When Governments Compete for Business, Do We All Win?

Hundreds of governments responded to Amazon’s search for a location for its second corporate headquarters (HQ2). The process exemplifies interjurisdictional competition in the U.S. par excellence. But not all interjurisdictional tax and subsidy competition is the same.

On the one hand, there is Tiebout-like competition among states (and cities) based on overall generally applicable tax rates. People and companies look at mixes of tax levels, corresponding levels (or not) of government services, and non-governmental amenities. Today, many Americans seem to like big cities, and like to live near the coasts. For now, given these preferences and the corresponding level of government services, those states seem able to extract higher taxes from their populations. Nonetheless, high taxes, particularly in California and in the northeast, do seem to be pushing some companies and some individuals away from those areas.

The competition for companies doing what Amazon is doing (and any number of sports teams) partakes of competition of a different sort. The inducements offered to Amazon include not only generally applicable tax rates, incentives and amenities, but firm-specific inducements made available only to Amazon. The inducement packages are not available to immobile capital within these jurisdictions, nor available to other mobile capital, unless they cut their own agreements.

A few things to note. First, Amazon (and other business who can establish interjurisdictional competitions) set up a Prisoners’ Dilemma (PD) among the jurisdictions. That means the governments don’t like having to compete, but they also can’t help themselves. Every few years the National Governors’ Association tries to facilitate agreements between the states not to compete with each other in this way. Governors agree to it, then head back to their state capitals to sign the most recent incentives packages.

A word of caution. While outcomes of PDs are suboptimal for the players, PDs need not necessarily be more broadly socially suboptimal. Recall the canonical expression of the game in which two prisoners rat each other out to the cops. This good for society. So, too, market competition reflects a Prisoners’ Dilemma structure among firms. In this setup, price competition dissipates economic profit and distributes it society. Market competition essentially socializes the benefits of production.

Analogous to price competition, incentives competition among political jurisdictions dissipates gains for the competing governments, even for the one that “wins.” If the competition is set up properly, the company can receive back via the incentive package the entire expected gain to the government of its location decision. All of it. This is not limited to revenue gain. Any benefit a locating firm would confer on a jurisdiction can be returned to it in the form of tax abatements or subsidies. The competition drives down any net gain of winning the firm to zero. This is the analogous outcome to market competition driving down economic profit to zero. (Keep in mind the difference between normal and economic profit.)

Secondly, in equilibrium, incentive programs don’t change where companies choose to locate relative to where they could locate if no government offered location incentives. This point is often misunderstood, so I need to explain. Countless policy studies show that incentives programs do not alter where business choose to locate. That consistent with the theoretical prediction. The next conclusion, however, is that one that’s not warranted: cities and states need not offer incentives programs.

Why would studies show that incentive programs don’t alter location decisions, yet cities and states still need to offer them? That’s the “dilemma” of the Prisoners’ dilemma.

Let me explain using a goofy example.

Alaska could offer a pineapple company various incentives to move production from Hawaii to Alaska. Given the cost advantage the company has when growing pineapple in Hawaii relative to Alaska, Alaska would need to offer a huge subsidy to entice the company to relocate.

Hawaii wants to keep the company in Hawaii. Does it need to match Alaska’s level of subsidy? No. All Hawaii needs to offer the company is a subsidy sufficient so the overall cost of production in Hawaii just equals (or is a smidgeon lower) the firm’s cost in Alaska. The pineapple company will stay in Hawaii.

No surprise there. While (intentionally) a silly example, the truth remains the same in less exaggerated circumstances: Locations with cost advantages for a firm can always beat locations that lack the same cost advantage.

Here’s what we observe: Companies locate in the same place they would have in the absence of any location incentives. In fact, we observe governments that offer lower incentives package (because the region already offers lowest cost production to the company) nonetheless attract the companies.

In equilibrium, nothing happens (except governments receive lower tax revenues). The trick is this, however: This is true only in equilibrium. If the low-cost region’s government did not respond to the other government’s incentive package, then the firm would in fact move to the higher-cost region. But that doesn’t happen in equilibrium. The low cost state has to respond to keep the firm from moving.

A third implication: If two or more regions offer the same costs of production to a firm, then any expected gain from the firm locating in its jurisdiction will be entirely dissipated in the incentives package. “Gains” from a locating business need to be limited to tax revenues. Whatever officials value, not only tax revenues, but employment to their constituencies and benefits beyond tax revenues, the company can ask that value to be returned via additional subsidies. The full value of the locating firm to the jurisdiction’s officials will be dissipated in the subsidies and tax abatements.

This makes sense of studies showing the tax revenue gains from a company locating to a jurisdiction are often less than the cost of the incentives package offered to the companies. Government officials aren’t making mistakes, however. They’re offering the firm the full value of its location decision; value for the jurisdiction that goes beyond the added tax revenue it brings in.

First-order effects of incentives competition between governments for specific firms appear to be minimal: While governments do lose tax revenues (which might well be a positive rather than a negative), their incentive packages don’t induce suboptimal location decisions in equilibrium.

There are, however, some second-order distortions from the sorts of special, one-off incentives programs that Amazon and sports teams try to entice: They create discriminatory tax and subsidy regimes in the jurisdictions in which they locate. To wit, specifically tailored incentive packages discriminate against immobile capital and against non-preferred mobile capital (such as companies that are too small to attract the attention of policy makers).

We might have normative concerns about the discriminatory tax and subsidy programs of the sort the Amazon HQ2 will receive, and this type of discrimination can distort economic outcomes by artificially inflating economic returns to firms that receive special incentive packages relative to firms that do not receive the special treatment. It’s difficult to gauge just how significant the distorting effects might be in an economy replete with taxes that distort optimal decisions. Nonetheless, these firm-specific types of incentive programs have normative and economic consequences that distinguish them from normal competition between taxing jurisdictions based on across-the-board policies that apply to all similarly positioned individuals and corporations.

Reader Discussion

Law & Liberty welcomes civil and lively discussion of its articles. Abusive comments will not be tolerated. We reserve the right to delete comments - or ban users - without notification or explanation.

on January 23, 2018 at 14:32:34 pm

[…] Read more[…] […]

read full comment
Image of When Governments Compete for Business, Do We All Win? | Top 100 Blog Review
When Governments Compete for Business, Do We All Win? | Top 100 Blog Review
on January 23, 2018 at 15:51:12 pm

Off topic, but since you chose to illustrate this discussion with Amazon’s HQ2 proposal, here’s a query: WHAT’S UP WITH AMAZON’S HQ2 PROPOSAL?

First, WHAT FIRM BUILDS TWO HEADQUARTERS? Sometimes firms end up with two headquarters as a result of a merger—and generally this proves to be a bad outcome. Why would any firm CHOOSE to saddle itself with these headaches?

Because this idea is just stupid, I figured the whole exercise is just a way for Amazon to shake down Seattle for more tax breaks—testing Rogers’s thesis about immobile capital’s ability to extract such benefits from government.

Second, when you look at the list of places Amazon is considering, WHY PICK DC? You get wildly inflated real estate, congested roads, congested airports, and a region that isn’t especially beloved by techies. What’s up with that? Yet Amazon didn’t merely pick DC—it picked DC THREE TIMES! It picked DC. And Montgomery County, MD. And Northern Virginia.

So now I’m entertaining a second hypothesis: This whole exercise reflects an effort to build some massive federal security apparatus disguised as a new Amazon HQ. It would look really conspicuous for the feds to start hiring all of the tech grads and housing them in some massive new facility around DC. But if we dress it up as Amazon’s new headquarters, maybe people won’t figure it out.

And it might just work—except that they seem intent on putting it in DC, which makes no damn sense, and so kind of gives the game away.

read full comment
Image of nobody.really
on January 23, 2018 at 16:25:16 pm


Well, the dolts at (or formerly at) the Boeing Company at an expense in the hundreds of millions of dollars. What a drain on resources that has become; not to mention the added expenses associated with communicating between the productive arm(s) of the company and the administrative.

C'mon, nobody! Do you think Bezos would assist The Trumpster in staging a new security apparatus?
Then again, it could be to spy on the Trumpster and his ilk. - Ha!

More than likely, it is simply an opportunity to save on travel costs for amazon's lobbyists at the Federal level. Now how is that for paranoid economics?

read full comment
Image of gabe
on November 20, 2020 at 09:26:48 am

This is an important topic and deserves a healthy public discussion. The topic also deserves closer editing and more careful writing. Both efforts best left to those more qualified than myself. Site selection, be it Amazon or Bob’s Widget Mfg. is an auction process with as many bidders as beneficiaries. A few may include labor and management at existing plants competing for an expansion or in danger of relocation, governmental entities seeking more or better jobs for their constituents and larger tax bases. Any business has choices available to it including a bounded choice of locations from which to select, all of which would serve their purpose as well as an ultimate go/no go decision. These decisions are generally driven by relative costs between potential location. These costs are, in part, a function of costs controlled by local government agencies r which can be subsidized by local government. Hence, the bidding process. Like any auction, one must take care not to overbid.

read full comment
Image of Jason

Law & Liberty welcomes civil and lively discussion of its articles. Abusive comments will not be tolerated. We reserve the right to delete comments - or ban users - without notification or explanation.