The picks adhered so closely to their criteria an algorithm could have done a decent job back in the fall and saved hundreds of towns, counties and metros a ton of work. But there were a few glaring omissions.
To weed out the winners and the losers, let’s do a quick experiment: Consider Amazon’s criteria, consider the actual attributes of the cities that were selected, and check off where the two intersect. It’s classic retrofitting, but now that we have the benefit of hindsight, why not make the most of it?
We’re using data from the Census Bureau’s 2012-2016 American Community Survey, because Amazon’s proposal makes it clear it’s thinking in decades, and so we’re looking for the longest-term, most robust data around.
Unfortunately, it also means we have to leave out Canada. Our 20 finalists will here be just 19. Sorry, Toronto!
For consistency and comparability, we based all measurements on Metropolitan Statistical Areas, even when bids came from smaller jurisdictions within them. We had to split the two metros that are home to multiple finalists. Greater New York is divided up between New York and Newark, while the D.C. metro’s cities and counties are split between Montgomery County, Md., the District itself, and Northern Virginia. See the graphic for full details on which county ended up where.
Before we even start our checklist, we can eliminate any metro area with fewer than 1 million people, as well as Seattle, where Amazon already has a headquarters. (Amazon's founder and chief executive, Jeffrey P. Bezos, is the owner of The Washington Post).
Once we read past all the hints at the sort of corporate perks Amazon might welcome from its new host city, one of the clearest criteria in the document is for “locations with the potential to attract and retain strong technical talent.”
It’s easy to measure that potential. Look at the cities which have already pulled that feat out, and compare them to the top 19 in the category. The majority of Amazon’s 19 finalists are also in the top 19 cities with the largest population of young people (ages 25-39) who have engineering or science degrees.
But that leaves nine Amazon finalists that don't have that particular talent pool. Why were they still appealing to the online retailer?
For that, we look to another Amazon requirement: public transit. Four of the cities that weren’t loaded with technical talent are in the top 19 for the highest use of public transit for commuting to work. This includes Pittsburgh, as well as three areas within commuting distance of some of the biggest American cities: Newark and the D.C. suburbs in Maryland and Virginia.
But to explain the inclusion of Austin; Columbus, Ohio; Nashville; Indianapolis and Raleigh, N.C., on Amazon's shortlist, we have to look at another big item on the company's wish list: the business environment.
That’s a little harder to quantify, but it turns out that an active (and probably younger) labor force is an excellent indicator. All the remaining cities are near the top of the list when it comes to labor-force participation, or the percentage of the adult population that is either working or looking for work.
Armed with quick and dirty gauge of what Amazon was looking for, we can now puzzle out which cities got hosed. What other metro areas fit the models, either of a well-educated metropolis with relatively strong transit infrastructure (Boston, Atlanta, and many other top cities on the list), or a rust-belt city with a strong labor market (Columbus, Indianapolis)?
The clear loser is Minneapolis. The Minneapolis-St. Paul metro area performed well in every measure and, according to the Star Tribune, offered incentives of between $3 million to $5 million, yet lost out while many others with weaker resumes moved on.
Houston missed its bid to create an “Innovation Corridor” anchored by Amazon, even as two other Texas cities made it in. Houston ranked well in education and transit, and it missed the top-ranking trifecta by just one spot on labor-force participation.
Baltimore also had a rough go. It met two of the criteria (and was top half in the other, labor-force participation) and submitted a bid based around a plan to redevelop South Baltimore’s Port Covingtonarea, but lost out to no less than four of its neighbors — Washington, Northern Virginia and Montgomery County, Md., as well as Philadelphia
The other cities that met two criteria yet didn’t advance are perhaps easier to explain. All three — Portland, San Francisco and San Jose — are on the West Coast which probably isn’t attractive to a Seattle retailer looking to expand its American footprint.
San Francisco and San Jose are already home to the most intense marketplace for tech talent in the country and may not have the room or stomach for the kind of growth Amazon will require.