Now Amazon wants to break the bank

On Thursday (October 19), the first stage of the Amazon sweepstakes for a second headquarters, dubbed “HQ2,” was completed as cities submitted their bids to the Internet giant. With a possible $5 billion investment and eventually 50,000 jobs at salaries over $100,000, this is one of the very best economic development projects to come along in a long time.

BUT IS IT?* I have a hard time understanding how this project even makes economic sense. Why does a company need two headquarters? Doesn’t that defeat the whole idea of having a headquarters as a centralized coordination site? Dean Barber has gone so far to suggest that Amazon will “A/B test” the two headquarters: See which one does better, and close the other one (h/t Greg LeRoy). I can believe it, but as long-time readers know, this is an area where we’re looking at the terrible information  asymmetry that bedevils governments in site selection bidding wars. In other words, we have no idea whether Amazon intends to close its existing headquarters in Seattle or close HQ2 if it doesn’t work as well as the current headquarters. This is kind of an important point for assessing the impact of HQ2 on the United States as a whole. And then there’s the effect that subsidizing Amazon will have on its competitors, which will reduce net job creation and net benefits for the country as a whole. On top of everything, unemployment is pretty low in most states, so why would they want to shell out billions of dollars for this project?

I’m not going to try to prognosticate the HQ2 location (I almost said “the winner,” but there is a good chance of overbidding in a public auction with so many competitors), because there is no way to wave away the information asymmetry problem. I think we can be confident that it will be a place that is already a tech hub. CNN offers eight plausible contenders (Atlanta, Pittsburgh, Toronto, Dallas, Austin, Boston, San Jose, and Washington, DC), while the New York Times predicts Denver. Instead, I want to highlight some specific bids that caught my eye.

One of the good stories is Toronto. Both the city and the province of Ontario were quick off the mark to say that while they would certainly do things like facilitating land assembly, there would be no cash subsidies for Amazon. The city’s application to the company emphasizes its abundance of tech talent. As a sign that this “no subsidies” talk is not empty, we can point to the announcement earlier this month that Ontario had secured a new investment from Thomson Reuters, adding 400 jobs over two years, scheduled to eventually grow to 1,500 jobs. This project received no financial support at all, only what provincial officials described as “concierge service” (h/t Carpentry of the Heart).

Another major city not getting involved in the bidding war is San Jose. Like Ontario officials, Mayor Sam Liccardo argues that talent availability is the key to attracting businesses. While this tells us nothing about incentives the state of California might provide, it is good to see two of the major potential sites taking an anti-subsidy position.

At the other end of the spectrum, New Jersey is offering $5 billion over 10 years, and Newark will provide another $2 billion over 20 years. This is crazy on a number of levels. First of all, it has a nominal aid intensity of 140%: New Jersey and Newark will pay all the costs of the project and give Amazon some more money on top of that. Second, this is 18 times bigger than the largest subsidy the state has given in the past ($390 million). (We saw the same situation with Nevada and Tesla, but that was only 13 times the state’s formerly largest incentive package.) Whether the state is really capable of managing a deal of that size is doubtful; indeed, the $390 million American Dream retail/entertainment project has been under construction for more than 10 years. Third, while not all of the approximately 50 bids have yet been revealed, currently the second-highest bid is $500 million from Worcester, Massachusetts (state support is currently unknown), 1/14 the size of Newark’s package. This is reminiscent of the disparity in bids between North Carolina and Virginia for a Dell manufacturing facility in 2004: $300 million (nominal) vs. $37 million! Fourth, and finally, this project requires the legislature to tear up the current geographic targeting of the state’s Grow NJ subsidy fund, currently restricted to the four poorest cities in the state, Passaic, Paterson, Camden, and Trenton.

Now, we wait. Amazon does not plan to announce its decision until next year. There is plenty of time to announce a list of finalists and subject them to more pressure. I am going to guess that when all first-round bids are revealed, the number 2 bid will be over $1 billion; indeed, there may be several $1+ billion bids. This is going to get ugly before we’re done.

*Bonus points if you recognize the literary allusion.