The gap you're describing is real and it shows up most often in the terminal cap rate assumption, which is where sponsors tend to bury optimism.
The mechanism that does the most work in those aggressive decks is the implied cap rate compression at exit. The sponsor models full NOI from abated taxes, exits at a market cap rate, and the buyer in that model is implicitly paying for income that includes a tax benefit with less than half its term remaining. A sophisticated buyer prices the remaining abatement term explicitly, often by capitalizing the tax benefit separately and applying a haircut for burn-off risk, then adding that to a yield-adjusted valuation on the stabilized NOI. That is a materially different number than the sponsor's model produces.
Your approach of lowering the acceptable exit cap rate to reflect diminished abatement term is a reasonable conservative framing, though the specific adjustment depends on how the jurisdiction structures the abatement, whether it transfers on sale, whether there is a recapture provision, and how liquid the buyer pool is for partially-abated assets in that submarket. Those terms vary enough that I'd treat your lender and local tax counsel as the place to confirm the mechanics in writing before they go into the model.
The risk I don't see named in your post is abatement transferability itself. Some programs do not transfer on a fee simple sale without reapplication or reapproval, which turns the remaining term into contingent value rather than underwritten value. If a buyer's counsel flags that risk at due diligence, it comes out of your exit price, not theirs.
In adaptive reuse deals specifically, the abatement is frequently the policy lever that makes conversion economics work, so the underwriting sensitivity to its treatment is higher than in a stabilized acquisition.
What does your exit cap rate adjustment look like quantitatively relative to the abatement burn-off period, and are you stress-testing transferability?