The 40-year economic mistake that let Google conquer (and enshittify) the world
I'm on a tour with my new book, the international bestseller Enshittification: catch me next in Miami, Burbank, Lisbon! Full schedule here.
A central fact of enshittification is that the growth of quality-destroying, pocket-picking monopolists wasn't an accident, nor was it inevitable. Rather, named individuals, in living memory, advocated for and created pro-enshittificatory policies, ushering in the enshittocene.
The greatest enshittifiers of all are the neoliberal economists who advocated for the idea that monopolies are good, because (in their perfect economic models), the only way for a company to secure a monopoly is to be so amazing that we all voluntarily start buying its products and services, and the instant a monopoly starts to abuse its market power, new companies will enter the market and poach us all from the bloated incumbent.
This "consumer welfare" theory of antitrust is obviously wrong, and it's the best-known neoliberal monopoly delusion. But it's not the only one! Another pro-monopoly ideology we can thank the Chicago School economists for is "industrial organization" (IO), a theory that insists that vertical monopolies are actually really good. This turns out to be one of the most consequentially catastrophic mistakes in modern economic history.
What's a "vertical monopoly"? That's when a company takes over parts of the supply chain both upstream and downstream from it. Take Essilor Luxottica, the eyeglasses monopoly that owns every brand of frames you've ever heard of, from Coach and Oakley to Versace and Bausch and Lomb. That's a horizontal lobby – the company took over every eyewear brand under the sun. But they also created a vertical monopoly by buying most of the major eyeglass retailers (Sunglass Hut, Lenscrafters, etc), and by buying up most of the optical labs in the world (Essilor makes the majority of corrective lenses, worldwide). They also own Eyemed, the world's largest eyeglasses insurer.
IO theory predicts that even if a company like Essilor Luxxotica uses its monopoly power to price gouge in one part of the eyeglass supply chain (e.g. by raising the price of frames, which Essilor Luxxotica has done, by over 1,000%), that they will use some of those extraordinary profits to keep all their other products as cheap as possible. If Luxottica can use its market power to mark up the price of frames by a factor of ten, then IO theory predicts that they'll keep the prices of lenses and insurance as low as possible, in order to make it harder for lens or insurance companies to get into the frame business. By using monopoly frame profits to starve those rivals of profits, Essilor Luxxotica can keep them so poor that they can't afford to branch out and compete with Essilor Luxottica's high-priced frames.
Like so much in neoliberal economics, this is nothing but "a superior moral justification for selfishness" (h/t John Kenneth Galbraith). IO is a way for the greediest among to convince policymakers that their greed is good, and produces a benefit for all of us. By energetically peddling this economic nonsense, monopolists and their pet economists have done extraordinary harm to the world, while getting very, very rich.
Google is a real poster-child for what happens to a market when regulators adopt IO ideas. "Google’s hidden empire," is a new paper out today from Aline Blankertz, Brianna Rock and Nicholas Shaxson, which tells the story of how IO let Google become the enshittified, thrice-convicted monopolist it is today:
https://arxiv.org/abs/2511.02931
The authors mostly look at the history of how EU regulators dealt with Google's long string of mergers. By the time Google embarked on this shopping spree, the European Commission had already remade itself as a Chicago School, IO-embracing regulator. The authors trace this to 2001, when the EC blocked a merger between GE and Honeywell, which had been approved in the USA. This provoked howls of disapproval and mockery from Chicago School proponents, who mocked the EC for not hiring enough "IO expertise," contrasting the Commission's staff with the US FTC, which had 50 PhD (neoliberal) economists on the payroll. Stung, the EU embarged on a "Big Bang" hiring spree for Chicago School economists in 2004, remaking the way it viewed competition policy for decades to come.
This is the context for Google's wave of highly consequential vertical mergers, the most important of which being its acquisition of Doubleclick, the ad-tech company that allowed Google to acquire the monopoly it was last year convincted of operating:
When Google sought regulatory approval in the EU for its Doubleclick acquisition, the EC's economists blithely predicted that this wouldn't lead to any harmful consequences. Sure, it would let Google dominate the tools used by publishers to place ads on their pages; and by the advertisers who placed those ads; and the marketplace in which the seller and buyer tools transacted business. But that's a vertical monopoly, and any (IO-trained) fule no that this is a perfectly innocuous arrangement that can't possibly lead to harmful monopoly conduct.
The EC arrived at this extraordinary conclusion by paying outside economists a lot of money for advice (that kind of pretzel logic doesn't come cheap). Two decades later, Google/Doubleclick was abusing its monopoly so badly that the EU fined the company €2.95 billion.
It's not like Google/Doubleclick took two decades to start screwing over advertisers and publishers. Right from the jump, it was clear that this merger was an anticompetitive disaster, but that didn't stop the EC from waving through more mergers, like 2020's Google acquisition of Fitbit:
Once again, the EC concluded that this merger, being "vertical," couldn't have any deleterious effects. In reality, Google-Fitbit was a classic "killer acquisition," in which Google bought out and killed the dominant player in a sector it was planning to enter, in order to shut down a competitor. Within a few years, the Fitbit had been enshittified beyond all recognition.
Despite these regulatory failures (and many more like them), the EC remains firmly committed to IO and its supremely chill posture on vertical monopolization. But as bad as IO is for regulating vertical mergers, it's even less well suited for addressing Google's main tactic for shaping markets: vertical investments.
Google Ventures (GV) is Google's investment arm, and it is vastly larger than the venture arms of other Big Tech companies. Google invests in far more companies than it buys outright, and also far more companies than any other Big Tech company does. GV is the only tech company investment fund that shows up in the top-ten list of VCs by deal.
In the paper, the authors use data from Pitchbook to create a sense of Google's remarkable investment portfolio. Many of these deals go through "Google for Startups," which allows Google to acquire an equity stake in companies for "in-kind contributions," mainly access to Google's cloud servers and data.
By investing so widely, Google can exert enormous force on the shape of the entire tech ecosystem, ensuring that the companies that do succeed don't compete with Google's most lucrative lines of business, but rather funnel users and businesses into using Google's services.
This activity isn't tracked by academics, regulators, or stock analysts. It's the "hidden empire" of the paper's title. 9556 companies that show up in Pitchbook as receiving Big Tech investments up to 2024. 5,899 of those companies got their investments from Google.
Combine Google's free hand to engage in vertical acquisitions and its invisible empire of portfolio companies, and you have a world-spanning entity with damned few checks on its power.
What's more, as the authors write, Google is becoming an arm of US foreign power. Back in 2024, Google made a $24b acquisition offer to the cybersecurity company Wiz, which turned it down, out of fear that the Biden administration's antitrust enforcers would tank the deal. After Donald Trump's election – which saw antitrust enforcement neutralized except as a tool for blackmailing companies Trump doesn't like – Wiz sold to Google for $32b.
The Wiz acquisition is an incredibly dangerous one from a competitive perspective. Wiz provides realtime cybersecurity monitoring for the networks of large corporations, meaning that any Wiz customer necessarily shares a gigantic amount of sensitive data with the company – and now, with Google, which owns Wiz, and competes with many of its customers.
Google has already mastered the art of weaponizing the data that it collects from users, but with Wiz, it gains unprecedented access to sensitive data from the world's businesses.
Google's consolidation of market power – power it has abused so badly that it has lost three federal antitrust cases – can be directly traced to the foolish notions of Industrial Organization theory and its misplaced faith in vertical mergers.
As the authors write, it's long past time we abandoned this failed ideology. The Google/Wiz merger still has to clear regulatory approval in the EU. This represents a chance for the EC to abandon its tragic, decades-long, unrequited love affair with IO and block this nakedly anticompetitive merger.
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
Before "disruption" turned into a punchline, it was a genuinely exciting idea. Using technology, we could connect people to one another and allow them to collaborate, share, and cooperate to make great things happen.
It's easy (and valid) to dismiss the "disruption" of Uber, which "disrupted" taxis and transit by losing $31b worth of Saudi royal money in a bid to collapse the world's rival transportation system, while quietly promising its investors that it would someday have pricing power as a monopoly, and would attain profit through price-gouging and wage-theft.
Uber's disruption story was wreathed in bullshit: lies about the "independence" of its drivers, about the imminence of self-driving taxis, about the impact that replacing buses and subways with millions of circling, empty cars would have on traffic congestion. There were and are plenty of problems with traditional taxis and transit, but Uber magnified these problems, under cover of "disrupting" them away.
But there are other feats of high-tech disruption that were and are genuinely transformative – Wikipedia, GNU/Linux, RSS, and more. These disruptive technologies altered the balance of power between powerful institutions and the businesses, communities and individuals they dominated, in ways that have proven both beneficial and durable.
When we speak of commercial disruption today, we usually mean a tech company disrupting a non-tech company. Tinder disrupts singles bars. Netflix disrupts Blockbuster. Airbnb disrupts Marriott.
But the history of "disruption" features far more examples of tech companies disrupting other tech companies: DEC disrupts IBM. Netscape disrupts Microsoft. Google disrupts Yahoo. Nokia disrupts Kodak, sure – but then Apple disrupts Nokia. It's only natural that the businesses most vulnerable to digital disruption are other digital businesses.
And yet…disruption is nowhere to be seen when it comes to the tech sector itself. Five giant companies have been running the show for more than a decade. A couple of these companies (Apple, Microsoft) are Gen-Xers, having been born in the 70s, then there's a couple of Millennials (Amazon, Google), and that one Gen-Z kid (Facebook). Big Tech shows no sign of being disrupted, despite the continuous enshittification of their core products and services. How can this be? Has Big Tech disrupted disruption itself?
That's the contention of "Coopting Disruption," a new paper from two law profs: Mark Lemley (Stanford) and Matthew Wansley (Yeshiva U):
The paper opens with a review of the literature on disruption. Big companies have some major advantages: they've got people and infrastructure they can leverage to bring new products to market more cheaply than startups. They've got existing relationships with suppliers, distributors and customers. People trust them.
Diversified, monopolistic companies are also able to capture "involuntary spillovers": when Google spends money on AI for image recognition, it can improve Google Photos, YouTube, Android, Search, Maps and many other products. A startup with just one product can't capitalize on these spillovers in the same way, so it doesn't have the same incentives to spend big on R&D.
Finally, big companies have access to cheap money. They get better credit terms from lenders, they can float bonds, they can tap the public markets, or just spend their own profits on R&D. They can also afford to take a long view, because they're not tied to VCs whose funds turn over every 5-10 years. Big companies get cheap money, play a long game, pay less to innovate and get more out of innovation.
But those advantages are swamped by the disadvantages of incumbency, all the various curses of bigness. Take Arrow's "replacement effect": new companies that compete with incumbents drive down the incumbents' prices and tempt their customers away. But an incumbent that buys a disruptive new company can just shut it down, and whittle down its ideas to "sustaining innovation" (small improvements to existing products), killing "disruptive innovation" (major changes that make the existing products obsolete).
Arrow's Replacement Effect also comes into play before a new product even exists. An incumbent that allows a rival to do R&D that would eventually disrupt its product is at risk; but if the incumbent buys this pre-product, R&D-heavy startup, it can turn the research to sustaining innovation and defund any disruptive innovation.
Arrow asks us to look at the innovation question from the point of view of the company as a whole. Clayton Christensen's "Innovator's Dilemma" looks at the motivations of individual decision-makers in large, successful companies. These individuals don't want to disrupt their own business, because that will render some part of their own company obsolete (perhaps their own division!). They also don't want to radically change their customers' businesses, because those customers would also face negative effects from disruption.
A startup, by contrast, has no existing successful divisions and no giant customers to safeguard. They have nothing to lose and everything to gain from disruption. Where a large company has no way for individual employees to initiate major changes in corporate strategy, a startup has fewer hops between employees and management. What's more, a startup that rewards an employee's good idea with a stock-grant ties that employee's future finances to the outcome of that idea – while a giant corporation's stock bonuses are only incidentally tied to the ideas of any individual worker.
Big companies are where good ideas go to die. If a big company passes on its employees' cool, disruptive ideas, that's the end of the story for that idea. But even if 100 VCs pass on a startup's cool idea and only one VC funds it, the startup still gets to pursue that idea. In startup land, a good idea gets lots of chances – in a big company, it only gets one.
Given how innately disruptable tech companies are, given how hard it is for big companies to innovate, and given how little innovation we've gotten from Big Tech, how is it that the tech giants haven't been disrupted?
The authors propose a four-step program for the would-be Tech Baron hoping to defend their turf from disruption.
First, gather information about startups that might develop disruptive technologies and steer them away from competing with you, by investing in them or partnering with them.
Second, cut off any would-be competitor's supply of resources they need to develop a disruptive product that challenges your own.
Third, convince the government to pass regulations that big, established companies can comply with but that are business-killing challenges for small competitors.
Finally, buy up any company that resists your steering, succeeds despite your resource war, and escapes the compliance moats of regulation that favors incumbents.
Then: kill those companies.
The authors proceed to show that all four tactics are in play today. Big Tech companies operate their own VC funds, which means they get a look at every promising company in the field, even if they don't want to invest in them. Big Tech companies are also awash in money and their "rival" VCs know it, and so financial VCs and Big Tech collude to fund potential disruptors and then sell them to Big Tech companies as "aqui-hires" that see the disruption neutralized.
On resources, the authors focus on data, and how companies like Facebook have explicit policies of only permitting companies they don't see as potential disruptors to access Facebook data. They reproduce internal Facebook strategy memos that divide potential platform users into "existing competitors, possible future competitors, [or] developers that we have alignment with on business models." These categories allow Facebook to decide which companies are capable of developing disruptive products and which ones aren't. For example, Amazon – which doesn't compete with Facebook – is allowed to access FB data to target shoppers. But Messageme, a startup, was cut off from Facebook as soon as management perceived them as a future rival. Ironically – but unsurprisingly – Facebook spins these policies as pro-privacy, not anti-competitive.
These data policies cast a long shadow. They don't just block existing companies from accessing the data they need to pursue disruptive offerings – they also "send a message" to would-be founders and investors, letting them know that if they try to disrupt a tech giant, they will have their market oxygen cut off before they can draw breath. The only way to build a product that challenges Facebook is as Facebook's partner, under Facebook's direction, with Facebook's veto.
Next, regulation. Starting in 2019, Facebook started publishing full-page newspaper ads calling for regulation. Someone ghost-wrote a Washington Post op-ed under Zuckerberg's byline, arguing the case for more tech regulation. Google, Apple, OpenAI other tech giants have all (selectively) lobbied in favor of many regulations. These rules covered a lot of ground, but they all share a characteristic: complying with them requires huge amounts of money – money that giant tech companies can spare, but potential disruptors lack.
Finally, there's predatory acquisitions. Mark Zuckerberg, working without the benefit of a ghost writer (or in-house counsel to review his statements for actionable intent) has repeatedly confessed to buying companies like Instagram to ensure that they never grow to be competitors. As he told one colleague, "I remember your internal post about how Instagram was our threat and not Google+. You were basically right. The thing about startups though is you can often acquire them.”
All the tech giants are acquisition factories. Every successful Google product, almost without exception, is a product they bought from someone else. By contrast, Google's own internal products typically crash and burn, from G+ to Reader to Google Videos. Apple, meanwhile, buys 90 companies per year – Tim Apple brings home a new company for his shareholders more often than you bring home a bag of groceries for your family. All the Big Tech companies' AI offerings are acquisitions, and Apple has bought more AI companies than any of them.
Big Tech claims to be innovating, but it's really just operationalizing. Any company that threatens to disrupt a tech giant is bought, its products stripped of any really innovative features, and the residue is added to existing products as a "sustaining innovation" – a dot-release feature that has all the innovative disruption of rounding the corners on a new mobile phone.
The authors present three case-studies of tech companies using this four-point strategy to forestall disruption in AI, VR and self-driving cars. I'm not excited about any of these three categories, but it's clear that the tech giants are worried about them, and the authors make a devastating case for these disruptions being disrupted by Big Tech.
What do to about it? If we like (some) disruption, and if Big Tech is enshittifying at speed without facing dethroning-by-disruption, how do we get the dynamism and innovation that gave us the best of tech?
The authors make four suggestions.
First, revive the authorities under existing antitrust law to ban executives from Big Tech companies from serving on the boards of startups. More broadly, kill interlocking boards altogether. Remember, these powers already exist in the lawbooks, so accomplishing this goal means a change in enforcement priorities, not a new act of Congress or rulemaking. What's more, interlocking boards between competing companies are illegal per se, meaning there's no expensive, difficult fact-finding needed to demonstrate that two companies are breaking the law by sharing directors.
Next: create a nondiscrimination policy that requires the largest tech companies that share data with some unaffiliated companies to offer data on the same terms to other companies, except when they are direct competitors. They argue that this rule will keep tech giants from choking off disruptive technologies that make them obsolete (rather than competing with them).
On the subject of regulation and compliance moats, they have less concrete advice. They counsel lawmakers to greet tech giants' demands to be regulated with suspicion, to proceed with caution when they do regulate, and to shape regulation so that it doesn't limit market entry, by keeping in mind the disproportionate burdens regulations put on established giants and small new companies. This is all good advice, but it's more a set of principles than any kind of specific practice, test or procedure.
Finally, they call for increased scrutiny of mergers, including mergers between very large companies and small startups. They argue that existing law (Sec 2 of the Sherman Act and Sec 7 of the Clayton Act) both empower enforcers to block these acquisitions. They admit that the case-law on this is poor, but that just means that enforcers need to start making new case-law.
I like all of these suggestions! We're certainly enjoying a more activist set of regulators, who are more interested in Big Tech, than we've seen in generations.
But they are grossly under-resourced even without giving them additional duties. As Matt Stoller points out, "the DOJ's Antitrust Division has fewer people enforcing anti-monopoly laws in a $24 trillion economy than the Smithsonian Museum has security guards."
What's more, Republicans are trying to slash their budgets even further. The American conservative movement has finally located a police force they're eager to defund: the corporate police who defend us all from predatory monopolies.
