed the new facilities to remain competitive in the physical sciences and engineering are likely to turn around and chastise the administration for spending too much on buildings and not enough on faculty salaries, new faculty positions, and graduate student support. … Many faculty members understand the tradeoff between buildings and other costs, but apparently only after their unit’s new building is finished.

Raising money for a new building at a college or university can be difficult, but being able to put the donor’s name on the edifice helps. Maybe with some creativity, donors could put their name on the renovation of a building: The Smith Renovation (Re-Imagining? Resurrection?) of good old Jones Hall. Or in some cases, the right answer will be for the college or university to rethink and shrink its use of physical space, and the best answer will be the Smith Quadrangle (or playing field, or garden) which will be sitting in the space where good old Jones Hall used to be.

Credit Suisse was founded in 1856, and then shut down earlier this month by Swiss bank regulators, who forced the sale of the firm to UBS. Thus, there is some irony and even poignancy in looking at the just published 2023 yearbook of the Credit Suisse Research Institute, titled “Credit Suisse Global Investment Returns: Leading perspectives to navigate the future,” written by Elroy Dimson, Paul Marsh, Mike Staunton. A summary edition of the report is freely available online.

Each year, a main emphasis of the report is on long-term returns going back to about 1900. Here’s a graph showing nominal and inflation-adjusted returns for US stocks, bonds and “bills” (short-term government debt). Yes, investing $1 in a diversified portfolio of US stocks back in 1900 and reinvesting all dividends since then would have led to a real gain by a factor of more than 2,000 since then. (Notice that the vertical axis is logarthmic, rising by factors of 10.)

In addition, this long-term perspective–using annual data–puts some prominent events into perspective. The authors write:

The chart shows that US equities totally dominated bonds and bills. There were severe setbacks of course, most notably during World War I; the Wall Street Crash and its aftermath, including the Great Depression; the OPEC oil shock of the 1970s after the 1973 October War in the Middle East; and four bear markets so far during the 21st century. Each shock was severe at the time. At the depths of the Wall Street Crash, US equities had fallen by 80% in real terms. Many investors were ruined, especially those who bought stocks with borrowed money. The crash lived on in the memories of investors for at least a generation, and many subsequently chose to shun equities.

The top two panels of Figure 10 set the Wall Street Crash in its long-run context by showing that equities eventually recovered and gained new highs. Other dramatic episodes, such as the October 1987 crash, hardly register; the COVID-19 crisis does not register at all since the plot is of annual data, and the market recovered and hit new highs by year-end; the bursting of the technology bubble in 2000, the Global Financial Crisis of 2007–09 and the 2022 bear market show on the chart but are barely perceptible. The chart sets the bear markets of the past in perspective. Events that were traumatic at the time now just appear as setbacks within a longer-term secular rise.

But it’s also worth remembering that the US investment experience is extraordinary. As the authors put it, it would be “unwise for investors around the world to base future projections solely on US evidence.” Here’s a figure with international comparisons. Although two tiny stock markets, South Africa (ZAF) and Australia (AUS), have outperformed the US stock market over time, the US market has dominated world returns. For the record, most of the abbreviation here are for countries, but WLD is the index for the entire world, WXU is the world leaving out the US, EUR is Europe, DEV is developed markets, and EMG is emerging markets.

The extraordinary growth in the US stock market since 1900 means that, when it comes to global equity markets, the US stock markets dominate the world. Here are the sizes of stock markets around the world in 1900 and in 2022.

The American public is in favor of less total government spending, but it would prefer to avoid reducing spending in almost every category. Here are two figures showing results from an AP/NORC poll (March 29, 2023).

The first figure shows results for whether people believe the government is overspending as a whole. Overall, 60% of the public thinks government spends “too much” and 16% says “too little,” with 22% in the “about right” category.

But when you ask about specific categories, the public wants to see expanded spending in most areas. The only area which has a clear majority for spending “too much” is assistance to other countries. Other surveys have typically found that the public vastly overestimates the amount spent in this category, usually thinking that it covers about 25-30% of total federal spending–when it actually is only about 1% of all federal spending.

The conflict between these results–which are from the same survey!–suggests that the public wants politicians who advocate both for less total spending and also more spending in many individual categories. On this subject, in other words, the public is wide-open to embracing demagoguery.

Many of us comment on China by reading the second-hand literature published in English. Yasheng Huang is Professor of Global Economics and Management at MIT’s Sloan Business School, who came from China to the United States in the 1980s and has thus had the freedom and a front-row seat to study China’s evolution since then. Tyler Cowen has one of his “Conversations with Tyler” with Huang “on the development of the Chinese state” (March 8, 2023, audio and transcript available). Much of the discussion is about how the tradition of China’s civil service examinations evolved over the centuries, and the effects on literacy, creativity, and commerce. Here, I’ll focus on some of Huang’s comments more related to current events:

What is a big misconception about China’s economy?

[O]ne of them is that they look at the Chinese R&D spending, and they look at, for example, some of the impressive technological progress the country has made, and then they drew the conclusion that the Chinese economy is driven by productivity and innovations. In fact, studies show that the total productivity contributions to the GDP have been declining in the last decade and even more. As China has begun to invest more in R&D, the economic contributions coming from technology, coming from productivity have been actually declining. In the economic sense, it’s not a productivity-driven economy. It is an overwhelmingly investment-driven economy.

I think that’s one of the biggest misunderstandings of Chinese economy. It entails implications about the future prospects of the country, whether or not you can sustain this level of economic growth purely on the basis of massive investments.

Huang also offers some thoughts on the nature of political protest in China and how the Communist Party shapes the form of protests in a way that helps the Party hold on to power.

There’s a difference between a civil society consisting of isolated individual actions and a civil society that consists of organized activities that have a program, that have financial support, that have the capability to operate independently. By the second criterion, China has none of that.

If you look at the recent protests against Zero-COVID controls, let’s keep one number in perspective. By various estimates, in 2022 there were probably 400 million people under some sort of long-term quarantine. And let me just concretize that word quarantine. That means you’re essentially locked up in your home, sometimes for weeks, and in some cases, for two months. That’s the level of the suffering, and sometimes you can’t get food. Sometimes you cannot get patients into the emergency room because the hospitals also shut down, refused to take in patients who are tested positive or who cannot show a negative test on COVID. Some people have died. There are suicides, there are fires, and all these collateral damages from the Zero-COVID control.

Relative to that, China experienced a wave of protests — by one estimate, in 17 cities. I don’t really have a good idea how many people were involved, but we are not talking about millions of people. We’re talking about maybe 10,000 people, or tens of thousands of people.

Contrast that with Iran. In the case of Iran, one woman died in the hands of the moral police. There were other grievances, but that was the trigger. The protests are still going on. Millions of the people took to the street. … If you look at the color revolution in Tunisia, it started with a peddler whose assets were confiscated by the government official, and then he committed suicide. That sparked the color revolution.

Those kinds of brutalities toward small peddlers happen almost on a daily basis in China. It’s very important to specify, relative to the grievances and the level of the misery . . . We’re not talking about large-scale social movements here. These are individual actions. …

If you look at what the CCP has been doing, it is actually quite clever. It’s not the case that they don’t take input from the society. They create portals, they create websites, and they create phone numbers for the citizens to call in. They also do surveys. What they want to do is, they want to solicit opinions and information from the citizens without creating conditions for the citizens to get organized. If you think about all these opinions expressed to the government through the government control portals, you are doing it as an individual. You’re not doing it as a member of a larger group. The CCP has no problem with that, and sometimes those opinions can be quite negative. The CCP has no problem with that. …

Yes, China has had a lot of protests, but those protests tend to happen in rural areas, in less urban settings, in isolated situations, and on single issues. Usually, in the 1990s, it was about the land that the government took away. And then it was about the salary, that employers were late in paying my salary, so there were protests about that — very single-issue, very focused.

This time around, you’re talking about people demanding the CCP to step down, demanding Xi Jinping to step down. That’s just something entirely different from what we saw before. …

The reason for that is, I think — although it’s a little bit difficult to generalize because we don’t really have many data points — one reason is the charisma power of individual leaders, Mao and Xiaoping. These were founding fathers of the PRC, of the CCP, and they had the prestige and — using Max Weber’s term — charisma, that they could do whatever they wanted while being able to contain the spillover effects of their mistakes. The big uncertain issue now is whether Xi Jinping has that kind of charisma to contain future spillover effects of succession failure.

This is a remarkable statistic: Since 1976, there have been six leaders of the CCP. Of these six leaders, five of them were managed either by Mao or by Deng Xiaoping. Essentially, the vast majority of the successions were handled by these two giants who had oversized charisma, oversized prestige, and unshakeable political capital.

Now we have one leader who doesn’t really have that. He relies mostly on formal power, and that’s why he has accumulated so many titles, whereas he’s making similar succession errors as the previous two leaders. Obviously, we don’t know — because he hasn’t chosen a successor — we don’t really know what will happen if he chooses a successor. But my bet is that the ability to contain the spillover effect is going to be less, rather than more, down the road, because Xi Jinping does not match, even in a remote sense, the charisma and the prestige of Mao Zedong and Deng Xiaoping. There’s no match there.

I always gain some additional useful perspective from reading Huang’s work. Back in 2012, he wrote  “How Did China Take Off?” for the Journal of Economic Perspectives, where I work as Managing Editor. He made a persuasive case that most of us tend to see China’s economic takeoff as a matter of foreign trade and exports. However, he argues that the early stages of China’s burst of economic growth, through the 1980s and 1990s, were actually led by rural industry in the form of “township and village enterprises” that were led by private entrepreneurs in the context of a high degree of financial liberalization. At this time, China’s economic growth was driven primarily by a rise in China’s domestic consumption, not by export sales. However, in the early to mid-1990s, Huang argues, China’s leadership switched from a rural to an urban focus, took over the financial sector, and essentially drove the rural-based township and village enterprises out of business in favor of expanding state-financed and -controlled urban enterprises.

Sometimes “industrial policy” is defined very broadly, as when people say: “Every country has an industrial policy–even not having an industrial policy is a kind of industrial policy.” But in a more specific meaning of the term, “industrial policy” doesn’t include, say, support of K-12 education or university research and development or a well-regulated banking system. Instead, it refers to when government targets the growth of specific industries with subsidies or trade protection, in the belief that these industries will repay the near-term government support by leading to stronger growth that benefits the broader economy in the future.

In the more limited use of the term, the United States is embarking on a major experiment in industrial policy. The Creating Helpful Incentives to Produce Semiconductors (CHIPS) and Science Act focuses $280 billion over the next decade on building a domestic semiconductor manufacturing industry. The Inflation Reduction Act (IRA) commits $579 billion over the next 10 yearswith a heavy focus on promotion of noncarbon methods of producing electricity from non-carbon sources (mainly solar and wind) and supporting energy users in switching to the use of such energy (including subsidies for electric cars). The Infrastructure Investment and Jobs Act (IIJA) commits $1.2 trillion over the next decade to standard infrastructure like roads, bridges, rail, and transit, which don’t fit into a narrow definition of infrastructure, but also includes less-discussed infrastructure like broadband, electrical power, and support for non-gasoline infrastructure for cars.

In an article from Deloitte Insights, William D. Eggers, John O’Leary and Kevin Pollari discuss “Executing on the $2 trillion investment to boost American competitiveness” (March 16, 2023). They emphasize that the new laws involve large amounts of money, with many different funding streams, all with different compliance standards, that need to be run and coordinated across a large number of federal agencies. In addition, the chances of success will often depend on interactions between these programs, not on the sum of the individual programs.

It stands to reason that industrial policy isn’t simple. If industrial policy was as simple as tossing a log on the fire and getting the desired heat, then every nation would be able to do it. Having a boom in manufacturing jobs, or union jobs, or strong industries related to semiconductors, green energy jobs, steel, or cars, would just be a matter of passing the legislation. But it’s obviously not that simple and easy for industrial policy to work, not in the United States and not in other countries either.

There are many examples of these complexities and constraints: I’ll just give a couple of examples here. The Deloitte authors write about the infrastructure act:

Under IIJA alone, more than 45 federal bureaus and 16 federal agencies and commissions are allocated funding for 369 new and existing programs. Grants fund more than 200 programs and represent 78% of the total funding. … These three new laws establish more than 160 entirely new programs. IIJA alone has created 129 new programs with more than $226 billion in funding. Seven existing programs worth $275 billion have been substantially revised or expanded. In the IRA, out of the total $228 billion appropriated across 18 federal agencies, more than $80 billion was appropriated for 34 new programs.

Thus, the basic workability of the new laws depends an ability to administer the money across these bureaus and agencies and commissions and grants–ranging across federal, state, and local government actors as well as universities and the private sector–and to do so in a way that actually boosts competitiveness and isn’t just a money trough for the politically connected.

As another example, “The CHIPS and Science Act, for example, has earmarked $10 billion for the Department of Commerce to create 20 regional technology hubs across the United States  in partnership with universities and private businesses.” I’m a supporter of funding for regional technology hubs, but I’m not fool enough to think that organizing them is easy.

It’s not just bureaucratic constraints, either. The Deloitte authors note an estimate that the “the country will need one million additional electricians for the clean-energy transition.” Maybe that estimate is overstated? Maybe we only need several hundred thousand more electricians. But all the plans for installing new public and home charging stations, as well as building new electricity charging facilities and transmission lines, are going to fall flat if there aren’t plenty of electricians to do the work. In turn, the electricians won’t be able to do their work without getting necessary permits, which in turn will have to pass zoning, land-use, and environmental regulations and lawsuits.

Moreover, all of this needs to happen in a way that is accountable and, if not fraud-proof, at least fraud-resistant. The authors call this the “thieving squirrel problem”:

The “thieving squirrel” problem: You put seeds into the birdfeeder, but clever, agile, and highly motivated squirrels manage to eat a big share. The only answer is a birdfeeder designed to limit access and frustrate raiders. With funding levels this large, the problem of waste, fraud, and abuse is real. Especially for agencies that are disbursing sizable grants for the first time, controls baked in up front will be critical. Governments need to ensure proper compliance, reporting, and transparency—or risk rewarding the squirrels and undermining overall trust in the process.

Around the world and over time, “industrial policy” narrowly understood doesn’t have a great reputation. There are a few successes, and a distressingly large pile of failures. It’s worth remember that the resources committed to industrial policy–including money, capital, and human talent–could have been spent on other uses. For example, a big chunk of the $200 billion per year or so being spent on these programs could have gone into supporting pregnant mothers and infant children, or rebuilding public schools, or training a few hundred thousand electricians. Perhaps setting up a steady increase in taxes related to pollution and carbon emissions over time, and then letting the incentive effects of such taxes percolate through the economy, would be more effective–but spending more money is always a more popular way of seeking change.

It’s impossible to prove that industrial policy can’t ever work, for the same reasons that it’s often very difficult to prove a negative. Thus, even if this particular US industrial policy experiment fails, I expect that its supporters will just explain that with more money or commitment or vision or energy or an improved structure, it could easily have succeeded. So this post is just laying down a marker: When these pieces of industrial policy legislation were passed, the comments of supporters often suggested that this iteration of industrial policy was nearly as simple as tossing a log on the fire–virtually certain to succeed. If the programs are only mild success, or a considerable failure, the supporters should have to eat their words.

I hope the supporters are correct. I would prefer to see public money well-spent. In a few years, we can evaluate the results. But the administrative, political and economic conditions for success of industrial policy are a difficult set of obstacles to cross. The Deloitte author do not predict success or failure, but they do say: “Once a law is passed, there is a temptation to assume that desired results will follow. But much will depend on how government actually executes its strategy.”