As the COVID-19 virus fueled a rise in anti-Asian sentiment in 2020, it affected economic activity on Airbnb, the world’s largest short-term rental marketplace.
That’s according to a working paper published by a Harvard researcher this month, which analyzed reviews and found fewer people booked homes with Asian-American hosts in 2020, compared to the year prior and to hosts of other ethnic backgrounds.
The findings, based on data from New York City, also suggest that Airbnb’s design choices made it “easier for users to discriminate” because the platform displays hosts’ names.
“If you give people information about race without being thoughtful about what it’s going to do, you are in some ways setting it up to facilitate discrimination,” said Michael Luca, co-author of the study and an associate professor at Harvard Business School.
Airbnb said in a statement that it condemns “all forms of discrimination” and is “committed to fighting it proactively.” The firm is “reviewing this research seriously to understand its basis” and plans to contact the researchers for further information.
The study is based on data from Inside Airbnb, a website that tracks Airbnb listings, from 2019 and 2020. The findings show hosts with “distinctively Asian names” experienced a 12 percent decline in guests in 2020 relative to hosts with “white-sounding names” during the same period. The drop became notable in the spring of 2020 and remained that way for the rest of the year.
Luca said that since Airbnb has a high review rate, the study was able to use reviews as a proxy for bookings. To determine the ethnicity of hosts, the researchers used a publicly available algorithm that can identify “how likely a name is to be one ethnicity or another.”
Airbnb said that “measuring race based on name does not paint a full or accurate picture.”
“Many hosts who would identify as Asian may not have ‘distinctively Asian names’ and therefore would not be included in the analysis,” the company said.
Airbnb also said analyzing just over 900 hosts in New York City is an “extremely small sample size and inappropriate to extrapolate to the entirety of a global platform with over 4 million hosts.”
Airbnb, though, is aware that its design choices could enable discrimination. The company has previously faced criticism over its booking processes, because of features that display the names and profile photos of hosts and guests.
“Discrimination is based on perception — and on Airbnb, people perceive race from things like first names and profile photos,” the company wrote in a press release announcing the Project Lighthouse initiative in 2020, which aims to study and overcome discrimination on the site.
Luca said he accounted for several limitations when analyzing the Airbnb data to show anti-Asian sentiment led to discrimination on the rental site.
For instance, the researchers only looked at data for hosts that allowed “instant booking,” to eliminate the possibility that Asian Americans accepted fewer guests at the start of the pandemic. They also found that Asian-American hosts did not increase the price of their homes more than hosts of other ethnicities.
Airbnb has taken several steps over the years to address discrimination concerns. In 2018, the company said it would no longer display guest profile photos before a host accepts a booking request. The names and photos of hosts, however, are still visible to guests. Several other short-term rental firms, including VRBO (which Expedia owns), also show users information about hosts, including their names.
Luca said he understands guests might need to know the name of a host to check in, but he doesn’t think that information is necessary until a transaction is completed. Airbnb might have “set out thinking that this was a tool to build trust,” he said.
Communities being scapegoated in a time of crisis is nothing new. But Luca, whose research focuses on the design of online platforms, said companies like Airbnb play a role in letting these sentiments affect activity on their marketplaces.
“The question to reflect on,” he said, “is how do you design a platform that’s not going to let whatever bias is occurring in society creep into the platform?”