Uncertain Tax Position Documentation Standards Under ASC 740

By Keira O’Connell, Senior Tax Technical Writer
Reviewed by A. Sorensen, CPA, Corporate Tax Practice
Jurisdiction focus: United States GAAP (FASB ASC 740-10) and IRS federal tax administration. This piece does not address IFRS (IAS 12) or non-US tax authorities.

A lot of controllers assume uncertain tax position documentation is only for companies running aggressive shelters or offshore structures. That assumption is wrong, and it gets expensive. Any company claiming a research credit, allocating income across states, or pricing an intercompany transaction is making a judgment call the IRS could challenge later. ASC 740-10 does not care whether the position is aggressive. It cares whether you can show your work.

Why ASC 740-10 Documentation Exists

FASB Interpretation No. 48 was issued in July 2006 and became effective for fiscal years beginning after December 15, 2006, for public companies (nonpublic entities got a two-year deferral to fiscal years beginning after December 15, 2008, under FASB Staff Position FIN 48-3). The board’s concern was straightforward: companies were recording tax benefits for positions they privately expected to lose on audit, and financial statement users had no consistent way to tell which reserves were solid and which were hopeful. FIN 48 was later folded into the FASB Accounting Standards Codification as Subtopic 740-10 when the Codification became the single source of authoritative US GAAP in 2009, but the substance never changed. Practitioners still call it “FIN 48 analysis” out of habit.

The standard forces a company to separate two questions that used to get blurred together: is this position even defensible, and if so, how much of the benefit gets booked. Before FIN 48, some companies answered both questions with a single gut-check estimate. That is no longer permitted. Each step needs its own written support, and the support has to survive contact with an external auditor who was not in the room when the position was taken.

The Recognition Threshold: More Likely Than Not

Step one asks whether a tax position clears a 50% bar. ASC 740-10-25 requires that a position be recognized only if it is more likely than not (meaning a likelihood greater than 50%) that the position would be sustained on examination, based solely on its technical merits. Two words in that sentence carry most of the weight. “Technical merits” means the analysis rests on the tax law itself: statutes, regulations, case law, rulings. It explicitly excludes the odds of getting audited in the first place. A position cannot clear the threshold because a company assumes the IRS is unlikely to look at its return.

I think this is the single clause most tax teams misapply in practice, not because they disagree with it, but because detection-risk reasoning creeps in unconsciously. A memo that says “this issue rarely comes up on exam” has already drifted off the technical-merits track, even if nobody intended it to.

Documentation at this stage typically takes the form of a technical memo. A good one states the position plainly, walks through the controlling authority, addresses the strongest counterargument, and reaches a conclusion with a stated confidence level. A weak one just cites a code section and asserts the company qualifies. Auditors reviewing the tax provision will ask for the memo, not just the conclusion, and an unwritten “we’re pretty sure” does not count as evidence.

Positions that fail the more-likely-than-not test are not automatically written off entirely. If the position is one the company would even claim on the return in the first place, and it does not clear 50%, none of the benefit gets recognized in the financial statements. This is sometimes called the all-or-nothing test for unrecognized positions. That is a harsh outcome, and it is exactly why the recognition memo needs to be built before the return is filed rather than reconstructed months later when an auditor asks hard questions.

Measurement: The Cumulative Probability Method

Once a position clears recognition, step two sizes it. This is where the cumulative probability table comes in. ASC 740-10-30 requires a company to identify the possible settlement outcomes for the position, assign a probability to each one, and then recognize the largest amount of benefit that is greater than 50% likely of being realized on cumulative basis, not the single most probable outcome and not an expected value average.

That distinction trips people up constantly. Expected value would average every outcome weighted by its probability. Cumulative probability instead stacks outcomes from most favorable to least favorable and finds the point where the running total first crosses 50%. In practice this usually produces a more conservative number than a simple average would, particularly when there is a wide range of possible settlement amounts.

Building this table requires real inputs: prior audit settlement patterns for similar issues, the strength of the technical position at each dollar level, and any relevant precedent from the company’s own examination history or industry experience. A table with three made-up percentages that were never tied to anything is not measurement. It is decoration. Examiners and auditors both know the difference, and reviewers increasingly ask where each probability number came from.

Documentation Requirements: What Reviewers Actually Expect

Contemporaneous documentation beats reconstructed documentation every time. A memo dated the same quarter the position was taken carries more weight than one assembled eighteen months later after the audit letter arrives, because the earlier version cannot be accused of hindsight bias. Reviewers know this, which is why the timing of the file matters almost as much as its content.

A complete uncertain tax position file generally includes the recognition memo with cited authority, the measurement table with probability sources noted, a record of any changes in judgment from prior periods and why they occurred, and a reconciliation showing how the reserve rolls forward from opening balance to closing balance each period. Interest and penalty accruals need their own line item and their own support, since ASC 740-10 requires accrual of both once a position is recognized as uncertain.

Research and development tax credit positions under IRC Section 41 are one of the most common sources of uncertain tax positions that require exactly this discipline. Qualified research expenses depend on subjective calls about what counts as a process of experimentation, and the four-part test leaves real room for disagreement between a company and an examiner. Teams that want a closer look at how that discipline works in practice for R&D claims specifically can review our companion piece on research credit documentation standards, which covers the qualified research expense substantiation trail in more depth.

Documentation quality gets judged retrospectively, often years after the position was taken, by someone who was not involved in the original analysis. Write for that reader. A memo that makes sense only to the person who wrote it has already failed its main job.

Interaction With IRS Schedule UTP

Schedule UTP is a separate filing requirement layered on top of the ASC 740-10 analysis, and it asks a different question. Instead of a probability estimate, it asks for a concise description of the position and a ranking by size. Corporations filing Form 1120, 1120-F, 1120-L, or 1120-PC must file Schedule UTP if total assets equal or exceed $10 million for the tax year, and the position must be one for which the company recorded a reserve (or decided not to record one specifically because it expects to litigate) in audited financial statements.

The asset threshold has moved over time. When Schedule UTP launched for 2010 tax years, only corporations with total assets of $100 million or more had to file. The threshold dropped to $50 million for 2012 tax years, then to $10 million for 2014 tax years and has stayed there since, pulling a much larger population of mid-size corporations into the filing requirement than the original design contemplated. That phase-in history is worth understanding because it explains why so many companies that never worried about Schedule UTP a decade ago now file it every year.

A position can sit comfortably under ASC 740-10 and still create a Schedule UTP disclosure obligation, because the two regimes measure different things. A reserved position with a well-documented recognition memo is exactly the kind of position Schedule UTP is designed to surface. Good ASC 740-10 files make the Schedule UTP concise description easier to write, not harder, since most of the technical explanation already exists.

I think the 2022 overhaul of the concise-description requirement did the profession a favor, even though it added work. Vague descriptions used to technically satisfy the form while telling a reviewer almost nothing about the actual position. Forcing a specific Code section, a specific regulation citation, and a specific dollar figure into the description exposed how many companies had been coasting on boilerplate language that checked a box without explaining anything real.

The IRS revised the Schedule UTP instructions effective for tax year 2022 filings, adding five new data elements to the concise description requirement: the applicable Code section, the relevant regulation citation, any governing revenue ruling or revenue procedure, the specific form and line number where the position appears, and the dollar amount of the position. Positions ranked as “major,” meaning their size represents 10% or more of the total size of all reported UTP positions, get flagged for closer review inside the schedule itself.

Position Type, Recognition Likelihood, and Documentation Depth

Not every uncertain position deserves the same file thickness. Some categories carry inherently higher technical risk and need correspondingly deeper support, while others are narrower judgment calls that can be documented more briefly without losing rigor.

The pattern across these five categories is not subtle. Positions that depend on someone else’s numbers, such as a transfer price, a comparable set, or a multi-state apportionment factor, need deeper files than positions built almost entirely on a company’s own contractual facts.

A Simplified Worked Example

Assume a mid-size manufacturer claims a $2,000,000 research credit under Section 41 for a tax year. The tax team’s recognition memo concludes the position is 65% likely to be sustained on technical merits, clearing the more-likely-than-not threshold, so recognition is permitted.

Measurement then builds a cumulative probability table for possible settlement outcomes if the position were examined:

• 100% of the credit sustained ($2,000,000): 30% probability
• 75% of the credit sustained ($1,500,000): 25% probability, cumulative 55%
• 50% of the credit sustained ($1,000,000): 20% probability, cumulative 75%
• 25% of the credit sustained ($500,000): 15% probability, cumulative 90%
• 0% sustained: 10% probability, cumulative 100%

Reading down the table, cumulative probability first exceeds 50% at the $1,500,000 outcome (30% plus 25% equals 55%). Under ASC 740-10-30, the company recognizes $1,500,000 as the tax benefit, not the full $2,000,000 claimed on the return and not a probability-weighted average of roughly $1,375,000. The unrecognized $500,000 becomes a liability for unrecognized tax benefits on the balance sheet, carried until the position is resolved, expires under the statute of limitations, or new information changes the analysis.

This is the step where the two approaches diverge most visibly. An expected-value calculation would have produced a lower number here, and using it instead of the cumulative method is one of the most common measurement errors reviewers flag.

Common Mistakes in Uncertain Tax Position Files

Some failure patterns show up repeatedly across audit cycles and industries.

Confusing detection risk with technical merit. A company that reasons “the IRS rarely audits this issue, so we’re fine” has answered the wrong question entirely. ASC 740-10 explicitly bars audit-risk reasoning from the recognition analysis.

Using expected value instead of cumulative probability. The worked example above shows why this matters: the two methods can produce materially different recognized amounts, and auditors check this specifically.

Writing the memo after the audit letter arrives. Reconstructed documentation invites hindsight-bias challenges. A memo dated close to the return filing date is far more credible than one dated close to an IRS notice.

Treating Schedule UTP and ASC 740-10 as the same exercise. They share inputs but ask different questions. A position can be well documented for financial statement purposes and still get an incomplete concise description on Schedule UTP if the team assumes one file covers both needs.

Letting probability tables go stale. A measurement table built two years ago, before a relevant court decision or a new revenue ruling, should get revisited. Facts and law both move, and the file needs to move with them.

In my view, the single biggest failure point is not the technical tax analysis at all. Most tax teams can build a defensible legal argument. What breaks down is the paper trail proving the argument existed at the time the position was taken, in the form it now claims to have existed in. I’d rather see a company take a slightly weaker position with airtight contemporaneous documentation than a strong position supported by a memo drafted from memory eighteen months later.

Practical Documentation Checklist

A working file for any uncertain tax position should be able to answer each of these before the return goes out the door:

• Is there a dated recognition memo citing the controlling statute, regulation, and case law?
• Does the memo address the strongest counterargument, not just the favorable authority?
• Is there a cumulative probability table with sourced inputs, not assumed percentages?
• Does the file show the rollforward of the reserve balance period over period?
• Are interest and penalty accruals calculated and supported separately from the principal reserve?
• Has someone checked whether the position triggers a Schedule UTP filing obligation given current asset levels?
• If it does, is the concise description drafted using the post-2022 five-element format?
• Has the file been reviewed by someone outside the original preparer, ideally before quarter close?

None of these questions are exotic. They are the questions an external auditor will ask anyway, so answering them first internally saves a round of scrambling during the provision review.

Frequently Asked Questions

What does “more likely than not” mean under ASC 740-10?
It means a likelihood greater than 50% that a tax position would be sustained on examination, judged solely on the technical merits of the tax law involved. The odds of actually being audited play no role in this determination.

How is the tax benefit measured once a position clears recognition?
Using a cumulative probability table of possible settlement outcomes. The company recognizes the largest benefit amount that is greater than 50% likely of being realized on a cumulative basis, working down from the most favorable outcome.

Does documenting a position for ASC 740-10 also satisfy Schedule UTP?
Not automatically. Schedule UTP requires its own concise description with specific elements (Code section, regulation citation, form and line number, and dollar amount), separate from the recognition memo, though the underlying technical analysis usually overlaps heavily.

What is the current asset threshold for filing Schedule UTP?
Corporations filing Form 1120, 1120-F, 1120-L, or 1120-PC must file Schedule UTP if total assets equal or exceed $10 million for the tax year. That threshold has applied since 2014 tax years, down from $100 million when the schedule launched for 2010 tax years.

What happens to the unrecognized portion of a tax benefit?
It becomes a liability for unrecognized tax benefits on the balance sheet. That liability stays until the position is resolved with the taxing authority, the statute of limitations expires, or new facts or law change the recognition or measurement conclusion.

Is expected value the same as the cumulative probability method?
No, and treating them as interchangeable is a common error. Expected value averages every outcome by its probability. The cumulative method instead finds the largest outcome amount whose running probability, stacked from most to least favorable, first exceeds 50%. The two methods frequently produce different numbers.

References

• FASB Accounting Standards Codification Topic 740-10, Income Taxes: Overall (originally FASB Interpretation No. 48, issued July 2006).
• Internal Revenue Service, Instructions for Schedule UTP (Form 1120), Rev. December 2022.
• Internal Revenue Service, “Uncertain Tax Positions: Schedule UTP,” irs.gov/businesses/corporations/uncertain-tax-positions-schedule-utp.
• Journal of Accountancy, “IRS Amends Plan for Uncertain Tax Positions,” November 2010 (secondary analysis of Schedule UTP asset threshold phase-in).
• Bean, Kinney & Korman, “FIN 48: Will It Be the Feared Road Map for IRS?” (secondary analysis of FIN 48 recognition and measurement mechanics).
• PwC Viewpoint, “15.3 Recognition of benefits from uncertain tax positions” (secondary Big Four technical guidance).