Today Airbnb released what it calls an “economic impact report” for the Seattle region, detailing the economic benefits it believes its business has delivered to Seattle. I went through it with a fine-tooth comb; there are interesting useful statistics, there are some misleading ones, and there are some conspicuously missing ones.
Airbnb is doing this because yesterday Council member Tim Burgess announced that he and his committee would be looking at whether Airbnb needs to be locally regulated to mitigate any adverse effects it is having on the local long-term housing market. Burgess’s announcement was no surprise, as he had hinted on this direction before, and Airbnb I’m sure has been working on this report for a while (they have published on their web site similar reports for other regions).
The slick 29-page report has all the things one might expect from a document produced by a corporation’s marketing group: lots of colorful graphs, happy quotes from satisfied customers, and a handful of numbers in VERY LARGE FONT SIZE that you’re supposed to pay attention to. The message they want you to take away is clear: Airbnb brings a lot of people, and money, into Seattle, and we should be oh-so-grateful to them and not spend a lot of time considering negative repercussions.
Most of the statistics cover an aggregate 12-month period from August 2014 to July 2015. Why that period? They don’t say. When reading documents like this, though, one should always assume that ranges (date, geographic, etc.) are always chosen for a reason. Airbnb has more recent data than that; if they chose those dates, it’s because it aligns with the story they want to tell. But putting that aside, let’s look at the numbers they present.
First, let’s start with some hard facts: in that period there were 2900 hosts in Seattle. They hosted 151,000 guests, earning $30 million in income.
Now some claims: based upon their surveys, they figure those guests spent an additional $108 million at Seattle businesses during their stays. They also claim that a “typical” (whatever that is) Airbnb host earns $8000 per year, renting out accommodations 79 nights per year.
But the big claim (in the biggest font) is that Airbnb’s total economic impact is $178 million, and that’s the headline in all the press coverage.
Ignore that number. It’s meaningless. It’s calculated by multiplying the total average spend of an Airbnb guest by the total number of guests. That would be meaningful if all 151,000 of those people would not have made their trip to Seattle if Airbnb didn’t exist — a claim that is surely not true. Airbnb helpfully gives us a comparison of an Airbnb guest and a “typical hotel guest,” and informs us that Airbnb guests stay longer (3.6 days vs. 3.0 days) and spend more per trip($945/guest vs. $687/guest).
If we take all of that at face value, then Airbnb guests are compelled to spend $258 more per trip. If we assume the opposite extreme — that all 151,000 were to come to Seattle and stay elsewhere (also clearly not true) — then that would mean Airbnb brought an additional $38,958,000 of income to Seattle. The truth about how many of them would make their trip anyway is somewhere in the middle of that range, and a big unknown.
But wait: there’s another assumption built in here: that it’s fair to compare a typical Airbnb guest with a typical hotel guest. According to Airbnb’s helpful pie chart, “82 percent of Airbnb guests visited for vacation and leisure or to see friends and family.”
Given the huge amount of convention and business travel in Seattle, that can’t possibly be true for typical Seattle hotel guests. To really understand how much of a difference Airbnb is making, we would need to see what the length of stay and spending rate is for a group that matched the demographic of the Airbnb guests. We would also need to understand the seasonality of the business: if the bulk of the stays are during the peak tourist season, when hotels are already full, then they are indeed increasing Seattle’s capacity in a way that brings additional dollars to the area — which would be great.
This is a great point to pause for a moment and make a general critique of the report: it relies too heavily on typical/average, when in fact for many of the statistics they report we really want to see what the range is, and how people spread across the range. What is the range of spending by an Airbnb guest? Is it a “normal curve” distribution clumped around a mean? Or is it a “power law” distribution where the vast majority spend little but a small number spend a lot? The same for host income: is everyone making nearly the same, or is there a wide distribution with some people raking it in and others making little? And most pertinent to Council member Burgess’s question: what’s the range of “days per year” that hosts are renting out their space — are many renting for enough days of the year that it legitimately competes with long-term housing rental?
Airbnb claims that rentals spread across 80 neighborhoods in Seattle, and they chose to highlight five: Capitol Hill, Columbia City, West Seattle, Ballard, and Queen Anne. Those five represent 826 of the 2900 hosts, $8,390,100 of the $30,000,000 of host income, and $44,100 of the 151,000 guests. That’s a large chunk of the volume, and the income figures are pretty close to the mean (which I calculated from the total Seattle figures they listed). But that suggests (as do their maps) that there is a strong concentration of listings in a small number of neighborhoods. Interestingly, they don’t list the ones with the greatest concentrations: downtown and South Lake Union; one might suspect that the rents are much higher there. And the corollary: that in many other neighborhoods — especially the economically disadvantaged ones — income is much lower. Again, seeing the ranges would be very enlightening, but Airbnb doesn’t want to share that. In a similar vein, Airbnb claims that “42% of Airbnb’s guests’ daytime spending is spent in the neighborhood in which they stay,” which they calculate to $526 per guest per trip in that neighborhood. But you would have to be foolish to believe that number is consistent across all Seattle neighborhoods: some are benefitting greatly and are pulling up the mean for the others.
I have no doubt that Airbnb is having an economic impact, and is providing a valuable second source of income for many citizens of Seattle. But their report does little to help us understand the size of the impact.
What information could Airbnb provide that would really help us to understand their economic impact?
First, we’d want to look at the ranges that characterize hosting:
- a histogram of the range of nights per year that hosts rent out their spaces;
- a histogram of the annual income of a host;
- rental-nights broken out by month (so we know how many are during peak season);
- these ranges broken out by neighborhood.
Second, we’d want to get a better view on the habits of hotel guests visiting Seattle on vacation or getting together with family and friends, to know how long they stay and how much they spend locally.
But none of that really answers what Burgess is going after: is Airbnb hurting long-term housing availability? To answer that, we’d need to do some basic math: armed with an understanding of the “going rate” for a given rental property and the average number of days per year that rental properties are rented out in that neighborhood (which is less than 365 due to vacancy between renters), plus the “going rate” for the same property through Airbnb (which will be higher), we can calculate how many days per year a landlord would have to rent a property through Airbnb to earn what they would through a long-term rental — and how likely it is that they could do that. The lower that number the more attractive it is to rent through Airbnb rather than long-term — and if the actual number of rental days is much higher, than it’s even more attractive because you’re making a lot more money.
So then we’d try to calculate how many people are actually doing that: how many hosts are making more money from Airbnb than they would if they were renting their property long-term. Because there are other factors too, including a landlord/host’s tolerance for having strangers in their house, privacy issues, the time and expense of cleanup, potential theft and bad behavior, and the general hassle of having people arriving and leaving. Even if there is the potential to make more money, many people may be intentionally aiming lower to avoid these issues.
I have little confidence that Airbnb will release any of that data (they have competitors, after all, and much of that is valuable confidential information) but in its absence, we’re unlikely to be able to truly understand the impact that Airbnb has on Seattle. And their economic impact report adds little of substance to the conversation.