The proposed merger of the Union Pacific (UP) and Norfolk Southern (NS) railroads would consolidate ownership and control of a significant part of the central arteries or “trunk lines” of the rail network in this country. Currently, four railroads control most of these key components of the rail network, especially for the east-west service, and each has a clear regional dominance. The other two major trunk line railroads are the Burlington Northern Santa Fe (BNSF) and CSX. A number of “branch lines” extend out from the trunk lines, and they are often operated by short line railroads. Although there are 600 short lines, a number of them have common ownership.

Proponents of the UP-NS merger argue that it would allow the combined railroad to operate trains from the west coast to the east coast without having to switch the cars between rail lines at some transfer point. It is speculated that if the UP-NS merger is allowed, the BNSF and the CSX would then seek to combine, creating even more concentrated control over trunk lines. The current position of the BNSF and CSX is that, even without a merger, they will jointly offer through service from west coast ports to various locations in the eastern half of the country.  This service would use the same train for the entire trip with crews changing as the train moves from one system to the other. Notably, the routes that they propose for this combined service will directly overlap and compete with the UP-NS lines. All the major trunk lines already share track usage in some places with each other or with short lines. Furthermore, Amtrak, which only owns a few lines in the Northeast, has operated trains on the trunk lines of all major railroads for its nationwide services since its creation in 1971. In addition, many of the freight railcars in use belong to third parties that use them for their own goods or lease them to others for use. All of this shows that ownership of trunk lines is not essential to the operation of freight or passenger service on those lines.

Increased consolidation of ownership of trunk lines might induce their rationalization and enhancement to facilitate more efficient movement of trains. But this would come at the cost of increased dominance of these essential transportation facilities. Moreover, much of the rationalization and improvement is rational conduct by each major trunk line without any merger. Hence, the proposed UP-NS merger should be forbidden as long as the resulting company both owns the tracks and controls their use.

Beware the Bottlenecks

Trunk lines can be understood as “bottlenecks” through which the bulk of rail freight must pass. Ownership of trunk lines confers the ability to control the use of the capacity—namely, whose trains at what price will operate on those lines. Because the owner also controls the dispatch of trains on the line, it has substantial capacity to affect the quality of service. For example, Amtrak, which enjoys a right to use trunk lines for its services, has a long history of problems of having its trains delayed so that the trunk line owner can send the latter’s freight trains ahead. Where freight lines share track use rights, similar disputes are not uncommon between the track owner and the other railroad sharing that track.

What is concentrated is the control over the track itself. Major interstate highways are a single system with multiple users. The same is true of the inland waterways. Should ownership of railroad trunk lines be separated from operation of freight service by multiple users? Amtrak’s and freight lines’ shared use of some trunk lines suggest that operating freight and passenger trains do not require owning the trunk lines on which those services operate. Moreover, while the capacity of any rail line is ultimately limited, with proper scheduling and dispatch these lines can accommodate substantial increased use. It follows that if ownership of the trunk lines were separated from the operation of freight service, it is probable that the number of competitors in providing such service would increase. Freight rates would likely decline, and assuming proper coordination of use, service itself would increase in efficiency. In particular, services such as the movement of container shipments would likely increase significantly, as trucking companies as well as new entrants would be able to develop a range of through services. The same might be true for passenger service, but given low total demand, this seems unlikely except in a few potentially high-volume routes 

The collaboration of BNSF and CSX to provide a through freight service from coast to coast absent any merger demonstrates the feasibility of such service. Not surprisingly, that proposed service will focus on competing directly with the potential (merged) UP-NS services. This highlights the importance of competition in stimulating innovation in transportation service. Notably, the other ports on the west coast that BNSF serves, but which do not have UP-NS alternatives, are not in the through services package. 

As noted, there are significant problems with the coordination in the use of trunk lines. There are ways to coordinate use to facilitate more efficient shared use of track. The railroad controlling the line, however, has limited incentive to resolve those problems because it has a conflict of interest. Its rational goal is to maximize its own use rather than facilitate the use by others whether they provide a distinct service like Amtrak or are competitors in freight service as in the case of shared track rights.

Experience with Separating Ownership from Operation

Some countries have experimented with separating ownership of the tracks and licenses for operators to use those tracks to stimulate competition in rail services. A leading example is the United Kingdom. In the 1990s, the publicly owned national rail system was terminated. Regional passenger services devolved into private ownership but retained a localized monopoly. In addition to the regional services, long-distance passenger services emerged that provided some competition. Freight service was similarly privatized but in a manner that encouraged competition. A distinct but private authority maintained the tracks and provided coordination of services.  

 By 2024, parts of that experiment failed. The government had retaken the responsibility for ownership and coordination of the rail system itself because of concerns about maintenance of the rail lines. It has similarly retaken ownership and operation of some regional passenger services that had persistent problems in both finances and operations. In 2024, the new Labor government determined that as the remaining contracts for the regional service expire, their passenger service will be absorbed into a state-owned entity that will provide service in the England and the parts of Wales and Scotland where those regions had not already recaptured public ownership of local passenger service. 

On the other hand, the independent freight services and long-distance passenger services will continue to operate. The distinction is that the regional services were local monopolies of rail passenger service, while both the freight services and the long-distance passenger lines operated in a competitive environment. There are five major freight services competing in the UK market. The American short line holding company, Genesee & Wyoming, owns one of the leaders, Freightliner, which is particularly strong in handling containers, including offering its own trucking service to deliver the container.

Remedy Design for the STB

If the UP-NS merger is to be allowed, there should be separation between ownership of the tracks and operation of the freight services on those tracks. Because the law exempts these mergers from standard antitrust review, the decision will come exclusively from the Surface Transportation Board (STB), whose authority to regulate the operation of the rail system extends beyond standard antitrust criteria to encompass a broader concern for the overall public interest in the rail system. If it adopts this strategy, the STB should also require the BNSF and CSX to make a similar separation of rail lines from their operation of freight services. Indeed, this separation should also apply to the other two Class I railroads—the Canadian Pacific Kansas City and Canadian National. Doing so would immediately create six freight service operators even if the STB did not require subdivision of the operating companies. Of course, new entry would have to be authorized as well. The Genesee and Wyoming Railroad is among the most obvious entrants given its UK experience and expertise. The STB would have to develop standards for entry and define the rights of such firms. It might be necessary to have criteria for use of trunk lines such as those into the Los Angeles area, which might have capacity limits.

A central question would be the form of ownership and management of the trunk system. Private ownership would risk exploitation, as such an owner would have monopoly power over the use of its track. The STB would have to impose rate regulation to avoid overcharges. Perversely, the track owners will then have the incentive to make overly costly investments if rate regulation is based on the assets devoted to the business. This is the history of regulated public utilities where rates are based on the capitalized value of the investment in the facilities. If rates are not based on capital investment, however, there would be an incentive to underinvest in the maintenance and improvement of the rail lines. A related risk is the control over dispatch along lines. An owner might have incentives to favor some users over others or use this control to demand additional compensation for priority. 

It is very hard to imagine that the American government would take ownership of the rail system even though it “owns” and maintains both the highway and inland waterway systems. Not only the UK, but many other countries have public ownership of their rail systems. Early in the development of America’s railroad system, states were often the owners of the lines that operating companies used and North Carolina still owns a major trunk line in that state. Indeed, Amtrak is a government entity reflecting the failure of private companies to maintain a successful national rail passenger system. But Amtrak’s problems with getting sufficient support to develop and operate its system cautions against government ownership. Indeed, the deferred maintenance for both the national highway system, especially bridges, and the failure to make timely investments in the inland waterways counsel against public ownership. All are dependent on Congressional appropriations. 

The source of these problems, whether privately or publicly owned, is the conflict of economic interest between rail users and the track owner. Darren Bush and I have suggested that one solution to dealing with the operation of bottlenecks is to have the ownership rest in either a cooperative owned by the users or by creating a condominium type of ownership in which owners of specific use rights share ownership but not management of the overall entity. Both strategies assume that there are a significant number of users whose primary interest is in using the bottleneck as a means to provide some valuable good or service. Hence, the goal of these owners is to maximize the utility of the bottleneck for all its users. This means providing as much access as is technologically feasible, ensuring proper investment in maintenance, and providing the kinds of coordination that would again maximize the use of the resources involved. Our article showed that such strategies have in fact been used successfully in a variety of contexts—from cooperative grain elevators at a time when farmers faced very local, monopsonistic markets for their grain to electric transmission that facilitated the opening of competitive markets for the sale of electricity to local utilities.

In the case of railroads, given the goal is to have actual and potential competition for freight service between all major markets, the logical option for the trunk lines is that of a cooperative in which the users would collectively own and manage that essential element of the business. Given a large number of participants, a necessary element to ensure that the incentives to exploit or exclude are minimized, the day-to-day operations would have to be run by a board of directors and managers. The key here is that the incentives of management would be the efficient and open operation of the trunk system. 

Remaining Challenges Are Big But Solvable

There are many other hurdles that would have to be overcome to make this transformation possible. During a substantial transition period, the STB would have to play a strong role in policing conduct and guiding the development of the new system. The lessons of path dependency tell us that there are almost certain to be major challenges from unexpected problems with such a transition, but there can also be unexpected positives. Among the issues that would need resolution are the ownership and relationship of the short lines and branch lines owned by the major truck lines to the system. While having all rail lines owned by one or more cooperatives is a possibility, given the limited and often focused use made of branch lines, creating a distinct ownership model for such lines might be a better strategy. A similar issue would arise with respect to the switching yards in which trains are made up, and cars are moved around either for unloading or to be dispatched to a new location. Who would own and operate those entities and how would they relate to the operating companies? The experience of the UK system should be instructive.

Yet another challenging issue is the development and operation of a dispatch system that would coordinate not only the freight services but also Amtrack and any other passenger service that might immerge. Interchange among carriers would require development of protocols. This is likely to be especially challenging for low volume freight service that smaller users require. With the technology from AI and other computer systems such as those currently coordinating air traffic, these problems are solvable but nonetheless challenging.

Finally, there is the major question of pricing. The cooperative presumably would have purchased the rail trackage from its prior owner. The price has to be set in some equitable way and the resulting access charges appropriately allocated among uses. Here there would be real questions about differentiation of prices for through service for trains carrying containers, passenger service, as well as unit trains carrying bulk commodities such as soybeans, coal, corn, or oil. 

As the foregoing review suggests a great leap into a cooperative trunk system seems very risky. Perhaps in addition to blocking the UP-NS merger, the STB could consider requiring access to existing trunk lines for other users given the CSX-BNSF and Amtrak examples. As Neil Averitt has suggested, such mandatory access could be limited to specific routes or kinds of freight such as containers where the Genesee & Wyoming as well as national trucking companies are potential entrants. For the reasons discussed earlier, the conflicts of interest between the railroads and such users would be a problem. Still, experimenting with opening some types of rail service to third party participation would provide a way to test the viability of this concept in the American context, as well as a useful experience about the challenges that would emerge.

Peter Carstensen is Emeritus Professor at the University of Wisconsin Law School.

One of the favorite jabs of the left-punching commentator community—from Matt Yglesias, Alex Trembath, and the abundance-flavored podcast “Everyone Gets Pie”—has been to accuse progressives of being “Degrowthers.” While degrowth as a framework has been around for a while, its extremely limited presence in the zeitgeist is a relatively recent phenomenon. Modern progressives are not anti-growth; they simply recognize that growth by itself can’t solve inequality and other social problems. (As Noah Smith wrote, degrowth’s only real influence is in Europe.) The Growthers are punching at shadows. 

A profile in the New York Times from last year is illustrative. Among the assortment of degrowth champions included, the only American mentioned (economist Herman Daly) has been dead for three years. Indeed, even among left-wing economists, there’s very little support for degrowth. Prominent economist Jayati Ghosh dissected the degrowth argument from Kate Soper—one of the prominent degrowth luminaries highlighted by the Times—in Boston Review in a piece whose title did not mince words: “Degrowth is a Distraction.” 

Even among environmental champions, hardly anyone is advancing degrowth-oriented economic policy. Even the much maligned Green New Deal spearheaded by Alexandria Ocasio-Cortez and Ed Markey was not remotely degrowth; it still envisioned a greened economy through the lens of a growing economy, including expanding jobs and opportunity and leaning into the “social cost of carbon,” a market-based mechanism to influence emissions via price pressure. 

That, though, is the beauty of boxing with shadows: you don’t have to worry about taking a punch.

Judging from how often the “degrowth” label is thrown around to malign progressives, however, you’d be forgiven for thinking that the entire progressive movement now proudly championed it. Particularly in light of the intra-left sparring over whether to lean into antimonopolist economic populism or abundance liberalism, degrowth is frequently invoked to paint those aligned with the antimonopolist/populist line as pro-scarcity. In reality, the word hasn’t even been used enough in news coverage to show up in Google’s trend analytics. 

Undeterred by reality, Matt Yglesias has characterized the constellation of abundance critics as “degrowth environmentalism and the many, many strands of progressive organization that claim to be something else (like economic populism) while in practice just serving as skin suits for degrowth environmentalism.” 

As Lee Hepner of the American Economic Liberties Project has pointed out, this (bad faith) critique presents a dichotomy where one needn’t exist. This characterization serves only to flatten a multi-dimensional topic into an either or proposition. Most of the time, someone arguing against maximal growth is not, in fact, pushing for degrowth. Rather, they generally are making a case for why foregoing some growth is an acceptable cost to achieve another policy objective. Or, why a particular instance of growth is actively harmful, or that growth by itself can’t solve many of society’s problems.

This shouldn’t be a controversial position; society makes these kinds of determinations all the time. Coca-Cola’s demand curve was probably a little bit more conducive to growth back when it had cocaine in it. We could probably increase bread production if we once again let bakers cut the flour with sawdust. Is banning those harmful inputs degrowth? Obviously not. We can acknowledge that maximal growth can go against other priorities we care about and cause human costs. 

All this line of argument boils down to is that externalities matter. 

In his essay first proposing an “abundance agenda,” Derek Thompson wrote that growth is an objective “not because growth is an end but because it is the best means to achieve the ends that we care about: more comfortable lives, with more power to do what we want, with more time devoted to what we love.” I think that’s precisely correct. But the implication of Thompson’s case is that growth is good, insofar as it makes people’s lives better. So it follows that when, at the margin, growth produces more harm than benefit, growth is no longer our aim.

Growth is responsible for the widely shared prosperity ushered in across much of the world in the twentieth century, leading to declines in infant mortality, longer life expectancy, and some of the best standards of living in human history. But growth has also driven mass deforestation, soaring income inequality, and the erosion of democracy as power became increasingly concentrated. When a handful of companies control local television stations, it’s easier for a president with authoritarian tendencies to punish his perceived enemies. Acknowledging that duality and seeking balance is different than monomaniacally pursuing economic growth or degrowth.

Demystifying Degrowth

I personally am not a believer in degrowth; I think that the objective ought to be to reorient economic growth away from extractive practices and towards types of production that can be sustained by renewable resources. Many ecosocialist critiques of capitalism center on the mismatch between the Earth’s finite resources and the presumption of infinite growth. I think that this gap can be mostly reconciled by gradually creating economies that emphasize things like knowledge production and the provision of social services, along with cultivating essential resources that can be replenished. That’s a far off pipe dream, but at this point, so is any meaningful implementation of a degrowth approach. 

But if degrowth is taken in the expansive sense used by some advocates of the abundance agenda, then degrowth is anything that moves away from the maximum possible economic expansion. Taken to its logical extreme, this would mean that any approach that is not as resource-intensive as possible would be degrowth, because you are reducing demand for whatever the relevant inputs are and thus not creating the incentive for more resource production. Curing a disease would be “degrowth” under this metric; treating the symptoms would sustain a consumer’s (inelastic) demand for pharmaceuticals and medical care, whereas curative care is finite. But we don’t go around asking people to live with chronic diseases in the name of juicing aggregate demand.

For example, opting to hang laundry outside on a clothesline rather than always using the dryer could be construed as “degrowth.” It decreases the wear on the drying machine and clothes, making them last longer, thereby reducing the demand for new dryers and clothing, and can result in less need for maintenance, reducing the demand for appliance technicians. If tens of millions of people shifted to using clotheslines, that would translate into lost jobs and income, which would then further reduce total consumption. But degrowth shouldn’t sound scary if it’s just hanging your clothes up to dry. 

In a bigger picture sense, degrowth is often about identifying ways to shift from production for profit, which makes allocation decisions based on maximizing returns, to production for use, which makes allocation decisions based on need. But degrowth generally doesn’t seek to simply slash living standards across the board. Rather, it would look for instances like our clotheslines, where it’s relatively easy to substantially reduce resource-intensiveness without immiserating everyone.

Barring a societal collapse, wholesale degrowth seems unlikely. But under the strawman that neoliberal pundits have constructed to try and tar antimonopolists, we do degrowth every day. Every time you wear a sweater an extra time before washing, every time a doctor cures a patient, or every time someone grows their own herbs or vegetables, they are engaging in “degrowth.”

Costs and Benefits

Often in economic analyses, costs and benefits are denominated in terms of how much growth would be stimulated. But that elides that there are costs and benefits to growth as well.

Far too often, growth has been allowed to eclipse non-monetary harms. In the name of economic growth, we have built polluting infrastructure that condemns entire communities to cancer, ignored soaring income inequality, and undermined the fabric of society. 

This is why distribution has to be considered alongside growth. If an economy creates $100 billion in value annually but 90 percent of that goes to the richest one percent of people, is that actually better than creating only $50 billion if that is then distributed evenly across the population? The authors of Abundance assume, in their sci-fi (or perhaps more high fantasy) opening vignette, things like the growth from AI will being equitably shared–but, c’mon, Andreessen and Musk and Altman and Thiel and Huang–NONE of them are going to share anything without state coercion against which they will fight like hyenas.

Relative well being matters a lot in social relations. Most of the time, we judge ourselves in relation to the other people we see in the world. Humans care about fairness; extremely pronounced inequality necessarily makes people feel worse off. 

The immense appeal of “growing the pie” or “a rising tide lifts all boats” is that if it’s true, economic policy can be oriented around a single overriding objective. The nitty gritty work of shaping market structure and tax codes becomes tinkering around the edges. 

But this single-mindedness overlooks the very real harms of widening inequality and becoming reliant on unsustainable production models.

Growth As a Means Not an End

The pursuit of economic prosperity in a real sense requires that growth be properly understood as a means, not an end. Indeed, many of the most cutting socialist critiques of a growth-oriented economy is that the drive for accumulation becomes totally detached from the real reasons we want growth in the first place. 

To overcome such obstacles will require interrogating market structure to question whether growth actually serves the common interest. Many times, economic development has been used to justify a race to the bottom, where communities mortgage their future for hollow promises of jobs and investment that may be fleeting or illusory. Take Iowa’s investment in John Deere. The state provided incentives in excess of $270 million and the company still opted to substantially reduce its workforce across the state. 

Rather than allow degrowth to continue to haunt the imagination of our political discourse, we should embrace an understanding of growth as an instrument of progress. As such, we can acknowledge that it interacts with other policy considerations, rather than overriding them. If we want growth because it makes people healthier, happier, and freer, then it is a public disservice not to weigh consequences that might undermine health, happiness, and freedom. 

Avenging Antimonopolists

Characterizing antimonopolists as degrowth is particularly risible, as the entire point of antimonopoly politics is actually about maintaining a competitive equilibrium in the interest of long term stable growth. The premise of fighting monopolies is that in markets that are not concentrated, firms have to compete against each other, which drives innovation and risk-taking, the engine of economic progress. 

If I had a dollar for every New York Times business story that attributed an industry-wide price hike in the post-Covid era to an outward shift in demand—that is, a story that blamed consumers for higher prices—I’d have, like, 25 dollars. Since the post-Covid era, the Times in particular, and the mainstream business press in general, have been feeding us a steady diet of demand-driven inflation explainers.

The latest installment of this neoliberal series came via a story on why rates for Vegas hotel rooms are surging. The Times reports that in August 2019, the average daily room rate in Vegas (before resort fees) was $120.96, but has since climbed to $162.38 by 2025, an increase of 34.2 percent or about five percent per year. (Another data series shows the average daily room rate rising from $133 in 2019 to $194 in 2024, an increase of about eight percent per year before considering the drop in rates in 2025.)

Per the Times’s telling, backed with an expert economist, Vegas hotel rates are soaring because … wait for it … tourists are engaged in “revenge travel:”

By mid-2021, the city was back with a vengeance. Demand for pool reservations, college basketball tickets and hotel bookings surged as travel rebounded. Prices rose. In August 2019, the average daily room rate was $120.96. This year, it was $162.38. But rates go much higher. The Venetian, for example, offers rooms as high as $1,112 — and that’s before resort fees. The surge in revenge travel said Mr. Woods, the U.N.L.V. professor, led the resorts to raise prices because of increased demand.

The story doesn’t explore any other hypotheses, especially those that might put the onus back on the hotels. Never mind there was significant antitrust litigation against Cendyn, maker of Guest Rev software, alleging that the Vegas hotels use a common pricing algorithm that facilitates coordination and thereby permits hotels on the Vegas Strip to reach the jointly profit-maximizing price. That this challenge was not legally successful based on a horrible ruling by the Ninth Circuit—even Herb Hovenkamp thinks it horrible—doesn’t mean that outsourcing one’s prices to a common pricing consultant has no impact on hotel rates.

(The Sling will have much more to say on this horrible decision in a bit!)

How can we rule out the demand-based explanations peddled by the UNLV economist? If hotel rates are going up because an outward shift in the demand curve, we’d observe both an increase in rates and an increase in rooms rented (output). The figure below shows what happens when demand shifts outwards, holding supply constant.

When demand shifts outward from Demand 0 to Demand 1, prices rise from P0 to P1 but also output increases from Q0 to Q1.

What else can explain an increase in prices? Let’s see what happens when instead the supply curve shifts inwards, holding demand constant, say as a result of a coordinated drawdown in room availability made possible through a common pricing algorithm.

When supply shifts inward from Supply 0 to Supply 1, prices again rise from P0 to P1 but also output decreases from Q0 to Q1.

Where could one go to find evidence of an output reduction? That’s right, in the same Times story peddling demand-based explanations:

Throughout those many versions, Las Vegas had largely endured as an affordable destination, with reasonable hotels and all-you-can-eat buffets. But in its latest iteration, the city is in the midst of a tourism downturn, with an 11 percent decline in visitor volume since last year, according to the Las Vegas Convention and Visitors Authority.

Now visitor volume might not exactly mirror hotel rooms rented but it’s a good proxy for that figure. If visitors believed that rates were too high, they would cut back on visiting Las Vegas, reducing hotel occupancy rates. Fortunately we don’t have to guess: Per the Las Vegas Convention and Visitors Authority, in 2019, city-wide occupancy was 88.9 percent, but by 2024, occupancy on the Las Vegas Strip had declined to 86.4 percent. In the first half of 2025, occupancy on the Strip was 81.9 percent. Because the supply-side contraction predicts a decline in output, the supply-side story is more compelling than the demand-side story peddled by the Times.

If this were just a one-off episode, I’d let it go. But since the onset of inflation in the post-Covid era, the Times and mainstream business press have been feeding us a steady diet of blame consumers or blame workers or blame just about anyone or anything but the folks who are actually setting prices. In a new study, co-authored with a rising Yale senior, Abla Abdulkadir, we examined a dataset of articles discussing price hikes across industries from nine publications: The Wall Street Journal, The Washington Post, The New York Times, The Economist, Politico, Financial Times, CNBC, Bloomberg, and Axios. We then manually analyze a smaller subset of these articles in order to determine how they discuss what motivates rising prices and their overall portrayal of seller behavior. From this manual analysis, we determine that in the majority of cases an outside factor is cited, to varying degrees, as the motivator of price hikes. Here’s a preview of what we found.

The figure above shows a breakdown of price hike causes by year. The blue category on the left, “Outside Factor,” represents the largest category. Stories in this category increased over time, growing from a little over 40 percent in 2023 to nearly 80 percent by 2025.

Our advice for business reporters is to speak with more than one economist, especially if your source has financial ties to the industry under study. Be especially dubious of public relations consultants who are paid to peddle industry talking points, lest you end up carrying that same water. (The Times coverage of the predicted price effects of the pending Union Pacific-Norfolk Southern railroad merger is a case in point.) And if your sources keep peddling demand-based theories, ask whether output has also increased; if it hasn’t, then first ask them why and then consider alternative explanations, including those that shift the onus back on the sellers.

Google’s August 2024 launch of AI Overviews has reshaped online search and upended the original bargain between publishers and Google. Google places its own AI Overviews at the top of its search results pages, providing the user with a synthesised answer to the query. That answer is provided by Gemini, Google’s large language model (LLM), which has been trained on Google’s search index data supplemented by its FastSearch product to ensure it is up to date. By presenting its own product at the top of the search results pages, Google’s AI Overviews takes up a third of the screen real estate, pushing rival publisher content out of sight and out of mind. Not only are publishers deprived of clicks, AI Overviews takes content from web pages of publishers that are made available on the world wide web for the purpose of being indexed in Google’s search index.

The changes to the search results have dealt a significant blow to news and other publishers, which has grave implications for both media plurality and, by extension, freedom of speech.

Our law firm, which represents the Independent Publishers Alliance, the Movement for an Open Web, and Foxglove Legal (a Community Interest Company), filed complaints with the European Commission (EC) and the UK Competition Markets Authority (CMA) in June 2025 (“The Foxglove cases”).

The complaints identify breaches of EU and UK competition law (abuse of dominance) and a breach of the Digital Markets Act (DMA) in the EU and the Digital Markets Consumers and Competition Act (“DMCCA”) in the UK. They are pleaded on the basis that AI Overviews excludes rivals and discriminates in favor of Google’s own AI-based products while exploiting publishers’ content that is made available for Google’s search index but is simultaneously being used to train Gemini. Publishers have no choice in this new bargain, and if they want to be seen on the web, they are forced to allow Google to use their websites for Gemini’s training. The target of the complaint is to stop Google’s interference with freedom of speech and consumer access to news and other content on the internet.

We are seeking to persuade the EC and CMA to take urgent action, and, if possible, impose interim measures to restrain Alphabet from continuing to use web content for AI Overviews without providing an opt-out for publishers, so they can be found in search without having to train Gemini. Our lawsuit has caught the eye of Reuters, TechSpot, and Antitrust Intelligence, among others.

In September 2025, a similar claim was filed in the United States under U.S. antitrust law for Google’s exclusionary and exploitative practices, accompanying a claim for damages, on behalf of Rolling Stone and other magazines owned by Penske Media.

The Armageddon of News Publishers

The Wall Street Journal dubbed AI Overviews the “AI Armageddon” for online news publishers, and it’s not hard to see why. The combination of (1) diverted traffic from a publisher’s sites, (2) a publisher’s coerced choice to appear in AI Overviews, and (3) Google’s use of a publisher’s content in their AI Overviews from this choice spell a doomsday scenario:

1. Websites are provided free of charge to users because they are funded by advertising—either advertising on the website itself (by display ads) or in response to users accepting cookies so that websites can generate a relevant ad. To generate income, websites need traffic from users and clicks on ads. The AI Overviews summaries inherently exclude relevant results from Google results pages and allow users to find answers directly without needing to click on any links. This alteration in search results has increased the rise of the “zero-click” searches or searches that end without clicking on any of the results presented. In the EU, for every 1,000 Google searches, only 374 clicks go to the Open Web. In the United States, the number falls even lower to only 360. The decline in click-through rate for news publishers directly translates into a severe erosion of their online advertising revenue.
2. Chrome’s default-setting agreement with Apple and telecoms companies mean Google has over 90 percent of the market for search in the EU and UK. This dominance means news publishers rely on Google to connect their work with readers via the Google search engine result page. As the news organisations’ primary source of online revenue, publishers depend on Google’s display page. In exchange for indexing publisher content, Google is requiring publishers to supply their content for other uses, such as prompting generative AI (“GenAI”) programs to summarize content responsive to user search requests. Google uses a publisher’s content without consent to train its AI models, including its LLMs and the AI assistant product, Gemini. A model of Gemini customized for Google Search then produces AI Overviews and reduces traffic to the same publisher’s websites that it used to train its model.
3. AI Overviews both push news publishers down the results page—to make space for Google’s own AI product to cover the results page—whilst also scraping the content of those same publishers to generate summaries using Google’s LLMs. The imbalanced bargaining power has been leveraged by the monopolist leading to a coerced choice: publishers must either surrender control of their content to then feature in AI Overviews or withdraw from Google entirely (sacrificing all search-driven traffic and, for many, their economic viability).

This position is a life-and-death predicament for smaller publishers. While the AI Overview answer is offered at the top of the results page in milliseconds, the work done by news publishers to provide this content can take months (alongside years of professional training) for which they are unpaid. Google is simultaneously exploiting the copyright of news organizations and the labor of their journalists without consequences, and taking part in an unlawful course of what the antitrust recognizes as reciprocal dealing—that is, where company A agrees to buy from company B only if the company B agrees to buy from company A.

Owen Meredith, CEO of the News Media Association, stated Google’s AI Overviews threatens the business model of those who invest in journalism and quality information. This changing landscape is evident in the actions of major publishers that are looking for licensing deals to underpin disappearing revenues, such as the New York Times dealing with Amazon to license its editorial content to train Amazon’s AI platform.

As a result, the availability of different voices and hence media plurality is being eroded by AIOs under the guise of offering a ‘convenient’ summary.

Google Is Becoming a Rival News Publisher

By scraping and repurposing publishers’ content, Google is positioning itself as a rival news publisher, now in direct competition with the organisations that created the journalism its platform distributes. In the United States, slightly more than three percent of trending news queries are now being answered by AIOs, which means that for every 100 trending news queries, more than three are answered solely by AI. To develop into a rival news publisher, Google is expected to have leveraged the algorithms it creates.

Google has been a fierce competitor to news and other publishers for years when it comes to advertising online. Indeed, the essence of the cases brought by the U.S. Department of Justice against Google is the search giant’s inherent “conflict of interest.” Google is a publisher of others’ results as well as its own, and it is a competitor to everyone in online advertising

The Mechanics of Google’s AI Overviews

Google’s AI Overviews are increasingly dominating the real estate for searches. The portion of the search results landing page that is visible to a user before ‘scrolling’ is termed as “above the fold.” Since May 2024, AI Overviews has occupied roughly 600 pixels or 100 percent above-the-fold area on a mobile display. And it has occupied 800 pixels or roughly 67 percent of the viewport of the 1,200 pixels available on a standard laptop display.  

In other words, AI Overview now appears in a user’s immediate results page for 67 percent of the global browser market. As of March 2025, AI Overviews are triggered by 19 million keywords—a 91 percent rise in the last six months, demonstrating a rapidly accelerating appearance.

The Future Is Not Bright for Publishers

Google is shifting from a search engine to an answer engine. The impact of its transformation of the results page with AI Overviews is not only reshaping how information is presented but also risking permanent changes to consumer behaviour and market dynamics.

Google’s AI Overviews’ anticompetitive effects on publishers are clear. Market forces will not stop Google from swallowing the Internet. Only an intervention by a court can save the news.

Tim Cowen is Antitrust Chair at Preiskel & Co. Abdea Coomarasamy and Uni Valdivieso Wooldridge are Paralegals at Preiskel & Co.  

We are at a crossroads today in American law. It’s not just about Trump and his cronies clearing out the agencies that keep consumers, workers, and everyday people safe from corporate abuse. It’s not just about Trump refusing to enforce safeguards for corporate accountability—or worse, weaponizing them against working families. It’s also about the many ways that our law has been shaped that too often make it impossible to protect working people and make a positive difference in their lives.

One of the most important of these problems sounds technical, but is profoundly important for understanding what has gone wrong with the law and how we can fix it. That is the way that our laws have created artificial silos between workers, consumers, and small businesses.

These divisions don’t just harm people individually or make enforcement more difficult—though they of course do both. They also weaken our ability to take the bold action that the present moment demands of us: to organize, build political power, and fight back against an economy and a government that every single day becomes ever more rigged against working people. This fight is the fight: to reclaim agency and dignity in the face of overwhelming corporate greed and unchecked power.

And make no mistake, this work is not only about economic justice—it is about the future of our democracy. To reclaim our democracy from oligarchs and billionaires, we need to build a system that working people can believe in. That means reshaping an economy that is insurmountably stacked against working people and dramatically reform a government that too often seems too incapable or indifferent to help. Breaking down the silos between workers, consumers, and small businesses isn’t just about changing the law—it’s about uniting those with shared interests and a common vision for a better future into a political movement that offers real hope, not the chaos and lawlessness we are living through today.

This effort boils down to a simple proposition: the family selling stuff out of their garage on Amazon has a hell of a lot more in common with the guy stocking the Amazon warehouse than they do with Jeff Bezos. And until our law, policies, and politics reflect that, we are never going to be able to build a future that actually works for working people.

Big corporations and the powerful few who run them dominate our economy, our politics, and our country. Their influence, their riches, and their rigged rules come at the direct expense of the people who actually make America great: workers, small business owners, farmers, ranchers, and everyday consumers.

Time and again, private equity, big tech, and big banks have shown they will bend every rule, break any law, and fight any shred of accountability to extract every last dollar—without any concern for who gets hurt in the process. We’ve seen it in how they treat their workers, how they treat the small businesses that fill their shelves and populate their platforms, and even how they treat their own customers.

We must build a new coalition—a new legal movement—capable of meeting the scale of corporate power that we’re up against. We must unite workers, small businesses, and consumers—groups too often pitted against each other—around a shared recognition: that the economy is not working for them, and that it’s not broken by accident. It’s been made this way, by design.

That means we need to be just as deliberate in pushing back. This is our moment to rethink the law according to how people actually live and the shared sets of harms they are facing, rather than how law schools sort their courses. We must stand together to reclaim fairness and dignity in our economy with sharper tools, a broader coalition, and a bolder vision for what’s possible.

Misclassification and Beyond

Over the last decade, we’ve watched corporations shift their business models in ways that too often maximize profits not by innovation, but by exploitation.

The remarkable work of David Seligman and Towards Justice has exposed the abuses of misclassification throughout the economy. Whether it’s Uber drivers fighting for basic protections or Amazon’s last-mile delivery drivers forced to urinate in water bottles to hit Jeff Bezos’s insane quotas, this tactic is being used to strip people of their rights—from fair wages to health insurance to workplace safety.

Yet misclassification is just the tip of the iceberg.

What we’re seeing is a systematic effort to offload risk and cost from corporate balance sheets onto the backs of working people. And to shift responsibility and risk from those with the deepest pockets to those already on the financial brink.

It’s not just gig workers. It’s also the small businesses, the contractors, and the countless foot soldiers in multilevel-marketing schemes that help these big companies make their billions. The truck driver, the franchisees, the family-run supplier. Too often under the law they are treated as purported “equals” to the most powerful corporations in the history of the world. In practice, these working people are often just as exploited as the driver laboring under the abuse of the Uber algorithm.

For example, when I worked at the CFPB, we talked to state-level franchisee regulators. They shared how local supermarkets used to be a source of well-paying, union jobs. Many of those jobs have now been converted into franchises—from the baker to the sushi counter. Gone are the benefits, the stability, and the good wages. In its place are the risk, junk fees, and predatory financing. The regulators recounted how working people were lured in with the promise of owning a business—of finally being their own boss. And how those deals turned out to be dead-end debt traps.

This is not an isolated example. It is part of a coordinated, calculated playbook to strip benefits and protections from working people.

A New Path Forward

To fight back, we must first understand the legal arbitrage that created this dark reality. The laws we rely on to protect people were often written for a world that no longer exists. Our legal system is fragmented—divided into rigid silos that no longer reflect the realities of how people work, live, and do business. And corporations know this. In fact, they’ve built entire business models around exploiting these gaps.

During my time as General Counsel at the CFPB, I saw firsthand how outdated legal categories and artificial distinctions make it harder to protect the people who need it most. One of the clearest examples is consumer financial protection laws, which are often limited to transactions for “personal, family, or household” use. That distinction may have made sense in another era—but today, it often breaks down. A rideshare driver taking out a small cash advance, a solo commercial truck driver financing her rig, a small supplier getting paid through an app, a small restaurateur using a digital point-of-sale system that steers him into a predatory loan, an online seller forced to pay fees for platform access and then more for “ads” that are nothing more than glorified kickbacks. If these are considered consumer transactions, the individuals involved would get significant protections. But there’s a different, weaker set of rules for business-to-business transactions. Given the power dynamics at play, does it really make sense to treat that family selling stuff out of their garage as an equal to Amazon? The answer is: obviously not.

And while we at the Bureau worked hard to go after predatory practices wherever we could justify doing so under the law, these distinctions often created unnecessary hurdles—especially in a judicial climate that is increasingly skeptical of anything counter to the US Chamber of Commerce’s ideological crusades.

After years of this work, I’ve become increasingly convinced that our laws are forcing us to ask the wrong questions. And the result is that working families—already bearing the brunt of endless economic exploitation—are often left vulnerable simply because the law hasn’t kept pace with how the modern economy actually functions.

The result is that we have a system where if you are financially stable enough to get a loan to purchase a car, you can drive your kids to school with a series of protections for that debt. But if you use that car loan to help you start driving another family’s kids to school so you can make money on a ride-sharing app, you don’t.

Corporations are fully aware of these blind spots. They know that if they can push a person just outside the definition of a “consumer” or an “employee,” they can largely avoid legal responsibility for their own misconduct. It’s not a coincidence; it’s a business strategy. And it’s one that regulators and policymakers must catch up to—fast.

Take ride share companies. Are they satisfied with exploiting potential legal loopholes in the employment context? Of course not. We now see them using the same playbook for predatory lending schemes. They offer their drivers credit that walks and quacks like a payday loan, but does not—they claim—come with even the most basic consumer protections—because the drivers are in a business relationship with them. For example, they claim that critical protections against people being locked into a particular payment account and prohibitions on offsetting one’s pay to pay off existing debt do not apply. And of course, they use their favorite trick—buried fine print—that purports to waive away these legal protections.

This is just one of many examples. It’s not just gray areas or differences in opinion—it’s entire products and even industries specifically designed to take advantage of people through the tool of the law.

The path forward is clear: to build real protections in today’s economy, we need to stop thinking in these narrow legal categories, where outdated ideas of markets—including the labor market—are catnip for corporate lawyers and manna for Silicon Valley’s obsession with regulatory arbitrage. We need to start designing systems that recognize the actual experiences of people in our economy—whether they’re workers, small business owners, contractors, or the ever increasing number of people who don’t cleanly fit into these boxes at all.  Because as long as the law stays stuck in the past, corporate power will keep racing ahead—unchecked.

In short, we must focus on power, and the imbalance of it, wherever we find it.

Where We Go from Here

So where do we start? If we are serious about taking on corporate abuse—for workers, consumers and small businesses alike—we need to act boldly, and we need to act together in three important ways.

First, we need to tell the story—and tell it well. We must shine a light on how today’s economy really functions, and who is paying the price. That starts with deep, sustained investigations and reporting—into franchise agreements, gig contracts, retail platform terms, and financial deals that trap small firms in cycles of debt and dependence. These are not theoretical harms. They are real, and they are documented. We need to bring them to the surface, build the narrative, and make sure journalists, lawmakers, and the public see what’s really happening. That means convening stakeholders like we’re doing today, compiling records, and publishing reports that expose abuses.

Second, we must take the fight to the courts and to regulators. There are legal tools already on the books that are underused—and often overlooked entirely—when it comes to the abuse of small businesses. We need to deploy state protections where they exist, the Equal Credit Opportunity Act, federal and state antitrust laws including the Robinson-Patman Act, and new commercial finance laws in creative, aggressive ways. That means supporting private litigation, but also pushing public enforcers—state attorneys general, banking regulators, franchising examiners—to step up. Many of these regulators already have the authority they need to begin getting to work. And we should recognize when legal silos are more a product of norms and habits than anything written in the statutes themselves.

To be clear, fighting misclassification remains essential. Companies should not be allowed to label employees as independent contractors just to dodge basic responsibilities like paying into Social Security, providing healthcare, or following labor protections. That fight continues to be vital.

But even as we hold corporations accountable for misclassification, we also need to recognize the economic reality: millions of workers have been pushed outside the traditional employment structure—not because they chose it freely, but because corporations designed the system that way. So beyond stopping companies from misclassifying employees, we also have to make life better, safer, and more sustainable for those who really are operating as independent contractors or small business owners today. In addition to being the right thing to do for those people, if we raise standards for those roles—through fair contract terms and protections against abuse—we weaken the incentive for companies to misclassify in the first place. Workers deserve protections and stability whether they’re driving for an app, running a franchise, or selling on a platform.

In other words: the goal is not just to close one loophole—it’s to change the playing field so that no one gets stuck in a system where the path to profit is squeezing working people and dodging responsibility. That means extending protections, enforcing against predatory practices, and making sure our laws reflect the real economy people are living in today.

Third, we need to push for stronger, bigger, and bolder policies. We’ve already seen important work get started, with an eye toward what’s possible. For example, in California, remarkable work by a broad coalition of small business advocates and consumer groups has resulted in new laws to rein in predatory commercial lending, require meaningful disclosures in small business financing, and crack down on unscrupulous practices in business debt collection. These are real, concrete steps that begin to recognize that small businesses are not on equal footing with the corporate titans. This bold effort should be a model for statehouses and city halls throughout the country.

And if we’re serious about building power for working people, one of the most foundational things every state must do is ensure they have a strong, modern UDAAP statute—laws that prohibit unfair, deceptive, and abusive acts and practices—and that those protections clearly extend to small businesses. Because when a small restaurant is pushed into a predatory financing deal, or a seller on a major platform is subject to hidden fees and manipulated terms, that’s not “just business.” That’s abuse. And until our laws are equipped to recognize and stop it, we will continue to leave millions of people vulnerable to the very forces hollowing out our economy.

Seizing the Moment

None of this will happen unless we build the coalition to make it possible. This is a moment for advocates—across labor, consumer, and small business—to come together. To align not just around a shared set of values, but a shared understanding of who holds the power in this economy—and how we’re going to take some of it back for the guy who runs the halal cart, the mom working a shift before picking up her kids, and the restaurant owner trying to make it work.

Big business has been playing by its own rules for too long. And those rules have crushed too many workers, too many small businesses, and too many dreams.

We are not powerless. In fact, for the first time in a long time, we are seeing the potential for a united front, standing together to demand fairness, accountability, and a better system.

This is a moment to build a new coalition, united by a common enemy—corporate abuse.

Let’s not miss the moment. Let’s use every legal, policy, and political tool we have. Let’s tell the story clearly, act boldly, and build the coalition we need to take on corporate power and win.

Let’s get to work.

Seth Frotman is a Senior Fellow at Towards Justice and UC Berkeley Center for Consumer Law and Economic Justice. Previously, he served as General Counsel and Senior Advisor to the Director of the Consumer Financial Protection Bureau. This post is adapted from remarks at a conference hosted by Towards Justice and Small Business Majority.

After nearly five years of intense litigation, the landmark United States v. Google antitrust trial is finally over.

Judge Amit P. Mehta ruled last fall that Google illegally monopolized internet search for more than a decade by paying smartphone makers, wireless carriers, and web browsers to be the default—and sometimes exclusive—search engine for users. The case then moved on to the remedies stage, where Justice Department attorneys asked for a sweeping order that ranged from the forced sale of Google’s Chrome web browser, to a ban on all illegal pay-to-play deals, to prophylactic relief that would prevent Google from re-using the same playbook to extend its power into the nascent generative AI sector. Google unsurprisingly asked for a far narrower fix.

On Tuesday, Judge Mehta issued his long-awaited remedy decision, to very mixed reactions. Although the Justice Department may still decide to appeal, it issued a triumphant press release, and DOJ Assistant Attorney General Slater described the agency’s mood as “glass half full.” One big reason for that optimism: the order’s scope will include some fencing-in protections for generative AI markets. Google can no longer leverage its popular Play Store or Maps to coerce smartphone makers into also preinstalling Google’s Gemini AI app on their phones. Google will also have to license its massive cache of data compiled by scraping most of the web, as well as its sought-after search results, to would-be rivals. The underlying idea, pushed by DOJ’s attorneys at the remedies hearing earlier this year, is that a disruptive new technology like generative AI might be our best hope for unleashing real competition in this space.

Other observers are far less sanguine. Criticism of the opinion immediately zeroed in on the fact that the remedy won’t actually address the conduct that landed Google in antitrust trouble. Laypeople and antitrust experts alike tend to view terminating illegal behavior as a bare minimum response to proven violations. Most of us learn this on the playground. Or, as a leading scholarly treatise on antitrust puts it, “relief properly goes beyond merely ‘undoing the act.’”

Yet this remedy order won’t undo, or even halt, the illegal acts. Google can keep paying Apple billions of dollars each year to be the default search tool on hundreds of millions of iPhones, keep paying Mozilla for prime placement on Firefox, and generally keep throwing a portion of its enormous monopoly profits around to maintain its present monopoly power.

Judge Mehta portrayed his decision as a model of judicial humility and restraint. “[C]ourts must approach the task of crafting remedies with a healthy dose of humility,” he wrote. “This court has done so.” As the decision explained, its author has “no expertise” in the business of search engines, online ads, or generative AI. “If judges could accurately chart the path [of] innovation,” the court wryly observed, “we would work on Wall Street (or the Las Vegas Strip).”

But does this decision actually reflect the judicial humility that its author extolled? A judge in this position faces a menu of options that includes stopping the illegal conduct, prohibiting conduct similar to the particular tactics that were already condemned, extending the decree to related markets, ordering the disgorgement of assets gained via the illegal conduct, and mandating structural spinoffs to restore competition. Judge Mehta’s decision combines a quarter-measure of one with a half-measure of another: it leaves most of Google’s current conduct untouched, while taking some limited steps to help spur competition in closely related generative AI markets.

Although the remedies opinion insists strenuously that it’s not gambling on an uncertain future, the justification it offers for this watered-down approach suggests otherwise. Judge Mehta appears to have anticipated a firestorm of criticism for his decision to let Google continue engaging in conduct that he had already deemed illegal. In an attempt to defang that narrative, the decision explains that generative AI startups “are in a better position … to compete with Google than any traditional search company has been in decades,” and these “new realities give the court hope that Google will not simply outbid competitors for distribution if superior products emerge.”

To recap, this is a judge who already found that Google’s payments for default status and exclusive placement are causing significant harm. By locking up the most important ways to reach users, Google cemented its search monopoly and deterred competition for over a decade. Letting Google keep right on engaging in most of its harmful conduct entails real costs for consumers and society at large. This remedies opinion consciously chooses to inflict those costs based on a “hope” of good things happening in the future.

That’s nothing if not a gamble, one with exceedingly high stakes. And this wasn’t just a bad policy judgment. It was also a legal error. The Supreme Court has set down clear rules of the road that—if followed—should’ve yielded a much stronger outcome in this case. A remedy order “must seek” to accomplish multiple goals, one of which is ensuring “that there remain no practices likely to result in monopolization in the future.” That instruction should have prompted, at minimum, a ban on all of Google’s current illegal payment deals.

Our antimonopoly laws are only as strong as the judges who enforce them. This decision puts far too much weight on the mere “hope” of a highly uncertain future. Telling a proven violator that it can freely continue breaking the law is a near-certain recipe for recidivism. Worse yet, this decision also sends a very dangerous message to other would-be monopolists: go ahead and roll the dice.

John Newman is a professor at the University of Memphis School of Law, a former deputy director of the Federal Trade Commission Bureau of Competition, and a former trial attorney with the U.S. Department of Justice Antitrust Division. 

Antitrust restructuring of major corporations is on the table in a way it has not been since the Microsoft case in the late 1990s. Indeed, the historic moment may be comparable to the breakup of Standard Oil in the 1910s and AT&T in the 1980s, when courts reorganized those companies and freed the market from their control. Today, judges face a similar opportunity to rein in today’s rapacious monopolists.

Amid rising public pressure to challenge concentrated corporate power, the federal government has, since 2020, filed antitrust lawsuits against Google, Amazon, Meta, and Apple, as well as dominant firms in the agriculture, debit payment, and rental pricing software industries. These cases aim to break monopolistic control of markets, and not merely to stop unfair practices. The outcome of this litigation wave will determine whether antitrust remains a meaningful check on concentrated private power or operates as regulatory theater.

Antitrust enforcers stand poised to secure favorable judgments in lawsuits that affect multiple sectors of the economy. These lawsuits could bolster farmers’ ability to repair their equipment, reduce the costs retailers incur on e-commerce platforms, increase the revenue software developers earn from their smartphone applications, and reduce the interchange fees businesses must pay to credit and debit card companies.

The poster child for the burgeoning possibilities of antitrust remedies involves the lawsuits against Google. In August 2024, Judge Amit Mehta ruled that Google was liable for monopolizing the search market. He found that Google illegally paid Apple and other manufacturers billions of dollars every year to be the default search provider in web browsers on desktops and smartphones, resulting in the foreclosure of critical distribution channels to competitors. In April 2025, Judge Leonie Brinkema held that Google was liable for monopolizing the ad-tech market, which produces the revenue stream that underpins much of our modern media system. In her decision, she found Google used its dominant control over digital advertising to lock in news publishers, impose supra-competitive take rates on Google’s exchange, and exclude competing digital advertising providers.

Other lawsuits against Google, such as the widely publicized lawsuit by game developer Epic Games, have also found Google liable for monopolization. Given this legal onslaught against Google, odds are in favor that the corporation will undergo some form of corporate restructuring. Due to Google’s immense size and scale, the result would fundamentally alter how the public uses and accesses the internet. Breaking up Google—by divesting Chrome or its digital advertising business—would strip the corporation of its control over access to information and online revenue. The shift would mean more competition, the viability of privacy-oriented alternatives, and greater power for journalists, creators, and users concerning how information is distributed and monetized.

A few federal judges and Gail Slater, the Assistant Attorney General for the Antitrust Division of the U.S. Department of Justice, will control the scope of the remedies to be imposed on Google. Regardless of any obstacles the DOJ staff may face in determining which remedies they should pursue, one thing is certain: federal judges are vested with all the authority they need to impose the government’s demands, and they are obligated to impose sweeping remedies on antitrust violators like Google.

Flexing the Structural Relief Muscle

Remedies convert violations of legal rights into actionable consequences. In his leading casebook, renowned remedies scholar Douglas Laycock succinctly asserted that “remedies give meaning to obligations imposed by the rest of the substantive law.” In other words, remedies are how democratic institutions prove their legitimacy: equipping the law with real force to answer public calls for action, rather than serving as a meaningless political gesture. Without effective remedies, the law merely functions as a speed limit sign with no police or cameras to enforce it.

The antitrust laws are, in the words of Senator John Sherman, namesake of the titular Sherman Act, “remedial statute[s].” To accompany the sweeping prohibitions on “restraints of trade” and monopolization, lawmakers were diligent to buttress these proscriptions with a panoply of robust remedial provisions.

The antitrust laws include the ability for harmed private parties to obtain treble damages and attorneys’ fees (a novel feature in American law at the time of their enactment). Lawmakers authorized federal, state, and private enforcers to initiate lawsuits, eventually establishing two separate agencies to advance the cause. Further supporting these profound legal tools were the equity provisions that empowered enforcers to seek and, critically, courts to impose structural changes to a business’s operations. The purpose of this vast and deep remedial landscape was to facilitate the “high purpose of enforcing the antitrust law[s].”

At the federal level, the thinking surrounding the purpose and necessary goals of antitrust remedies has been lost for some time. With the notable exception of the antitrust litigation against Microsoft in the late 1990s, market restructuring has simply not been seriously contemplated by enforcers since the 1970s, when the DOJ was litigating its antitrust lawsuit against telecommunications giant AT&T for willfully stifling competition in, among others, long-distance services. According to historian Steve Coll’s book The Deal of the Century, the prospect of breaking up AT&T and imposing other remedies permeated the government’s legal strategy.

After the breakup of AT&T, however, in concert with the purposeful decline in federal antitrust enforcement, the intellectual and institutional muscles supporting ambitious remedies quickly atrophied. Decades of underenforcement drained both the doctrinal imagination and the human talent needed to design and implement structural relief.

For the federal government, a new opportunity to implement robust structural remedies almost presented itself in its lawsuit against Microsoft in the late 1990s. A structural breakup was interrupted, however, due to improper judicial conduct and changes in political administrations. Enforcers ultimately abandoned a breakup in favor of paltry restrictions on Microsoft’s business practices. Subsequent academic literature criticized the handling of the lawsuit for the lack of critical thinking regarding the remedies that enforcers wanted. Not since the antitrust lawsuit against Microsoft has another opportunity of similar magnitude presented itself to federal enforcers.

A New Opportunity Arises

Now, more than a quarter century later, the public has another chance to witness the full thrust and potential of the antitrust laws. What is critical is that enforcers and judges need to be reminded that restructuring businesses to restore competitive conditions and prohibiting dominant corporations, like Google, from engaging in unlawful behavior has always been at the heart of what the antitrust laws can and must do.

In one sense, the ability to restructure the economy provides the clearest visible indicator to the public that justice has been served. For too long, the public has witnessed instance after instance of corporations engaging in often blatant lawbreaking and walking away with no more than token penalties, accompanied by the boilerplate legal phrase “this is not an admission of guilt” on the settlement form.

It’s no secret that the public’s confidence in our political system was shattered after the 2008 financial crisis, when only one low-level bank manager was jailed, and the biggest financial institutions only got bigger. Meanwhile, President Obama refused to use his executive authority to prevent millions of Americans from losing their homes and livelihoods. The precipitous collapse of white-collar crime prosecution after 2012 only intensified this perception of corporate impunity, a trend that continues today. It should go without saying that a well-functioning democracy of the kind the antitrust laws were meant to buttress requires punishing wrongdoers.

Structural remedies also bring clarity to the purpose of the antitrust laws. They ensure that corporations are adequately incentivized to and remain subservient to the public, and must adhere to established norms concerning what constitutes lawful means of operating in the marketplace. Remedies, in this sense, serve dual purposes: they deter future violations and, when applied to offenders, reinforce institutional legitimacy by ensuring meaningful consequences rather than merely symbolic fixes.

None of this was accidental. Congress wrote the antitrust laws with ambition. In both the statute’s text and its legislative history, Congress codified deeply held moral and ethical norms, grounded in principles of fair competition, non-domination, and democratic control of the economy. To facilitate these principles, Congress gave the public the tools for broad economic reordering in the event of a violation. It was the Supreme Court’s ideological shift, which began in the late 1970s and was subsequently adopted by the Reagan administration in 1981, that neutered the institutional will to enforce and interpret the laws as Congress intended. Nevertheless, once liability is established, structural change is not merely justified by law—it is mandated by it. Indeed, the Supreme Court has been uncharacteristically clear that liability obligates, not just authorizes, the courts to impose structural change.

In a decision from 1944, the Supreme Court stated, “The Court has quite consistently recognized…[d]issolution[s]…will be ordered where the creation of the combination is itself the violation.” In another decision, the Court opined about the absurdity that would arise if weak remedies were imposed on antitrust lawbreakers. “Such a course,” the Court stated, “would make enforcement of the Act a futile thing[.]” In another decision, the Supreme Court stated that “[c]ourts are authorized, indeed required, to decree relief effective to redress the violations, whatever the adverse effect of such a decree on private interests.” The jurisprudence is replete with many more judicial directives commanding the lower courts to impose sweeping remedies to effectuate Congress’s legislative command.

Not only is there a duty to impose structural remedies, but the Supreme Court has been straightforward that in all but the most wholly unwarranted situations, a district court judge—like Judge Mehta or Judge Brinkema presiding over their respective lawsuits against Google—is afforded broad discretion on what remedy to impose both to “avoid a recurrence of the violation and to eliminate its consequences.” As long as the remedy is a “reasonable method of eliminating the consequences of the illegal conduct,” judges operate with expansive discretion and face virtually no doctrinal constraints on what can be ordered.

If the desired outcomes are realized, a breakup of Google could fundamentally reorganize the structure of the internet and our experience with it. Requiring Google to spin off its digital advertising platform could enable journalists and content creators to diversify their revenue sources through new competitors, giving them greater autonomy over their income. Divestiture could also enable them to have more control over the distribution of their work products and reduce the constant risk of censorship and algorithmic manipulation that Google has deployed to maintain its monopoly over search and advertising. Moreover, a spinoff could also erode the surveillance advertising model, making privacy-friendly alternatives more viable competitors. For the public, more competition in search could expand options for finding and presenting information on the internet. Likewise, a divestiture of Google’s Chrome browser could open new pathways to access the internet and lessen dependence on a single dominant provider.

Corporate Allies Spring into Action

In an attempt to get ahead of the litigation game, executives and ideological friends in the legal academy have churned out scholarship and opinion pieces designed to deter enforcers and courts from imposing remedies deemed “too harsh” to Google’s operations. In a recent paper on structural remedies, Professor Herbert Hovenkamp, a leading establishment antitrust scholar (and a Big Tech sympathizer), provided a hierarchical schematic outlining how remedies should be considered and administered. One of his points stated that “Even with market dominance established, alternatives to structural relief are often superior, and simple injunctions are often best; for any problem, they should be the first place to look.” The International Center for Law and Economics, a member of Google’s “army of paid allies,” submitted an amicus brief to the district court overseeing Google’s antitrust lawsuit, erroneously stating that “structural remedies are disfavored in Section 2 cases[.]”

Naturally, too, Google’s business executives have ardently defended the company’s business practices. In April 2025, Google’s CEO Sundar Pichai testified that any breakup of Google would be “so far-reaching” that it would be a “de facto divestiture” of its search engine. Pichai also decried the forced sharing of the data that underpins Google’s search engine as a remedy that would leave the company with no value. It is revealing to hear the highest-ranking corporate executive at the company admit that Google’s success is dependent on a select few unlawful practices, rather than its business acumen, and that Google is apparently incapable of deploying lawful methods of competition to succeed in the marketplace. Since the filing of both federal lawsuits, Pichai has also embarked on a marketing tour to tout the company’s operations, defend the benefits of its business practices, and detail the potential unintended consequences of the government’s lawsuits. Such alarmism is a standard defensive tactic, deployed to influence judges and sap public support for real solutions.

But from the earliest days of antitrust law, the Supreme Court has consistently affirmed that breakups, divestitures, and other corporate restructuring remedies—though often described as “harsh,” “severe,” or inconvenient by violators—are time-tested, necessary, and appropriate for restoring competitive market conditions. In a forthright statement, the Court stated that antitrust litigation would be “a futile exercise if the [plaintiff] proves a violation but fails to secure a remedy adequate to redress it.” In fact, the Court has lamented that prohibitions on specific conduct—rather than breakups or other corporate reorganizations—are “cumbersome,” delay relief, and position the court to operate in a manner for which it is “ill suited.”

Rising to Meet the Moment

The vigorous enforcement of the antitrust laws in the post-World War II era compared with the drastic decline that began in the late 1970s is clear evidence of the changing “political judgment” (as Professors Andrew Gavil and Harry First call it) concerning what remedies should be imposed. Judges today are far different from their historical counterparts, who viewed antitrust as a facilitator of economic liberty, a bulwark against oligarchy, and fundamental to protecting our democracy. Punishing antitrust violators was not just a legal formality but also a moral imperative.

Even though many judges have not considered these issues in decades—or, in some cases, ever in their careers—during this profoundly important moment in American history, judges should be cognizant of what the jurisprudence plainly mandates them to do. If the rule of law retains any meaning, it demands that courts decisively address the harms the government has been litigating for a half-decade.

The question now is not whether courts can impose structural remedies; it is clear they can. It is whether they will rise to meet the moment. The remedies that judges will impose on Google and the other alleged monopolists in the government’s lawsuits will be a defining test of judicial integrity and democratic accountability to the rule of law. A failure to act calls into question the very legitimacy of our legal system to hold the powerful accountable. As the jurisprudence makes clear, anything less than structural relief results in the public “[winning] a lawsuit and [losing] a cause.”

Daniel A. Hanley is Senior Legal Analyst at the Open Markets Institute.

On December 3, 2024, Chris Salinas officially entered a nightmare that would make Freddy Krueger proud—a nightmare in the medical industry known as prior authorization. Even I, as his gastroenterologist, didn’t know at the time that this one would become my biggest nightmare yet. Chris has given me permission to share the details of his experience, including his medical ailments.

Chris has ulcerative colitis, a chronic autoimmune disease of the colon that can cause flares of bloody diarrhea, abdominal pain, and fatigue, among many other symptoms. I have been treating Chris and his ulcerative colitis for over 15 years. Over those years, for various reasons, he had ultimately failed multiple medications for his illness. By December of last year, the next one we wanted to try was Entyvio, a drug under patent approved by the FDA for ulcerative colitis in 2014. Entyvio is considered in the industry a “specialty pharmacy” drug. In theory, nobody in the industry knows exactly what that means. In practice, what it means is big money, the money that pays for all those ads you see on TV for Ozempic, Wegovy, Skyrizi, and the like (rule of thumb: if the brand has a “z” or “v” or “y” in the name, it’s likely a moneymaking specialty drug). And in practice, what it meant for Chris and me is that I needed to get prior authorization for approval of Entyvio.

Just the phrase “prior authorization” sends a chill down every physician’s spine. On its face, prior authorization has a functional purpose: to control utilizing drugs that can be quite costly for a health plan to cover. Yet imagine your worst call climbing up the giant sequoia of customer service phone-trees. A call for a prior authorization is worse.

You think I’m exaggerating? Every year the American Medical Association conducts a nationwide survey of a 1,000 practicing physicians on prior authorization. In 2024, 40 percent of physicians had staff who worked exclusively on prior authorizations and spent 13 hours weekly on them; 75 percent reported that denials of prior authorizations had increased; 80 percent didn’t always appeal the denials, due to past failures or time constraints. And almost 90 percent of physicians reported burning out from the process. I know physicians who closed down their practices and retired early solely because of prior authorization.

According to the survey, the impact on patient care is no less nightmarish. Nearly 100 percent of physicians felt that prior authorization negatively affected patient outcomes; 23 percent noted that it led to a patient’s hospitalization; 18 percent said it led to a life-threatening event; and 8 percent said it led to permanent dysfunction or death. Fortunately, Chris wasn’t at death’s door. But even with a philosophical approach that most of the time in medicine, less is more, I felt that getting on Entyvio was Chris’s best shot at improving his quality of life, free of the extended flare of ulcerative colitis he was enduring.

So last December, I called Express Scripts, Inc. (“ESI”) to get the prior authorization for Entyvio. ESI is the pharmacy benefit manager (“PBM”) of Chris’s employer-based health plan. For those who haven’t studied the FTC’s damning 2024 interim report on the industry (“FTC report”), PBMs serve as the middlemen between the pharmacies that dispense drugs, the manufacturers that make them, and the government/employers/insurance plans that pay for them. I gave ESI all the required clinical information for approval of both the initial intermittent intravenous infusions for the first six weeks and the subsequent biweekly maintenance injections Chris would be giving to himself subcutaneously. As Chris had by then failed multiple other therapies, it was a relatively easy call for ESI to grant the prior authorization for Entyvio. I then called in both the intravenous and subcutaneous Entyvio prescriptions to Accredo, the administering specialty pharmacy, who assured me that we were good to go.

We were not good to go. At first, Accredo told me that Chris could get the infusions at home through a service set up by Accredo. Then it told me it couldn’t. Then it told me it would find a local infusion center to administer the drug. Then it told me I had to find one. At every twist and turn, I had to initiate the call to Accredo to find out why the infusions hadn’t started. So did Chris, who said about his Accredo calls: “There were many circular conversations where it just went nowhere. Every time they would tell me, you have to give us a week and then call us back; and then when I call them back, it’s just like starting the process over again. And then again, and again, and again.”

The breakdown in the prior authorization process that Chris and I were weathering is an endemic breakdown in healthcare quality. That breakdown in quality is the result of the incredible horizontal concentration and vertical integration the last few decades have wreaked in the healthcare industry. This chart from the same FTC report tells the tale.   

ESI appears in the second column under The Cigna Group, as the associated PBM. The 23 percent below its name indicates that ESI controls 23 percent of all prescriptions filled in the United States.

On horizontal concentration overall, the three biggest PBMs—CVS Caremark, ESI, and Optum Rx, manage nearly 80 percent of all prescriptions filled in the United States; in terms of the standard Herfindahl-Hirschman Index (“HHI”) measurement of horizontal concentration, the mean HHIs were pushing 4,000 and climbing in state and local geographic markets.

On vertical integration, the three biggest PBMs are owned by the three of the five biggest health insurance companies in America—CVS Caremark under Aetna, ESI under Cigna, and Optum Rx under United. The vertical integration doesn’t stop there: the healthcare conglomerates that own the biggest PBMs also increasingly own private labelers that manufacture the drugs, the providers who prescribe them, and the pharmacies that dispense them. This includes Accredo, the specialty pharmacy owned by Cigna—which owns ESI, the PBM!

To any reasonable person thinking through the inevitable conflicts of interest, beware. Your head may eventually explode—for which you’ll probably need a prior authorization for something. The FTC report catalogs many of the adverse effects of horizontal concentration and vertical integration involving PBMs: (1) excluding generic drugs from formularies in exchange for higher pay-to-play “rebates” from the manufacturers; (2) recasting group purchasing organizations into rebate aggregators, often headquartered offshore, that still enjoy safe harbor from anti-kickback law; (3) steering patients exclusively to the conglomerates’ own PBMs and specialty pharmacies while crowding out independent pharmacies, especially when high-profit specialty drugs are involved; (4) turning contracts with independent pharmacies that have little bargaining position effectively into opaque adhesion contracts with “clawbacks” that make it possible for the pharmacies even to lose money, in the end, from a sale; and (5) abusing prior authorization and other utilization management tools to preference the conglomerates’ financial interests over the patients’ best interests.

The FTC report was criticized by a dissenting Commissioner (Holyoak) for being prematurely released without “rational, evidence-based research” to show that horizontal concentration and vertical integration of PBMs raised consumer prices. Accordingly, in 2025, the FTC expanded on its 2024 report, by analyzing additional data received from PBMs, to show that significant markups on numerous specialty generic drugs—some exceeding 1,000 percent—made PBMs and their affiliated specialty pharmacies huge revenues. That, it followed, cost government and commercial plan sponsors significantly more money, along with the subscribers/patients who shared the increasing costs. 

Likely backroom political gamesmanship notwithstanding, the FTC should be praised and pushed to continue focusing on the effect of PBM consolidation on consumer drug prices. But with the focus on consumer prices, neither the 2024 interim report, nor its dissent, nor the 2025 update fully captures why horizontal concentration and vertical integration of PBMs are so bad for healthcare. Take it from an on-the-ground, practicing physician who has been at bedside, for over a quarter-century, observing policymakers afflicted with an evidence-enslaved (rather than evidence-informed) econometrics delirium rotting the core of what matters in healthcare.

It’s not just the price effects—it’s also the quality! Horizontal concentration and vertical integration of PBMs, and healthcare at large, are destroying quality in healthcare. Surely the FTC is aware of these non-price effects, as the over 1,200 public comments received by the FTC on PBMs’ business practices likely revealed. The 2023 Merger Guidelines published jointly by the FTC and the DOJ emphasized the same, adding a “T” to the end of the “SSNIP” of the Hypothetical Monopolist Test: “A SSNIPT may entail worsening terms along any dimension of competition, including price (SSNIP), but also other terms (broadly defined) such as quality, service, capacity investment, choice of product variety or features, or innovative effort.”

Yes, SSNIPTs may be harder to quantify than SSNIPs. But when it comes to PBMs, SSNIPTs galore smack you in the face.

Many phone calls later, and over a month after ESI had granted the prior authorization of Entyvio, Chris finally received his first of three infusions. It cost him around $300. He was then told the second infusion would cost him $1,800. That triggered another set of phone calls involving Accredo, with Chris noting, “They kept telling me that my insurance company denied [the Entyvio], and I said no, my insurance company did not deny it . . . I was actually getting ready to write an $1,800 check.” Fortunately, as Chris explained, the pharmacy and PBM then found the co-pay assistance program information Chris had sent, which they had apparently misplaced. The price of his second and third infusions? Five dollars each. 

Chris’s ulcerative colitis symptoms responded well to the Entyvio infusions. But the healthcare nightmare wasn’t over. Chris was supposed to transition to the subcutaneous Entyvio shots in late April, eight weeks after his last infusion. When he called Accredo in March to confirm, he was told that Accredo didn’t have the prescription for the shots. I then called Accredo and was told the opposite. Chris then called and learned that Accredo had two accounts in his name, one of which was dormant (supposedly from the change of calendar year) and where the prescription was hiding. Accredo assured him that it would merge the accounts, and all would be good. Chris called again a couple of days later, when Accredo told him it was waiting on approval from his insurance company—even though the Entyvio had already been approved. More phone calls the following week, and the prescription was lost again because the accounts had not in fact been merged. Late April came and went, with no continuation of drug.

Chris estimates that between March and May, he made over 20 phone calls to Accredo. During the lapse in treatment, his cramps, diarrhea, and urgency returned. “In the meantime, I’m having to deal with my diagnosis . . . which is not fun, right? I’m having to kind of manage my travel that I do for work around that situation, which is difficult. And it’s a lot of stress . . . being on airplanes and all that kind of stuff.” It was a lot of stress for me as well, knowing that the longer Chris was off Entyvio, the more potential he would have to develop antibodies to the drug that would render it ineffective when resumed.

It all came to a head on May 21, a month after Chris was supposed to start the Entyvio shots. I spoke to Paul at Accredo, then Jeanine at ESI, then Mark, Jeanine’s supervisor, while I made Paul stay on the line. ESI informed me for the first time that a fax was allegedly sent to me on January 3 denying the prior authorization for the subcutaneous Entyvio. When asked whether it had sent the denial by regular mail, ESI said no. The reason for the denial? According to ESI, I had not answered some clinical criteria question—a question I had certainly answered when the infusions and injections were initially approved back in December. When I insisted on clearing up this nonsense once and for all over the phone, ESI informed me that a policy change beginning in 2025 had eliminated approvals over the phone; they could only be done via fax. The very tech that had failed us in communication of the purported denial. All this, with ESI, as a PBM in the business of administering healthcare, knowing full well the impact this was having on Chris’s health.

I finally demanded to speak to a peer, for a peer-to-peer review. In 2024 and 2025, that no longer means a gastroenterologist, or even a medical doctor, but often a nurse or pharmacist. No offense to nurses or pharmacists; they are sacred to healthcare. But when it comes to prior authorization for a specialty drug like Entyvio, they are not my peers. Nonetheless, the peer here was a pharmacist named Stefan—the first person at ESI I felt wasn’t carbon-based AI. He acknowledged the injustice of the situation but said my only recourse was to fax clinical information again to seek approval.

And then, magically out of the blue . . . Stefan said he found some “override” to approve the subcutaneous Entyvio. Just like that, a personal-record-breaking two hours and 45 minutes into the call, the prior authorization was over.

I give you all these gory details (believe it or not, with many left out) not to be melodramatic, but because when it comes to going through the prior authorization gauntlet with these highly concentrated, vertically integrated PBMs, this is now the norm, not the exception. As chronicled in the public comment by the American Economic Liberties Project, past lapses in enforcement by the FTC and DOJ created these consolidated beasts. And if you download the appendices of that comment, you will read one similarly abominable experience after another sampled from anti-PBM Facebook groups with names like “DOWN WITH Express Scripts and Accredo!”. On the Better Business Bureau website, ESI has a 1.03/5-star rating, with 1,555 complaints in the last three years. Other PBMs fare no better. In other words, horrible quality is a feature, not a bug, of the biggest PBMs and the healthcare conglomerates who own them—and own you, should you be dependent on them. In the words of former FTC chair Lina Khan, these PBMs have become “too big to care.”

Cory Doctorow says it best: “Enshittification is a choice.” He was talking about Google search. But he may as well have been talking about too-big-to-care healthcare. The lobbyist arm of the big insurance companies recently took notice of their “enshittification” and vowed to simplify prior authorization. But insurance companies have vowed that before, in 2018 and again in 2023.

I’m a GI doctor. I know good shit from bull. And now Chris does as well. I asked him how he would clean it up. “I have to go through [ESI and Accredo],” he lamented, “because of the plan that I’m on with my company, right? But if I had the choice because of customer service, I’d never deal with them again . . . I think that people aren’t looked at anymore as patients; they’re looked at as a business. There’s no personal side to it.”

Chris wants choice in healthcare. The only way he’ll get choice—and not the choice PBMs have made over and over again—is to break these behemoths up. Only then will these companies have skin in the game again and have to compete for customer loyalties. And only then will we renew the “care” in healthcare.

Venu Julapalli is a practicing gastroenterologist and recent graduate of the University of Houston Law Center.

There is much discussion about what Hal Singer has dubbed “Gangster Antitrust,” the extraction of payments, bribes, or other concessions to allow passage of an otherwise anticompetitive merger. Gangster Antitrust can also take the form of conditioning the approval of a procompetitive merger on a seemingly unrelated remedy that advances the political interests of the administration. “Nice merger—be a shame if anything happened to it!”

David Dayen of the American Prospect correctly wrote that there is a law that is supposed to prevent such skullduggery, the Tunney Act, to assure that consent decrees by the Department of Justice (DOJ) are in the “public interest.” The Tunney Act of 1974 was drafted to prevent judicial “rubber stamping” of consent decrees. As explained here, the Tunney Act and its 2004 amendment, prohibiting continued judicial rubber stamping, have yielded just more vigorous rubber stamping.

I’ve written on the Tunney Act twice before, once with John J. Flynn, who happened to assist in its drafting, about the misuse of the Tunney Act in the Microsoft cases to compel the district court to accept a weakened settlement. I wrote a second time, after the D.C. Circuit continued to engage in activist and blatant disregard for the 2004 Tunney Act amendment.

Unfortunately, the Tunney Act’s purpose has been stifled by D.C. Circuit caselaw, and even the Act’s most fundamental purpose of destroying corruption has been neutered.

A brief history of the Act

The Tunney Act, named after Senator John V. Tunney, emerged from scandal surrounding backroom dealings to settle a DOJ merger challenge. It first became a major issue during hearings on Richard Kleindienst’s nomination to be attorney general. Senator Tunney expressed outrage at such closed-door discussions.

In 1969, the DOJ sued to prevent ITT’s acquisition of three companies under Section 7 of the Clayton Act. The DOJ lost two of the three suits. In 1971, the DOJ and ITT agreed to a settlement of the remaining suit. ITT was allowed to retain Hartford Fire Insurance Company but was required to divest several Hartford subsidiaries. The DOJ made no public statement as to the underlying reasons for the settlement. Instead, as was common practice at the time, only the proposed decree was made public.

Two significant events occurred that made people suspicious. First, President Nixon nominated Richard Kleindienst to be attorney general. Kleindienst had been involved in the ITT litigation in his capacity as deputy attorney general, and questions arose concerning his participation in the settlement of the case. Second, ITT offered to help finance the 1972 Republican National Convention. While no quid pro quo was proven, the appearance of impropriety sparked significant debate. (If you want to hear President Nixon ordering a DOJ official to back off the merger, you can listen here.)

Moreover, Kleindienst’s confirmation hearings revealed to the public for the first time the underlying rationale for the DOJ settlement with ITT: Kleindienst asserted that one reason for the settlement was DOJ fear that divestiture would cause ITT’s stock price to fall, causing hardship to shareholders. Another DOJ concern was apparently that the plummeting stock price would ripple throughout the U.S. economy.

All of this seems tame by today’s standards. But at the time, it was a massive scandal. The Supreme Court typically deferred to the DOJ traditionally with respect to consent decrees.

How the Act lost its teeth

Despite the Tunney Act’s prohibition against rubber-stamping, with rare exception, courts have continued to serve as rubber stamps, and the D.C. Circuit caselaw has played an important role in the rubber stamping. The basic standard laid out by the D.C. Circuit appears in the first Microsoft case, in which Judge Sporkin rejected the DOJ’s mealy-mouthed remedies (and eventually led to Microsoft II). The D.C. Circuit wrote: 

A decree, even entered as a pretrial settlement, is a judicial act, and therefore the district judge is not obliged to accept one that, on its face and even after government explanation, appears to make a mockery of judicial power. Short of that eventuality, the Tunney Act cannot be interpreted as an authorization for a district judge to assume the role of Attorney General.

Subsequent cases in the D.C. Circuit cling to this standard to assure that courts don’t bother with the “public interest” determination.

Congress reacted, and in 2004 changed the Tunney Act to compel a public interest determination. The legislative history expressly and in detail decried the D.C. Circuit’s caselaw (and cited my work with John Flynn, thank you very much).

The D.C. Circuit and its district courts flat out ignored the amendment, choosing to resurrect its “mockery of judicial function standard.” As the D.C. Circuit explained in a 2016 Speedy Trial Act case:

As we have since explained, we “construed the public interest inquiry” under the Tunney Act “narrowly” in “part because of the constitutional questions that would be raised if courts were to subject the government’s exercise of its prosecutorial discretion to non-deferential review.” Mass. Sch. of Law at Andover, Inc. v. United States, 118 F.3d 776, 783 (D.C.Cir.1997); see Swift v. United States, 318 F.3d 250, 253 (D.C.Cir.2003). The upshot is that the “public interest” language in the Tunney Act, like the “leave of court” authority in Rule 48(a), confers no new power in the courts to scrutinize and countermand the prosecution’s exercise of its traditional authority over charging and enforcement decisions.

The basis of the Court’s decision was a misguided notion that failing to enter a consent decree—inherently a judicial function—trampled the DOJ’s prosecutorial discretion under Separation of Powers. It did not consider that forcing a consent decree down the throat of the court also presented separation-of-powers problems. Nor did the Court explain why it is permissible for courts to reject criminal plea bargains without separation of powers problems, yet they must accept consent decrees for the rich and powerful. And while not all of the cases are in the DC Circuit, the vast majority are, and other circuits rely on DC Circuit caselaw and experience. Only one Tunney Act consent decree rejection. Ever.

So, here we are.

The question arises, then, about what exactly would it take to create a mockery of the judicial function?

We don’t know, quite frankly. There does not appear to be much, if anything, out there to suggest what mockery of the judicial function would look like sufficient to reject a consent decree under the Tunney Act.

Prior deals under Tunney Act review have been rubber stamped

Nothing raises eyebrows with the courts when it comes to the Tunney Act. Consider a couple of examples.

In 2008, the DOJ brought a broad and sweeping complaint against American Airlines’ acquisition of U.S. Airways. But politics played a role, according to Propublica: “People were upset. The displeasure in the room was palpable,” said one attorney who worked on the case. “The staff was building a really good case and was almost entirely left out of the settlement decision.” One of the reasons they might have been upset is that President Barack Obama’s former Chief of Staff was now Mayor of Chicago and advocating for the merger at the White House.

In another airline merger, an attorney representing the merging parties became DAAG after the merger won DOJ approval in August of the same year. Sometimes the revolving door in antitrust just spins just that fast.

Even if the courts did awaken to such questions, there is little interest in doing anything about it. One might claim that Judge Leon did a heroic Tunney Act review in CVS-Aetna, but I do not think that the D.C. Circuit precedent left him in  a good position to do anything other than accept the decree.

In other circuits, it is possible (but not likely) for a court to reject a consent decree. For a rare (and non-merger) exception, see U.S. v. SG Interests I, Ltd., 2012 WL 6196131 (unpublished opinion rejecting entry of consent decree in a Sherman Act Section 1 collusive bidding case as settling the case for nothing more than “nuisance value”). 

Parties have also been known to close deals even before the Tunney Act review has been completed.  Judge Leon complained of this practice in CVS-Aetna, but again, the D.C. Circuit caselaw leaves little in the way of judicial action. 

Thus, I imagine that courts will continue to do what they have always done—ostrich-like abdication of their powers.  

The Tunney Act won’t save civilization, democracy, or even antitrust

Is there a problem with Paramount making a major settlement with Trump and firing Colbert and then having its merger with Skydance approved? We’ll never know, because the courts will only review the complaint, the competitive impact statement, and the proposed final judgment. And rubber stamp.

The HPE-Juniper deal also raises serious questions related to the role of lobbying and whether the DOJ’s acquiescence has precious little to do with separation of powers and prosecutorial discretion and more to do with gangster antitrust. As the Wall Street Journal reported, “Hewlett Packard Enterprise made commitments, not disclosed in court papers, that called for the company to create new jobs at a facility in the U.S., according to people familiar with the matter.”  This, if true, ought to be sufficient to reject the consent decree. But I doubt it. While SCOTUS is hard-core killing Chevron and administrative law, it seems totally fine with the extreme level of deference the DOJ gets under the bastardized interpretation of the Tunney Act.

As David Dayen pointed out, there’s a friendly district court judge in the HPE-Juniper matter, who is a former labor lawyer. And both the HPE-Juniper and the UnitedHealth Group-Amedisys matters are outside the D.C. Circuit, which is a reason for hope. Yet other cases have had friendly judges and there are still no cases rejecting a consent decree. And the reason for that is the D.C. Circuit caselaw, regardless of circuit.

How about American Express? According to the Wall Street Journal,

American Express GBT hired Brian Ballard—a longtime Trump backer, who raised $50 million for his 2024 election—to lobby the Justice Department on antitrust issues for the company, according to lobbying disclosure forms. The Justice Department last week dropped a lawsuit it had filed seeking to block American Express GBT’s acquisition of a competitor, CWT Holdings.

This raises another point. The Tunney Act is only involved when we’re dealing with consent decrees with the DOJ. There is zero transparency with respect to merger investigations that have been dropped due to Gangster Antitrust. Decades ago, there was a push for greater transparency for when the DOJ was investigating a matter, and reasoning behind closing a matter without more. That went nowhere, and we are living with the consequences of that as well. And, as administrative law falls for independent agencies like the FTC, there isn’t much to suggest that courts will get in the way of settlements at DOJ’s sister agency, either.

The future looks grim. Sure, Congress reformed the Tunney Act once already. How’d that turn out? And now, it seems unlikely that Congress (in its current sycophantic posture to the Executive Branch) would dare attempt to correct the unbridled power of the Executive Branch to sell out on the cheap or engage in Gangster Antitrust.

The wealthiest man in the world was President Trump’s largest campaign contributor, thirteen billionaires were selected for positions in the administration, and the fourth wealthiest man in the world announced that the third largest newspaper in the country would no longer publish any opinion pieces critical of free markets. As if this weren’t an already alarming indication of the dangerous connection between economic might and socio-political power, next year, the Supreme Court will hear National Republican Senatorial Committee, et al. v. Federal Election Commission, et al., a case with the potential to erode some of the last remaining campaign finance limits. If the Court embraces the petitioners’ argument that broad categories of political spending should be freed from expenditure limits, it could intensify the transformation of American democracy catalyzed by Citizens United v. FEC and related decisions.

This looming decision arrives amid a broader, heated debate about the relationship between economic concentration and political power. Neo-Brandeisians, concerned with the corrosive effects that “bigness” of dominant firms and concentrated markets may have on democracy, increasingly find themselves at odds with “abundance” liberals and traditional antitrust centrists, who argue that dominant firm size, in and of itself, may not be as large a threat to democracy as alleged.

Yet this debate takes place atop some flawed empirical foundations. At the core is an assumption that lobbying expenditure data provides a reliable proxy for measuring concentration of political power. In a new working paper, I challenge that assumption, arguing instead that the apparent lack of correlation between rising economic concentration and lobbying market concentration obscures the rise of a more diffuse, opaque, and powerful influence ecosystem, enabled and financed by the wealth created through economic concentration. In short, what we see in lobbying data is not the absence of political capture, but its concealment.

Beyond Lobbying: A Complex Architecture of Political Influence

Quantitative analyses often treat political influence as a transactional marketplace where dollars spent on lobbying translate into policy influence. But this market analogy is conceptually flawed and empirically misleading. Lobbying, as captured by the Lobbying Disclosure Act, represents only a narrow slice of the influence economy. A growing share of influence is exercised through “dark money” groups, strategic litigation, media ownership, academic funding, “astroturf” campaigns, and campaign contributions by ultra-wealthy individuals.

These alternative channels of influence are not only substitutable with traditional lobbying; they are often more effective and less transparent. For instance, wealthy actors can achieve their policy goals by funding academic research that shifts public discourse, or by supporting litigation strategies that circumvent Congress altogether. These efforts rarely show up in lobbying disclosures, making them functionally invisible to traditional metrics.

The Empirical Illusion of Stable Lobbying Markets

Studies observing low or relatively stable concentration in lobbying expenditure patterns suggest that economic concentration does not lead to disproportionate political power. For example, a recent study by Nolan McCarty and Sepehr Shahshahani, comparing two decades of Lobbying Data Act expense reports to economic concentration and revenue, found that increasing corporate revenue does not lead to a disproportionate growth in the corporation’s lobbying spend and that an industry’s market concentration does not lead to a corresponding concentration in the industry’s lobbying “market,” suggesting there is little relationship between economic concentration and concentration of lobbying power. The reliability of this conclusion rests on the assumption that measurement error in lobbying data, though incomplete, is at least stable over time. If the magnitude of underreporting or misclassification is consistent year over year, then trends in lobbying concentration should still be informative about broader political economy dynamics, even if the data are imperfect.

This assumption has intuitive appeal. Longitudinal trends in a flawed dataset can, under some conditions, reveal meaningful shifts in behavior or structure. But this defense falters upon closer inspection, because it ignores the dynamic nature of political influence and the fungibility of influence-seeking behavior across channels. In practice, actors in political influence markets routinely engage in cross-channel substitution—dynamically shifting resources between lobbying, campaign finance, judicial advocacy, and public relations—in response to legal, political, and reputational shifts. These substitutions are not random; they are deliberate adaptations to maximize influence under changing constraints.

For example, following the 2007 Honest Leadership and Open Government Act (HLOGA), which imposed stricter disclosure requirements on lobbyists, many influence professionals rebranded themselves as “strategic consultants,” continuing their work without triggering Lobbying Disclosure Act reporting thresholds. Even more significantly, after Citizens United lifted restrictions on independent political spending, wealthy individuals moved significant resources into super PACs and dark money vehicles in ways that do not appear in lobbying disclosures. These shifts render the assumption of time-invariant measurement error implausible. As the legal and regulatory landscape changes, so too does the composition and visibility of political influence-seeking behavior.

Dynamic Substitution and the Post-Citizens United Shift

The 2010 Citizens United and related rulings drastically altered the strategic calculus of political influence. It enabled unlimited independent expenditures, allowing ultra-wealthy individuals to create expansive political infrastructures outside traditional corporate channels. These donors now routinely bypass firm rent-seeking allocations and operate instead through super PACs, nonprofit advocacy groups, and partisan media.

This shift from firm-based lobbying to capital-based influence is critical. A capital-holder with a diverse portfolio of shares across multiple industries may find it more efficient to invest in ideological advocacy and policy environments that favor their overall economic position rather than through individual firms’ budget allocations for political rent-seeking. This strategy is especially potent when coordinated across issue advocacy, electoral influence, and thought leadership.

Notably, while aggregate corporate lobbying plateaued post-2010, independent expenditures by individuals and non-corporate entities skyrocketed. Consider the graph below constructed from data obtained from opensecrets.org. Using federal election years when campaign related influence expenditures are likely highest, I show that the previous exponential growth in lobbying prior to its sudden plateau may be accounted for by the explosion of alternative influence channels utilized by new forms of influence purchasing organizations made possible by Citizens United and its progeny. This divergence suggests a strategic reallocation of political investments, not a reduction in influence-seeking. It is not that political capture had stalled; it evolved beyond ostensibly observable lobbying to opaque new influence channels and beyond the firm as the principal influence market actor.

Note: All data from OpenSecrets.org

Market Concentration, Shareholder Wealth, and the Feedback Loop of Influence

The link between economic concentration and political capture is most visible in the distribution of extraordinary shareholder wealth. Dominant firms—particularly in sectors with significant barriers to entry like tech, finance, and pharmaceuticals—generate supra-competitive profits. These profits flow disproportionately to a small group of shareholders, often billionaires with significant stakes in multiple dominant firms. For example, just five companies account for over 50% of the Nasdaq index. Collectively, their market cap is worth over $16 trillion, and their profit margins range from 10% to over 50%.

Moreover, the top 1% of households now control over 50% of U.S. corporate equities, the top 10% own nearly 90%, and dominant large-cap firms drive the returns from that equity. This extreme skew ensures that shareholder value maximization, often invoked as the corporation’s principle purpose, effectively channels wealth to a small elite. To wit, Elon Musk just secured a $29 billion payment from Tesla, despite the electric vehicle company’s sliding sales, in part due to Musk’s politics that are antithetical to electric vehicle consumers. This elite, in turn, finances the ideological and institutional infrastructure that resists regulatory or redistributive reforms, reinforcing both economic and political concentration.

For example, Elon Musk, Jeff Bezos, Mark Zuckerberg, and the Koch brothers derive the vast majority of their wealth from dominant firms in their respective sectors. Through a mixture of campaign contributions, media control, and think tank funding, these individuals have become central players in shaping the policy landscape, far beyond what traditional corporate lobbying would suggest. For example, the over $290 million spent by Elon Musk on influencing the 2024 election included direct campaign contributions as well as $240 million funneled into Musk’s own America PAC and $50 million toward political ads through Citizens for Sanity PAC. Musk has leveraged his acquisition of X (formerly Twitter) to support conservative and ultraright political forces worldwide, with the Associated Press finding that Musk’s engagement can result in political hopefuls gaining  millions of views and tens of thousands of new followers on his platform. When Jeff Bezos acquired The Washington Post in 2013, he pledged journalistic independence. However, in early 2025 he directed that the Opinion section “write every day in support and defense of” free markets. In conjunction with Meta, Mark Zuckerberg has donated hundreds of millions of dollars to colleges and universities like MIT and UC Berkeley leading to concerns about how such largesse may influence research agendas. The billionaire Koch brothers fund a network of  91 think tanks and organizations including the American Enterprise Institute and the Heritage Foundation that often promote “pro-business” viewpoints.

Reassessing Political Influence in a Multi-Channel Ecosystem

Researchers must move beyond single-channel metrics like lobbying data and adopt frameworks that capture the full architecture of political influence. This includes recognizing the strategic complementarities and dynamic substitution between campaign contributions, dark money issue advocacy, judicial influence, media ecosystems, and think tank networks. Influence is now wielded across a portfolio of channels, often in coordinated fashion and backed by extraordinary wealth.

Just as antitrust scholars recognize the importance of cumulative advantage and non-price effects in assessing market power, political economists must account for the structural and cumulative nature of influence. Political power is not merely purchased in discrete transactions; it is cultivated over time, embedded in relationships, and reinforced through systemic advantages in access, ideology, and information.

From Misdiagnosis to Reform

The forthcoming Supreme Court decision in NRSC v. FEC threatens to further erode transparency in an already opaque political economy. To understand and confront the risks this poses, we must discard the illusion that political power can be adequately assessed through lobbying data alone. Political capture in the post-Citizens United era is no longer primarily the domain of corporations, which were only ever a proxy for the profit interests of their owners—it is the domain of capital.

Abundance liberals risk sleepwalking into this moment. By focusing only on the consumer-welfare concerns of prices, output, and innovation, they miss the broader political implications of concentration. They treat economic and political power as separate, when in fact they are intertwined in a dangerous feedback loop.

Economic concentration produces wealth inequality; that wealth finances multi-channel influence; that influence protects the structures that maintain concentration. This isn’t a conspiracy; it’s a rational response to a regulatory environment that allows wealth to become political power. Addressing this risk does not mean we have to choose between bigness or democracy, a false dichotomy suggested by some of the debate between Neo-Brandeisians and supply-side abundance liberals.

Policy tools exist that can break the cycle of capture by the wealthy and minimize the democratic harms associated with concentration, while insuring that economies of scale actually deliver abundance. We need, among other things:

• Stronger antitrust enforcement to disperse economic power at its root.
• Wealth taxation that curtails the accumulation of influence capital.
• Campaign finance reform that restores public trust in electoral fairness.
• Media ownership rules that prevent consolidation of ideological gatekeeping.
• Enhanced transparency across all channels of political influence, not just lobbying.
• Academic disclosure standards that flag financial conflicts in policy research.
• Civil society watchdogs with the resources and access to investigate hidden influence.
• Legal reforms that broaden the definition of lobbying to include consulting and strategic advisory work.

The debate between Neo-Brandeisians, abundance liberals, and consumer welfare centrists is not merely theoretical. It reflects competing visions for the future of American democracy. If policymakers and scholars continue to underestimate the evolving architecture of political power and ignore the direct evidence of political capture by concentration-enabled capital, they risk providing rhetorical support for even further degradations to democratic responsiveness. The real threat of “bigness” isn’t just economic inefficiency. It’s the quiet capture of democracy itself.

Randy Kim is a municipal government attorney whose research and advocacy interests include economic justice, labor rights, the political dimensions of concentrated economic power, and their intersections. Opinions expressed herein are the author’s own and do not reflect the positions or opinions of his employers.

Railroad mergers haven’t happened in a while, and that’s a good thing. During the Reagan era, the country witnessed a rapid consolidation of its railroad industry. In the two decades following the 1980 Staggers Rail Act, the number of Class 1 freight railroads in the country fell from 39 to seven. The new millennium saw an industry dominated by four major railways that collectively controlled around 90 percent of the total domestic rail operating revenues.

In light of this rapid consolidation, regulators feared an eventual transcontinental rail duopoly. The Surface Transportation Board (STB) stymied railroad mergers in 2000, and issued new merger guidelines the following year, requiring that future deals would have to “enhance” competition. With the exception of Canadian Pacific-Kansas City Southern (CPKC), blessed with a waiver to be considered under the old merger guidelines, no Class 1 mergers have happened since the 2001 STB merger guidelines were released. That merger hiatus is set to be interrupted: In July, Union Pacific (UP) announced its intention to acquire Norfolk Southern (Norfolk) in an $85 billion deal.

The Economist touted the benefits of a UP-Norfolk tie-up, suggesting that the merger could lead to the “big four” railways becoming a “bigger two” railways. The magazine noted that UP-Norfolk merger would all but guarantee that BNSF and CSX combine as well. This potential rail duopoly presents a series of issues. Moving an already consolidated industry towards a duopoly means that shippers seeking to send traffic on overlapping routes will have fewer choices and will likely face higher prices. And railroad workers will have fewer employment options within an industry dominated by a duopoly. The STB will have to weigh the promised efficiencies against the harms to workers and shippers when considering this merger application.

Faster Speeds Are Not a Merger-Based Efficiency

Proponents of the UP-Norfolk merger argue that the arrangement will lead to faster trains, fewer delays, and better reliability. One purported benefit is that trains won’t need to interchange if they travel with the same railroad company along their entire route.

Here is The Economist’s merger-efficiency rationale, presumably helpfully shared by an industry lobbyist:

Avoiding interchanges between networks would mean faster trains and fewer delays. According to Oliver Wyman, a consultancy, the share of intermodal goods in America that travel by rail on journeys longer than 1,500 miles increases from 39% to 65% when served by a direct line. 

There’s one problem with that rationale: It doesn’t depend on the merger. Instead, such benefits could be achieved in many cases via contract without the associated merger harms. (And that stat from Oliver Wyman, is not impressive: The causation might run the other way, in the sense that demand for transport on a route might cause a railroad to acquire or invest in a direct line).

The STB’s merger guidelines require prospective merging parties to answer the question of “whether the particular merger benefits upon which they are relying could be achieved by means short of merger.” If the claimed benefits from the UP-Norfolk tie-up could be achieved absent a merger, then it follows that those benefits should be given no weight in the STB’s adjudication.

An examination of the rail industry today shows that railroads can already streamline their connections between networks. Nothing prevents two railroads from contracting to facilitate deliveries between separate lines. Indeed, Jim Vena, chief executive of Union Pacific, has acknowledged that UP has a track-sharing arrangement with BNSF in the Northwest. Track-sharing arrangements like the UP-BNSF agreement allow a railroad to move freight across the rail lines owned by another railroad.

After the UP and Southern Pacific (SP) merger in 1995, UP gave trackage rights to over 3,800 miles of track to BNSF. A press release announcing the agreement stated that “BNSF will be able to serve every shipper that is served jointly by UP and SP today.” The agreement “guarantees strong rail competition for the Gulf Coast petrochemical belt, U.S.-Mexico border points, the Intermountain West, California, and along the Pacific Coast.” If UP can survive and thrive with the BNSF trackage-rights agreement, what’s stopping them from creating one with Norfolk?

A 2001 STB report analyzing the UP-SP merger stated that “BNSF has competed vigorously for the traffic opened up to it by the BNSF Agreement,” and that it had “become an effective competitive replacement for the competition that would otherwise have been lost or reduced when UP and SP merged.” The report also cites enhanced competition for shippers who previously had only one rail carrier option. A potential UP-Norfolk track-sharing agreement could allow UP to service customers in Ohio and Kentucky without needing a merger.

Track-sharing agreements can also reduce the number of interchanges necessary to deliver rail cars to customers. A railroad company with a track-sharing agreement can allow another company’s trains to move along its tracks to service customers across two different rail lines without needing to swap engines. If UP and Norfolk made such a trackage agreement, a shipper in Minnesota could already ship their product to Pennsylvania without needing to swap train engines on the way, obviating the need for a merger.

The merging parties might argue that expanding these agreements will place a strain on their train dispatching systems. When two companies make a track-sharing agreement, the “host” railroad generally handles train dispatching on its own line (see the “operations” section of this Southern Pacific trackage rights agreement as an example). The host railroad has incentive to put the tenant railroad’s train at the end of the line for dispatching. (Anyone who has ridden Amtrak knows the experience of waiting for a freight train to pass ahead of you.) Hence, track-sharing agreements might lead to sub-optimal scheduling that would be avoided if all of the trains were owned by the host railroad. Rather than making investments in their train-dispatching systems to address these issues, railroads would rather merge.

It may be the case that some coast-to-coast routes cannot be streamlined through contracts, and that there are merger-specific efficiencies. In any event, the modest gains in efficiency will have to be carefully weighed against the anticompetitive effects felt by shippers and workers.

Past Mergers Have Caused Service Disruptions

Another weakness with the purported merger efficiency is that such benefits are undermined by the experiences of prior merger. Past mergers could not guarantee service improvements in the medium term. These operational failures directly contradict the recent promises that further consolidation of the railroad industry will improve speed and efficiency.

The aforementioned railroad merger, CPKC, was completed in 2023. During May and June of 2025, the combined CPKC firm experienced widespread service disruptions after merging the two legacy IT systems. The CPKC rail network experienced elevated delays, slower average velocity, and decreased on-time performance. In a report to the STB describing the situation, CPKC explained “Unfortunately, despite intensive efforts by CPKC over more than two years to prepare for a smooth transition, the Day N IT systems cut-over encountered unexpected difficulties.”

These issues aren’t isolated to the CPKC merger. When the Class 1 railroad Conrail was split between Norfolk Southern and CSX in 1999, both Norfolk and CSX experienced service disruptions. Shippers also experienced service disruptions in 1997 from the UP and Southern Pacific merger. For these major mergers, degraded performance in the medium term seems to be the rule rather than the exception. The STB’s 2001 merger guidelines recognize this history and place less weight on merger efficiencies that could take years to be realized.

Shippers Could Face Higher Prices on Overlapping Routes

Most of the business press coverage on the merger has ignored or handwaved the overlapping routes between UP and Norfolk. The New York Times has this to say on the subject:

Because Union Pacific and Norfolk Southern do not operate in each other’s regions, the tie-up would not reduce choice between railroads in those areas. Still, the companies together accounted for 43 percent of all rail freight movements last year, according to an analysis of regulatory carload data by Jason Miller, a professor of supply chain management at Michigan State University. (emphasis added)

Yet the map in same New York Times story (inserted below) shows that overlapping UP-Norfolk routes exist throughout Missouri and Illinois. Shippers who need to send products from, say, Kansas City or St. Louis through Chicago will have one less option.

Note: Norfolk Southern lines are shown in orange. Union Pacific lines are shown in maroon.

The eventual rail duopoly that would result from this merger would give many shippers either one or two carriers to choose from. Some shippers may become “captive shippers” who are beholden to a single railroad (aka monopoly) for their shipping needs. These captive shippers would likely face higher prices and worse service quality.

Recently, regulators have attempted to use reciprocal switching to solve the issue of captive shippers. When a shipper only has access to a single Class 1 railroad, a reciprocal switching agreement allows another outside railroad to compete for that shipper’s contract. If the outside railroad wins the contract, the incumbent railroad facilitates the freight pickup in exchange for a fee. These agreements enhance competition for shipping contracts while also compensating the incumbent railroad for using their tracks. It’s a win for both parties.

Yet the STB has only mandated these reciprocal switching agreements when a Class 1 railroad failed to meet performance standards regarding consistency and reliability. The competition standard for captive shippers should not be whether they reach a minimum level of service quality. When considering the possibility of creating a rail duopoly in this country, the STB should contemplate expanding their use of reciprocal switching agreements to enhance competition. But this reform is not a panacea for shippers facing a railroad monopoly from a merger. A recent Seventh Circuit decision overturned the STB’s 2024 reciprocal switching rule, finding it “inconsistent with the Board’s statutory authority.” This decision shows that ensuring competition cannot be achieved solely through rulemaking.

Workers Are Right to Be Skeptical of This Deal

Per The Economist, unions might “lie on the tracks” when it comes to the UP-Norfolk merger. The resulting rail duopoly from this merger would reduce the number of prospective rail employers, and prevent the bidding up of rail worker wages by rival railroads. The likelihood of layoffs and lower wages has already caused the Transport Workers Union to oppose the merger. The Sheet Metal, Air, Rail and Transportation Workers (SMART) union has also announced its opposition.

Despite the promises from UP’s and Norfolk’s management that they will preserve union jobs, the latest Class 1 merger should make rail workers skeptical. The recent CPKC post-merger service disruptions necessitated the temporary loaning of rail crews to address personnel shortages on their system. Yet one local SMART union alleges that the CPKC loaned crews reduced the number of yard jobs available for union employees. The local union alleges that CPKC took advantage of the service crisis to make these job changes. If that post-merger scenario is any guide, then the rail workers unions are justified in their opposition.

The Business Journalism Regarding the Deal Is Unbalanced

Despite the skepticism to the deal voiced by rail workers and shippers (the Freight Rail Customer Alliance representing over 3,500 businesses has criticized the proposed merger), business journalists seem enthusiastic about the deal. The New York Times and The Economist both trumpeted the deal, even citing the same industry analyst, Tony Hatch, in support of claimed efficiencies. Hatch’s support for the deal was widely shared throughout the business press, as he was also quoted by NPR, PBS, and the Bloomberg Podcast. Yet no message of skepticism of the merger efficiencies was widely shared in the media.

This is not to say that consulting a favorable industry expert is an issue. Yet instances like PBS gathering quotes from two analysts touting the benefits of the deal leaves readers with an unbalanced view. Our humble suggestion is that a business journalist, whether explaining a price hike or evaluating a merger, upon receiving a statement from an industry insider, should pick up the phone and seek an alternative opinion from a consumer or labor advocate (or, as a last resort, an economist not working for the merging parties).

The Surface Transportation Board Should Examine This Merger with a Skeptical Eye

The STB should ignore the glowing cheers from the business press and consider the harms to shippers and workers. This deal would likely create a transcontinental rail duopoly in this country, which would lead to even less choice for rail customers. The enhanced market power of the rail duopoly could lead to higher prices for consumers and less bargaining power for workers. Meanwhile, many of the promised service efficiencies could be gained without this merger. In light of the real costs and elusive benefits from this deal, the STB should examine the Union Pacific merger application with a skeptical eye and ensure that freight rail competition is maintained.

In a live discussion on Substack in June with Derek Thompson, neoliberal pundit Noah Smith called Dan Wang’s Breakneck a “companion volume” to Thompson’s and Klein’s recent bestseller Abundance. Wang is slated to speak at the upcoming Abundance 2025 conference, which is headlined by Klein and Thompson. Given the furious fight between the abundance faction and progressives for the soul of the Democratic party, I figured I should read it.

With Breakneck: China’s Quest to Engineer the Future (W.W. Norton 2025), the Hoover Institution’s Dan Wang makes a powerful authorial debut, deftly tracing the promise and pitfalls of China’s “engineering state” model of governance, which Wang presents as a foil to the American “lawyerly” society. In many ways, Breakneck is an ethnography of a government. And the portions of the book that lean into that frame are generally the best (largely chapters 3, 5, and 6). 

The book is fascinating and surprisingly fun—with Wang’s piercing dry wit interspersed at a near perfect frequency. It’s also frustrating. While the character of the engineering state is superbly developed, the points where its American “lawyerly” counterpart is brought into the mix are more tenuous. Wang leans heavily on a reiteration of Paul Sabin’s characterization of the American Left, from his book Public Citizens: The Attack on Big Government and the Remaking of American Liberalism (W.W. Norton 2021), with Wang’s explanation feeling paper thin next to the detailed portrait he paints of the engineering state. 

Some of this is entirely understandable; Breakneck is, first and foremost, a book about China, not about the United States. The comparatively flimsy explanation of the lawyerly state, however, is largely taken from a footnote in earlier chapters to a central point of framing in the book’s conclusion. 

While an overall terrific read, the book is somewhat uneven, great in most places but merely good in others. Perhaps the greatest criticism of the book is, in a way, a compliment to Wang as a writer: it should be at least 50 pages longer. 

About Abundance?

Wang makes no secret that he is sympathetic to the abundance movement, though he seems to indicate that he may not count himself as a part of it. He writes on page 50:

Under banners like “abundance agenda,” “supply side progressivism,” and “progress studies,” various movements are trying to loosen American supply constraints. These are excellent ideas that I hope are broadly adopted.

All of those movements are really better understood as constituents of the broader abundance movement. After all, people within the movement will identify them as such, plus they share a common network of funders, conferences, and prominent personalities.

Wang, though is not wedded to the abundance “lens,” as Ezra Klein and Derek Thompson term it, and instead is extremely clear in calling attention to some of the downsides such an approach entails when not balanced against other concerns.

In Wang’s parlance, the abundance movement seeks to make the United States less lawyerly and more engineering. And while Breakneck is supportive of that aim, it also cautions against overdoing it and promotes seeking to strike an appropriate balance between lawyers and engineers. 

This is one reason why even many critics of abundance may enjoy the book. One of the more frequent criticisms, both of Abundance and the movement sharing its name, is that it fails to engage in the hard tradeoffs that can accompany the quest for efficiency. Proponents will laud China’s infrastructure without considering the limitations on political freedoms or repression of ethnic minorities that enable its monumental building. Or, as Klein and Thompson do, they will lament the old times when Americans built things like the transcontinental railroad, with no acknowledgement of the genocide, racialized exploitation, or corruption that enabled it. 

Wang, largely, avoids this pitfall by confronting those questions head-on. He acknowledges, repeatedly, the environmental damage, persecution of ethnic minorities, and disastrous social engineering projects that came with high-speed rail, monumental bridges, and dams. The question presented by this acknowledgment, which could have been answered more substantially, though, is how possible it is to have an engineering society that steers around the worst of these harms. 

First Kill All the Lawyers?

The single biggest flaw in Breakneck is its chronicling of the role and influence of lawyers in the United States. The topline argument here is simple enough: we used to be more of a nation of engineers before becoming extremely lawyerly in the back half of the twentieth century. 

In particular, Wang points to Paul Sabin’s Public Citizens and blames this development substantially on the New Left, spearheaded by figures like Ralph Nader who continually sued the government to enact progressive ends. Yet Breakneck barely addresses the necessity of the lawyerly society to preserve and defend the public from powerful interests or corrupt political regimes. Red tape has its downsides, but it also can be an impediment to the worst impulses of those in power. Particularly in the age of Trumpian power-tripping and influence peddling, with the Supreme Court playing dead (at times quite convincingly), many people would probably like it if lawyers were able to slow things down more at this particular point.

At one point Wang proxies the level of lawyerliness on the number of lawyers per capita the United States has (p.14). Even by that metric, Israel, the Dominican Republic, Italy and Brazil are more lawyerly. Do they all struggle to build more than the United States?

There is also an issue of how one counts the lawyers. In the European continental civil law system, many nations separate out “lawyers” from “magistrates.” In the United States, by contrast, everyone who practices law is a lawyer, whether they are a defense attorney, a prosecutor, or a judge. In some places like France, for instance, prosecutors and judges are excluded. So when Ezra Klein and Derek Thompson make a similar point to Wang and cite the United States as having four times the lawyers per capita as France (Abundance p.92-93), it is, at best, misleading. 

Additionally, there are other metrics that would indicate the United States has become less lawyerly. If you look at presidents before and after the turn of the 20th century, the latter group has a much lower share of lawmen and includes the only two engineers (Hoover and Carter). Of the first eight presidents, six were lawyers. Of the eight most recent, only three went to law school. (Wang’s exploration specifically cites how many recent presidents were lawyers on page 4.) 

The greatest builder of all American presidents, Franklin Delano Roosevelt, was also a lawyer. The transcontinental railroad, one of our nation’s most defining megaprojects, was begun during the administration of Abraham Lincoln, famously a lawyer.

Wang quotes Alexis de Tocqueville calling lawyers the “American aristocracy” all the way back in 1833 (p.13). Breakneck’s conclusion would have been strengthened by an exploration of when America transitioned from an engineering society to a lawyerly one and how to determine the threshold where legalism becomes overburdening.

While there is some truth to the argument that the United States is a procedure-obsessed nation, Wang’s under exploration of this point results in the book concluding on what is probably the weakest link of his analysis.

Embracing Engineers

The limits of Wang’s description of the lawyerly state are stark because of how they stand in contrast to the detailed and nuanced articulation of the engineering state. To wit, Breakneck’s third chapter is a superb crash course on the cast of engineers who made China into the nation it is. Wang strikes an excellent balance here, keeping the conversation approachable to readers without much knowledge of modern Chinese history while still getting into the weeds enough to provide valuable commentary and insights. 

In the third chapter, Wang outlines a way of thinking about technology that sharply diverges from its American counterpart. He argues for what is in some ways a very worker-centered model of technology, noting that “Viewing technology as people and process knowledge isn’t only more accurate; it also empowers our sense of agency to control the technologies we are producing” (p.75). 

Reasoning from this model, Wang goes on to criticize the “elite consensus” that viewed American manufacturing as dispensable, and the dismissal of unions and heterodox economists who warned against deindustrialization (p.76, 78). Wang also points to financialization and corporate consolidation as major drivers in America’s transformation into a country that doesn’t build anymore (p.77). This is another key point where Wang aligns with many critics of abundance.

It is also at this point, however, where Wang begins a pattern of praising Tesla, calling it “America’s great hope in auto manufacturing” (p.77). While he does rebuke Musk’s shredding of the federal government towards the end, there are several other instances where Tesla is cited as an exemplar of American industry. That Elon Musk is gambling the company on a robo-taxi gambit and has directed Tesla away from its successful sedan models and towards the disastrous cybertruck is largely undiscussed. Also undiscussed are longstanding arguments that Tesla is actually a case in point of financialization driving manufacturing rationales and the company’s reliance on selling regulatory credits as a critical source of income.

The chapter also introduces the critique of America’s emphasis on software and intangible technology at the expense of what Xi Jinping calls the “real economy,” or making physical things. “Too many people,” Wang contends, “have argued away the strategic importance of manufacturing” (p.91). 

After establishing the importance of being able to build and transform the real world, the next two chapters explore the downsides of the engineering state that have accompanied China’s building prowess. The fourth chapter discusses how the cadre of engineers in the Deng Xiaoping-era tried to engage in social engineering, and the chaos that such policies sowed. Interspersed with stories of forced sterilization, infanticide, and coercive abortion, Wang’s telling of the one-child policy is chilling. As he explains, “the one-child policy could only have been implemented in the engineering state. While the state possessed a bureaucracy to enforce controls of such extraordinary scale, there wasn’t a sufficiently developed civil society to fight for legal protection against it” (p.118, emphasis in original).

The hardline positions of affixing social policy to clear numerical benchmarks is demonstrated again in the next chapter’s telling of Wang’s firsthand account of the zero covid policy.  This account is part of the dire warning embedded in Breakneck about what happens when a country is not lawyerly enough. 

Talking Tradeoffs

Wang argues that China and the United States would both benefit if they were a little more like one another. China should invite some of the lawyerliness in order to protect and serve its citizens, while we should do away with some procedural obsession to be able to build more. How exactly, and how much, to do that are left mostly open.

This is probably the point on which I most wish Wang had elaborated. I would like to know more how he envisions the United States becoming more engineering-focused without sacrificing the procedural protections provided by lawyerliness. 

One point completely absent from the book is how important the rule of law is at preventing authoritarianism. Particularly in light of January 6th, 2021 and the more recent attempted coup in Brazil, along with an attempted coup in South Korea and redistricting arms races at home, it seems that legal bulwarks’ ability to defend themselves is extremely topical. And yet Wang does not offer any thoughts on whether lawyerliness is important to stave off democratic backsliding.

With just eight pages left in the book, Wang declares that “For various American ideals to be fully realized, the country will need to recover its ethos of building, which I believe will solve most of its economic problems and many of its political problems too.” This is reflective of the last chapter, which tries to parlay the complex comparative tapestry Wang so intricately wove into a series of straightforward, broad conclusions. Those conclusions aren’t wrong, but are dependent on the context from which they are taken. Thus, the more sweeping general conclusions feel somewhat hollow coming after six chapters of richly developed contextualization. 

The Bottom Line

Breakneck will be of great value to anyone interested in modern China, geopolitics, infrastructure, or industrial policy. Wang’s perspective is invaluable, grounding debates on how to have a society that builds with his own lived experiences, interweaving historical analysis and personal reflection, and fleshing out the Chinese “engineering state” model, which is often invoked but rarely interrogated in contemporary American discussions of industrial policy. 

There is no shortage of debatable points the book makes, but there is absolutely no denying that it is an excellent read. Breakneck is handily one of the best-written popular nonfiction books in recent years.