May 25, 2021
Tech monopolies and the insufficient necessity of interoperability
A preview of my July 2021 Locus Magazine column
I care about monopolies for exactly one reason: self-determination. I don’t care about competition as an end unto itself, or fetishize “choice” for its own sake. What I care about is your ability to live your life in the way you think will suit you, to the greatest extent possible, and taking into account the obvious limits when other people’s needs and wants conflict with you realizing your own desires.
We live in a world of vast and increasing monopolization, with one, two or a few companies controlling everything from the arts (publishing, movies, music, streaming, comics, bookselling, movie theaters, talent agencies, games, wrestling) to finance (banks, investment funds, auditors, bond-rating agencies) to agribusiness (seeds, livestock, tractors, fertilizer, pesticides, precision agriculture) and everything in between (radio stations, cruise lines, cheerleader uniforms, pharmaceuticals, glass bottles, airlines, eyeglasses, athletic shoes, fast food, food delivery, pet food).
When just a few people have the ultimate say over what you can read, or where you can work, or how our food is grown, or even what you feed your cat, you’d better hope that they value the same things as you!
Of course, the executives who run these corporations don’t assume you want the same things as they do. When Mark Zuckerberg set out to buy Instagram, he unwisely sent a series of emails to Facebook’s CFO setting out the case for the acquisition, namely, that younger users no longer liked Facebook, so much so that they were quitting his company’s service and signing up for Instagram. Zuckerberg wanted to buy Instagram so that users who had the gall to live their lives according to their own priorities (and not his shareholders’ priorities) would be thwarted. He wanted to make sure that wherever Facebook’s defectors found a new online home, he’d be able to control them, spy on them, and make money from them.
Even corporations whose choices coincide with your own today might cease to suit your needs tomorrow. By definition, you know more about your needs than any product designer. The data-security measure that works to prevent third parties from accessing your backups might block your own file-recovery if you forget to transfer a critical file when switching to a new device; or a company’s decision to discontinue device support might strand you with no way to get your photos or notes off an old tablet. The paternalistic partnership that entrusts a company to protect its customers’ needs by overriding their choices works great, but fails badly.
The reason to want more choice isn’t grounded in consumerism. Self-determination isn’t about the superficial desire for a different shade of blue, or moving your desktop menu icon from the top right corner to the bottom left — it’s about you (and not a corporate exec) having the final say over how you live your life.
Now, all those monopolized industries got to their sorry state by the same route: they bought their small future competitors, merged with their large rivals, and used their market power to crush competitors. Monopolization is the story of investors directing their fortunes to companies who use that capital to buy other companies. It’s the story of success that’s purchased by eliminating alternatives, not by making the best thing that people want the most.
The modern epidemic of tolerance for monopolies began 40 years ago, in the Reagan years, and every administration since has waved through blatantly anticompetitive mergers and turned a blind eye to grossly anticompetitive acts. The story of how this came to pass is tawdry and oft-told: it’s the tale of how switching competition law’s enforcement to focus on “consumer welfare” (low prices) destroyed labor markets, national resiliency, and the credibility of democratic institutions. It’s the story of how control over industries dwindled to a handful of powerful people who captured their regulators and got themselves deputized as arms of the government.
That’s the story of how we got here. Now let’s talk about where we’re going.
If I had my way, we’d return to the muscular, aggressive form of antitrust enforcement of the pre-Reagan era. We’d shift the burden of proof so that mergers would be blocked unless the merging parties could demonstrate that the result wouldn’t harm competition. Hell, we could even go back to pre-emptively investigating the leading companies in any industry where the same companies lead the pack for more than a couple of consecutive years, on the presumption that such a situation is usually reliant on anticompetitive behavior as its underlying stabilizer.
We’re a long, long way off from such an arrangement. Public sentiment is more alive to the problems of monopoly than at any time in decades, but still most voters don’t see monopolism as a great evil. They may worry that all their beer comes from two companies or that the internet has turned into five giant websites filled with screenshots of text from the other four. They might resent the deep rake that the tax-prep giants take to prepare their returns and fume at the arbitrary and lethal conduct of their health insurer. But most people haven’t (yet) put all this together and named the common thread running through all of it: monopolism.
For once, the political class is getting out ahead of the voters. Canny politicians understand that fighting corporate abuse is an easy way to look good in front of the base, and many seem to have tired of carrying water for increasingly emboldened lobbyists who demand more and more special favors that are harder and harder to sell to the folks back home.
What’s more, politicians who take up the anti-monopoly cause can serve one set of corporate interests while fighting another: for example, the telecoms industry is happy to fund vigorous trustbusting of their arch-enemies in Big Tech. Monopolists are fine with trustbusting for other monopolists, assuming (perhaps unwisely) that antitrust can be awakened to shatter the power of a hated rival monopolist within their supply chain, and then it can be returned to its long slumber.
Without a huge public antimonopoly movement, trustbusting politicians are fighting an uphill battle. Today’s antitrust investigations and legislative proposals are very bold by recent standards, but they fall far short of the hard-line, radical trustbusting we need. That is a situation that calls for tactical thinking, finding ways to score wins and leverage them for bigger wins.
What tactics might we use to score those wins? I can tell you one thing: we’re not going to get support from the business world. It might seem like Google and Facebook are bitter rivals and one might be turned on the other, but the difference between the two companies’ “character” is paper-thin. That’s why top former Google execs like Sheryl Sandberg find themselves perfectly at home in Facebook’s upper echelons, and why searching LinkedIn for current Google employees shows that many of them once worked at Facebook and vice versa.
Same goes for Google and Apple’s famous rivalry: from the way that partisans for the two companies talk about them, you’d think that Apple’s obsession with consumer privacy and locked-down devices would represent a serious obstacle to any kind of collaboration with the surveillance- and open systems-based Google. But both companies found it easy enough to collude — along with Intel, Intuit, eBay and others — to enter into criminal “no poach” agreements that suppressed the wages of tens of thousands of tech workers.
Think of the “bitter rivalries” between US phone companies, with T-Mobile’s loudmouth badboy John Legere declaring himself the “un-CEO” and his company the “un-carrier,” a totally different venture from all of his competitors. Those differences were apparently a little, uh, oversold. Legere began these declarations of difference in 2012, when he was named CEO following T-Mobile’s blocked merger attempt with AT&T. Legere — who had held numerous senior executive roles at AT&T — stopped wearing suits, donned a t-shirt, grew out his hair, and declared that T-Mobile was nothing like any phone company in the sector…right up to the moment that he merged his company with Sprint (where he’d also held numerous senior roles), pocketed $137,000,000 in bonuses, and walked away.
Corporate personhood is obviously a sham. In his dissent in Citizens United, Supreme Court Justice Stevens wrote that corporations have no claim to free speech rights because “corporations have no consciences, no beliefs, no feelings, no thoughts, no desires.” Companies may project a set of “corporate values,” but these values are a marketing strategy, not a set of deeply held convictions. That’s especially evident during mergers, when companies that have styled themselves as “bitter rivals” are able to consolidate into a single company, like Disney-Fox, whose “brand identities” were at significant odds with one another. From their longstanding rivalry, you’d think they were the Montagues and Capulets, but if that’s so, then that means Rupert Murdoch and Bob Iger were Romeo and Juliet, star-crossed lovers whose desire for one another was so deep and sincere that they brought peace between their warring great houses.
Or, you know, maybe it was all bullshit. Maybe large companies all have the same ideology (“profit”). Maybe the distinctions between their character are as meaningful as the “flavors” of the different marshmallows in a box of Lucky Charms. Maybe the reason John Legere worked at AT&T and Sprint before going to T-Mobile is that they are interchangeable monopolies whose top ranks all came up together, know each other, take vacations together, and are godparents to one-another’s children. Maybe they aren’t really rivals.
Maybe monopolists have class solidarity, is what I’m saying.
If we can’t count on big companies to turn on one another, we’ll need popular support to advance the cause of trustbusting. To get that popular support, we’ll need to proceed stepwise, scoring small wins that we can build and build upon.
It will not surprise you to learn that I think we should start with tech.
Tech is in very bad odor right now. Few voters’ first priority is Doing Something About Tech, but likewise, few voters trust or like the tech industry. There’s already antitrust action pending against several major tech companies in multiple US states, at the federal level, and abroad. Tech was the first new industry to grow up without meaningful antitrust enforcement (the Apple ][+ hit the market the same year that Reagan hit the campaign trail) and there are people alive today who watched the industry go from this vibrant, dynamic, fast-moving thing to an ossified and chummy boys’ club run by a half-dozen CEOs who fit around a single board-room table at the top of Trump Tower in 2017.
But it’s not just the fact that tech is in trustbusters’ crosshairs that makes it the ideal first step in a path to total antitrust victory. The most important thing about tech, from a competition perspective, is the underlying, technical nature of networked computers.
For while it’s true that every industry took the same path to monopolization (runaway mergers followed by anticompetitive conduct), industries all have different paths out of monopoly, because each industry’s technical characteristics provide us with a different set of handles we can yank on to change their course. Think of this as antitrust’s answer to Anna Karenina: “All happy families are alike; each unhappy family is unhappy in its own way.”
Here’s the amazing thing that makes tech the poster child for a new anti-monopoly movement: interoperability. Interoperability is when one thing works with another: your shoelaces interoperate with your shoes, your AAA batteries interoperate with your TV remote, your coffee-maker interoperates with your electrical outlet.
Interoperability is an inescapable fact of life, but when it comes to digital systems, interop gets profound. That’s because all digital computers are “Turing-complete Von Neumann machines,” which means that all computers are capable of running all programs. The laptop I’m typing this on can run the programs I wrote on my first, Reagan-era Apple ][+, and the programs I’m running on this laptop will also run on that old Apple ][+, wherever it is (albeit so slowly that we might experience the heat-death of the universe before it finished loading the program).
This foundational flexibility bequeaths an intrinsic degree of latent interoperability to digital systems that physical ones can’t hope to match. It’s been 150 years since a lack of coordination among new Australian states created the “middle-gauge muddle”: a nation whose railroad tracks are laid at different widths depending on which part of the country you’re in. In 150 years, no one has figured out how to make a rail car that can change its wheelbase midway through its journey and after hundreds of attempts, Australia is giving up on interoperable rolling stock. Instead, they’re tearing up thousands of kilometers’ worth of rail and putting down new ones.
That’s not how it works in digital. Despite all the handwringing over the inaccessibility of old digital data, the reality is that new computers can emulate old computers and run the programs that were used to create and read that data in the deep past of computing (getting the data off of old storage media that is physically deteriorating is another story). If Australia’s middle-gauge muddle were a matter of digital incompatibilities, some programmers could whip up a “translation layer” that mediated between different tracks and cars and unify the system. If we can connect billions of devices running millions of versions of scores of operating systems to each other via the internet, getting six Australian states’ railcars to connect to each others’ (digital) tracks is a piece of piss.
To understand what interop has to do with digital monopolies, we need to understand the role that “network effects” play in the growth of these monopolistic services. A system has “network effects” if it gets more valuable as it adds more users. No one would use Facebook if they were the only user the system had, but once everyone you need to talk to is on Facebook, you can’t afford not to use it. Network effects can lead to a “winner take all” situation, where once a system attains “critical mass,” it grows on its own, signing up new users who become bait for still more new users.
Those network effects certainly contributed to the growth of several kinds of tech giant, and not just social media services. iPhones grow more valuable whenever software vendors create new apps for them. The more apps there are, the more reasons there are to buy one, and the more people who own iPhones, the more potential customers there are for new apps.
“Network effects” come up a lot when economists talk about competition in digital markets. But an even more important concept gets a lot less attention: “switching costs.”
Switching costs are whatever you have to give up to go from one situation to another. The switching costs of moving include movers, boxes, realtor fees, a moving van, the time it takes to enroll your kid in a new school, the cost of finding a new job, the losses you incur from the penalties for late bills you missed when your mail forwarding goes awry…the whole package.
When it comes to digital monopolies, switching costs are more important than network effects.
Here’s why. Today, people struggle to leave Facebook because doing so involves leaving behind their friends. Those same friends stuck on Facebook for the same reason. People join Facebook because of network effects (which creates a monopoly), but they stay on Facebook because of the high switching costs they face if they leave (which preserves the monopoly).
That’s where interoperability comes in. There’s no technical reason that leaving Facebook means leaving behind your friends. After all, you can switch from T-Mobile to Verizon and still stay in touch with your T-Mobile subscriber friends. You don’t even need to tell them you’ve switched! Your phone number comes with you from one company to the other and, apart from the billing arrangements, nothing changes.
Why can’t you switch from Facebook to a rival and still stay in touch with your friend on Facebook? It’s not because of the technical limitations of networked computers. It’s because Facebook won’t let you.
Interoperability lowers switching costs. Switching costs reverse network effects. If you can talk to people on Facebook without using Facebook, then Facebook becomes a reason to sign up for any service, not just Mark Zuckerberg’s walled garden. Likewise for app stores (if you can take your apps with you, that’s one less reason not to switch from Android to iPhone or vice-versa). Likewise for any technology that relies on network effects.
Companies have a funny relationship with interop. When companies are small and trying to build up their customer-base, they love interop, love the idea of selling ink for someone else’s printer or a way to read your waiting messages on someone else’s social media giant. Facebook once had a whole suite of interoperability tools to make it easy to plug Facebook into other services, but it has whittled these away over the years and today it routinely threatens and even sues rivals that try to interoperate with it.
That’s where antitrust comes in. The tech companies have no appetite to litigate antitrust suits to the bitter end. These suits take a long time (IBM’s antitrust case — AKA “antitrust’s Vietnam” — ran for 12 years), cost a lot of money (IBM’s annual legal bills for 12 years running were higher than the salaries of the DoJ’s entire antitrust division combined), and they’re incredibly painful (Bill Gates’s videotaped antitrust deposition went analog-viral, and it’s hours of agonizing, barely controlled rage and peevish obstructionism).
That’s why antitrust cases usually end in settlements, not judgments. Normally, the focus of these settlements is fines (and when the fine amounts to less than the profit for the illegal conduct, it’s not a fine, it’s just a price). But settlements can come with conditions, “conduct remedies,” spelling out what the companies must do as a condition of continuing to operate.
We could make interop one of these conditions. We could say, “Fine, Zuck, we’re not going to force you to sell off Instagram (yet), but we will require you to expose a digital interface that your competitors plug into so their users can exchange messages with your users.”
A lot of careful planning would have to go into such a settlement — to preserve privacy, fight spam, and guard against harassment — but that’s the basic shape of it.
We don’t even need to wait for a settlement: we could just pass a law requiring some kind of interop. Senator Mark Warner’s ACCESS Act of 2020 does just that, and more modest versions are contemplated in the EU’s pending Digital Market Act and the recommendations from the UK’s Competition and Markets Authority’s report in 2020.
There are other handles we can yank on, like procurement. The US federal government buys a lot of tech, and it could enact a rule saying that it will only buy tech from companies that allow for interoperability that lowers switching costs.
But mandatory interoperability won’t lower switching costs on their own, not in the long run. These mandates are easy for tech giants to sabotage, and regulators take a long time to prove that sabotage has gone on, and even longer to do something about it. If all we get are interop mandates, the tech companies will run circles around their regulators.
That’s how it’s worked out in the real world. In 2012, Massachusetts voters passed a ballot initiative that forced auto-makers to supply independent mechanics with the data they’d need to interpret the diagnostic information streaming around modern cars’ internal wired networks. Car-makers had spent years systematically monopolizing independent car service, and Bay Staters had had enough of it.
Even before the Massachusetts law came into effect, car makers started to redesign their cars so that all the data needed for independent repair traveled over cars’ wireless networks, which were exempted from the law.
It took eight more years before another ballot initiative was put before Massachusetts voters, in late 2020, updating the law to close this loophole. In the meantime, independent repair was widely suppressed, and many independent mechanics were forced to sell their businesses and go to work as employees for the car-makers’ authorized service centers.
The mismatch between the time it takes to subvert a mandate and the time it takes to amend it is why we can’t rely on mandates alone.
For mandates to work, they need to have a counterweight, a consequence that befalls companies that subvert them that hurts worse than obeying the mandate in the first place. We can get that counterweight through another kind of interoperability: adversarial interoperability, or “competitive compatibility” (comcom), the ad hoc, reverse-engineer-style, permission-free version of interop that has always existed alongside the more formal, standardized forms.
Some familiar examples of comcom: third party ink for your printer; remanufactured spare parts for your phone and programs like Apple’s Pages that read and write Microsoft Word files; VCRs that could record broadcast TV and TiVos that could do it digitally.
Every one of the tech giants relied on comcom in their rise to the top, and all of them have supported new laws and new interpretations of existing laws that transformed comcom into a legal minefield. That’s not surprising — the companies want to make sure no one does unto them as they have done unto others, and they’ve got the money and resources to make sure if anyone tries it, it’ll be a crime.
Comcom makes a great counterweight to interop mandates. If Facebook settled an antitrust case by promising to make it possible for its rivals to exchange messages with its users, and then turned around and broke or downgraded the interface they used for that, comcom could come to the rescue. The services that Facebook had just frozen out could switch to using comcom techniques, like running bots that scraped their users’ Facebook feeds and pushed their replies back to the service.
Facebook could fight these bots with technical countermeasures — updating their firewall rules, say — but they hate that. That’s why they’ve sued, rather than out-engineered, everyone who’s tried to use comcom to interoperate with the service so far.
After all, technological warfare presents a messy, unpredicatable risk to the business. Every rule that blocks a bot ends up blocking real users, too. Even when the blocks work, Facebook users who are accustomed to staying in touch with off-service friends will complain to Facebook about flaky delivery of their messages. It’s exactly the kind of ugly, unquantifiable X-factor companies dread.
Comcom makes the mandates look good by comparison, in other words — and if a company is unwise enough to sabotage their mandated interop, comcom offers the users who’ve come to depend on interoperability with a fallback they can keep using until a regulator can intervene.
Imagine if the carmakers had to worry about comcom when they were subverting their repair-code mandate in Massachusetts. When they switched their service messages from their cars’ wired networks to the wireless ones, a couple smart MIT kids could have gone into business selling a $20 gadget that cost $1 to make that could grab that data from cars’ wireless services.
Anything the car manufacturers did to freeze out these gadgets would mean retooling all their authorized service centers and dealing with the inevitable upgrade problems. Meanwhile, independent mechanics would have a new business to supply them with diagnostic tools — the MIT kids’ startup — and that business could offer other services to them, services that made the auto-makers even less central to automotive repair.
Comcom is the stiffener that turns structurally unsound mandates into sturdy, pro-competitive solutions. But getting comcom back is a tough fight. It would mean reforming all the laws that the tech giants have twisted into an impenetrable thicket around their products and business-models: copyright, patent, non-compete, binding arbitration, terms of service, cybersecurity, trade secrecy, and more.
The sheer number of legal reforms necessary to restore comcom is daunting, a legal-political project of years, if not decades. But there’s another way to get there: antitrust settlements.
Recall how settlements could be a lever to force interop mandates on tech monopolists. Unlike laws, which affect all companies and can be derailed or defanged on their long journeys through the legislative process, settlements are bespoke, drafted by government lawyers at a moment when companies are trying to avert traumatic, multi-year legal battles that expose their internal workings to public scrutiny and subject every new initiative to soul-searching legal analysis to ensure they aren’t going to get the company into even hotter water.
If we’re going to use settlements to force companies to allow interop, we can also use settlements to take away companies’ power to subvert interop. We can do that by appointing a “special master” — a legal babysitter whom the company has to get sign-off from when they want to bring a legal action against a rival. If the special master is satisfied that the lawsuit is based on a real legal grievance (say, a rival company that is trying to steal user data, rather than a competitor whose users have given it permission to get their data) then the suit goes forward. Otherwise, the monopolist loses its legal right to block interoperators.
An interop mandate means that new services can connect to a monopolist’s platform and give its users more self-determination in how they live their digital lives. A special master in charge of the monopolist’s legal remedies means that if the monopolist sabotages the mandate, the new service can use comcom to maintain the connection by other means.
But actually having to use comcom is a worst-case scenario. Big companies hate technological asymmetrical warfare, hate being forced to patch and re-patch their systems to block adversarial interoperators. And an interop mandate gives companies a way to avoid having to fight that unpredictable and messy guerrilla action: all they have to do is maintain the mandated interface in good order, and all those competing services will use it. Why wouldn’t they? They’d rather use engineering resources to improve the service they provide to their users, not trying to outsmart Facebook’s security team.
Ultimately, the answer to market concentration isn’t interoperability: it’s de-concentrating markets. But breaking up monopolists is hard. They have tons of money, they are structurally important to the lives of billions of people and many businesses.
Though we’re closer to having our long overdue monopoly reckoning than we’ve been in decades, there’s a long way to go. To sustain and increase the momentum of recent years, we need to score wins, and turn those wins into bigger wins, and then bigger ones still.
Tech is the logical place to start, not just because everyone is fed up with tech, but because tech is so central to everything else we do — it provides the communications and coordination layer that are at the heart of every mass movement. And tech’s flexibility — that protean, foundational ability to plug everything into everything else — means that tech trustbusters have a uniquely suitable tool for prying apart monopolies: interoperability.
Forcing interop back into tech won’t be the end of the anti-monopoly fight, but it’ll be the end of the beginning — the necessary but insufficient step we’ll take before moving on to far more ambitious projects.
Cory Doctorow (craphound.com) is a science fiction author, activist, and blogger. He has a podcast, a newsletter, a Twitter feed, a Mastodon feed, and a Tumblr feed. He was born in Canada, became a British citizen and now lives in Burbank, California. His latest nonfiction book is How to Destroy Surveillance Capitalism. His latest novel for adults is Attack Surface. His latest short story collection is Radicalized. His latest picture book is Poesy the Monster Slayer. His latest YA novel is Pirate Cinema. His latest graphic novel is In Real Life. His forthcoming books include The Shakedown (with Rebecca Giblin), a book about artistic labor market and excessive buyer power; Red Team Blues, a noir thriller about cryptocurrency, corruption and money-laundering; and The Lost Cause, a utopian post-GND novel about truth and reconciliation with white nationalist militias.
