When Amazon announced last year that it wanted to spend more than $5 billion to build a second headquarters with as many as 50,000 employees, it sparked a flurry of studies as to which metro areas had a chance to win the business that would transform most economies.

Now that the Seattle-based company has released its unranked list of 20 locations that will move on to the next round of competition, it is possible to take a look at how well the analysts did with their predictions, which were primarily based on criteria that Amazon laid out, starting with metro areas with a population of at least one million. Other criteria included business environment, workforce, costs, quality of life, and transportation.

The Inquirer and Daily News reviewed three studies by the Anderson Economic Group, the Brookings Institution, and Moody's. Differences in approaches affected the outcomes, though Philadelphia, Atlanta, Boston, and New York made all three lists.

Anderson, which is based in East Lansing, Mich., for example, ranked 35 metro areas that have at least a million people and airports that fulfilled Amazon's desire for both international flights and direct flights to Seattle, New York, the San Francisco/Bay Area, and Washington.

"Because of the current status of their airports, Columbus and Pittsburgh were not included among our 35 cities, but they were included in the finalists," said Jason Horwitz, Anderson's director of public policy.

Fifteen of Amazon's 20 locations ranked among the top 20 for Anderson, which based its analysis on metro areas. In sixth place in Anderson's ranking, Washington's metropolitan statistical area includes two of the spots on Amazon's list, Montgomery County, Md., and Northern Virginia.

To Mark Zandi, chief economist at Moody's Analytics, the fact that Amazon included Washington and two surrounding areas on its list means that region must be the front-runner.

"D.C. is a great choice for them. I wouldn't be surprised if they picked D.C., even before this list," Zandi said.

Moody's picked what it saw as the top 10 metro areas for Amazon's H2Q. Its list included three places that did not make Amazon's first cut. They were Rochester, N.Y.; Portland, Ore.; and Salt Lake City.

Zandi said he was surprised to see Indianapolis and Columbus, Ohio, on Amazon's list. "They are smaller. They don't have international airports. They are not very globally oriented," he said.

Another surprise for Zandi was the inclusion of Los Angeles, which did not make Moody's list. "L.A. is a reasonable choice, but I just don't think they want to be on the West Coast."

At Brookings, Joseph Parilla did a straight statistical analysis of what areas met the criteria Amazon mentioned in its request for proposals. That meant he did not take into account the widespread belief that the second headquarters will not be on the West Coast. Because of that, his top-20 list included San Francisco, San Jose, Los Angeles, San Diego, and even Seattle.

Parilla's analysis favored places with large labor pools, leading him to miss some smaller Midwestern areas.

"Places like Indianapolis, Nashville, and Columbus, partly because of scale, partly because of transportation assets in those communities, aren't as robust as in some of the other finalists," Parilla said.

On the other hand, those cities have a "civic collaboration ethos that I think probably was appealing to Amazon," Parilla said.