Amazon is looking for a city for its second headquarters (HQ2) and it recently released a list of 20 finalists. Many of the expected cities made the list—New York, Boston, Washington D.C., and Chicago—along with a few dark horses like Columbus, Indianapolis, and Newark.

Previously I argued that Amazon should consider Tampa-St. Petersburg, but since it didn’t make the list I’m going to examine the case for a few of the underdog cities that did: Columbus, Indianapolis, Pittsburgh, and Miami. Miami is the only finalist from my adopted home state of Florida, and Midwestern Columbus (where I once lived), Indianapolis, and Pittsburgh make for interesting dark horse contenders from so-called fly-over country.

Each finalist met Amazon’s basic criteria of a population over one million and an international airport. Amazon also prefers access to mass transit (e.g. subway), but as I wrote previously, mass transit is overrated—especially the dilapidated systems in New York, D.C., and Boston. And since Columbus, Indianapolis, Pittsburgh, and Miami lack robust mass transit it won’t be a factor in this analysis.

The other primary criterion desired by Amazon is a high-skill labor pool. Amazon’s headquarters will employ executives, managers, programmers, analysts and other high-skill occupations that typically require a college degree. Let’s start by looking at the percentage and number of adults age 25 and over with a bachelor’s or advanced degree in each metro area.

The percentage of adults is measured by the light blue columns and uses the left axis while the number is measured by the dark blue columns and uses the right axis. The first thing that jumps out is how much bigger Miami is than the others. Miami has over 1.2 million adults with a bachelor’s degree or higher compared to around 400,000 in Columbus and Indianapolis and 500,000 in Pittsburgh.

That said, Columbus’s workforce is relatively more educated than the other three, with about 35% of adults having a bachelor’s degree or higher compared to 32% in Indianapolis and Pittsburgh and 30% in Miami. Still, the differences in relative workforce educational attainment are small and may not be enough to sway Amazon one way or the other.

Another way to get at the skill level of a city’s workforce is to look at the number of people working in high-skill industries. The figure below shows the number of employees and the percentage of all employees in the finance and insurance; information; and professional, scientific, and technical services industries. Amazon will need to hire some people away from other companies and these data give us some insight into the labor pool it will be drawing from.Columbus has the largest percentage of employees in these high-skill industries, but again the differences are relatively small. Miami’s size advantage is once again apparent as well: It has more than double the number of employees in these industries than the three Midwestern cities. If overall workforce size is more of a factor than relative skill, Miami is clearly a better choice.

I imagine Amazon is a forward-looking company so they probably care about population growth in addition to the size of the current population since growth is evidence that an area is attractive to people and firms. Each of these MSAs is growing, but recently Miami and Columbus have grown faster. From 2013 to 2016, Miami’s adult population 25 and over grew 6%, Columbus’s grew 5%, and Indianapolis’s grew 3.7%. Meanwhile, Pittsburgh’s only grew 0.9%.

Miami’s growth is unsurprising considering its natural amenities and the general ongoing shift of the country’s population southward and westward. Columbus’s growth on the other hand, and to a lesser extent Indianapolis’s, should be an encouraging sign for its residents and local officials in light of the experiences of other Midwestern and Northeastern metro areas, regardless of what Amazon does.

The last measure of the skill of an area’s workforce we are going to look at is earnings and wages. Wages reflect worker productivity, so differences in wages across cities can tell us something about the productivity of each city’s workforce. As shown in the figure below, median individual earnings (yellow bars) are highest in Pittsburgh and lowest in Miami.

On the other hand, when it comes to average wages in the high-skill industries of finance and insurance; information; and professional, scientific, and technical services (gray bars) Miami has the advantage. So in the industries Amazon is most likely to draw its workers from, there’s evidence that Miami’s labor force has a slight productivity advantage over the Midwestern cities.

Based on the data here I have to give the edge to Miami. The relative skill levels of each city are so similar that Miami’s size advantage is hard to overcome. Miami’s geography and climate are also pluses, as is its location in Florida, which is one of the most economically-free and fiscally sound states in the United States.

As for the Midwest’s chances, Columbus and Pittsburgh have an edge over Indianapolis. Both are larger than Indianapolis, and Columbus is relatively more skilled while Pittsburgh has the highest productivity (wages) of the three. Pittsburgh’s slow growth, however, is a strike against it.

However, we don’t know what incentives each city is throwing at Amazon. Given the similarities between the Midwestern cities examined here, tax goodies may be the difference, even though the academic research fails to find significant benefits for cities overall due to such handouts. Cities would be better off in the long run if they created a favorable economic environment for all businesses, not just a lucky few, but unfortunately the siren song of an Amazon is hard for local politicians to resist.

Even with incentives, these cities are longshots. Boston, Austin, and Atlanta are the early favorites according to at least one oddsmaker. Still, the fact that Columbus, Indianapolis, Pittsburgh, and Miami made the top 20 over other cities such as Portland, Houston, and New Orleans is a good sign for them going forward. They may not get Amazon, but their futures are bright.