It’s the liability, stupid: Airbnb, platform economics, and regulatory bedrock
(An early version of this essay appears on my personal blog.)
To kick off the Fair Housing Berlin blog (editor’s note — this later became the Domi Labs blog!), I wanted to write about two legal developments of crucial interest to our project, and to the world of tech and business generally. First, a maverick Republican senator threatened to pull the plug on the legal statute that made social media a liability-free zone in the otherwise hyperbolically litigious United States. Second, AirBNB asked the EU to write an analogous statute that would protect them from two growing backlashes (popular and industrial) in Europe. If those two things seem unrelated to you, I have some sunny, undeveloped land in Florida to show you, in which you might want to consider investing.
History lesson: Free-as-in-beer speech, Free-as-in-Freedom-Fries speech
The way people talk about tech is often very similar to how they talk about “magic”: technological and/or informational disruption is mystified into the form of a “black box,” into which a few hundred hours of coding and speculative investment are put, and out of which giant (or even global) markets or business processes emerge. People are sorted into pro-tech and anti-tech camps to preserve the myth of creation ex nihil, of unknowable magick at the heart of tech, of it all being mAtHeMaTiCs too complex for laymans’ brains. I think the best corrective for this dangerous obfuscatory rhetoric is simply to approach tech from the point of view of industrial history — which markets, institutions, and public goods were changed by (or replaced by, or cannibalized by) the arrival of every “new” market? Who wins and loses, how are the rules changed for older markets when new markets debut?
Much of the “magic” disappears when you pull back the curtain in this way, but I am not trying to kill all magic. In fact, if anything, I think the true magic at the heart of the internet is that kernel of anarchic humanity most resistant to monetization and ownership. I’m referring here to the “user-generated” internet: craigslist, crowdsourcing, user reviews, podcasters solving cold cases and other “open source intelligence” (#osint for short), reddit AMAs, comments sections, all that beautiful chaos. The more chaotic and unmonetized the better, as far as I’m concerned.
Most of it is highly monetized, however, and the massive industry-wide shift towards doing so was called “Web 2.0” when I first started working in software. Back then, older corporations and investors were salivating over the prospect of extracting value from the torrent of content users donated to their privately-owned databanks. One illustrative example from this era is the “yelp playbook”: crowdsourcing restaurant or product reviews, then charging restaurants and product-vendors for the right to distort it, unbeknownst to the vast majority of end-users happily writing, reading, and trusting reviews, all “for free”. Google maps, facebook, amazon reviews, all of these pillars of our digital economy still work on this model of magically aggregating and number-crunching user-generated content that is, most of the time, freely given (donated). Curiously, the developments of the decade since and even the current media “techlash” don’t seem to be slowing down people’s enthusiasm to trade their data and their content for “free” web services or slightly-discounted pieces of hardware locked in to them. If anything, corporations are just getting savvier in how they incentivize influencer accounts and user forums and/or reviews as “sock puppets” to carry out corporate agendas through “organic” (i.e., deregulated) marketing.
Of course, all that freedom has an uglier side that no one likes to dwell on: spam, trolling, and other bad behaviors are a small (but perpetually-renewing, seemingly permanent) cost borne by the vulnerable for existing on the open-door internet. User-generated content is generally about as immature, hateful, selfish, cruel, irresponsible, and evil as the average human is, or a little more, probably, because anonymity and an easy escape route brings out the worst in most people, at least at this stage in world history. On the commercial web, the labor of separating the noise from the signal can be farmed out to massive call-center-like moderation farms, and on the quirky, global websites like Wikipedia, craigslist, and reddit, a robust community of unpaid careworkers do similarly draining work to keep the internet looking at least a little better than the species that made it. Either way, in all the talk of free speech and free content, the industry as a whole has historically waved its hands and changed the subject when asked about the real, permanent costs of all this moderation, refereeing, and data cleaning.
230 Today: Cracks in the dam
But lately it’s gotten very hard to wave away things like election influence and total surveillance, and the questions that critics, regulators, and even activist shareholders ask are getting harder and louder. On the one hand, traditional media is drying up on the vine as a steadily increasing proportion of the total advertising dollars goes to adtech middlemen (google, facebook, salesforce, and their partners and subcontractors) instead of to content-producers. On the other hand, variously tech-literate populists from Elizabeth Warren to Josh Hawley have been gunning for the economic foundations of all this free speech and free content: article 230 of the 1996 Communications Decency Act, the law that made Web 2.0 possible on an industrial scale.
Without going into too much detail (I’ll save that for a future article elsewhere), 230 was like a giant dam holding back most of the risks, liability, and regulatory issues inherit to publishing, and to the public sphere more generally. It came to pass as a kind of compromise during a high-profile moral panic over one of those issues, childrens’ expsore to obscenity. (Considering the recent scandals over Youtube’s inability to shield it’s rapidly-growing child userbase from dangerous content and contact, little has changed at the heart of the issue.) Unlike traditional publishing, where printing or disseminating dangerous medical misinformation could open up any publisher in any media to crippling criminal and/or civil liability, 230 established a blanket immunity by making internet companies and services “platforms” rather than “publishers” — they were not “publishing”, just “distributing” their content, particularly if that content were user-generated.
Not only are they less liable for content published online, they also have freer rein to govern by their own standards, moderating according to whatever business logic and/or community standards they see fit. In a nutshell, 230 freed the internet from inheriting most of the case law on “publishing” and the public sphere — online publications were… special. By setting a threshold for liability way lower than, say, the classified ads or the letters to the editor section of a newspaper, internet companies could extract whatever value they wanted from an ocean of free content, which they had not paid for and the consequences of which they could not be made to pay for in court.
This might seem like a clever lawyer’s sleight of hand (how is distributing content online so categorically different from disseminating it on paper?), but it would be more historically accurate to characterize it as a carefully and knowingly planned industrial experiment: what if we gave the cyberlibertarians what they want, a litigation- and regulation- free zone for the internet industry to develop in? The question that concerns us most in 2019 isn’t why the experiment was undertaken or on what terms, but how long the experiment was expected to last and what the return policy is if continuing it further is starting to look like a net loss. If the experiment isn’t working as originally conceived, what can be done to tweak or reverse the terms?
From all over the political spectrum and from the tops of various industries, people are starting to clamoring for just such a renegotiation, often with the threat of total abolition hovering in the background for rhetorical effect. SESTA/FOSTA, a complex law renegotiating internet corporations’ immunity from obscenity, sex work, and pornography laws, was the first law to chip away at 230, with a severe chilling effect across many corners of the internet. Reading between the lines of Elizabeth Warren’s rhetoric of “breaking up” the rogue adtech monopolies, one can detect two crucial pieces of Rooseveltian logic: 1.) that these companies have grown too powerful to be effectively regulated exactly because of the industrial advantage and network effects that 230 liability granted them, and 2.) that the revocation of 230 is an executable threat with which to cow the offending monopolies (and only them) into cooperation. (It’s worth mentioning that neither Warren nor Nancy Pelosi, whose constituency includes all of the relevant megacorporations, have directly threatened to revoke 230 as such; rather, the threat is implicit in ways worth parsing closely.)
One particularly dumb crack in the 230 dam
Last week, however, an unrelated, and I can’t stress this enough, far dumber attack was launched on 230, building on months of grandstanding and performances of a nationwide persecution complex on the part of Ted Cruz and other Breitbart-approved populists from the US right. Annoyed at a perceived “bias” against conservative thought in the moderation of tweets, knitting patterns, and other user-generated content that I would describe as too misguided to even debunk, Missouri Senator Josh Hawley proposed a law that would require the Federal Trade Commission to start doing what the Federal Communications Commission used to do with radio: laboriously monitor all online speech for “fairness”, defined in an even dumber “both-sides”/two-party framework than during the heyday of broadcast television and radio. Senator Ron Wyden, co-author of article 230 and generally tech-attentive legislator, has had to do another round of press (for the second time in a year) reminding the American public how indisputably irrelevant the concept of “fairness” is to his 26-word-long statue, and why the latter is worth keeping.
Regardless of what you think about 230, or about the validity of A.) scrapping it, B.) chipping away at it with more specific laws like SESTA/FOSTA, or C.) drastically reforming the related industrial framework around internet publishing, I just want to underline two points here. Firstly, in our topsy-turvy, post-GDPR moment, all three of those options are both actively being researched by influential American politicians and all three seem feasible outcomes.
Secondly, I want to insist on how foundational this liability shield has been for the modern internet. The drastic, consequential difference it made between traditional publishing and internet publishing allowed trillions of dollars to accrue to a tiny group of people that we now jokingly refer to as our “tech overlords” in reference to the feudal level of wealth and income inequality their industry has ushered in. I want to call 230 “bedrock regulation” because in defining the internet as a lawless zone of freeplay outside of the highly-regulated and litigious realm of publishing, it changed the course of both industries forever, and of the related bedrock regulations around copyright.
AirBNB and Immunity from Real Estate Consequences
What’s any of this have to do with housing, you might be asking, much less fair housing in Berlin? To make a long story short, AirBNB has longed for just such a regulatory bedrock since before they launched their site and began disrupting real estate worldwide. To put it more concretely, they have tried at every turn to lobby the courts of every jurisdiction in which they operate that renting an AirBNB is neither a real-estate transaction nor a hotel transaction. In many high-profile lawsuits, they have lost the latter argument; in a very consequential recent case, however, they seem to be winning the former, for all of Europe. If they succeed, this will be their 230: a “bedrock regulation” that exempts AirBNB from the laws governing those sectors of the economy where it has so indisputably wreaked havoc and “disruption”.
I have already gone on for too long to wade into the polemic waters of exactly how disruptive AirBNB has been to rental markets or real estate prices, or the methodological quandaries of how definitively we can arrive as measurements of those impacts. It is quite a rabbithole, and more appropriate to a dissertation than a blogpost (although, that said, a comparison of a few key cases analogous to Berlin is planned for this very blog!) In keeping with the focus of this article, though, it might suffice to say that however disruptive it is, US cities seem to think it’s disruptive enough to justify a wave of policy proposals to hold that disruption in check. While some of the specific measures proposed might be accused convincingly of doubling as tax grabs or defenses of entrenched hotel-industry interests, they share a common urgency against dismal numbers about housing affordability nationwide, particularly in cities with healthy job markets.
On the whole though, this groundswell of popular and political backlash has been going strong for years, particularly in the highly touristic cities where the disruption is definitively felt by all, and where the stratified but organized hospitality industry is a major player in local politics. Painting in broad strokes, the back-and-forth between AirBNB and specific regulators in Europe has been high-profile and contentious, with the strongest interventions in left-leaning tourist cities like Berlin and Barcelona. (This topic deserves an entire dissertation unto itself as well, but we will overview some relevant highlights in another blog post!)
Paris, however, has been a little late and uneven in its attempts to regulate AirBNB, perhaps because AirBNB has more listings there than any other city on Earth, or because of its already skyhigh prices, or because of a unique legal division of labor between city, state, and federal government on housing matters. Early in 2019, however, France started imposing invasive reporting requirements and new regulations, inspired by Barcelona’s relatively-succesful 2018 program to curb AirBNB’s inflationary effect on prices. AirBNB fired back, investng millions in fighting the aggressive French laws on the grounds that they were implemented too quickly for AirBNB to react and without adequate EU mediation. This latter point got the case all the way to the EU’s top court, which ruled in AirBNB’s favor in a non-binding resolution protecting the Ireland-based “online service” from the French government’s sovereignty over its pricey capital’s real estate. Perhaps there are other emergencies or exemptions where France might spring this kind of legislation on an international business concern, but the court ruled this was not one of them.
In my humble opinion, AirBNB’s real innovation (other than convincing middle-class people to pay half the price of a hotel to couchsurf) was finding a profitable economic niche that wasn’t technically illegal under existing laws and growing fast enough to hire armies of lawyers and PR professionals to legalize their practice after the fact. Mejor pedir perdón que permiso, as we say in Spanish: better to apologize later than ask permission (when you know it won’t be easily given). This perhaps uncharitable interpretation of their history as a whole is born of my decades of experience as a user of the site on both subletter and renter sides — I have watched closely as their terms of service, customer support pages, and hosting instructions have evolved as the legal climate changes and as I’ve moved and traveled between jurisdictions. Everywhere they go, they strike a delicate balance between the business they want to be doing, and the current local groundrules by which they are allowed to play.
AirBNB happily called the recent EU court decision “a clear overview of what rules apply”. I think they are seeing what they want to see: every time there is a contradiction or ambiguity between two sets of rules, they lobby hard and publicly for the more lax of the two, which offer the greatest liability from legal or civil impediments to their taking a cut of the maximum number of informal sublets.
What rules apply: trustless transactions and community justice
Here is the heart of matter: Which rules, exactly, should apply? Is AirBNB an “online service” that plays matchmaker between free agents in a libertarian “sharing economy,” or does connecting renters and leasers make them fundamentally equivalent to a rental agent and/or a property management company? Are rental arrangements made between French subletters and previously unknown tenants subject exclusively to EU law because the two parties met on and negotiated through an Irish website, or does France’s territorial sovereignty extend even to virtual arrangements? How severe must the local impacts be in one city or national jurisdiction before an emergency can be declared, or an exemption expected from international trade agreements and the legal hierarchies of the EU? These are very 2019 questions that require us to tease out many controversial assumptions and ideological precepts.
I think the overlaying of so many different jurisdictional disputes can distract us from the fundamental issue, and we should avoid letting AirBNB (or tech generally) from framing this as tech versus regulation, or international commerce versus local overreach. My fundamental view of government (and, for that matter, of governance generally) is that laws and regulations and governments exist primarily to give communities some measure of sovereignty, however flawed, over the economic activities that directly affect them. Too often, though narratives of tech “magic”, business innovation, or free speech can move our focus away from that fundamental question of sovereignty, and introduce unquestioned assumptions about economics which sound political and authoritative.
Lest I be accused of being too hard on the disrupters, I will admit that there is some genuine user-generated magic to what AirBNB enables: by crowdsourcing trust and reputation, peer-to-peer subletting transactions between mostly good-faith players can create a new market. This isn’t inherently a zero-sum game: each airBNB transaction isn’t necessarily robbing one equal transaction from the hotel industry (or hostel industry), and we should neither assume every 5 or 10 AirBNB listings in a city equate to one apartment taken off the rental market. End-users benefiting from this new option is a net good for society, and for the economy; the collateral damage to the rental market and adjacent industries, however, need to be addressed by AirBNB cooperating with local authorities to address those “externalities” of which AirBNB wants to wipe its hands. The costs cannot be offloaded onto local authorities by a jurisdictional sleight of hand, nor a blanket immunity from local consequences.
Alarmed by the EU court siding with disruption or self-determination and economic self-defense, ten of the European cities where AirBNB has wreaked the most havoc (excuse me, “disruption”) have begged the same court for jurisdiction over the main driver of their respective rental crises. I worry this flashpoint could get lost in the shuffle of so many other flashpoints and crises worldwide, but it is a fundamental one in the evolution of tech’s legal and regulatory position. We cannot let this get buried in on page six — this is crucially important to the fairness of housing markets everywhere.
And the strategic importance of this legal battle cannot be understated. Not only is AirBNB lobbying to be protected from the angry citizens and industry stakeholders in every city they make an easy fortune, they are also lobbying to substantially limit the legal options available to 10 cities, each experimenting with a homegrown regulatory solution. Fighting ten very different battles simultaneously, and more generally the long-term regulatory uncertainty this precedent could usher in, is a fate AirBNB will obviously invest millions towards avoiding. Ten laboratories each trying a unique solution to its unique problem is an outcome this blog much prefers, because ten chances at getting it right is ten more than we have otherwise. Besides, I for one love having more to blog about.