A proposal asks what the work costs. It rarely asks what running the proposal costs. The numbers in the budget are the easy ones to count: the training, the supplies, the staff days on paper. The expensive parts hide in the margins. The weeks spent writing the bid that lost. The systems built to track one reporting format. The finance team that exists because every grant counts differently. We call these overhead and try to shrink them, when they are the true price of accepting money the way we accept it.
So we win the grant and lose on the math. A short, restricted, heavily reported award can cost more to manage than it delivers in reach. The shortfall does not vanish. It is paid quietly from reserves, from unpaid hours, from the next proposal written to cover the gap left by the last one. The activity looks funded. The organization beneath it is running at a loss it was never allowed to name.
The build starts with one honest line. Before accepting an award, we cost the whole thing: the bid, the setup, the reporting, the closeout, the share of the people and systems it consumes. If that fully loaded cost is higher than the value the award creates, we say no, or we negotiate the terms until the numbers hold. We track win rate against bid cost, so we stop spending three months chasing money that takes four months to manage. We price the institution, not just the activity, and we ask funders to fund what delivery actually requires.
This is not a complaint about the people who write budgets or the people who approve them. It is a flaw in the format we all inherited, a template that makes the visible costs easy and the real ones invisible. The fix is to change the format. Add the hidden lines until the budget tells the truth, then make decisions against that truth. A budget that names its full cost is the first control that protects the mission. The rest is just accounting after the fact.