When the Jobs Number Is Small and the Check Is Large

The economic development subsidy is an old instrument, and like most old instruments it was calibrated for the world in which it was designed. That world was industrial. The paradigm case was a manufacturing plant: enormous capital investment, yes, but also a large permanent workforce — hundreds or thousands of people drawing steady wages, spending them locally, filling the schools and the shops. The subsidy logic followed naturally. A community forgoes some tax revenue up front; in exchange it gains a durable base of good jobs and, eventually, an expanded tax base once the incentives expire. The cost-per-job, the fundamental ratio, came out reasonable because the denominator — the jobs — was large.

Now apply that same instrument to a different kind of facility: one that is capital-heavy but labor-light. Enormous investment in the building and its equipment, but a permanent workforce of dozens rather than thousands. The category is broad — large automated distribution and fulfillment centers, and above all the new generation of computing facilities that house racks of servers behind a small operations staff. These facilities can represent some of the largest single capital investments a region has ever seen, which is exactly why they attract some of the largest subsidy packages. But the permanent headcount is a rounding error against the investment. Run the ratio and the cost-per-permanent-job can climb to figures — in the hundreds of thousands of public dollars per job, sometimes well beyond — that no official would willingly say aloud.

This is the mismatch at the heart of the current backlash, and it is worth being precise about why it is a genuine problem and not merely a talking point. The subsidy's original justification rested on the permanent jobs. That is what the foregone revenue was supposed to buy. When the permanent jobs are few, the justification has to migrate elsewhere, and you can watch it migrate in the language of the announcement. Suddenly the emphasis falls on construction jobs — the temporary employment of building the facility, which is real but ends when the ribbon is cut. It falls on the capital investment itself, as though the building were the public benefit rather than the means to one. And it falls hardest on multiplier math: the modeled indirect and induced jobs, the tax revenue projected years out, the vague promise of anchoring a regional cluster. Each of these may hold a grain of truth. But notice that the argument has quietly moved off the one number that clawbacks can actually enforce — permanent, on-site, direct jobs — and onto numbers that no exit clause has ever recaptured a dollar against.

There are further wrinkles specific to these facilities that sharpen the concern. They are heavy consumers of two public resources — electricity and, for the cooling of computing facilities, water — which means the community may be subsidizing an operation that also strains the very infrastructure the community depends on. This is where the domain's organizing idea earns its keep: fiscal policy is the second power grid. A large incentive package is a flow of public value every bit as real as a flow of electrons, and a facility that draws heavily on both the fiscal grid and the literal one concentrates two kinds of public cost in a single site with a small payroll. The permanent jobs, the thing the subsidy was meant to secure, are the smallest quantity in the equation.

None of this makes such facilities bad investments per se. A computing facility may bring genuine value — property tax after the abatement burns off, a strategic asset, a foothold in a growing industry. The point is narrower and sharper. When the permanent headcount is small, the traditional cost-per-job justification collapses, the announcement compensates by leaning on temporary and modeled benefits, and the gap between what is promised and what is enforceable grows to its widest. That is exactly the condition under which the exit terms matter most. If a facility is going to be justified by a small number of permanent jobs, then the clause protecting that small number had better be airtight — because it is carrying the entire weight of the public case. Read the exit terms first, Gil says, and read them hardest when the jobs number is small and the check is large. That is not cynicism. It is proportion.