What Are Commitments and Contingencies in Accounting?

In accounting, commitments and contingencies are two categories of future obligations that sit largely outside the recorded assets and liabilities on a company’s balance sheet but can still move the needle on its financial health. A commitment is a binding contractual obligation to do something in the future, like a signed ten-year supply contract, where the obligation itself is certain and only the timing of cash flows remains. A contingency is a potential obligation whose existence depends on a future event that may or may not happen, like a pending lawsuit. Under U.S. GAAP, commitments are handled primarily through footnote disclosure under ASC Topic 440, while contingencies follow ASC Topic 450, which uses a probability-based test to decide whether the item is accrued as a liability, disclosed in the footnotes, or left out entirely.

How Commitments Differ From Contingencies

The line between the two comes down to certainty. With a commitment, the company has already signed the agreement. It knows what it owes, to whom, and roughly when. A ten-year contract to buy raw materials at a fixed price is a commitment. So is an authorized capital expenditure that has not yet been incurred, or a take-or-pay arrangement with a supplier.

A contingency is different in kind. The obligation itself is uncertain. The company might owe nothing, or it might owe a great deal, and the answer only becomes clear when some future event occurs or fails to occur. The accounting standards define a loss contingency as an existing condition involving uncertainty about a possible loss that will be resolved by one or more future events.¹

That difference in certainty is what drives the different accounting treatment. A commitment, because it exists but has not yet been performed by either side, generally waits for actual delivery before creating a liability. A contingency, because the loss may already exist in substance even though it cannot be measured with confidence, may need to be recognized before any cash changes hands.

The Three Probability Levels for Loss Contingencies

ASC 450 sorts every loss contingency into one of three probability buckets, and each triggers a different response.

Probable. The future event is likely to occur. SFAS 5 defines this as “likely to occur,” generally interpreted in practice as roughly a 70 percent or higher likelihood. A probable loss must be accrued on the balance sheet when the amount can be reasonably estimated.¹
Reasonably possible. The chance of loss is more than remote but less than likely. These contingencies are not accrued, but they require footnote disclosure covering the nature of the exposure and an estimate of the potential loss or range of loss.
Remote. The chance of loss is slight. Remote contingencies generally require no disclosure, with narrow exceptions for certain guarantees.

The standard does not attach numerical thresholds to these categories. SFAS 5 kept the definitions qualitative on purpose, which means management has to exercise judgment on each contingency, often relying heavily on legal counsel for litigation matters.¹

When a Probable Loss Gets Accrued

Two conditions have to be met at the same time before a loss contingency is recorded on the balance sheet: the loss must be probable, and the amount must be reasonably estimable. If a loss is probable but genuinely cannot be estimated, no accrual is made. The company discloses the contingency in the footnotes and states that an estimate cannot be determined.²

Estimating a Range of Loss

Loss estimates rarely produce a single figure. When management identifies a range of possible outcomes and no single amount within the range is a better estimate than any other, the company must accrue the minimum. If the probable loss falls somewhere between $2 million and $8 million with no point more likely than another, the recorded liability is $2 million.² The footnotes then have to disclose the full range so readers can see the exposure above what was accrued.

If management does have a best estimate within the range, that specific amount gets accrued instead of the minimum. The minimum rule is a fallback, not a default.

Insurance Recoveries

Companies facing a probable loss often expect an insurance carrier to cover some or all of it. The recovery is not netted against the loss. The loss liability is measured on its own. If recovery from the insurer is itself probable, the company records a separate receivable asset, capped at the amount of the recognized loss, and evaluates whether an allowance for uncollectibility is needed based on the insurer’s ability to pay.

Any expected recovery above the recognized loss becomes a gain contingency and faces a higher recognition bar. When the recovery claim is in litigation, there is a rebuttable presumption that realization is not probable, which makes it harder to book the receivable until the dispute resolves.

Why Gain Contingencies Are Treated Differently

Potential gains are not recognized until they are realized or realizable. Even a lawsuit the company is confidently winning cannot be recorded as revenue before the proceeds arrive or the remaining uncertainties are substantially resolved.¹

This asymmetry is deliberate. The standards set a lower recognition bar for bad news than for good news because pulling a gain into the financial statements prematurely can mislead investors into thinking earnings have been secured. A gain contingency can be disclosed in the footnotes when realization is likely, but the language must stop short of suggesting the outcome is settled.

How Commitments Are Reported

Commitments arise from executory contracts, meaning neither party has yet performed. A five-year fixed-price purchase agreement is a commitment, but no liability appears on the balance sheet until the supplier actually delivers goods. Until then, nothing has been received and no payable has been triggered.

Off the balance sheet does not mean unimportant. A company locked into billions of dollars of future purchase obligations has a very different risk profile from one with flexible procurement, even if both balance sheets look similar. The accounting treatment centers on disclosure. Material commitments require footnote disclosure covering the nature, term, and amount of the obligation. For unconditional purchase obligations that meet certain criteria, the disclosure has to include the fixed and determinable portion of the obligation for each of the five fiscal years following the balance sheet date, along with the nature of any variable components and the amounts purchased under the obligation for each period presented in the income statement.

Analysts use those payment schedules to model liquidity needs and to estimate the present value of a company’s total commitments. Adding that present value to recorded debt often gives a more complete picture of leverage than the balance sheet alone.

Leases Moved Onto the Balance Sheet

Leases used to be one of the biggest categories of off-balance sheet commitments. ASC Topic 842 changed that by requiring both operating and finance leases to be recognized on the balance sheet, where the prior standard only required capital leases to appear there.³ Companies now record a right-of-use asset and a corresponding lease liability for essentially all leases with terms longer than twelve months. Footnote schedules of future lease payments remain important because analysts use them to model liquidity and compare leverage across firms with different lease structures.

Guarantees and Product Warranties

Guarantees sit in a category of their own. ASC Topic 460 requires a guarantor to recognize a liability at the inception of the guarantee, measured initially at fair value, even when the likelihood of ever having to perform is remote. For a guarantee issued in a standalone transaction, that fair value is typically equal to the premium received. When the guarantee is bundled into a larger transaction, the company has to estimate what premium it would have charged to issue the same guarantee on its own.²

Product warranties are treated differently. Warranties tied to the functional performance of a product the buyer owns are carved out of ASC 460’s initial recognition rules and follow the general loss contingency model under ASC 450 instead. A manufacturer estimates warranty costs at the time of sale from historical claims data, records the full estimated expense against revenue in the same period, and credits a warranty liability. Actual claims later reduce the liability rather than hitting the income statement a second time.

What the Footnotes Have to Say

The footnotes carry the information the balance sheet cannot. On the contingency side, any loss classified as reasonably possible needs disclosure of the nature of the contingency and an estimate of the potential loss or range of loss. If management cannot make a reasonable estimate, the footnotes have to say so explicitly.² Probable losses that have been accrued also require footnote disclosure describing the nature and circumstances of the accrual.

Generic language about litigation risk does not satisfy the standard. The SEC staff has reminded registrants in comment letters that when there is at least a reasonable possibility of a loss exceeding amounts already recognized, the company must disclose an estimate of the additional loss or range of loss, or affirmatively state that an estimate cannot be made.⁴ “The outcome cannot be predicted” without further analysis is not enough.

How IFRS Handles the Same Items

The framework described above is U.S. GAAP. Companies reporting under IFRS follow IAS 37, and the differences can produce materially different financial statements for identical facts.

Probability threshold. Under IFRS, “probable” means “more likely than not,” or greater than 50 percent. U.S. GAAP reads “probable” as “likely to occur,” generally taken to mean roughly 70 percent or higher. More contingencies get accrued under IFRS than under U.S. GAAP.⁵¹
Measurement when a range exists. When each amount in a range is equally likely, U.S. GAAP accrues the minimum. IFRS accrues the midpoint. For a $2 million to $10 million range, U.S. GAAP records $2 million and IFRS records $6 million.
Discounting. U.S. GAAP generally does not require loss contingencies to be discounted. IFRS requires provisions to be measured at present value when the time value of money is material, using a pretax discount rate.
Onerous contracts. IFRS requires a provision when a contract becomes onerous, meaning unavoidable costs of fulfilling it exceed the expected economic benefits. U.S. GAAP has no general onerous contract provision, though specific guidance exists in narrow areas such as leases.

A contingency that lives only in the footnotes under U.S. GAAP can appear as a recognized liability under IFRS, affecting equity, leverage ratios, and net income. That gap is worth keeping in mind whenever you compare financial statements across the two frameworks.

1
Financial Accounting Standards Board. Statement of Financial Accounting Standards No. 5
2
Financial Accounting Standards Board. Contingencies Topic 450 – Disclosure of Certain Loss Contingencies
3
Financial Accounting Standards Board. Leases
4
U.S. Securities and Exchange Commission. CORRESP – AU Optronics Corp
5
IFRS Foundation. IAS 37 Provisions, Contingent Liabilities and Contingent Assets