When to Kill a Deal You've Already Spent Time On
The hardest go/no-go decision in affordable housing development isn't the easy ones — the sites that obviously don't work, the capital stacks that clearly don't close, the entitlement paths that are clearly blocked. Those decisions are uncomfortable but not actually difficult.
The hard decision is when you've already invested significant resources in a deal and new information has made it significantly less likely to work. Stopping at that point requires overcoming the sunk cost pull in a way that stopping early doesn't.
This is a decision that most development teams face regularly and that most handle worse than they should. Here's a more useful framework for thinking about it.
Sunk cost is real but it's not the right input
The resources already spent on a deal are gone regardless of what you decide next. The relevant question is whether the expected value of continuing — the probability of the deal closing, multiplied by what it produces if it does, minus the additional resources required to get there — is positive.
This sounds obvious but it's genuinely hard to apply in practice because the resources already spent create a psychological anchor. A team that has invested three months and $50,000 in predevelopment on a deal is not going to make a neutral cost-benefit calculation about whether to continue. The sunk cost distorts the analysis.
The most useful protection against this is establishing decision criteria before you've invested heavily — defining in advance what would constitute a go/no-go trigger at the next decision point, so that the decision is evaluated against a pre-established standard rather than against the sunk cost.
The specific triggers that should end a deal
Some conditions should end a deal regardless of how much has been invested:
Program eligibility is resolved negatively. If the site was being evaluated for 9% LIHTC and it's now clear it won't score competitively in the relevant QAP, the program path has closed. Other program alternatives should be evaluated explicitly, but the default program assumption being off the table is a significant trigger.
The capital stack gap can't be closed. If soft debt assumptions have been tested and the realistic availability is materially lower than what the capital stack requires, the deal has a structural financing problem that isn't going away. The question is whether there's a different structure that works — not whether to continue pursuing the current one.
The land basis can't be resolved. If the seller's price expectation has been clarified and it's above what the program can support by a margin that no realistic structural fix closes, the deal is over.
The entitlement path has become untenable. If a required rezoning has been denied, if the entitlement timeline has extended past the point of compatibility with the financing structure, or if the political environment has shifted in a way that makes approval unrealistic, the deal needs to be reassessed.
What "pausing" actually means
Many deals that should be killed get "paused" instead. Pausing is sometimes right — a deal that doesn't work today because of a seller's price expectation or a program cycle timing issue may be worth revisiting in 18 months. But pausing also becomes a way to avoid the decision.
The useful discipline is to make pausing a documented decision with a specific condition for reactivation. "We're pausing this deal until the seller's asking price comes down to X or until the city housing trust fund opens its next application cycle" is a real pause. "We're pausing this while we figure out what to do" is a delayed no.
Documenting the specific condition for reactivation serves two purposes: it protects against the deal quietly dying in the queue without a formal decision, and it ensures that if the condition is met, the deal gets a real second look rather than getting overlooked.
Alpha Deal helps development teams make consistent go/no-go decisions at each stage of site evaluation — so resources stay concentrated on deals with realistic paths forward.