The combined state and local tax incentives that are being pitched to Amazon for their second headquarters are out—reported by both Fran Spielman of the Sun-Times and Greg Hinz of Crain's—and the numbers are, unsurprisingly, real big.

Start with $1.4 billion in tax credits via Illinois's Economic Development for a Growing Economy program, about 28 times the biggest EDGE credit doled out in recent years (to the electric-vehicle maker Rivian). The next biggest chunk is "site-specific infrastructure improvements" from some combination of the Illinois & Chicago Departments of Transportation, CTA, and other agencies, for $450 million, the equivalent of how much it cost to rebuild the Wilson and 95th CTA stops. Then "$250 million worth of investments in education, workforce development, and 'Neighborhood Opportunity Funds'."

Those are the biggest elements of a package that adds up to $2.2 billion dollars, or about $45,000 per job that Amazon promises to create.

The interesting thing about the package, at least from the outline that's been reported on, is that not all of that would go directly to Amazon's bottom line. The infrastructure improvements would provide a more general benefit (though they could, in theory, come at the expense of those elsewhere); same with the education and development credits.

So let's just take a look at the EDGE credit: $1.4 billion for 50,000 jobs, assuming that the full credit is dependent on that figure. If so, it's above average for how much the state has promised companies in the past couple years, but far from the highest amount of money per job.

The state's Department of Commerce and Economic Opportunity lists all the EDGE credits given out from January 2015 through April this year. Excluding the instances when some or all credits were given to retain jobs, there were 108 credits given to create 14,850 jobs at a cost of $286,629,352.

That's an average of $19,301 per job (the median is $21,000). A $1.4 billion credit for 50,000 jobs is $28,000 per, so it's above average—but 32 EDGE credits have been given at a higher price per job. The Rivian deal, for instance, is for 1,000 jobs and a tax credit of over $49 million for 1,000 jobs—so almost $50,000 per job for a company with no track record.

Amazon's EDGE credit is closer to that offered to Glassdoor, an online job-search site: $6.3 million for 240 jobs, or about $27,500 per job, which makes sense—EDGE is tailored to the salaries created by the jobs. According to a 2016 filing, the company's first batch of 104 employees was hired at an average salary of $70,000. Doing some admittedly primitive back-of-the-envelope calculations, 240 people at that salary paying about $3,000 in state income tax every year would be $612,000 a year, so it would—assuming a whole lot, like whether the presence of thousands of more workers would not substantially increase demand on public services that a company without tax credits would be paying toward—break even after 11 years. 

Of course, part of the reason we assume a whole lot about tax incentives for job creation is that their effects are not terribly well understood, as seen in this analysis that cites a local example:

Using a wide range of assumptions and methodological approaches, research has thus led to mixed results. A few studies have reported a positive relationship between tax incentives and job growth, particularly when companies are deciding where to locate within a specific region. Most, however, have found that tax incentives are a zero sum game for the states in aggregate and have little to no effect on aggregate job creation or other metrics of economic development.

A 2008 review of research on business location decisions, for example, reported that most research shows incentives not to be worth the money spent by state and local governments. One example describes Chicago’s effort to attract the headquarters of Boeing, the world’s leading aerospace company and the largest manufacturer of commercial jetliners and military aircraft. Included in the $63 million incentive package were $41 million in state tax credits and local property tax abatements…. There was, in fact, no evidence of local job creation associated with the Boeing relocation to Midway and the multiplier effect—the creation of additional jobs and income as spending from the Boeing headquarters rippled through the economy—was not significant.

(Boeing got $63 million for 500 jobs, or about $120,000 per job, in 2001 dollars.)

One reason business tax credits are not terribly well understood is that a lot of states aren't good at tracking them. Take, say… Illinois. Two years ago the Tribune published a brutal analysis of the state's track record with the EDGE credits, finding that the state couldn't actually say how many jobs had been created and that 79 recipients had actually cut over 23,000 jobs.

As an example, Mitsubishi got tax incentives back in 1988 to open a joint plant with Chrysler in Normal, Illinois. It got an additional EDGE credit to retain its 1,250 jobs—just retain existing jobs—in 2011. It closed in 2015. Now the aforementioned electric-car maker Rivian is the recipient of the largest EDGE credit of recent years… which will build cars in the old Mitsubishi plant. (After being briefly mothballed, a seemingly improved EDGE program will be used to target Amazon.)

So while the raw numbers are eye-popping, they translate to something fairly typical. And they certainly translate to less than New Jersey's offer—seven billion dollars—much less what it took for Wisconsin to land Foxconn. At $3 billion for fewer jobs at a lower average salary, they're looking at $15,000 to $19,000 in tax incentives per job… per year.