This article originally appeared in The Washington Post on December 11, 2019.

In September 2017, Amazon announced that it would build a second corporate headquarters, where it would offer 50,000 potential new jobs. The announcement asked for proposals for the location, suggesting that it hoped local governments would offer economic investment incentives. (Amazon founder and chief executive Jeff Bezos owns The Washington Post.) Bids came in from 238 sites across North America, including Los Angeles, Green Bay, Wis., and Orlando.

For a year, observers speculated about which city would be chosen. The day before the midterm elections, the Wall Street Journal reported that Amazon would split HQ2 between New York City and Arlington, Va. New York offered incentives valued at roughly $3 billion, based on existing state and local economic programs. Virginia offered about $1.5 billion in incentives, much of which included investments in education, infrastructure and workforce development. These were new initiatives but hardly giveaways just to Amazon.

Both Arlington and New York were obvious locations for Amazon’s investment, located close to major cities with highly educated workforces and other significant advantages. Many cities had proposed bigger incentives, including an $8.5 billion incentive from a Maryland site just a few miles from the Virginia location. Some observers suggested that these were the front-runners from the beginning.

Amazon canceled HQ2 in New York after unions, activists and politicians objected to the deal. But last week, less than a year later, Amazon announced that it would nevertheless expand operations in New York, adding 1,500 jobs and 335,000 square feet of office space.

Does this mean that Amazon didn’t need incentives to expand in New York City?

Incentives are rarely pivotal for investment decisions

Economic development incentives are benefits given to particular firms, often in the form of cash grants or reduced taxes. The goal is to encourage a firm to make an investment that it might not have made otherwise to create jobs, expand the tax base or achieve some other development goal.

Amazon’s newly announced expansion does not appear to involve public incentives — meaning that at least some of Amazon’s operations don’t need to be lured there and that New York can sometimes compete on its own strengths as an international business hub. Of course, this is just one event. And even when Amazon canceled its HQ2 location in New York, the company said it would nevertheless increase its New York presence.

But there’s extensive academic research on whether and how much public incentives influence corporate decisions. When economist Tim Bartik summarized recent studies of government incentive schemes in the United States, he found that they influence only between 2 and 25 percent of all business investments. Research suggests that major firms such as Amazon aren’t really swayed by public offers when they decide on locations. For example, the Maryland and Virginia HQ2 sites were only a few miles apart — and Maryland offered billions more in incentives. If billions aren’t enough to move an investment a few miles away, such incentives probably aren’t very effective policy tools.

We don’t know how many jobs Amazon would have created with HQ2 in NYC

Amazon’s original proposal for HQ2 suggested that it would create 50,000 jobs at an average salary of $100,000. When Amazon announced its decision to split its second headquarters, it announced that this would locate 25,000 jobs in Queens with an average salary of $150,000.

By contrast, the newly announced location in Manhattan is to include 1,500 jobs without a firm commitment on the average wages. If we go by these projections, this Amazon investment will apparently be less valuable to New York. But the original announcement may have been overly optimistic, as is often true.

For example, just last year, Apple announced that it would build a new corporate campus in Austin that would start with 5,000 employees and eventually scale up to 15,000. However, in Apple’s incentive contract with the county, which includes tax abatements, the fine print commits the company only to a total of 4,000 jobs by Year 12, with no penalty from the county unless Apple fails to create 70 percent of these promised jobs. Apple also will receive state incentives on a sliding scale; the company will be able to cash in on many of those well before hitting 4,000 jobs.

Amazon’s HQ2 deal with Virginia is similar. In the first year, Virginia requires only that Amazon occupy 65,000 square feet of office space to reap the benefits the state is offering. That’s much less than the newly announced 335,000 square feet of office space in New York. Virginia may eventually host 25,000 Amazon jobs as compared with New York’s 1,500, but that may not happen any time soon.

What’s more, as I’ve shown in research with Calvin Thrall, companies are often allowed to amend their incentive contracts, often reducing jobs requirements and even redefining what counts as a “job.” For instance, one company listed its CEO as among the “new jobs” it was creating in the area — which pushed up the average wages the company was paying there. Other companies have counted contractor and subsidiary positions as new jobs.

Promises are only meaningful if someone is willing to enforce them.What might happen when Amazon moves to your city?Will Amazon’s HQ2 be a good neighbor or a nightmare? Editor’s note: An earlier version of this video contained outdated population and incentive figures. (Video: Daron Taylor/Photo: Daron Taylor/The Washington Post)

Investment announcements are strategic

It’s hard to evaluate the jobs benefits that companies promise when they receive incentive checks. Politicians often attend ribbon-cutting ceremonies to take credit for new investments. Some states such as Texas require that firms coordinate any official announcements of the investment and the state incentives with the governor’s office. Politicians want to announce as many jobs as possible because it makes them look good; companies want to lock in as many incentives as possible for as little as they can deliver. Both politicians and companies want to throw out big job numbers.

In contrast, when a company announces an investment decision made without incentives, as Amazon did last week, there’s no need to detail how many jobs may or may not be involved. If it does make bold jobs predictions, that may just be big talk to get more publicity. In other cases, it may downplay plans, hoping for concessions in the future.

Like any company, Amazon had every reason to make big promises about jobs to win concessions from New York while negotiating flexibility in its incentive contract. Now it has strong reasons to downplay its plans to create jobs so it can avoid negative publicity for expanding in New York after all, after having backed out of its previous $3 billion deal.

Nathan M. Jensen (@natemjensen) is a professor of government at the University of Texas at Austin, a senior fellow at the Niskanen Center and the co-author (with Edmund J. Malesky) of Incentives to Pander: How Politicians Use Corporate Welfare for Political Gain (Cambridge University Press, 2018).