Airbnb, as with some other of its fellow peer-to-peer “disruptive” tech solutions, has come under fire from a few directions over the past years. Along with concerns that it could be causing needed housing to be taken off the market in areas where prices are already tight, and essentially allowing commerical hotel operators to operate unregulated and untaxed, it has also faced such strong accounts of racial discrimination that entrepreneurs saw a market niche for Airbnb alternatives for Black travelers.
In response, Airbnb has recently announced a set of anti-discrimination policies, including a “community commitment,” which basically involves having hosts promise not to make decisions based on race, ethnicity, national origin, or sexual orientation. As a statement of principles, it's a fairly clear and explicit one, and that's a good thing. Using your leadership role to set a tone for your organization does matter. But I think other proposals are likely to matter more, and Airbnb could go further.
One limit to the community commitment is that discrimination happens against hosts too, with Black hosts commanding lower prices, but Airbnb cannot make its travelers promise not to discriminate where they choose to stay (though it could give them an exhortation to be open-minded when they sign up), since every location is unique.
The authors of the study on hosts' experiences suggest downplaying host photos—which is in fact one of Airbnb's stated goals. This is good news, though just how much change it will make is unclear. I wonder how far it will be able to get with “downplaying” the photos in a world where not having a photo associated with a profile has become strongly associated with being a “spammer,” “troll,” or “cheater.” I don't think it will get rid of photos, but substanially downplaying them might reduce cases that are based on the more mild ends of implicit bias.
When it comes to discriminating against potential guests, Airbnb says it is increasing its “enforcement” in some unspecificed way, and is also pushing hard to increase the use of the “Instant Booking” option, where hosts do not have to pre-approve a guest. That latter move should have a practical impact, and is probably worth more than a million community commitments.
But presuming there will be substantial remaining listings that are not instant book (and that the percent of hosts who don't acccept instant booking who want to be discriminatory might be higher) there is something else Airbnb could do, if it chose to be seriously proactive: It could take a page from the fair housing playbook and do paired testing. This would involve sending requests from guests with otherwise similar profiles and needs, but different races or other categories of concern, to the same host (even for the same days, with the first “guest” canceling if they were accepted). If one was accepted and the other not, there could be warnings issued, and follow-up testing, with an eventual option to deactivate an account if there was a consistent pattern.
I don't doubt it's a massive project. It might well be too much for underfunded advocates without access to Airbnb's data to undertake, but I would like to suggest that if Airbnb wants to prove that it means what it professes, it could voluntarily institute paired testing itself, in response to complaints as well as at random, with transparent results. For a company with such sophisticated tech at its disposal, I'll bet that if it wanted to make paired testing work, it could. (Perhaps Airbnb could even tap into the expertise of some fair housing organizations for the non-tech design of the program and third-party oversight, and render itself a little less underfunded in the process.)
Of course having affordable housing taken off the market is likely to also have a racially disparate impact, so if Airbnb really wanted to think big about its equity impact they should also crack down on commercial hosts offering multiple listings and stop fighting against being held the same rules as other hotel operators. I won't hold my breath on those.
(Image: By smile_kerry, via flickr, CC BY-NC 2.0)
[Correction: An earlier version of this article repeated an incorrect statement from a footnote in this law journal article stating that the National Fair Housing Alliance does not consider paired testing feasible in this scenario. That is not NFHA's position. In fact, NFHA says large volume testing is practical. The law journal's footnote was a misinterpretation of a statement the NFHA had made about the feasibility of lodging hundreds of individual administrative complaints against people who post discriminatory advertisements via online rental or sale sites versus holding the sites liable for such illegal ads.]
About the Author
Tags
Comments
3 thoughts on “Airbnb, Test Your Hosts for Bias”
Leave a Reply
This site uses Akismet to reduce spam. Learn how your comment data is processed.
I think this is a thoughtful piece but, like nearly all pieces I see on this topic, it misses out on some points that are really key from a host perspective. I’ve been a host for 4+ years and I’ve hosted travelers of all races and from all around the world, and yet, I am not a hotel!! I don’t have a security staff; it’s just me there in my pajamas with my cat and a hook-and-eye latch on my bedroom door. So I care a lot about the personality of the person who comes through my door. Imagine if a stranger rang your doorbell and asked to stay the night. They are willing to pay, they just need a place to stay. What kind of questions would you ask them? Maybe, tell me about yourself. Who are you? What brings you to this place? Are you on vacation, business, between apartments? For me, guests who understand that I’m a person (and, specifically, a young woman living alone) and who volunteer information about themselves so we can be friends, at least for the duration of their stay—they get accepted. Lots of guests of all races understand that intuitively, and lots (of all races) don’t. But in my experience, African Americans are less likely to tell me about themselves.
I always start out by asking questions and encouraging that approach of treating each other as individuals, not like this is just a transaction. As Airbnb becomes more commercial and mainstream over the years, many guests need help making this mindset shift because they are accustomed to the hotel experience.
But if people come across as cagey, I’m not super stoked about them coming to stay in my home with me. That’s the person who’s going to be there when I’m yawning and just woke up and haven’t had my coffee. That’s the person hanging out at my place while I’m at work.
And I could spend some time justifying and qualifying—I am a “minority white person” in my neighborhood; here’s how many black guests I did host, etc. And I also fully agree that racial bias IS a problem in our society and I try to be very cognizant of it in myself. And it is, I have no doubt, affecting black guests and hosts.
But I just think this more sort of “cultural issue” of “tell me about yourself” needs to be part of the conversation. (A recent category on SNL’s “Black Jeopardy” segment was “I don’t know you!” — Not giving out personal information to strangers seems like a practice in the African American community and it might not be serving travelers well here.)
At the end of the day, as a host I don’t accept or deny categories of people, I welcome individuals (whom I have never met) into my home.
So, update. I wanted to go back and stress-test what I said above with the data evident in my messaging history with potential guests. I went through about half of my messaging history (~300 potential guests, plus their traveling companions, over 2 years out of the 4+ I’ve been hosting). I looked at the messages with any potential guest who was black and American (so many of my guests from around the world including the USA are non-white, but it seems like this is the group experiencing the worst discrimination).
I actually surprised myself. It was true that a couple people were cagey and got declined—but only ONE was in that 2-year period which means the other was earlier. I think those two individuals stand out in my memory disproportionately. (Maybe because I am overly sensitive of being implictly accused of racism. In fact the older situation I remember vividly because I was trying so hard NOT to be racist that I initially accepted the booking and then freaked out because not only was she cagey but I couldn’t shake the feeling this college-age girl who lived in my own city and wanted my whole place for just one night was planning to throw a huge party. So I eventually cancelled—my only cancellation ever.)
Out of the remaining dozen or so messages, they fall nearly evenly into three categories:
1) People whose desired dates were unavailable (because they wanted the whole apartment but the guest room was already booked, or vice-versa—one of the idiosyncrasies of the Airbnb system is that you need to create separate listings for these situations and the calendars don’t automatically update). These people did not get “declined” because I explained the situation to them and they didn’t attempt a booking.
2) People I hosted!
3) People who chose for their own reasons not to book with me. (One I asked why; she didn’t want to take care of the cat—a requirement for booking the whole-apartment listing. One of the remaining two was a filmmaker looking for a place to do a shoot so presumably he found a better site.)
I did notice that younger people (college-age, 20s) were more likely to be friendly and open than 30s+-ish. (You can’t tell age for sure.) Maybe growing up with social media impacts how much you tell strangers about yourself?
So, maybe if Airbnb were testing hosts for bias, I would do well. Or maybe I would do poorly. Or maybe as the Avenue Q song says, everyone’s a little bit racist.
But I still think that not getting to see someone’s face before they show up on your doorstep is very weird. And having all kinds of people opining on the situation who seem like they’ve never hosted a houseguest or their kid’s friend, much less a stranger from the interwebz, opining on what kind of factors you should take into account (or be permitted to take into account)… is still annoying.
And as you can imagine, I would NEVER use Instant Book unless, again, I started running an actual hotel with 24-hour staff. But maybe that’s just me.
——
(Not to beat a dead horse, but just as a fun thought exercise, imagine how many questions your ask your child if he wants to invite a friend for a sleepover:
“What’s your friend’s name?”
“Bobby.”
“Do Bobby’s parent’s say it’s okay?”
“Yes.”
“Are you going to stay up late watching movies?”
“No.”
“Does he have any allergies?”
“No. Also not ‘he’; Bobbi is a girl.”
Oh!
—and now imagine that Bobbi and Andrew are 16 instead of 8…. details matter. You care about who’s coming to your house and what they’re going to do there! —and yes I did once have a pair of Brooklyn high schoolers try to book my guest room for prom night… haha! ohhh dear. Yes, I will never ever use Instant Book!)
Deirdre,
I’m glad you checked your own data and your initial sense of what was going on to find that your first generalization didn’t actually hold. Props for taking that step. And I certainly appreciate the desire to be able to make judgment calls about who enters your home—it’s reasonable but also exactly where implicit bias can step in.
In any case, I figure that paired testers could easily address that initial concern by being equivalently forthcoming or not and and following up to find out reasons for refusals.