Airbnb And The Problem Of Data

Housing activists, San Francisco policy makers and Airbnb are in the midst of another showdown over short-term rental regulation that will stretch into next month and potentially into the November election. The debate centers on enforcement, data and caps on the number of days a host can put a unit out on a platform.

If a new set of rules isn’t formalized soon, the battle will head to the ballot. That would mean that voters will get to decide on how they want to manage short-term rentals in San Francisco, which is in the midst of one of the worst housing shortages in the city’s history. If regulation goes to the ballot on a citizen’s initiative, that will mean that the city’s rules will be more inflexible than they would be if a bill passed through the normal legislative process, since amendments will have to be ratified by the voting public.

Ultimately, how San Francisco, Los Angeles, New York and other economically prosperous and land-constrained cities decide to regulate Airbnb will impact how much the company can grow its inventory in its most lucrative markets. While Los Angeles city council members are currently considering new rules, New York Mayor Bill de Blasio’s administration does not yet have a position on short-term rentals.

In any case, all of these discussions may affect how much the company is worth in an eventual IPO. CEO Brian Chesky said last month that the company is on track to host 1 million guests per night. With more than 1.2 million rooms on its platform, Airbnb’s inventory is now larger than branded chain hotels like Intercontinental, Marriott and Hilton. If it was successful in raising additional funding at a $20 billion valuation as we and Bloomberg have reported, that implies that its latest investors are probably betting on a 2 or 3X return or more.

The key questions here are:

  • Should there be a cap on how many days a host can rent out their unit on concern that short-term rentals will erode the existing housing stock? If so, what should it be?
  • Should hosting platforms be compelled to share data with the city? Should it be in an aggregated, anonymized way or at an individual level?
  • How should permitting work? Should hosting platforms be penalized if they list units that aren’t registered with the city?
  • How should complaints and enforcement work? Should neighbors have the right to file complaints? Should they be entitled to financial relief if they win and find that a nearby host is violating city law?

Screen Shot 2015-05-28 at 1.49.36 AM

Screen Shot 2015-05-28 at 1.49.46 AM

Caps on a maximum number of days:

A huge point of debate is whether there should be a cap on the number of days a unit can be rented out. The suggested numbers range from 60 days in supervisor David Campos’ proposal to 75 days in the November ballot initiative to 120 days in the legislation backed by Mayor Ed Lee.

It’s incredibly hard to identify an appropriate cap. All of the studies — from the San Francisco Planning Department, the legislative analyst’s office and the city economist — operate off of website scrapes and data subpoenaed by other governments, because Airbnb and the other hosting platforms do not share aggregate, anonymized data with the city government. The other platforms like Homeaway don’t collect this data because their revenue model isn’t based on bookings.

Airbnb disclosed a select amount of data on Monday about the 9,692 units that were listed from May of last year until April on the platform. They argued that a person would need to rent out a room for 211 days to make it more lucrative than taking on a long-term tenant and that only 6 percent of their listings crossed that threshold.

Screen Shot 2015-06-10 at 1.04.55 PM

But it’s problematic to price breakeven rates off of market-rate rents, because only 10 percent of the city’s housing stock is market-rate. Almost half of the city’s housing units are rent-controlled and then another third are owner-occupied. As such, it’s not clear what people are paying and when the financial difference between what a landlord could earn on Airbnb is strong enough to forgo renting out an extra bedroom or even to try and illegally evict a tenant. The city hasn’t studied the demographic composition of the rent-controlled stock in more than a decade, as there are concerns from tenants groups about how the data would be politicized. (Yes, the need for data goes many ways.)

One study from the legislative analyst’s office, commissioned by Supervisor David Campos, estimated that between 925 and 1,960 units citywide were removed from the housing market through Airbnb listings. A separate subpoena on New York City by the state’s attorney general last year found that six percent of the city’s hosts were generating 37 percent of the revenues and that 38 percent of total revenues in 2013 were going to hosts who rented their units out for more than half of the year.

The concern here is that a long tail of casual hosts is concealing the fat head of Airbnb’s profitability with more permanent operators, who may be removing housing units from supply. There is a whole cottage industry of professional Airbnb managers, including Pillow Homes, which just raised $2.7 million in January and offers a guaranteed monthly income stream to its hosts.

San Francisco’s city economist Ted Egan acknowledged that is impossible to how many units, if any, are taken out of the city’s permanent residential housing stock without accurate data. Still, in a separate report, he argued that all of the proposed caps are more than conservative enough to protect housing. His breakeven calculations ranged from 123 days a year in Russian Hill to 241 days a year in the Inner Sunset.

Because Airbnb launched a pricing recommendation algorithm last week, you can actually check the hypothetical breakeven rate on your own space. The recommended prices on my own personal bedroom in a shared flat have fluctuated from $90 to $160 per night. So the annualized breakeven rate on my room is sometimes below Supervisor Mark Farrell’s suggested 120-day cap, and is sometimes a month or more higher. (I’m a frequent Airbnb guest, but I do not host because I’m a tenant and the current city laws around short-term rentals do not supersede existing tenant-landlord agreements. Because most tenant agreements do not explicitly mention short-term rentals, the city’s legislation in practice is probably more favorable to landlords.)

Airbnb responded to my point, saying that evicting tenants to use a unit for short-term rentals is illegal. But in practice, it’s difficult to identify and then penalize these cases when they happen. For example, city attorney Dennis Herrera settled a $276,000 case last month with two landlords who had put their unit on VRBO and Homeaway after evicting a disabled tenant. But it took six years from the time the defendants started listing the unit in 2009 to get the case settled. That’s because enforcement is complaints driven, rather than automatically matched by unit to Ellis Act eviction or property records.

Iterative regulation and data transparency:

The data problem speaks to a larger issue as technology companies facilitate more and more online-to-offline behavior, which at scale starts to impact public systems. In Los Angeles, driving app Waze became so popular that it began moving large flows of traffic through side streets and residential neighborhoods instead of main thoroughfares and highways.

Waze, which has a lot less to lose than Airbnb, chose to share aggregate, anonymized driver data with Los Angeles Mayor Eric Garcetti’s administration to design better traffic flows. Of course, you might argue that roads are public infrastructure while most housing is privately owned. However, the practice of residential zoning has been upheld at the U.S. Supreme Court level for almost a century.

Supervisor Jane Kim, who represents the Tenderloin and Central SOMA, hinted that the city policy makers might push for more data down the line at Tuesday’s Board of Supervisors meeting.

She said:

I don’t think there’s an equivalent between the example that I brought up about sharing people’s private emails and the content of that email to sharing the number of nights that a host rents out to a tourist is the equivalent type of data sharing. That being said, if we as a city government are going to dive in to that policy area of data sharing, I would feel more comfortable knowing that we had a larger policy discussion in the long-term about when we do that and when we don’t, so we set parameters to make sure we are really protecting the valid privacy and security rights of our residents.

These are not ideas that only regulators or politicians are talking about. Technologists are discussing them too.

Nick Grossman of Union Square Ventures, the high-profile East Coast venture firm that has backed Twitter, Kickstarter and SoundCloud, penned a series of essays on what he calls ‘Regulation 2.0’.

Grossman says that is:

The idea that, given the huge volume of real-time data produced by modern web services and the potential for radical transparency based on that, there is an opportunity to explore completely new regulatory approaches.  Approaches that, rather than make up-front decisions about an activity (say, ride-sharing or peer-to-peer apartment renting), as we do in a “1.0” regulatory regime, we can instead be more permissive on the front-end, while at the same time introducing increased accountability through transparency.

At its core, Airbnb is a form of regulatory arbitrage between hotel and permanent residential zoning. It takes one kind of physical space, previously available only to one kind of demand — permanent residential demand — and makes it efficiently available to an additional type of short-term or itinerant demand, which potentially resets the housing stock at a higher equilibrium price at scale. While there are many larger underlying contributors to the housing shortage in San Francisco, it is possible that Airbnb adds stress on the margins.

But we wouldn’t know that without data.

While I’ve written extensively on how the region has underbuilt housing relative to population growth and urbanization, it has also underbuilt hotel space. Hotel occupancy rates in San Francisco have actually increased even despite Airbnb’s growth. That’s why hotels haven’t been visible political opponents to the company.

Screen Shot 2015-06-10 at 1.17.30 PM

Screen Shot 2015-06-10 at 1.17.19 PM

San Francisco city hotel occupancy rates and tax revenues continue to rise even with the emergence of Airbnb and other hosting platforms.

But unlike hotel zoning, where the city allocates set heights and densities and developers build hotels on fixed pieces of land, short-term rental inventory fluctuates.

Cities need to understand whether short-term rental inventory is expanding faster than they can produce housing stock and whether that is impacting the availability of housing to long-term residents. Put another way: you would not A/B test changes to a software platform or interface without data, so why would you ask city policy makers to A/B test regulatory changes in the dark when it comes to the “platform” of their finite physical space?

I know that the counter-argument is that the city should just build more housing. But it is building housing at the fastest pace in two decades, along with pushing for a $300 million affordable housing bond and creating subsidies for middle-class housing for the first time in its history.

Even if the San Francisco Bay Area had as permissive an entitlement process as Houston, which was approving more units than the entire state of California in certain quarters last year, it is just easier to put an existing housing unit onto a short-term rental website than it is to go through the motions of building an entirely new unit by raising capital, acquiring land, hiring an architect, going through the entitlement process, addressing environmental impacts, hiring construction workers and marketing and selling the units. The developer who did the Vara project at 15th and Mission told me earlier this week that it’s about $650,000 in raw costs per new housing unit at this point in the economic cycle.

On top of that, there are many ways in which Airbnb and municipal governments are just not fundamentally aligned. While the identity, reputation and review systems of a hosting platform can supplant some of the governmental licensing meant for safety, Airbnb is a privately held, for-profit company that is incentivized to maximize revenues and bookings, while city governments are accountable to their voting public, who are long-term residents who live and work in the city and need housing.

City economist Ted Egan estimated that for every housing unit that disappears, there is a loss of $250,000 to $300,000 per year in impact to the city’s economy. Those losses can quickly outpace the more than $1 million per month that Airbnb remits to the city in transit occupancy taxes, plus whatever hosts earn and visitors spend, he argued.

To be clear, none of this is an argument to ban or eliminate short-term rentals. It’s just to have accurate and timely data as the market scales, so that negative impacts on housing can be managed. That a privately held company touches an increasing share of the residential housing stock in the world’s most economically productive, culturally vibrant and land-constrained cities is a matter of public concern.



Airbnb’s public policy department made this graphic to criticize San Francisco’s registration process.

While most of this post has been harsh on Airbnb, there are areas in which the city government has made things too complicated.

The San Francisco Planning Department decided to require in-person meetings to register as a host. Of course, this has resulted in only about 625 out of thousands of hosts setting up an appointment with the city as of yesterday. The rationale is that the planning department wanted to make sure that the host is the person who owns or leases a unit and that it’s their primary residence with proper documentation.

Again, I do think this goes back to the lack of data.

If you think about the situation, you have two or three dedicated, full-time planning staff members, who are paid government wages, collecting self-reported forms once a year from thousands of hosts.

Somehow they have to replicate the database that a $20 billion company, which can hire the best engineers in the industry with stock options, can probably call up on internal dashboards instantaneously.

Unlike Airbnb, they are also responsible for making sure that all of these listings follow building safety, fire and health codes and they also need to make sure that none of the units are income-restricted, or affordable.

Civil suits and injunctive relief:

If Airbnb is not careful amid what is a very politically fractious situation, this is going to go to ballot in November.

And I can tell you that the November ballot initiative, which is backed by a bunch of citizens’ groups and housing activists, allows ‘Interested Parties’ or neighbors to win financial relief in civil court after a Planning Department investigation. That would raise the spectre of neighbor-versus-neighbor lawsuits, which could have a chilling effect on the total number of listings. That would be way worse.

The other legislative proposals are more conservative on this. Farrell’s proposal only allows the city to obtain civil penalties, for example.

Technological change and “post-tourism”:

Lastly, it’s important to remember that technological change breeds changes in the way that we travel.

The ocean liners and railroads of the late 19th century gave rise to grand hotels like the Waldorf Astoria in the U.S.’s Gilded Age and Europe’s Belle Epoque. Then, the private automobile cemented the road trip as a quintessential American traveling experience, with scores of motels planted along the sides of highways. Jets and plane travel enabled large, international chains with franchised hotels offering the standardized experiences that Airbnb now criticizes for being sterile.

The web, with its identity and reputation systems, is facilitating something altogether different. Much of the current Airbnb opposition in San Francisco criticizes the company for turning homes into tourist hotels.

But I don’t think it’s really accurate to call what is happening “tourism.” It’s an entirely new market that isn’t really competitive to traditional hotels.

In a New York Magazine article touching on how expats and Airbnb are changing Berlin, Johannes Novy, an urbanist at BTU Cottbus-Senftenberg, talked about the idea of post-tourism:

the notion of “post-tourism” doesn’t represent a change in the way we travel so much as a change in the way we think about travel – and it primarily shows the extent to which technology and heightened mobility have helped obliterate categories like “work” and “vacation,” “local” and “visitor.” “The idea of this new tourism puts into question a lot of fundamental things about the kinds of lives people live these days,” he says. “We are in a world of endless mobility, and so we are all tourists all of the time.”

I personally know several people who don’t really live anywhere, and perma-float from Airbnb to Airbnb listing globally.

So at some level, the Airbnb regulatory debate might be the beginning of a longer-term discussion about the way we think about long-term residents versus an emergent and more globalized and mobile population.

Indeed, the most interesting struggles in urban real estate today have to do with increasingly globalized flows of capital and labor. They have to do with how overseas capital from non-resident home buyers and increasing flows of short-term rental tourists are subverting the protective scaffolding the U.S. originally built half a century ago around residential property to stimulate home ownership for the middle-class.

And that is a very interesting conversation.