While the Board of Supervisors debates revisions to regulations on short-term rentals, expected next month, Airbnb—the $20 billion behemoth that dominates this industry—is busily marshaling U.S. Census data to defend its business.

Its recent 10-page report on its impact on the city’s housing calls accounts of thousands of housing units being removed from the rental market “both baseless and mathematically impossible.”

“The Airbnb community has no material impact on housing availability in San Francisco,” the report insists.

In fact, say independent statisticians and demographers, U.S. Census data on vacancy rates doesn’t address this issue. One statistician with the U.S. Census said that short-term platforms arguably fly under the radar of the questions.

“It may be something we have to send out instructions on,” the statistician said, adding that new developments often call for adjustments in the way Census data is collected. Officially, the U.S. Census said it is not revising its questions.

Moreover, the data is unreliable—it’s not current, it’s gathered in different ways and there’s a large margin of error, said other statisticians.

“One of the problems is the data lag and when we would expect to see the impact,” said Dr. Carolina Reid, an assistant professor in the Department of City and Regional Planning at the University of California, Berkeley. The data come from the decennial census and the American Community Survey.

Any impact, she said, was unlikely to be evident in the data from the American Community Survey, which has “a very large margin of error because it’s a small sample.”

Moreover, statisticians note, although Airbnb has disrupted the hotel market since its inception in 2008, much larger forces played on the post- 2008 housing market – including the worst recession in decades.

Michael Orr, the director of the Center for Real Estate Theory and Practice at Arizona State University wrote simply: “The census does not realistically measure the effect of Airbnb.”

An Airbnb spokesperson declined to comment on the reports’ reliance on Census data.

The other glaring weakness in Airbnb’s report is its assumption that San Francisco landlords are motivated entirely by money. This assumption allows Airbnb to conclude that both of the caps being considered by the Board of Supervisors – 60 and 120 days – are far below the 211 nights a landlord would have to rent out a unit on Airbnb to make the short-term rental a better financial deal than a long term tenant, according to its analysis.

Alex Marqusee, who produced a 52-page report for the City Planning Department as part of his Master’s Project for the Goldman School of Public Policy, came up with a similar number – 213 nights.

But, while Airbnb’s analysis stops with the profit motive, Marqusee offers a much fuller picture of why a landlord might be motivated to move into the short-term rental market and why a cap is necessary in places like San Francisco where housing shortages have jacked up prices.

“There are so many pressures on landlords to evict tenants and [Airbnb] is an opportunity on top of” those pressures, said Marqusee.

Landlords’ unhappiness with rent control and regulations is well known. Some 39 percent of the landlords surveyed in 2003 said they kept vacant apartments off the market because of regulations, his report notes.

Airbnb gives landlords new options. Take the Bernal homeowner who has a legal unit attached to her house. She has a modest income and purchased her house a decade before the last dot-com boom.

She’s kept the unit occupied, but would now like to have it for her family’s use. If she does an owner move-in eviction, however, she must pay her tenants a moving fee.

“Why should I have to give people money to get back into my own home?” she asked. Thus, when the current tenants move out, she will keep it off the market and use Airbnb on the days when the family doesn’t need it.

Another Mission landlord has already turned a long-term studio rental into an Airbnb for similar reasons – the ease of being able to make the space available to family and friends while also being able to make money on Airbnb.

Marqusee’s report goes on to show how easily landlords can recoup the  estimated $4,500 a year it generally costs to operate a unit in San Francisco, by comparing those costs to a 2012 Airbnb claim that its average host generated $6,772 income annually.

His bottom line: landlords need to rent an apartment for only 24 days to cover their operating costs. Moreover, if landlords don’t want the hassle of dealing with short-term tenants, they can use one of the concierge services like Pillow to manage their properties.

These pressures and services explain why, contrary to Airbnb’s claim, Marqusee and the city’s legislative budget analyst both concluded that Airbnb is reducing the available housing stock.

While a cap of 120 days would only make 1.7 percent of the city’s apartments vulnerable, the number goes up to 7.4 percent at 150 days and to 21.6 percent at 180 days, according to his report.  There is no impact at 75 days and .2 percent impact at 90 days.

Marqusee found that 10 percent – or some 500 units—have been taken off the market and are essentially commercial Airbnb units.

These, he found, are concentrated in the downtown and central neighborhoods including the Mission District.  “There is clearly a link between areas that are gentrifying and attractive to tourists and commercial Airbnb units,” he said.

His estimates are admittedly conservative; the city’s budget analyst found a higher number, some 23 percent or 925 units citywide with scenarios in some neighborhoods like the Mission totaling 29 percent or 1,960 units.

Marqusee said he thought about the possibility of geographic caps – allowing more nights in some neighborhoods and fewer in other neighborhoods – but was more focused on addressing questions raised in the proposals at the Board of Supervisors.

His report recommends a cap of 120 days instead of the alternative proposal of 60 days. His reasoning gets into the policy decision in all the proposals to eliminate the current distinction between so-called hosted visits, in which the tenant or owner is present, and un-hosted visits, when a tenant or owner is absent. The current legislation has no cap on the first and a 90-day cap on the second.

Because the distinction is impossible to monitor, the two proposals have come up with either a 60- or 120-day cap for any type of visit.

Given the need to combine the two kinds of visits, Marqusee said, the 120-day cap allows tenants renting out a bed or room while they are there to continue to benefit from sharing their space. The money these hosts earn, he said, helps them stay in the city and his study found that it is unlikely they are taking rooms off the market. A cap of 60 days would deprive these hosts of significant income.

Regardless of the cap, he stressed, the final policy needs to have teeth. “The enforcing agency needs to be able to require short term rental housing platforms to regularly provide non-anonymized data and/or to fine hosting platforms each day for listing illegal short term rentals.”

As it is now, any cap is “completely unenforceable ” and an unregulated market will “lead to the loss of more long term rentals. ” That outcome, the city’s reports show, ends up hurting everyone and negating the benefits to host, city’s tax coffers and the local economy.

Earlier Stories

Airbnb and Tax Collector Wrangle Over Payments,  June 18, 2015

Airbnb Cuts into Shortage, Should Share Data, May 14, 2015

The Supes Try (again) to Regulate Airbnb, June 9, 2015

Airbnb Hosts,  ‘It’s about the Economy and the Experience’ ,  August, 2014