Your capital reapprovals ask the wrong question
The following is a guest post from Sergey Kyunttsel, an economist and independent researcher. He founded and led an engineering company that delivered more than 200 industrial energy-efficiency projects. Opinions are the author’s own.
Most companies with formal capital-approval procedures already put a clock on capital. Capital authorizations have validity periods, delegation-of-authority manuals specify when approvals lapse and variance rules send a project back to committee when its cost or scope crosses a threshold. If a company approves material capital spending, some version of these controls probably already appears in its policy manual.
The gap is not the date. It is the question the company asks when review begins.
The standard reapproval asks: Does this project still clear the hurdle? That tests the project against itself. It establishes that its net present value remains positive or its projected return still exceeds the minimum. But it can miss the question capital allocation is supposed to answer: Is this still the best available use of the money?
A project can remain acceptable in isolation and still lose its place in the ranking. Capital allocation is a comparative decision under scarcity. A reapproval that does not revisit relative priority is less a control than a formality.
When a reapproved project is still the wrong project
Months can pass between committee approval and an irreversible commitment. In the meantime, engineering continues, vendors are qualified, quotes are refreshed, scope is renegotiated and operating plans change, sometimes materially.
In the second-quarter 2026 CFO Survey from the Federal Reserve Banks of Richmond and Atlanta and Duke University, finance executives raised their forecasts for 2026 unit-cost and price growth by 1.1 percentage points in a single quarter. That result does not show that approved projects became stale; it shows how quickly the inputs to a capital case can move. A case built in spring may rest on materially different assumptions by autumn.
Consider a modernization project approved in March on the basis of a nine-month implementation schedule, a vendor quote, an assumed production volume and projected labor savings. Procurement reaches a final contract in September. The quote has risen, the production plan has softened and the installation window has moved beyond the shutdown on which the schedule depended. None of those changes is necessarily fatal. Run the numbers again and the project may still clear the hurdle.
But the competitive set can move too. A smaller project ranked second in March may become more attractive, a new proposal may enter the pipeline or the operating constraint that made the original project urgent may disappear. The approved project can still be good and no longer be first. A reapproval that reruns only its own case will miss that change by design.
I saw this while my engineering company was competing against an incumbent lighting manufacturer. At the start of a long assessment and tender process, a new line of our luminaires was still in final development and could not be included in the bid. By the time the customer was ready to commit, the equipment was available and offered 10% to 15% greater energy efficiency, depending on the fixture type. The customer reopened the comparison and reversed the original selection, even though the better-known incumbent offered more favorable financing. The incumbent’s proposal had not become unacceptable. It had been overtaken. That customer reopened the comparison on its own initiative; the mechanism worked. The same logic applies with more at stake when what may have been overtaken is not a bid but a project’s place in the capital plan.
What reapproval should actually test
When a date or threshold triggers a review, finance should first ask whether the assumptions that determined the ranking have changed. The review should end in one of four decisions: reaffirm, resize, pause or change the project’s priority.
The comparison must be bounded. Reapproval should not reopen every possible use of corporate cash; debt repayment and share buybacks run on different clocks and often involve different decision makers. Compare the project with alternatives competing for the same capital in the same cycle and capable of advancing on the same approval calendar, including those ranked just below it and those that have entered the pipeline since.
Most reviews should be short. If no decisive assumption has crossed its policy threshold, the project proceeds. If the economics have changed but the order has not, finance updates the record and continues. Only a material change in relative priority needs to return to the committee.
Make the ranking reviewable at approval
Reranking is possible only if someone records why the selected project ranked first. Six months later, a team may not reliably remember which two or three variables decided the choice, and a large model is too unwieldy to reconstruct the decision from scratch.
The approval should therefore carry four things:
1. A latest commitment date. The date by which the contract must be signed or the order placed—not the expected completion date.
2. The assumptions that determined the ranking. Not every model input, but the few variables capable of changing the choice: price, volume, timing, service life or a specific operating cost.
3. A revalidation threshold set by policy. A sponsor allowed to set the trigger will be tempted to choose one that never fires.
4. A named revalidation owner in finance. Project management can report that something changed. Finance must determine what the change does to the ranking.
A compact approval record could contain four lines:
Valid until [date]
Rerank if [thresholds]
Ranking assumptions [variables]
Decision owner [name or role]
That is not new bureaucracy. It is the memory that a reapproval needs to be worth running.
Revalidation also needs a stop rule. It should test the recorded ranking assumptions against alternatives eligible in the same cycle. It should not reopen every technical input, invite perpetual redesign or give losing vendors repeated opportunities to rewrite their bids.
A validity date has a predictable side effect: It gives sponsors a reason to sign early.
If review occurs only when the clock runs out, the cheapest way to avoid it is to commit before the deadline. A policy designed to preserve optionality would instead hasten the irreversible decision it exists to test.
Two safeguards reduce that risk. First, a threshold must override the date: If a decisive assumption crosses its limit, authorization pauses even if the validity window has not ended. Second, do not penalize a sponsor for releasing capital when the case weakens. The money should be promptly reallocated, but returning it should not count as failed planning or diminish the sponsor’s standing in the next cycle. As long as releasing capital signals failure, sponsors will have reason to spend it to protect it.
Changing the answer is not admitting error
The people who approved a project often staff its review. That is the mechanism’s real weakness, and no form eliminates it. The goal is not to pretend reversals are frictionless. It is to make a changed ranking a legitimate outcome rather than an admission that the original approval was wrong.
Consistency means applying the same standard when the facts change, not preserving the same answer. A project should not keep its place in the capital plan simply because it reached the committee first.
I have argued in these pages that modernization projects can pass the payback test and still lose value after they are built, through costs that never entered the original case. This is the parallel exposure at the other end of approval: not value lost after commitment, but priority lost before it.
A project can remain good and no longer deserve the next dollar. Putting a date on the approval is the easy half. The harder question is the one to ask when review begins — not whether the project still clears, but whether it still wins.
