Airbnb has been under fire globally and in Australia from housing advocates who have accused the short-term letting platform of forcing up rents and limiting availability for people seeking long-term lets.
In its defence the platform has released a report finding non-hosted short-term rental accommodation (STRA) has “no consistent impact” on housing affordability, only accounting for 1% to 2% of dwellings in each state in June 2022 – a bit more than 100,000 properties in total. (A rental is “non-hosted” when the host is not present during the stay, often meaning the entire property is listed.)
But 1% to 2% of all dwellings is about 3% to 7% of rental properties, depending on the state, according to Guardian analysis. And the figure of 100,000 properties is close to the housing shortfall the government expects over the next couple of years as construction fails to keep up with demand.
So what does the data say about the effect of Airbnb on long-term rental availability?
How much housing stock is tied up in short-term rentals?
“In a good year we’d be lucky to produce 2% of the housing stock in new supply,” says Prof Nicole Gurran, chair of urbanism at the University of Sydney.
“And when you consider it against rental vacancy rates, which at the moment are under 2%, it’s quite a significant amount.” The vacancy rate hit a record low of 1.09% in January.
The share of dwellings that are rentals in Australia has been relatively stable on a national level over time. About 3m properties – 30.6% of occupied private dwellings – were occupied by renters in the 2021 census. This is about the same share as in 2016 (30.9%) and slightly higher than 2011 (29.6%).
Airbnb declined to share its data on short-term rentals by location, so we estimated the number using the Airbnb report and the dwellings in the 2021 census.
Our estimates produced figures roughly similar to those in a report released by the Real Estate Institute of Australia last year. The institute also found more than 80% of properties listed for short-term accommodation were for the entire dwelling – making them broadly comparable to those on the long-term rental market.
How many short-term rentals could be long-term?
A 2022 study found that most short-term rentals in Tasmania used to be longer-term lease, but not all Airbnbs would convert to housing or rental stock as many were previously holiday homes and could become so again.
The New South Wales government has estimated there are 100,000 dwellings not being used for long-term housing in the state, including 45,000 holiday homes and 33,000 registered non-hosted short-term rentals.
The share of entire dwellings listed for short-term rental in NSW was 1.8% of all dwellings in June 2022, according to the Airbnb report. This is almost 6% of rented dwellings in NSW in the 2021 census. The share is about 1.6% of all dwellings in Victoria – also almost 6% of rented dwellings.
But Gurran cautions that the share of housing that is on the short-term rental market is not the only factor affecting housing.
“Most housing scholars would read those figures as demonstrating a not insignificant proportion of housing stock. But that doesn’t mean, of course, it’s the only factor that is impacting on the housing market,” Gurran says.
“And nor does it mean that all of those dwellings would have been available as permanent rental stock [prior to platforms like Airbnb]. We can’t go that far. But it’s certainly not insignificant.”
Are the figures reliable?
Sean Brosnan from Urbis, one of the authors of the Airbnb report, said Airbnb stock could not be accurately calculated as a percentage of total rental stock.
The Airbnb report calculated the share of short-term rental properties using the ABS estimate of dwellings by local government area in June 2022. This dataset is based on the census and is updated regularly to reflect construction and demolition.
The most similar, comprehensive data on rental properties comes from the 2021 census, but it has several drawbacks, according to the report’s authors. There is a lag between when the census was completed and when the total number of short-term rentals was calculated. The census also only represents a point in time – census night in August 2021 (which was also affected by the pandemic and border closures). And it may not accurately capture all short-term rental properties as the census has to be completed by occupants, who may not have identified the property as a rental.
“In contrast, total dwelling stock is accurately recorded at an LGA level by the ABS and the number of dwellings only changes as a result of construction or demolition (whereas the number of rental dwellings is much more fluid and can change more quickly over time simply due to the whims of the owners),” Brosnan says.
“Ultimately, our analysis has not focused on non-hosted STRA as a proportion of total rental stock because there is no way to accurately estimate total rental stock (comprising both long term rentals and STRA) across Australia at a local government area (LGA) level.
“I cannot comment on whether considering non-hosted STRA as a percentage of total rental stock would impact the findings as we have not undertaken that analysis.”
