Underwriting the Tax Package as Infrastructure

The final move is to stop admiring the incentive and start underwriting it, with the same rigor and the same time horizon you bring to a power-purchase agreement. A tax package is a stream of value and risk that plays out over the entire life of the asset, and it should be modeled as such — instrument by instrument, year by year, with the political and legal exposures priced rather than assumed away. This is what it means, concretely, to treat the fiscal grid as infrastructure.

Begin by disaggregating the package into its instruments, exactly as we did earlier, and giving each one three attributes: a value stream, a duration, and a probability. The statutory equipment sales-tax exemption is a large, near-certain, one-time value at the moment of purchase — carry it at close to full weight. The twenty-year property abatement is a long annual stream, durable but exposed to reassessment risk and clawback conditions — carry it at high but not perfect probability and model the phase-down if there is one. The discretionary grant is contingent — carry it at a real discount reflecting the chance it shrinks or evaporates. The PILOT is a contract with an escalator — model the escalator honestly. Do not collapse these into a single headline; the whole point is that they behave differently and fail differently.

Then extend the model to the full asset life, and pay special attention to the year the abatement sunsets. This is the single most underappreciated number in incentive underwriting. An abatement does not vanish quietly; on its sunset date the full assessed value rolls onto the tax rolls, and the annual tax bill can jump by an order of magnitude in a single year. If you have modeled only the abated years and treated the deal as a permanent condition, you have mispriced the asset and possibly stranded a future owner. The sunset is also, as noted earlier, the community most persuasive number, so it serves double duty: model it for your own economics and disclose it for the community case. A package underwritten with the sunset priced in is a package that will not surprise you or the buyer in year twenty-one.

Next, reserve explicitly against the conditional exposures. If the package carries job or investment covenants with proportional clawbacks, build a reserve sized to your realistic probability of a shortfall and the recovery it would trigger. This is ordinary risk practice, but incentive clawbacks are routinely left out of models because they feel like remote contingencies. They are not remote when your operating assumptions are aggressive, and the Foxconn episode is the reminder that the gap between promise and delivery can be wide enough to convert a celebrated package into a liability. Price the clawback you might actually trigger, not the one you hope never fires.

Underwrite the political and legal risk as a real line as well, even though it resists precise quantification. The relevant questions are concrete: How durable is the statute — has anyone moved to cap or repeal it, as with the Georgia exemption in 2024? How exhausted is the local political tolerance? How exposed is the deal to a bad-headline year? You will not get a clean number, but you can assign a durability grade to each instrument and stress-test the model against the loss of the most politically fragile ones. A deal that only pencils if every discretionary and politically contingent piece holds for twenty years is a fragile deal, however attractive the base case looks.

Finally, integrate the fiscal underwriting into siting at the same stage as power and water, not after. The recurring failure this whole course has traced is the team that solves the physical infrastructure first and treats the tax package as a closing formality. By then the advantage a jurisdiction offers to an early, courteous applicant is gone, the calendar has hardened, and the community conversation is starting on the back foot. The teams that win capacity run the fiscal map in parallel from the first week: statutory versus discretionary, the bodies and their calendars, the competitive room, the political tolerance, the asset-specific value, and the full-life model with the sunset priced in. Do that, and the tax package stops being a surprise and becomes what it always was underneath — a piece of infrastructure, negotiated years ahead, as load-bearing as the substation and the water main, and every bit as decisive about where the capacity finally lands.