Do we really want Amazon’s HQ2?
The winner’s curse in economics suggests that winning bids are very often too high.
Twenty American cities, including Los Angeles, have been busy in recent weeks wooing visiting executives from Amazon, hoping to become the location for the company’s second headquarters. Amazon has promised that up to 50,000 jobs and $5 billion in investment will f low to the new HQ2 city.
The winner of this competition, however, may find itself a big loser.
Despite the long list of requirements — including a “a compatible cultural and community environment,” a skilled workforce and close access to transportation options — Amazon has essentially pitted these cities against one another in an auction. The cities are bidding with tax incentives, infrastructure investments and the like hoping to win HQ2 and presumably the jobs and riches that will flow from it. A city that estimates, for example, that HQ2 will produce $5 billion in new tax revenue from all the economic activity may be willing to spend, say, $4.5 billion to get the headquarters. Another area may assess the benefits are even higher, and so offer $7 billion in incentives. (That's what the state of New Jersey is offering for the Newark location.)
Each city, however, does not really know what the benefits of the headquarters will be; it can only ask various experts to estimate them. Even with deliberation, some guesses will be too high, and some too low. But the city that offers the greatest incentives, and gets the headquarters, is likely one that overestimated the benefits.
This problem, called the winner’s curse, was well described by Nobel Prize-winning economist Richard Thaler some years ago. He was building on work by engineers from Atlantic Richfield Oil Co. showing why the winner of an auction for drilling rights may, on average, bid more than the value of the drilling rights. Follow-up studies found that oil firms really do suffer from the winner's curse. For example, in auctions for oil tracts in the Gulf of Mexico, oil companies earned profits so low they would have been better off taking their money to the local credit union.
Today state and city officials hope to land Amazon’s HQ2. In the 1970s the big prize was the first automobile assembly plant in the United States by a foreign carmaker, Volkswagen. States competed to get the factory, which VW promised would create as many as 20,000 jobs. States didn’t have to believe that number, but the most credulous states would be the ones willing to offer the richest incentives.
The state of Pennsylvania provided $75 million (more than $300 million in today’s dollars) to land the plant in Westmoreland County. The state also spent $30 million ($135 million in today's dollars) building a railroad spur and a highway to serve the plant. It spent smaller sums helping to screen 45,000 job applicants. In addition, the U.S. Environmental Protection Agency compromised on air pollution enforcement, and local governments granted fiveyear property tax breaks.
But the Volkswagen plant never hired anywhere near 20,000 people; the peak was about 6,000 jobs. By 1984 employment dropped to 1,500. The plant shut down in 1988.
As another example of overestimates, General Electric moved its headquarters to Boston in 2017, after looking at 40 cities. Boston and the state of Massachusetts offered an incentives package totaling $145 million, expecting to get 800 jobs and an investment of $200 million. But a few months afterward, GE delayed opening its new building by two years and announced layoffs.
When Amazon announces its HQ2, 19 cities are likely to, at first, feel the sting of disappointment. But they should pause and consider that the winning city probably gave away excessively high tax breaks and incentives, and their own city officials did not.