Skip to main content
Menu
Airbnb Probably Isn’t Driving Rents Up Much, At Least Not Yet

Airbnb and other short-term rental platforms have been a boon to travelers in recent years, offering lodging that is cheaper, better located or more interesting than traditional hotels. But as such platforms have grown, housing advocates have become increasingly concerned that what’s good for visitors is bad for residents – that as more landlords rent out their units to out-of-towners on Airbnb instead of offering them to locals to live in, they are exacerbating already severe housing shortages in cities such as New York and San Francisco, driving up rents. Airbnb and its backers dispute that charge, arguing that most spaces offered on its platform would otherwise be vacant. Figuring out who’s right has been difficult, because Airbnb has released only extremely limited reservation data.

But a FiveThirtyEight analysis of Airbnb booking and revenue data provided by consulting firm Airdna gives the most rigorous look to date at how many units Airbnb could be taking off the rental market nationwide. It shows that Airbnb’s impact is probably still small in most cities, but it also shows that a disproportionately large share of the company’s revenue comes from the listings that most worry its critics — homes that are rented out for a large portion of the year. That could give the company an incentive to focus on increasing such listings as it grows — something some experts believe may already be happening.

Airbnb disputed the analysis but declined to provide its own data.

Airbnb rose to prominence as a way for renters and homeowners to list spare rooms or even couch space to out-of-town visitors, or to rent out their apartments when they left town. Such “shares” still account for the majority of all listings on the site. But a growing number of the rooms on Airbnb are so-called commercial listings — whole units rented out full time, like hotel rooms. It is these commercial units that have drawn the most scrutiny from critics such as Sens. Elizabeth Warren, Diane Feinstein and Brian Schatz, who last month wrote a letter to the Federal Trade Commission calling for an investigation into the practice. Critics argue commercial listings take off the market units that would otherwise be available to local residents, thereby reducing the supply of housing and driving up rents.

According to FiveThirtyEight’s analysis, there were nearly 9,000 commercial listings in Airbnb’s 25 largest U.S. markets between June 2015 and May 2016.1 That’s just 10 percent of Airbnb’s total listings in those markets. But those commercial listings make up about a third of host revenue.

CITY ACTIVE LISTINGS SHARE COMMERCIAL
Total across markets 91.2k 9.7%
Honolulu 1.4 21.1
Portland, OR 2.7 15.6
Los Angeles 8.9 15.5
Seattle 3.4 12.5
New Orleans 2.8 11.9
Nashville, TN 1.8 11.6
Washington, DC 3.4 10.8
Denver 1.8 10.6
Dallas 0.6 10.6
Salt Lake City, UT 0.6 9.7
Las Vegas 1.4 9.2
Phoenix 0.7 9.2
Boulder, CO 0.9 9.1
Atlanta 1.3 8.9
San Francisco 6.9 8.9
Orlando, FL 1.0 8.9
Boston 1.4 8.6
San Diego 3.6 8.4
Miami 2.5 8.3
Houston 1.0 8.2
New York City 30.8 8.0
Austin, TX 4.6 6.9
Chicago 4.9 6.7
San Jose, CA 0.7 6.1
Philadelphia 2.1 5.4
What share of active Airbnb listings are commercial?

Commercial rentals are defined as units that are rented at least 180 days each year.

Source: airdna

Experts recommend looking at revenue, rather than raw listings, as a measure of the company’s impact on the housing market because listing numbers alone don’t reveal how often a unit is booked. A spare couch and a professionally managed apartment both count as a “listing” on Airbnb, but the couch, which might only be booked a few nights a year, has no practical impact on the local housing market. An apartment that is reserved by Airbnb users 250 days a year, by contrast, is a unit that isn’t available for local residents to rent.

In some cities, commercial listings make up a large share of Airbnb bookings. In Los Angeles and Portland, Oregon, nearly half of all the money travelers spent on short-term rentals went to commercial hosts. In New York and San Francisco, two cities where Airbnb has faced particular scrutiny, commercial listings accounted for about 30 percent of revenue. In every one of Airbnb’s top 25 markets, commercial listings accounted for at least 20 percent of host revenue.

CITY HOST REVENUE SHARE COMMERCIAL
Total across markets $1,133m 32.7%
Los Angeles 121 46.4
Honolulu 26 45.0
Portland, OR 32 40.8
Washington, DC 37 36.2
Seattle 42 33.7
New York City 384 32.5
Denver 19 32.3
San Francisco 110 32.1
Dallas 6 31.5
Atlanta 12 31.3
Boston 21 31.2
San Diego 40 28.8
New Orleans 45 28.5
Las Vegas 18 28.4
Orlando, FL 9 28.3
Boulder, CO 9 27.6
Salt Lake City 5 27.4
Phoenix 6 27.0
Houston 8 26.6
Nashville, TN 29 25.9
San Jose, CA 8 25.5
Chicago 46 25.3
Miami 27 25.3
Austin, TX 57 21.6
Philadelphia 15 20.6
What share of Airbnb host revenue comes from commercial rentals?

Commercial rentals defined as units that are rented at least 180 days each year.

Source: airdna

The revenue numbers alone don’t necessarily mean Airbnb is having much effect on rents in the cities where it operates. That is because the raw number of commercial units in most cities remains low, under 1,000 in all but the biggest cities. Even in New York, Airbnb’s biggest U.S. market, the service had only about 2,500 commercial listings, a sliver of the city’s roughly 2.2 million total rental apartments.

Stockton Williams, executive director of the Terwilliger Center for Housing at the Urban Land Institute, said that as a percentage of total rental units in America’s big cities, the commercial units don’t add up to much. He said other factors, such as the increasing demand for urban living, have played a much larger role in driving up prices in big cities.

“It’s a very valid concern,” Williams said. “But I’m not sure there is evidence that Airbnb has had a significant effect on either price or supply.”

That could change as Airbnb grows, however. Airbnb reports it currently lists over 2 million spaces for rent worldwide. Analytics firm 7Park Data pegs the exact number at 2.3 million, up from 3,000 in 2009, a compound annual growth rate of 153 percent. Airbnb added about 229,000 new units in North America alone in 2015, according to 7Park Data. (Airbnb is by far the largest such service. Its main rival Homeaway lists about 1.2 million units worldwide, according to its website. FlipKey, owned by TripAdvisor, lists about 300,000 units, according to its site. Many units are listed on more than one platform.)

The available data suggests commercial use on the site is growing as well, despite the company’s oft-repeated statements decrying de facto hotel operations. A 2014 report by the New York attorney general based on subpoenaed Airbnb reservation data going back to 2010 found that the proportion of host income earned on commercial listings had grown steadily over that period. Williams said there is “no question” that commercial listings are becoming a larger share of the company’s business. He cited recent discussions between Airbnb and a group of major national landlords exploring a potential revenue sharing deal wherein the landlords would allow their tenants to openly rent their units through service, while taking a cut of the revenue.

Estimating the scale of the commercial-listings market, and its growth, is difficult, however, because Airbnb has generally refused to release bookings data, citing users’ privacy. Instead it has released only aggregate data illustrating Airbnb’s economic impact or the behavior of typical users. “Airbnb shares data with the public they think supports their position and withholds data they think opposes their position,” said Ben Edelman, a professor at Harvard Business School who recently published a widely cited study on racial discrimination by hosts using the service.

Airbnb spokesman Nick Pappas disputed that characterization, reiterating what he called the company’s commitment to sharing data, and citing Airbnb’s recently released Community Compact and its new by-city data dashboards.

FiveThirtyEight’s analysis is based on data scraped by Airdna, a company unaffiliated with Airbnb that provides market reports and data services to hosts and real estate investors. The company’s computers visit every Airbnb listing roughly once every three days, recording the unit’s price and whether it is available for rent. The company then uses that information to estimate how many days a year a unit is booked and how much revenue the host received from it.2

FiveThirtyEight’s analysis isn’t the first to use such scraped data. Hotel industry groups, universities, housing advocates and city governments have studied various aspects of Airbnb’s impact, including commercial listings. Data for those studies, which generally focused on individual cities, usually came from Airdna, or from Inside Airbnb, a website maintained by anti-Airbnb activist Murray Cox that publishes raw listing information. Other studies were based on scrapes performed by the analysts themselves.

There is no universally accepted definition for “commercial” units. FiveThirtyEight used a standard developed by New York’s attorney general in the 2014 report, which is one of the most conservative definitions used in previous studies. To be considered “commercial” in our analysis, a listing had to be for an entire home or apartment (not just a room within a larger unit) and had to have been booked for 180 days or more between June 2015 and May 2016. 3

Airbnb disputed the analysis. Company spokesman Christopher Nulty said some of the units FiveThirtyEight identified as “commercial” might instead be boutique hotels or guest houses. He also called the 180-days-booked standard “arbitrary,” and said an approach based on days-booked could never accurately identify units that would otherwise be long-term rentals.

“This analysis doesn’t provide a complete and accurate picture of our community,” Nulty said in a statement.

FiveThirtyEight’s analysis provides fodder for both sides of the Airbnb debate. It suggests the absolute number of commercial units on Airbnb is relatively low, representing well below 1 percent of all rental units. Any impact on rental supply is therefore small, and limited to a handful of neighborhoods where such services are most popular. (Airdna provided data at the city level, making it impossible to estimate the impact Airbnb has on specific neighborhoods.) That undercuts claims by some of the services’ harshest critics, who say Airbnb and the rest pose a clear and present danger to city dwellers.

At the same time, the analysis shows that listings that are operated commercially, though a small minority, represent a major chunk of all host revenue. That’s especially important in cities such as New York, Los Angeles and San Francisco, where low vacancy rates, rising rents and progressive populations make Airbnb a prime target for activists and politicians seeking tougher regulations.

Dependence on commercial operators complicates Airbnb’s claims that it represents homeowners and renters making extra money “sharing” their living spaces, and strengthens the hand of critics who say that cities should treat the services more like traditional hotels, which are generally far more tightly regulated. And it could give Airbnb an incentive to try to attract more commercial investors as it seeks to grow.

George McCarthy, president of the Lincoln Institute of Land Policy, a think tank, said the prospect of new investors — including massive, national real estate firms — is real. “Landlords might see it as a better option in the market than conventional renting,” he said.

“I’m not convinced it’s that big a deal yet,” McCarthy said, “but the growth rate would be of concern.”

Footnotes

  1. The 25 “principal cities” of the Metropolitan Statistical Areas with the highest number of active Airbnb units during the period.
  2. Airbnb previously allowed hosts to indicate on their calendar when a listing was booked, allowing scrapers access to that information. Airbnb removed that option late last year. As a result, Airdna identifies instances when a unit becomes unavailable, then uses an algorithm based information collected during the previous period to estimate the likelihood that event amounts to a booking. Airdna calculates revenue by adding up the per-night cost for all the bookings it identifies, plus cleaning costs.
  3. We also excluded new listings and, at Airbnb’s request, units only available for bookings of 30 days or more. These listings are often for traditional sublets, or are operated by corporate housing firms that predate Airbnb.

Ariel Stulberg is a writer and data analyst based in New York.

Comments