A sunk cost is a cost that has already been incurred and cannot be recovered, so it is the same under every alternative still open and cannot change which alternative is better. Managerial accounting states it as a cost that cannot be avoided because it has already occurred, one that will not change regardless of the alternative that management chooses, so sunk costs have no bearing on future events and are not relevant in decision-making (OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, 10.1). Leaving it out of a decision is not the same as taking it off the books.
The Boundary Is Recoverability, Not Age or Size
What makes a cost sunk is that nothing still available brings the money back. Its counterpart is an avoidable cost, meaning a cost that can be eliminated in whole or in part by choosing one alternative over another (OpenStax, Principles of Accounting, Volume 2, chapter 10 key terms). Economics draws the same line from the other side, defining sunk costs as costs that were incurred in the past and cannot be recovered, which should not affect the current decision (OpenStax, Principles of Microeconomics 3e, 2.1).
The date on the invoice decides nothing, and neither does the amount. A deposit paid this morning that the vendor refunds on request is not sunk, because canceling recovers it. A non-refundable license bought three years ago is sunk, because nothing available now recovers it. Sorting a whole set of costs into relevant and irrelevant, and running the comparison itself, is the work of differential costing, which sits inside managerial accounting.
Where Sunk Costs Still Count on the Books
A sunk cost drops out of the forward-looking analysis and out of nothing else. What was spent was recorded when the obligation attached, the timing question set out in incurred, and it stays in that period’s results. Where the spending bought an asset, the amount stays on the balance sheet, because the book value of an asset, historical cost minus accumulated depreciation, is a sunk cost regardless of whether a business keeps the asset or disposes of it in some manner (OpenStax, Volume 2, 10.1). One number, two readings, and only the decision drops it.
One future consequence does travel with the recorded amount, which is why “ignore it” is too blunt an instruction for tax work. Selling or otherwise disposing of the asset measures gain or loss against its adjusted basis under 26 U.S.C. 1001(a), the subtraction worked through in disposition, and the tax that follows is a future cash flow that belongs in the analysis even though the historical cost standing behind the basis does not.
The pattern repeats inside a firm. A tax software license already paid for the season is sunk when the firm weighs switching platforms, while the renewal ahead of it is not. Hours already written off on a fixed-fee engagement are sunk when a partner decides whether to finish the file, and only the hours still to be spent and the fee still to be collected carry that decision. The same question reaches the advisory side whenever a client asks whether to keep or replace equipment, where the book value of the old machine is the figure hardest to leave out.
Sunk Cost and Opportunity Cost
The two run in opposite directions in time. A sunk cost is a past outlay no current choice can change, while opportunity costs are the costs associated with not choosing the other alternative (OpenStax, Principles of Accounting, Volume 2, chapter 10 key terms), so they are created by the choice being made now and move with it. The two profits that opportunity cost separates, accounting profit and economic profit, are worked through in explicit vs implicit cost.
Common Questions
A non-refundable retainer already paid to a consultant. Implementation and training fees on software the firm has stopped using. Development spending on a service line that never launched. A signing bonus with no clawback clause, paid to someone who has since left. In each case the money is gone whatever happens next, so the identical amount sits under every option on the table.
The sunk cost fallacy is the name given to treating money already spent as a reason to keep going. The behavior is described plainly without the label, since dealing with sunk costs can be frustrating because it often means admitting an earlier error in judgment, and many firms find it hard to give up on a new product that is doing poorly because they spent so much money creating and launching it (OpenStax, Principles of Microeconomics 3e, 2.1).
There is no formula for it in the way there is one for marginal cost. Totaling a sunk cost means adding up what has already been spent on the thing being decided, then taking out whatever part of that is still recoverable. Say a firm has paid the amounts below on a practice management system it is now considering abandoning.
Amounts already paid
License fee, non-refundable 9,000
Data conversion, work already delivered 2,400
Staff training, already delivered 1,600
Deposit on unused seats 2,500
Total paid 15,500
Less deposit refundable on request (2,500)
Sunk to date 13,000That closing total is identical under every option still open, so it belongs in the record of what the system has cost and nowhere in the choice about what to do next.
Neither accounting nor economics has settled on a name for one. What the literature does name is the discipline the fallacy departs from, since the lesson of sunk costs is to forget about the money and time that is irretrievably gone and instead to focus on the marginal costs and benefits of current and future options (OpenStax, Principles of Microeconomics 3e, 2.1).
A firm carrying the cost of an offshore attempt that did not work is looking at a sunk cost, and Accountably’s trial task lets the firm’s own reviewer grade real work before any client file moves.
