Airbnb regulation in Santa Monica

Airbnb Regulation: How is new legislation impacting the growth of short-term rentals?


Executive Summary

Berlin, Barcelona, San Francisco and Santa Monica have imposed tough regulations on Airbnb hosts and the platform, in an attempt to curb the short-term rental market. By utilizing differences-in-differences and log regression methodologies, this report analyzes to what extent this Airbnb regulation has been effective. It finds that punitive Airbnb regulation, in particular steep fines for hosts, has the most impact on Airbnb. In Berlin and Santa Monica, listings fell by 49% and 37% after fines of $100,000 and $500 respectively were introduced. By comparison, in San Francisco, a 90 day cap on renting entire homes on Airbnb led to a modest 5% decrease in listings.

Conversely in Barcelona, limiting the licensing of all tourist accommodation (including hotels and hostels) actually increased the number of Airbnbs in the city. In the case of targeting professional Airbnb hosts, in San Francisco, the 90 day cap caused a 16% drop in professional host listings compared to a 5% drop in the overall group. Professional hosts in both Berlin and San Francisco, however, were less affected by Airbnb regulation, as they readjusted their nightly prices to compensate.

 

Legislating for Airbnb: How Has Regulation Made an Impact on Airbnb Around the World?

Many cities around the world have grown frustrated and intolerant to hosts putting up their homes on Airbnb. The short-term rentals platform has been blamed for unruly tourists, rising house and rent prices, as well as a weakening of the local community fabric. As a result, many cities have turned to Airbnb regulation, imposing laws that are tough on both hosts and the platform.

Each city has taken its own approach towards Airbnb regulation. Some have decided to fine hosts, others have placed caps on how many days hosts can rent out their homes. But has this regulation had an effect?

By using Monthly Property Performance Data from Airdna, we can track the evolution of listings and other data over time. Airdna counts a listing as ‘active’ if it has been live on the site, or has had a booking, in the last month. The results therefore do not include Airbnb listings that are no longer active but haven’t been removed from the site, but do include bookings in the last month that have since been removed from the Airbnb website.

 

Tough fines for Airbnb hosts in Berlin, Germany

Airbnb regulation in Berlin

 

Berlin has perhaps the strictest Airbnb regulation in the world. Starting on May 1, 2016, Berlin banned hosts from renting out their homes to short-term visitors unless the hosts occupied at least 50% of the homes themselves. Currently, the fine for hosting entire homes on Airbnb is an eye-watering $100,000. This doesn’t affect, however, Airbnb hosts renting out private or shared rooms in their own house.

Data collected by Airdna is a good reflection of how the ban in Berlin has made an impact on Airbnb in the city, using a difference-in-differences methodology. The graph below shows that the rise of the number of entire home listings flattened after May 1, 2016. By comparison however, the number of shared home listings continued to rise. We can assume that if the ban had not been introduced, the number of entire home listings could have grown at the same rate as the number of shared home listings.

 

Number of Airbnb Listings by Entire Home and Shared Home Listing Types in Berlin, November 2014 – January 2017

Entire Home and Shared Home Airbnb Listings in Berlin Nov-14 to Jan-17

 

This common trends assumption is based on an analysis of the trajectories of entire home listings and shared home listings are before and after the ban in Berlin. Before May 1, 2016, the trajectories of entire home listings and shared home listings were 99% correlated. After May 1, 2016, the two series were only 22% correlated.

To determine the impact of banning entire home listings in Berlin, differences-in-differences methodology calculates the difference between what really happened after the ban and what would have happened had the ban not been introduced [1]. The graph below shows what the entire home listings trajectory would have looked like if it had followed the growth trajectory of shared home listings.

 

Number of Entire Home and Counterfactual Airbnb Listings in Berlin November 2014 – January 2017

Number of Entire Home and Counterfactual Airbnb Airbnb Listings in Berlin Nov-14 to Jan 17

 

Other reports that have looked at the effect of the ban in Berlin have found a 40% decrease in the number of Airbnb listings in a month. Our regression results show that indeed, monthly listings did fall by approximately 49% after the ban, which also corresponded to a 55% decrease in monthly revenue.

Perhaps due to the drop in supply, or because hosts were keen to shift the burden of a higher risk of renting to their guests, the Average Daily Rate (ADR) charged by hosts also increased by approximately 5%. As a result of these higher prices and with fewer listings on the site, monthly reservation days fell by a significant 59%.

For professional Airbnb hosts (hosts who listed more than one property in a month), the number of listings fell by 60%, which is over 10% more than the overall group. Professional Airbnb hosts, who are typically a little savvier at monitoring changes in the market, raised their prices by 9%, almost double the overall group. Their number of their reservation days still fell by 63%, however, and their monthly revenue by 55%.

 

Suspending tourist licenses in Barcelona, Spain

Airbnb regulation in Barcelona

 

The cosmopolitan Catalan capital Barcelona has taken some relatively extreme steps in its push for Airbnb regulation. The council is keen to curb tourism in its most popular neighborhoods by slowing the growth of short-term rentals.

All holiday rentals in Barcelona require a tourist license. In the past, the council has taken to freezing the issuing of tourist licenses in an attempt to control tourism. They have also threatened to fine Airbnb for publishing unlicensed listings. It is unclear, however, whether Airbnb have taken any steps to remove illegal listings from their website. In fact, Airbnb said they would appeal the fines.

On July 2, 2015, Barcelona’s new mayor Ada Colau suspended all new tourist accommodation licenses until the city could reach a development plan. This suspension did not just apply to Airbnb and home-sharing accommodations, but also to hotels, hostels and the like. Eight months later, in March 2016, a new development plan was published and included a complicated set of licensing restrictions by city zones.

Before July 2015 the previous mayor of Barcelona had already suspended licenses for a subset of neighborhoods in the city that were particularly congested with tourists: L’Eixample, Vila de Gràcia, Poblenou, Camp d’en Grassot and Gràcia Nova, Poble Sec, Sant Gervasi-Galvany, Putget-Farró, Clot-Camp de l’Arpa and the nearby areas of Hospital de Sant Pau and Sants station.

In the graph below, these neighborhoods make up the Already Frozen series. All neighborhoods that experienced a license freeze starting on July 2, 2015 are represented by the Newly Frozen series in the same graph.

 

Number of Airbnb Listings by Already Frozen and Newly Frozen Properties in Barcelona, November 2014 – January 2017

Number of Airbnb Listings by Already Frozen and Newly Frozen Properties in Barcelona, November 2014-January 2017

 

The graph above paints a clear picture of how Airbnb regulation has affected Barcelona. Before Colau’s tourist license freeze in July 2015 came into effect, the trajectories for the number of listings in all neighborhoods in the city almost completely overlap – they are almost 100% correlated. Yet after July 2015, the number of listings in the Newly Frozen neighborhoods begins to look very different.

We can see in the graph below that the percentage difference between the number of listings in Already Frozen and Newly Frozen neighborhoods is significantly lower before July 2015, averaging at 20%. The percentage difference then shoots up in July 2015 before reaching a new equilibrium which averages around 27%.

 

Percentage Difference between Newly Frozen and Already Frozen November 2014 – January 2017

Percentage Difference between Newly Frozen and Already Frozen November 2014-January 2017

 

Perhaps surprisingly, regression results show that the freeze led to an increase in available days by 7% and a 6% increase in listings. One suggestion is that the freeze prevented hotels from obtaining licenses, which encouraged Airbnb hosts to try to capitalize on the shortage in supply. In fact, international hotel chains that did not have their management projects approved and licensed would have also been affected by the freeze. However, this increase in listings did not seem to translate into more reservation days. There was no concomitant increase in demand. In fact, the freeze led to a marginally significant decrease in revenue—7%—which remains a puzzle.

The story with professional Airbnb hosts is much the same as with the overall group. They increased their listings by 5%, and monthly reservation days fell about 4%, which contributed to a decrease in revenue of around 9%. Similarly, however, the puzzle persists as to why listings increased, but reservation days and revenue fell.

 

Strict regulations for short-term rentals in San Francisco, CA

Airbnb regulation in San Francisco

 

In October 2014, San Francisco passed a law that legalized short-term rentals, but only under a number of conditions. Properties had to be: offered by permanent residents, registered with the city, pay a hotel tax and carry $500,000 in liability insurance.

The law further stated that entire home rentals would be capped at 90 days per year, but owner-occupied rentals – Airbnb hosts that rent out a private or a shared room — would not be limited. We are going to look at how the 90 day cap on entire home rentals, which came into effect on February 1, 2015, has affected the city of San Francisco.

This issue is particularly pertinent as it comes on the heels of a measure to restrict all short-term rentals — entire or shared homes — to 60 days per year. Mayor Ed Lee ultimately vetoed this measure in December 2016. But the threat of Airbnb regulation in the form of additional caps remains.

The graph uses differences-in-differences methodology to analyse of the number of entire home listings and shared home listings. Prior to the February 2015 cap, the entire home trajectory and shared home trajectory are 99% correlated. Yet after February 2015, the two lines are still closely aligned, but visibly less so. In fact, they are 93% correlated after February 2015 — suggesting that the cap made a dent in the number of entire homes listings in San Francisco. Regression results show that number of listings altogether in San Francisco decreased by 5% [3].

 

Number of Airbnb Listings by Entire Home and Shared Home Listing Types in San Francisco, October 2014 – January 2017

Number of Airbnb Listings by Entire Home and Shared Home Listing Types in San Francisco, October 2014-January 2017

 

This 5% drop in the number of listings had additional effects. The ADR increased by 12%. As a result of the fall in supply and increase in price, there was also a 15% decrease in reservation days.

The story for professional hosts is more severe, whereby the 90 day cap seems to have limited the operations of the professional Airbnb hosts. With a 16% decline in listings — more than three times the overall group — and a 7% increase in price, they faced a 35% drop in reservation days and a 30% decrease in revenue.

 

Santa Monica, CA imposes fines on Airbnb ‘entire home’ hosts

Airbnb regulation in Santa Monica

 

On June 12, 2015, the beach-front haven Santa Monica enforced strict Airbnb regulation, banning the short-term rental of entire homes. The council also restricted home-sharing to hosts that obtain a business license and pay a 14% hotel tax. Hosts who violate the law in Santa Monica can face fines of up to $500.

By comparing entire home listings and shared home listings, we can determine how the new law impacted Santa Monica’s entire home listings. Just like in the other examples, we can assumed that entire home listings would have faced the same trajectory as shared home listings had the law not come into effect.

This is not a perfect comparison, however, because shared home listings faced its own regulations after the law make it a little less reliable as a counterfactual. Having to obtain a business license and paying a 14% hotel tax would certainly deter some Airbnb hosts. This means we can’t be absolutely sure whether the impact shown by the figures was caused by the ban on entire home listings, or because shared home hosts weren’t willing comply with new regulations.

The graph below shows that prior to June 12, 2015, entire home listings and shared home listings track closely, with a correlation of 94%.  After June 12, however, the two series move in opposite directions. Shared homes continue to rise and entire home listings fall sharply, only -39% correlated.

 

Number of Airbnb Listings by Entire Home and Shared Home Listing Types in Santa Monica, August 2014 – January 2017

Number of Airbnb Listings by Entire Home and Shared Home Listing Types in Santa Monica, August 2014-January 2017

 

Regression results show that the change in law also decreased the number of entire homes by 37% [4], and negatively impacted monthly revenue by 41%. Perhaps due to the shortage in supply of listings or the heightened risk of listing entire homes, hosts increased their ADRs by 15%. As a consequence of both decreased listings and higher prices, monthly reservation days fell by 51% in Santa Monica after the ban.

The story for professional Airbnb hosts is very similar. In fact, the pricing structure used by professionals did not deviate from the rest of the hosts as with other cities. Professional hosts’ ADR, for example, increased by 13% compared with 15% in the overall group. Their revenue fell by 43% and number of entire home listings fell by 40%, which is in the same ball park as the overall group.

 

So just how effective is Airbnb regulation?

The results seem to indicate that the impact on Airbnb is highest when hosts are targeted as the principal offenders. In the case of Berlin and Santa Monica where both cities fined short-term vacation rental hosts, listings dropped by 49% and 37% respectively. Over in Berlin, the phenomenal impact of the ban on entire home listings may be tied to the punitive $100,000 fine. But, even with a $500 fine, Santa Monica also brought about a significant drop in the number of Airbnb listings. For both cities, the drop in percentage of listings was also reflected in a sizeable drop in revenue.

Interestingly, in Barcelona, limiting the licensing of tourist accommodations presents somewhat of a puzzle. The freezing of tourist accommodation licenses for vacation rentals — including hotels, bed and breakfasts, hostels, and the like — actually increased the number of Airbnb listings in the city. One interpretation is that Airbnb and hotels are substitutes for each other. If the threat for a hotel violating the freeze is more than the threat for an Airbnb host violating the freeze, there might be a surge in listing on Airbnb. However, it does not explain why monthly revenues marginally fall.

For San Francisco, the 90 day cap on renting entire home listings led to a 5% decrease in listings. Compared with fining Airbnb hosts in Berlin and Santa Monica, this is a modest decrease. Furthermore, there was no significant decrease in revenue in San Francisco because prices adjusted upwards and compensated for the drop in listings and reservation days.

Airbnb regulation in Berlin and San Francisco had the biggest impact on professional Airbnb hosts. In Berlin, the fine led to a 60% decrease while the overall group posted a 49% decrease. Not surprisingly, professional hosts were able to adjust their price to compensate for the decrease in Airbnb listings and their monthly revenue fell by 55%, which is the same amount as the overall group. In comparison, in Santa Monica, where hosts are also subject to a fine, the regulations had almost the same effect on professionals and the overall group. The one factor that sets Berlin apart from Santa Monica is the sheer magnitude of the fine, which could certainly be a differentiating factor.

San Francisco’s 90 day cap on entire home listings also served to deter many professional hosts. They faced a 16% decrease in monthly listings compared with a 5% decrease in the overall group. The number of reservation days fell by 35%, which is double the overall group. This contributed to a 30% decline in monthly revenue.

It seems that cities that enforce regulation targeting individual hosts with fines, like Berlin and Santa Monica, have the biggest impact on the number of Airbnb listings in those cities. Airbnb professionals, on the other hand, are most affected by steep fines and caps. Some cities have attempted to target Airbnb itself. It seems inconclusive, however, whether this approach has a real impact on the number of Airbnb listings.

If you want to understand more about how we put this report together don’t hesitate to get in touch with hello@airdna.co.

 

Effects of Airbnb regulation

 

[1] The regression equation estimated is Y_et=α+β〖Treat〗_e+γ〖Post〗_t+δ_rDD (〖Treat〗_e × 〖Post〗_t )+ϵ_dt, where 〖Treat〗_e equals 1 if listing is for an entire home, 〖Post〗_t equals 1 if listing is after May 1, 2016. Standard errors are clustered by yearmonth.

[2] The regression equation estimated is Y_et=α+β〖Treat〗_e+γ〖Post〗_t+δ_rDD (〖Treat〗_e × 〖Post〗_t )+ϵ_dt, where 〖Treat〗_e equals 1 if the neighborhood is Newly Frozen, 〖Post〗_t equals 1 if listing is after July 1, 2015 and before March 31, 2016. Standard errors are clustered by yearmonth.

[3] The regression equation estimated is Y_et=α+β〖Treat〗_e+γ〖Post〗_t+δ_rDD (〖Treat〗_e × 〖Post〗_t )+ϵ_dt, where 〖Treat〗_e equals 1 if listing is for an entire home, 〖Post〗_t equals 1 if listing is after Feb 1, 2015. Standard errors are clustered by yearmonth.

[4] The regression equation estimated is Y_et=α+β〖Treat〗_e+γ〖Post〗_t+δ_rDD (〖Treat〗_e × 〖Post〗_t )+ϵ_dt, where 〖Treat〗_e equals 1 if listing is for an entire home, 〖Post〗_t equals 1 if listing is after Jun 1, 2015. Standard errors are clustered by yearmonth.