When Amazon discloses its plans to build a new warehouse, the news release is predictable.

An Amazon executive cites the number of jobs that the company plans to fill. Elected officials thank Amazon and praise the vibrancy of their business community. And, sometimes, they predict benefits that will flow from Amazon's investment.

Middletown, Del., Mayor Kenneth Branner, for example, said an Amazon facility announced in 2012 would "be a jump start to our economy and bring additional employment growth to our area."

A new study challenges that premise, finding that counties that house a new Amazon depot show no growth in the number of total jobs in the wider economy during the two years after a facility's opening.

"Amazon, when it opens a fulfillment center, does add warehousing jobs," said Ben Zipperer, an economist at the Economic Policy Institute (EPI) and a co-author of the study released last week, using Amazon's term for its warehouses, which tend to employ between 500 and 1,500 people each. "But those don't really translate to any sort of broad-based economic growth in the county that they open the centers in."

In the absence of an obvious hiring boom, the study's authors contend that mayors and state officials should refrain from spending public dollars and providing tax breaks to secure Amazon's commitment to build warehouses in their region.

The report is the latest fodder for the debate about Amazon's growing role in the U.S. economy, and the use of public funds to offset the costs of the company's rapid expansion.

The Seattle company's warehousing footprint has expanded dramatically in recent years, as Amazon placed depots closer to major population centers to deliver packages quicker and cheaper. The company had fewer than 10 centers through the mid-2000s, the study says, and was nearing 100 by the end of 2017.

In the process, the company has become the second-largest U.S.-based employer, trailing only Walmart. Critics have scrutinized the high-pressure working conditions present in its warehousing arm, as well as the public subsidies spent to lure warehouses in the first place.

"They are clearly opening a lot more fulfillment centers, this is a big development in our economy," Zipperer said. "Local governments are more or less falling over themselves to attract Amazon. That raises the question, what are you actually getting when you basically sacrifice future tax revenue? How big is that trade off?"

In one case cited by Good Jobs First, a corporate subsidy watchdog that has criticized payments to Amazon, Chattanooga, Tenn., offered Amazon $30 million in property tax rebates, 80 acres of free land and $4 million in state payments to prepare the site for a warehouse built in the area that came online in 2011. .

In a statement last week, Amazon pushed back against the EPI study.

The company says it employed more than 200,000 people in the U.S. in 2016, and estimates that its spending led to the creation of another 200,000 jobs outside the company, including temporary construction jobs. (Amazon didn't provide a breakdown of how many of those jobs were created by high-paying corporate and software development positions, versus the lower-paying roles in warehousing and logistics.)

"These data points are not demonstrative of our current network, community impact, and both the direct and indirect job creation near fulfillment centers," the statement said.

Michael Mandel, an economist with the Progressive Policy Institute who contends that e-commerce jobs of the sort Amazon creates reduce income inequality, criticized the EPI study in a blog post. He noted that the study didn't include in its data set the case of Kenosha, Wisconsin, home to an Amazon warehouse since 2015 that has seen stronger employment growth than the state as a whole.

Zipperer said EPI limited its study to counties with warehouse employment data in a Bureau of Labor Statistics database available for the period between 2001 and 2015.

The study examined 54 Amazon fulfillment centers located in 34 U.S. counties, which the authors say covered about three-fourths of Amazon's warehousing network as it stood in 2015. The authors limited their sample to warehouses opened by that year, an effort to evaluate employment changes in the years following the opening.

"We used the data we had, and tried to compare apples to apples," Zipperer said. "We will almost surely update this, the more data we get."

EPI, based in Washington, D.C., describes itself as a nonpartisan think tank that advocates for worker-friendly policies. About 30 percent of its funding comes from labor unions.

Zipperer said he had two theories as to why Amazon didn't show up as a larger contributor to employment in his data.

"The most charitable explanation that would mesh with our findings is Amazon is creating jobs, but they're not that many to actually detect as a (meaningful) addition to overall employment," Zipperer said. Warehousing has historically been a small employer, he said, accounting for about 1 percent of total employment in most areas covered by the study.

The other possibility, he said, is that Amazon warehouse "are just pulling people away from other sectors of the economy," he said. "Some retail worker is now going to take a job at the warehouse."