In a forthcoming book this spring, the philosopher Elizabeth Anderson has advanced a critique of “private government,” the concept used to express the power wielded by large corporations in the labor market. For many workers, the force that most actively restricts their liberty is not government regulations but their employer’s rules; even gig economy laborers such as Uber drivers, who unlike traditional employees have the flexibility to set their own hours, risk being put out of a job for committing relatively innocuous sins such as canceling too many rides, driving with a car other than the one registered to you, or complaining about Uber on social media. There is little doubt that modern technology allows employers enhanced surveillance power, but hypothetically, the voluntary nature of employment should limit the abuse of this power; mistreated workers can always exercise their right to “exit” the employment relationship and get another job.

This works great if you’re a superstar athlete or hotshot coder; employers need you as much as you need them, and you can always take your talents to South Beach and/or Facebook (they’re hiring!). But contemporary conditions make simply up and quitting unfeasible for a significant slice of the population — namely, those who are already the most vulnerable. If you need a job as a condition of probation, or if your visa is tied to your employer, or if you’re an immigrant working without documentation, “exit” is impossible without imprisonment or exile. Even the average low-income worker might not have the savings to ride out a short gap in employment without getting evicted or going hungry. The frictions in the employment marketplace force many of us to put up with terrible treatment by giving bosses the leverage that knows that we can’t afford to just walk out. (Anderson’s preferred solution is to beef up employee participation within a company, like in the German “codetermination” model, which would make the shitty treatment that necessitates exit less common.)

I think that the idea of private government is also useful for understanding the power that the largest tech companies have over businesses that use their platforms. When Republicans say that government should be “run like a business,” they’re asserting that businesses are more competitive and efficient, and better at giving the people what they want. Because they’re all convinced that this is a priori true (and because they don’t know how the internet works), they haven’t noticed that all sorts of companies, mostly in the tech world, have become private government unto themselves. As David Graeber makes abundantly clear in his book Bullshit Jobs, the ���bureaucracy” we’ve been trained to hate and to associate with the state is at least as common in the corporate world. Going to the DMV kinda sucks, but is dealing with Time Warner’s customer service any better?

An increasing share of online commerce takes place within the walled gardens of Amazon, Facebook, Apple and Google. To manage their platforms, these companies have had to set up processes for creating rules (legislatures), systems for resolving disputes (courts), and leaders that set the agenda for the company (Tim Apple, Mark Facebook, Jeff Amazon, Sergei and Larry Google) — in short, their own private governments.

For the sake of argument, let’s concede the question of how an ideal government should function to the economists and chamber of commerce types. How do these private governments stack up? Central criticisms of government regulation are that it creates an uncertain business environment that reduces risk-taking and investment (consider automakers’ dislike of changing emissions regulations), or that it increases the incentive for businesses to profit by taking advantage of artificial conditions rather than providing better value in the marketplace. The US sugar industry, for example, can only exist by seeking protection from sugar produced in places that actually have the right climate.

The right and the left tend to agree that this is a problem, although their proposed solutions are quite different. One thing that hasn’t been proposed is to create private, capricious marketplaces governed only by corporations, but that’s where we’ve ended up.

The most prominent example is Amazon Marketplace, the e-commerce giant’s network of third-party sellers — that is, of small businesses. In 2018, Amazon accounted for over 49 percent of all online retail, over two thirds of that Amazon Marketplace. The potential for the abuse of monopoly power due to Amazon favoring its in-house brands has been well-discussed, but the inefficiencies caused by its private government control of the third-party marketplace are at least as serious.

Again per economics orthodoxy, good government should be predictable, neutral, and transparent, with due process at both the level of lawmaking and dispute resolution. Amazon’s performance on these dimensions is dismal, creating confusion and incentivizing fraud. An article in The Verge details the dirty competition on Amazon's Marketplace, in which framing a competing vendor for violating Amazon’s frequently-changing rules is a surefire way to get ahead. Buying obviously fraudulent product reviews for a rival will cause Amazon to suspend a vendor’s account, which “can send a seller’s business into bankruptcy, with few avenues for appeal.” The most dramatic examples from The Verge are fraudsters who buy their rival’s products, set them on fire, and post a picture with one-star reviews saying they spontaneously combusted. This seems especially common in high-margin electronics of dubious quality; in hindsight, maybe hoverboards weren’t actually that dangerous?

And just like the Heritage Foundation thinks that state intervention encourages wasteful rent seeking and lobbying, Amazon’s governmental power makes bribery and corruption more profitable. A lengthy Wall Street Journal investigation last September revealed that Amazon employees will literally take bribes to remove negative reviews and share internal sales figures. Although company brass want to crack down on this, the money offered by vendors is a serious enticement. A sign of a healthy marketplace would be that there is greater incentive for vendors to to re-invest that money in their firms, winning market share by outcompeting their rivals, than to cheat the system and try to out-bribe each other.

Whatever the flaws of Amazon’s private government, in an economically ideal world, they would be tempered by the possibility of setting up online shop on another platform. Just as in Anderson’s framework, however, there is no guarantee that such a thing is possible. The issue with Amazon is that it’s simply too big. In real-world terms, it’s fine to take your wares out of Wal-Mart and sell them only in independent retailers, but the problem is that at least half of your potential customers only go to Wal-Mart. The fact that Amazon has its own suite of house brands further increases their power; even if all of the battery vendors somehow organized a boycott, Amazon has a line of perfectly acceptable replacements.

The economic logic behind the tendency towards natural monopoly is outlined by the Nobel Prize-winning economist Jean Tirole in Economics for the Common Good. Sellers want to be where buyers are, and vice versa, so everyone ends up in the same place. Furthermore, different sectors of these integrated tech companies cross-subsidize each other. Amazon’s massive cloud storage business allows it to both manage user data and pay for legions of top data scientists to leverage that data to power its recommendation engine. This multi-billion-dollar infrastructural head start means that new firms are unlikely to compete with Amazon head-on — the lack of “contestability” creates precisely the kind of monopoly power that Tirole warns against. To use another real-world example, imagine a farmers’ market where anyone can sell whatever they want, but the person who runs the farmer's market has their own stand and already knows what everyone wants because they also own the local private spy-organization-for-hire. Seems bad, right?

Skeptics might say that big companies come and go; Facebook overtook Myspace and Google defeated AltaVista, so just give it a few years and exiting to a new marketplace will be easy. This argument is deeply, hilariously misguided. Facebook and Google’s investments in proprietary hardware and software, their massive databases, and server farms make them hundreds of times more entrenched and powerful than Myspace.

In The Internet Trap, communications scholar Matthew Hindman explains how the digital nature of infrastructural investments makes them more difficult to appreciate, but no less real. Microsoft’s effort is illustrative: the company “absorbed an astonishing $12.4 billion in cumulative losses to establish Bing as a credible competitor [to Google search]. It is cheaper and easier to build a manned space program than it is to build a modern search engine.”

The analogy to government is again useful. People have been into roads for a long time, and road-building is now (to libertarians’ eternal consternation) seen as a basic governmental function. Roads are a public space that everyone can use, subject to (in theory) equal treatment. But the internet was developed by private firms, who have since built the all of the online “roads.” There are clear advantages to this — we certainly wouldn’t have TikTok if Al Gore were still in charge of the internet — but we need to acknowledge the danger. The rollback of net neutrality under Trump’s FCC removes one check on tech corporate power; competition, the other potential check, doesn’t seem to really work.

Online marketplaces do not operate by standard economic rules. In Hindman’s words, “The internet breaks our mental models.” I have focused on Amazon Marketplace because of its recent explosion in popularity, but also because the consumer-facing nature of retail makes the unpredictability and power of their private government more transparent. The competition to rank high in a Google search, i.e., search engine optimization, is a much older example. Once people realized there was serious money to be made through SEO gamesmanship, an entire industry of Mountain View Kremlinologists competing to reverse-engineer the Google search algorithm sprang up. If the entire SEO industry were to disappear the net effect on the consumer would be virtually zero.

Facebook’s impact on the market for online news is another troubling example of the market-distorting power of private government. In 2015, it touted the effectiveness of video advertising on their platform, causing many media firms to “pivot to video” and then lay off hundreds of journalists when what turned out to be a gamble didn’t pay off. Since then, there has been good evidence that these viewership numbers were seriously inflated; in any case, Facebook’s power to shape the online news market is undisputed. (Facebook’s ability to regulate speech on its platform is another function of its private government status; the challenges of designing and enforcing this regulation have been extremely well-discussed so I’m not going to go into them here.)

Businesses thrive under stable, predictable regulation and rule of law. According to conservatives, governments pose the biggest threat on this dimension, but that is just clearly no longer the case, at least for online commerce. Private governments run by unaccountable bureaucrats inside tech companies are a serious and growing problem for millions of small businesses across the country.

The solution is far from obvious. At the intellectual level, Tirole warns that traditional antitrust models are ill-suited to this context, so we don’t even have a theory to try and implement. Elizabeth Warren recently announced a plan to break up Big Tech, correctly reading the winds of public opinion. But getting the details right is going to extremely difficult, as each platform creates unfairness in its own unique way. But we’ve reached the point that no one, whatever their attitude towards business and government, can deny the magnitude and the novelty of the power wielded by the tech companies as private governments.