Sunk Cost

A sunk cost is a cost that has already been incurred and cannot be recovered through the decision now being considered. Because the amount will not change whether the decision-maker continues, stops, sells, or switches, it should be excluded from the forward-looking comparison of those alternatives.

Past spending can still matter for accounting, taxes, performance review, legal obligations, and learning. Calling it sunk means it does not change between the available choices; it does not mean the history should be erased or that accountability no longer matters.

Key Takeaways

• A cost is sunk only to the extent that it has been incurred and is unrecoverable at the decision date.
• Recoverable sale proceeds, cancellation refunds, transferable rights, or alternative uses mean part of an earlier outlay may not be sunk.
• Continue-or-stop decisions should compare future incremental cash flows, risks, and Opportunity Costs.
• Book value, tax basis, and historical purchase price are not automatically sunk-cost measures.
• A future payment can be committed or unavoidable without being a past sunk outlay. It may nevertheless be irrelevant if every alternative requires the same payment.
• The fact that a project has consumed substantial resources does not establish that continuing will recover them.
• Past results can update forecasts, reveal execution quality, or trigger legal and governance review even when the past cash outlays are excluded from NPV.
• Abandoning a project can require exit costs and can generate salvage proceeds; both are future decision-relevant amounts.
• Ignoring sunk costs does not mean ignoring reputation, customer obligations, employee effects, contractual duties, or strategic consequences that will occur in the future.
• The related Sunk Cost Fallacy is a behavioral error, while sunk cost itself is a cost classification.

The Forward-Looking Decision Rule

For a continue-or-stop decision, compare the future net value of each alternative:

The prior unrecoverable expenditure appears in neither expression because it is the same under both alternatives. The preferred choice under this simplified financial rule is the one with the higher future net value, subject to risk, legal duties, financing constraints, and nonfinancial objectives.

This is not the same as asking whether the project has earned back its total historical cost. That question can be useful for performance evaluation, but it does not change which future action creates more value now.

Worked Example: Continue or Stop a Software Project

Assume a company has already spent $4 million developing a software platform. The $4 million cannot be refunded or recovered by selling completed work.

At the review date, management estimates:

The forward-looking values are:

Stopping is estimated to preserve $450,000 more value than continuing:

The $4 million already spent is important when reporting the project’s total loss and reviewing how the original decision was made. It does not become recoverable merely because another $1.5 million is spent. Under the stated assumptions, including the $4 million on both sides would not change the ranking, but it would obscure the incremental choice.

The forecast may still be wrong. Management should test launch probability, future operating benefits, additional delays, legal obligations, cybersecurity requirements, alternative uses of the team, and the option value of pausing rather than stopping permanently. This example is educational, not a recommendation about an actual project.

When Is a Cost Truly Sunk?

Use four tests:

1. Has the cost already been incurred? A planned budget is not a sunk cost.
2. Can any amount be recovered? Refunds, resale, insurance, indemnification, or transferable rights can reduce the sunk portion.
3. Does the amount change across alternatives? If stopping avoids a future payment, that payment is relevant even if the original contract was signed earlier.
4. What is the decision date and scope? A cost can be sunk for today’s operating decision but still relevant to a later tax, legal, performance, or reporting analysis.

Suppose a machine cost $1 million and can now be sold for $300,000 net. The entire $1 million historical purchase price is not the useful measure for a keep-or-sell decision. The $300,000 sale alternative is recoverable and therefore represents an Opportunity Cost of keeping the machine. Depending on the purpose, the unrecoverable portion of historical outlay may be described as sunk, but current cash flows and alternatives drive the decision.

Sunk Cost Versus Related Cost Concepts

The classification is decision-specific. A lease payment due next month may be unavoidable under both alternatives, making it irrelevant to the ranking, but it is not literally a past expenditure. A refundable deposit was paid in the past but is not fully sunk because recovery is possible.

Sunk Costs in Capital Budgeting

Capital Budgeting should use cash flows that change because a project is accepted, continued, expanded, delayed, or abandoned. Past feasibility studies, research, or application costs may be sunk at the approval date if they cannot be recovered.

However, several nearby items remain relevant:

• additional construction or development spending;
• working capital that can be invested or released;
• sale or salvage proceeds;
• shutdown, remediation, severance, or contract-termination costs;
• taxes caused by sale, closure, write-off, or future income;
• cash flows displaced from another use of assets or staff;
• financing or liquidity constraints; and
• the value of waiting for more information.

A proper Net Present Value analysis separates those future amounts from past unrecoverable spending. It should also avoid treating accounting depreciation as an incremental cash outflow while still recognizing any related future tax effects.

Sunk Costs in Investing

Once an asset is purchased, its historical price does not determine its expected return from today’s market value. A hold-or-sell decision should compare expected future after-tax cash flows, risk, liquidity, transaction costs, portfolio role, and feasible alternatives.

Historical cost can still matter because:

• tax basis may affect future tax consequences;
• legal, mandate, or accounting rules may use carrying values;
• transaction records support performance measurement and controls;
• selling can crystallize contractual, tax, or reporting effects; and
• the original investment thesis can provide evidence about forecast quality.

These are future consequences or review uses, not reasons to assume the market price must return to the purchase price. This distinction is educational and does not determine whether a reader should hold or sell an investment.

Sunk Costs in Lending and Restructuring

A lender that has already advanced funds cannot recover them by making an uneconomic additional advance. The next decision should compare expected recovery under funding, restructuring, enforcement, sale, or other feasible alternatives, including legal cost, collateral value, priority, timing, and execution risk.

Past exposure remains central to accounting, provisioning, capital, documentation, and recovery analysis. Calling the advance sunk for a narrow incremental decision does not mean the claim has no value or that contractual rights should be ignored.

Similarly, a company should not continue a weak project solely because debt was raised to fund it. Financing obligations and operating choices interact, but the funds’ source does not make future negative-NPV spending value-creating.

Sunk Costs in Pricing and Operations

Past development or equipment spending can explain why a product exists, but short-run pricing decisions often depend on future incremental cost, capacity, demand, strategic effects, and legal constraints. In the long run, a business must recover enough total value to justify remaining in a market and replacing assets.

“Ignore sunk costs” therefore does not mean price below relevant cost or disregard long-run economics. It means do not force today’s customer, product, or project decision to recover a past amount that cannot be changed by that specific choice.

Information From Past Spending

Sunk cash flows are excluded from the incremental calculation, but the experience that produced them can be valuable evidence. A project that missed every milestone may justify lower confidence in the next forecast. A successful prototype may increase the probability of technical completion. Supplier disputes may reveal new legal or execution risk.

Good analysis separates:

• the amount already spent, which is sunk if unrecoverable; from
• what the spending revealed, which can change future probabilities, costs, benefits, controls, or decision rights.

This distinction prevents two errors: continuing merely to justify the past and ignoring genuine new information because it arose during a failed phase.

HM Treasury’s Green Book 2026 states that appraisal should exclude costs already incurred and unchangeable while considering the opportunity cost of resources already paid for. The principle is useful beyond public appraisal, but company, investment, tax, and accounting decisions require context-specific rules and evidence.

How to Review a Sunk-Cost Classification

1. State the current decision. Continue, stop, sell, replace, expand, delay, or restructure are different comparisons.
2. Fix the decision date. Classify each cash flow based on what can still change at that date.
3. Separate paid from committed. Identify future obligations even if the contract already exists.
4. Test recoverability. Estimate refunds, sale proceeds, insurance, indemnities, transferable rights, and salvage value.
5. List future alternatives. Include exit, pause, staged funding, sale, partnership, or redeployment where feasible.
6. Measure incremental cash flows. Use cash flows that differ among alternatives and discount them consistently.
7. Include opportunity cost. Value alternative uses of assets, people, capacity, capital, and time.
8. Model exit effects. Include closure, remediation, severance, tax, legal, customer, and reputational consequences that occur in the future.
9. Use updated evidence. Revise probabilities and forecasts based on what prior phases revealed.
10. Keep separate records. Report total historical performance and the forward-looking recommendation without mixing their decision purposes.

The U.S. Government Accountability Office’s Cost Estimating and Assessment Guide emphasizes defined scope, technical baselines, assumptions, data, risk analysis, documentation, and updates with actual costs. Those practices improve the forecast inputs but do not remove uncertainty from a stop-or-continue decision.

Risks and Limitations

• Recoverability uncertainty: an asset may have a buyer in theory but little net value after time and transaction costs.
• Exit-cost omission: stopping can trigger penalties, remediation, customer support, severance, or legal obligations.
• Tax and accounting effects: historical amounts can affect future taxes, impairment, reporting, or covenants even when excluded from project NPV.
• Strategic dependencies: abandoning one project can affect another product, contract, platform, or regulatory approval.
• Option value: delaying, staging, redesigning, or selling a project may be better than a binary continue-or-stop choice.
• Forecast error: future benefits and costs can be more uncertain than the known historical spend.
• Reputation and trust: honoring a commitment can produce future value even though past spending itself is sunk.
• Organizational incentives: managers may classify costs strategically to support continuation or cancellation.
• Scope mismatch: a cost can be sunk for one unit’s decision but relevant to the enterprise or another stakeholder.
• Nonfinancial requirements: safety, law, ethics, and public obligations cannot be overridden by a narrow financial calculation.

Common Mistakes

• Calling every historical cost sunk without testing whether any amount is recoverable.
• Treating book value as cash that must be recovered by the next decision.
• Omitting future cancellation, closure, remediation, or tax effects.
• Including the same historical expenditure in both alternatives and presenting it as decision-relevant.
• Ignoring what past failures reveal about future forecasts.
• Treating an unavoidable future commitment as if it has already been paid.
• Assuming an owned asset has no cost when it could be sold or used elsewhere.
• Using “sunk cost” to avoid accountability for the original decision.
• Continuing only to reach an accounting break-even point.
• Assuming that stopping is always rational once any forecast deteriorates.

Authoritative Sources

• HM Treasury: The Green Book 2026
• U.S. Office of Management and Budget: Circular A-4
• U.S. Government Accountability Office: Cost Estimating and Assessment Guide
• Arkes and Blumer: The Psychology of Sunk Cost

The public guidance supports forward-looking appraisal and disciplined estimation; the research article addresses behavioral response to prior investment. None provides a universal decision rule for a specific company, security, contract, or jurisdiction.

Related Terms

• Opportunity Cost: Value of the best feasible alternative forgone by the current decision.
• Sunk Cost Fallacy: Behavioral tendency to continue because of past unrecoverable investment.
• Net Present Value: Present value of expected cash inflows minus present value of cash outflows.
• Capital Budgeting: Process for evaluating and selecting long-term investments.
• Cost-Benefit Analysis: Structured comparison of an alternative’s expected benefits and costs.
• Economic Profit: Profit after deducting explicit and implicit economic costs.
• Agency Cost: Monitoring, bonding, and residual loss created by delegated decision-making.

FAQs

This article provides general economic and financial education. It is not a project recommendation, valuation conclusion, impairment analysis, lending decision, or individualized investment, tax, legal, accounting, or regulatory advice.