Left and Right Both Hate It -- a Good Sign for the Amazon HQ Deal

posted by Michael Lewis on November 19, 2018 - 12:00am

FMI’s commentary on the use of state & local incentives in the Amazon headquarters deals.

Amazon HQ Flawed But Still Good Deals for Winning States

When Socialist flavor-of-the-month Representative-elect Alexandra Ocasio-Cortez and conservative Fox host Sean Hannity both denounce a policy as an outrage, it must be truly outrageous. Or they must both be very confused. In the case of the incentives given Amazon for its new headquarters, it is the latter.

There are valid issues with the deals, but not the ones being criticized by the left and the right.

Critics are grossly exaggerating the costs of the deals to taxpayers, both in terms of the direct tax incentives and negative impacts posed by new employees to the communities. They are also greatly overstating the windfall to Amazon and its multi-billionaire owner.  

Amazon HQ Incentives
Amazon Jobs 25K 25K 5K
Incentives - Direct $1.2B $0.55B $0.065B
Incentives - Others $1,.6B $1.4B $0.036B
$/Job Stated $48,000 $22,000 $13,000
$/Job Actual $109,000 $76,000 $20,200





The taxpayers of these states are not being ripped off; they will benefit nicely. There is no shortage of public projects that are unmitigated disasters for taxpayers -- witness every sports stadium ever -- projects that politicians undertake to pay off favored constituents and, sometimes, to take a cut for themselves. The Amazon HQs are not even close to such boondoggles.

While critics focus on the billion-dollar price tags (over ten years!) of the incentives, they downplay the substantial taxes that will be paid by Amazon and its 25,000 new employees in both the New York and Virginia headquarters, plus 5,000 at a new hub in Tennessee. The state credits will not be paid until the new employees are on the job.

Amazon will be paying property taxes on the new facilities plus sales taxes on all items purchased for them. The three states will increase their collection of corporate taxes from Amazon as more activity is shifted there.

Most important, the bulk of new employees will be living in these states (or they will be drawn from other employers in the state and replaced there with new residents). Those employees will pay income, sales and property taxes, not to mention minor levies, ranging from cell phone and cable TV taxes to real estate title transfer fees to vehicle stickers to taxi and ride-share charges. All will add to state & local government coffers.

The incentives end in ten years but, presumably, the facilities and the associated taxes will continue indefinitely.

The incentives are proportional to the expected tax windfalls. Tennessee offered only a fraction of the per-job incentives of New York and Virginia. Rather than shrewd bargaining, this reflects the fact that, because it has no state income tax, Tennessee had a much lower prospective tax collection to rebate.

A small share of the NY incentives and the bulk of the Virginia package do involve actual cash outlays for infrastructure improvements on the project sites -- road extensions and such. These are also predicated on Amazon achieving certain benchmarks to fulfilling its jobs pledges. While these may not be the best public investments, they do improve the sites. Even without Amazon, the projects add value and encourage future development. They would generate some new tax revenue over time.


Critics also fault policymakers for not accounting for the costs that new employees will bring to those communities, but it is these critics who are failing to do the math properly. They cite new public sector outlays that will be needed, such as $18,000 for every Amazon employee’s child enrolled in NY public schools. (These complaints come from the right; liberals seem congenitally unable to criticize the cost of public schools in any context).

In fact, that hefty figure is the average cost per student. The relevant factor is the marginal cost for each additional student. Since the schools, teachers, administrators et al are already there, that marginal cost is essentially zero, provided one more desk can fit in a classroom. Even if thousands of children of new employees end up in public schools, the cost of a adding a relatively few more teachers is modest compared to the property and other taxes that will be generated by those new families.

Critics on both the right and left do complain about congestion from the new Amazon HQs. This is a more valid but still minor problem. Like the proverbial restaurant that is so crowded no one goes there, congestion is a secondary factor. By definition, it cannot overwhelm the primary factor, burgeoning Amazon-led growth.


Critics of the Amazon deals ask why the states should be throwing money at Amazon owner Jeff Bezos, one of the richest men on earth. However, Bezos’ true takeaway is only a fraction of the incentives. Undeniably, the NY and VA locations carry far higher costs than the alternatives. The incentives moderate those costs in part but not in total. Without the incentives, Amazon would have gone elsewhere. In world of perfect information, the dozens of competing cities would have offered effectively identical packages, including external factors, and Amazon would have been indifferent to the choice.

Since information is never perfect, New York and Virginia may have offered more than necessary, but probably not dramatically so.

New York and Virginia are taking risks on the deals. Ten years is a long time; Amazon could well be much smaller in a decade, about the time it took Borders bookstore to go from category-killer to bankruptcy. Since the incentives are linked to actual employment, both states are protected in large part. And Amazon and its HQs could end up being even bigger. New York is calculating/hoping that Amazon eventually doubles its Long Island City footprint to eight million square feet and 50,000 workers.

Amazon is also taking some sizeable risks in making these deals. Notably, the incentives are delineated in dollars per job, not percentage of taxes paid, so increases in state and/or local taxes could make these HQs much less desirable.(This no doubt ruled out fiscal basket-case New Jersey, which offered far bigger incentives for a site five miles from Long Island City).

Ultimately, how good the deal is for all sides depends on the next ten years. Presumptively, however, an agreement entered into voluntarily by all parties will be beneficial to all parties.


By FMI’s accounting, the Amazon deal is not a bad one for New York or Virginia (or Tennessee) taxpayers. The question for voters and for future policymakers is, was it the best deal?

New York will “spend” $109,000 for each added Amazon job over the next decade, a job that will generate even greater tax revenue over that period and beyond. It seems virtually certain, however, that there are other companies, from small start-ups to established performers, that could generate even greater jobs growth if offered a similar incentive. They lack only Amazon’s lobbyists and PR operation.

For that matter, with some creative financing, New York and the other states could channel ten-years of tax incentive payments into upfront “micro-loans” to 25,000 prospective entrepreneurs (or not-so-micro-loans to 10,000). This would create jobs for those people now with the potential to create more jobs.

Some, perhaps many, of those companies or entrepreneurs would fail, but some would succeed. A few even spectacularly. On average, they could meet or exceed Amazon’s likely performance (and allow for a greater geographic distribution of benefits). Transaction costs would multiply along with the number of beneficiaries.

This alternative would no doubt cheer liberal critics. The problem, of course, is that government is not good at picking winners (especially excluding those who are never weaned from public support or sweetheart contracts). Amazon is surely not the worst bet governments could make, but it may not be the best.

Accordingly, the free-market prescription (and Free Market Inc.’s position) is that the best course would be to make government as efficient as possible with the lowest necessary taxes and regulations for everyone. States would compete based on general preferences, not specific giveaways.

Then let the winners, well, win.