The siren song of industrial policy

“New Megafab Will Create Nearly 50,000 Jobs” proclaimed Gov. Kathy Hochul and Sen. Chuck Schumer at an Oct. 4 press conference announcing an agreement reached with Micron, a U.S.-based memory and storage manufacturer and one of the largest producers of semiconductors worldwide.

Building on the federal CHIPS and Science Act of 2022 (CHIPS being the acronym for “Creating Helpful Incentives to Produce Semiconductors”—oh, how clever…), New York has committed $5.5 billion in “performance based” tax credits under the state’s “Green CHIPS” program plus $200 million in infrastructure spending and promised access to low-cost power through the New York Power Authority. Onondaga County promises an additional $25 million for workforce development and other supports and commits to a 49-year payment-in-lieu-of-tax agreement affecting property and sales tax liability.

For its part, Micron promises 9,000 jobs and $100 billion in new investment over the next 20-plus years and agrees to use 100 percent renewable power “with preferences for in-state sources.” It will be, the company says, the biggest private investment in New York history—and the largest semiconductor fabrication facility ever built in the U.S.

That’s heady stuff. Gilding the press release lily, the state cites a study (using a REMI model) that predicts “transformative growth” over the first 31 years of operation involving nearly 50,000 total jobs, $5.4 billion in annual disposable income to New York residents and a total of $25.6 billion in revenue to state and local government. (Wow! State and local taxpayers actually MAKE money on the deal! What’s not to like?)

Central New York (and most of the upstate region) has been languishing economically for years. Of the 100 largest MSAs in the country, the Syracuse MSA ranked 93rd in job growth from 2000 to 2019. Buffalo and Rochester hardly did better, ranking 89th and 91st, respectively.

Nor has performance through the COVID-19 era been any better. From the 2019 average through August 2022, all three ranked in the bottom 10, losing more than 4 percent of total jobs.

By piggybacking on the federal CHIPS and Science Act, New Yorkers are taking advantage of incentives funded by federal taxpayers. This is one of New York’s core competencies, to be sure.

Too good to be true?

Breathless pressers from elected officials deserve careful scrutiny, particularly when they justify a substantial payout by taxpayers.

Nine thousand jobs in semiconductor manufacturing is impressive, were it to happen. These would be good jobs. The state Department of Labor reports employment of 16,307 in 258 establishments in 2021, with an average annual wage of $95,558 (QCEW for NAICS code 3334). The global demand for semiconductors is likely to be strong for decades to come and New York has demonstrated its ability to attract and retain firms and workforce in this sector. It isn’t fanciful to believe that the state can increase its market share.

Are the incentives a “good deal” for the state’s taxpayers? Leaving aside the federal subsidies provided through the CHIPS and Science Act and considering just the 9,000 jobs that are “countable” at Micron, the subsidy per job is well over $600,000.

That’s a lot of money. Timothy Bartik, economist at the Upjohn Institute in Michigan, has spent his long career exploring this question. His analysis of the CHIPS and Science Act is worth reading. Writes Bartik: “Prior studies suggest that well-run economic development programs that provide services and infrastructure for business development can create jobs at a cost per job of around $55,000.” As he implies, the question is not whether public spending to stimulate local economies is appropriate, but whether there are less costly ways to invest our tax dollars.

That’s why the “nearly 50,000 job” figure is important. Swapping 9,000 with 50,000 in the “cost per job” calculation lowers the figure to $115,000 (still twice the figure quoted by Bartik, however).

Yet this figure is highly speculative. Having conducted economic impact analyses for over 20 years, I know well that results of such analyses are heavily driven by assumptions.

The model cited in the press release estimates offsite or “spillover” impacts. These are the more than 40,000 jobs that are not directly connected to Micron. Spillover jobs (and associated payroll) includes jobs at firms that are supplying Micron with goods and services plus employment at firms supplying goods and services to Micron employees. As these jobs cannot be directly observed, these spillover estimates depend on changing and unpredictable future events.

■ Jobs at supplier firms may or may not occur within the state—or the nation, for that matter. Contracts with firms receiving incentives can force “in-state” sourcing, but even this does not guarantee that the jobs will be in state unless the program implements a costly (and likely inaccurate) tracing regime.

■ No one even tries to monitor or direct the spending of employees. The continuing expansion of online commerce erodes an important connection between economic development incentives and regional impact.

The REMI analysis claims that the investment will spin off 4.6 spillover jobs for every direct job. Without descending too far into the wonky details, this figure is about twice what would be generated from a simpler model that assumes that the structure of the economy remains largely unchanged. This larger impact depends on the Micron facility spurring synergistic growth in the state’s economy. That’s possible, of course, but by no means assured.

The personal income figure—$5.4 billion—is surely inflated. At $108,000 per job, this is higher than the average wage in New York semiconductor manufacturing in 2021 (again, $95,558). The spillover jobs would have to pay better than the “core” jobs at Micron. Yet the jobs spurred by Micron employees would be in the retail and service sectors, typically paying about half that of the semiconductor manufacturing jobs. Are jobs at firms supplying Micron making up the difference? That is improbable.

Moreover, I find the use of 31 years peculiar. When assessing the economic impact of incentives, it is typical to match the time frame of the impacts to that of the incentives, which is 20 years, not 31. Did the model need to run that long to ensure that the “return on investment” for the public sector turned positive?

Finally, let’s remember that we are basing an enormous commitment of state resources on a forecast that extends far into the future for a technically sophisticated product in a fiercely competitive global marketplace.

Oh, but we’ve been assured that the incentives are “performance based”—Micron benefits only if they meet their commitments. First, remember that Micron’s influence on off-site jobs is limited, so performance metrics can only address the 9,000 jobs, even though the magnitude of the incentives can be justified only with the spillover jobs included in the calculation.

Second, there’s a long history of high-profile economic development deals being renegotiated when job creation falls short. The state comptroller’s office issued a report in 2020 that is highly critical of Empire State Development oversight of major projects. Of the RiverBend Hi-Tech Manufacturing Hub in Buffalo, for example, the comptroller’s office notes that job creation agreements were amended several times, reducing the number of jobs required for the incentives provided.

In 2018, the Center for American Progress posted an overview of what can go wrong when we use incentives to attract jobs, noting the temptation to elected officials when “short term political optics often outweigh reality.” Other questions about the Micron incentives are addressed in an analysis by the Empire Center for State Policy.

We can hope that the Micron investment will truly be our “Erie Canal moment,” as Schumer declared. History suggests that the claims made on the deal’s behalf are wildly inflated, however. The incentives offered to Micron by New York may have been needed to secure the deal, given how the CHIPS and Science Act has encouraged many states to pursue chip fabs.

Let us not forget, however, that the fabulous sums being promised could have been used for other worthy purposes or simply left in the pockets of taxpayers.

Kent Gardner is Rochester Beacon opinion editor. The Beacon welcomes comments from readers who adhere to our comment policy including use of their full, real name.