Reading a Deal Like a Skeptic: A Practitioner's Checklist

We end where a practitioner has to live: in front of an actual deal, with an actual set of documents, needing to reach a defensible judgment. This lesson assembles the course into a working checklist. It is not a script to recite but a discipline to internalize, run in the order given, because the order encodes the argument that a deal is a sequence of documents read in reverse.

Begin with the exit terms, before you let the headline anchor you. Locate the clawback clause. Ask the four questions from Module Two. Does the trigger fire on a measurable, documentary event — payroll records, permits, tax filings — rather than the company's own attestation? Does the recapture prorate to the size of the shortfall, avoiding both the unusable all-or-nothing design and the negotiable no-formula design? Is the recovery secured — a letter of credit, a bond, a lien, a parent guarantee — rather than resting on the government's willingness to sue an entity that may by then be gone? And is there a party with the statutory duty and the actual staff to enforce it? If the clawback fails on any axis, downgrade everything downstream, because the deal's optimism is uninsured.

Second, isolate the enforceable number. Set aside construction jobs, which end; set aside multiplier and induced jobs, which no clause recaptures; set aside the capital investment, which is a means and not the benefit. Find the permanent, direct, on-site jobs the agreement actually requires — not the announcement's projection but the agreement's minimum — and find the wage floor those jobs must clear. That number and that floor are what the public is actually buying. Everything else is context or decoration.

Third, measure the gap. Compare the announcement's headline figure to the agreement's required minimum. A wide gap does not by itself condemn a deal, but it tells you how much of the public presentation was aspiration rather than obligation, and it calibrates your trust in every other claim the announcement made. Where the announced number is many times the required number, treat the announcement as marketing and the agreement as the deal.

Fourth, check the maintenance terms. A job requirement met on a single compliance date and then abandoned is a subsidy for a photograph. Confirm that the jobs must be sustained across a maintenance period of years, verified continuously or on an average rather than a snapshot. Facilities with small permanent headcounts make this especially acute — with few jobs to begin with, erosion after the audit is both easy and costly to the public case.

Fifth, assess the transparency regime. Does the jurisdiction disclose foregone revenue, as the governmental accounting standard on tax abatement disclosures requires? Does it maintain a public database of awards and progress reports? Does it require independent evaluation with sunset provisions? A strong regime lets you extend some default trust and gives you a record to check against. A weak one means the announcement may be the only document you can see, and every protection must be independently verified — often by turning to the watchdog databases that exist precisely because official disclosure fell short.

Sixth, run the cost-per-permanent-job ratio yourself, using the enforceable job number as the denominator and the full subsidy value as the numerator. Do not let construction jobs or multipliers into the denominator; they will flatter the ratio into meaninglessness. If the resulting figure is one no official would say aloud, you have found the deal's pressure point — and probably the reason the announcement leaned so hard on everything except permanent jobs.

There is a temptation, running this checklist, to slide into reflexive opposition, and Gil would caution against it. The skeptic's discipline is not to conclude that every subsidy is a giveaway. It is to insist that the deal be judged by what it enforces rather than what it announces — to hold the number and the clause to the same standard, and to trust the clause. Some deals survive this scrutiny handsomely: real jobs, real wages, a proportional clawback secured by a letter of credit, a jurisdiction that discloses and evaluates. Those deals are conditional public investments, and they are defensible. Others announce a great deal and require almost nothing, protected by a clause that performs accountability without delivering it. The checklist does not tell you which kind you have. It tells you where to look, and in what order, so that when you reach a judgment it rests on the executed terms rather than the ribbon-cutting. Fiscal policy is the second power grid. Judge it the way you would judge the first — at its failure points, by whether the safeguards actually hold when the load comes on.