When a Tax Position Gets Challenged: What Practitioners Need to Know

The IRS closed 505,514 audits in fiscal year 2024 and recommended over $29 billion in additional tax off the back of them. That's the scale practitioners need in their heads before any conversation about process, because the process only matters once a client's return lands in that pool.

The position standards practitioners must have satisfied before a challenge can be effectively defended

Before anyone drafts a response to a notice, there's a prior question that determines how much room there is to maneuver: did the position meet its required confidence level when the return was filed? Tax law runs on a hierarchy of these thresholds, and knowing where a position sits on that ladder tells you, almost immediately, which penalty defenses survive contact with an examiner.

At the top sits "more likely than not," meaning better than a 50% shot at holding up if challenged. That's the bar for tax shelters and it's also the standard preparers need to meet under IRC § 6694(a) to avoid personal penalty exposure. One notch down is "substantial authority," an objective standard under Treasury Regulation § 1.6662-4 that is around a 40% likelihood of being sustained. It's a lower bar than "more likely than not" but a higher one than "reasonable basis," which is around 20 to 25% and is the floor for avoiding accuracy-related penalties, provided the position gets disclosed.

The AICPA layered its own standard on top of these in Statement on Standards for Tax Services No. 1.1, effective January 1, 2024: a CPA shouldn't recommend a position without at least a 33% ("realistic possibility") chance of it holding up when the taxing authority hasn't set its own standard, or when its standard is lower than 33%. Whichever standard is higher, that's the one the member has to follow.

One thing that doesn't factor into any of this, and this trips people up more than it should, is how likely the return is to get audited in the first place. Treasury Regulation § 1.6662-4 makes clear that audit odds are irrelevant to whether a position meets its standard. A practitioner who leans on "the IRS probably won't look at this" has already given up the argument before a notice even shows up.

The preparer penalty under IRC § 6694 is the greater of $1,000 or 50% of the income the preparer earned from that return, and it attaches when the preparer knew, or should have known, that a position would understate the client's liability. That standard climbs to "more likely than not" for shelters and reportable transactions. Circular 230 § 10.34 backs this up on the ethics side: a practitioner can't sign or advise a position that lacks substantial authority unless it has at least a reasonable basis and gets disclosed. Willful or reckless violations bring § 6694(b) penalties and can draw scrutiny from the oversight body that disciplines practitioners.

Documentation matters here in a way that's easy to underweight. Preparers have to document that they gave the client advice on likelihood of success and potential penalty exposure, and a boilerplate disclaimer at the bottom of an engagement letter doesn't satisfy that requirement. So when a notice does arrive, the first internal question a practitioner should ask isn't "how do we respond." It's "did this position meet its standard when we filed it." That answer decides which defenses are still on the table, and whether the practitioner is now exposed right alongside the client.

Disclosure as a pre-challenge tool that changes the penalty calculus before any notice is issued

Disclosure is probably the most underused tool in the whole risk-management kit. Filed correctly, it can knock out several accuracy-related penalties before they ever attach, and it shifts the penalty calculus before any notice is issued.

Form 8275 covers positions that aren't contrary to a specific regulation but need disclosure anyway, typically because the position clears the reasonable-basis bar but not much higher. It shields against the accuracy-related penalty for disregard of rules or substantial understatement on items that aren't tax shelters. Form 8275-R is the one to use when the position actually conflicts with a Treasury regulation, and using the wrong form doesn't get you adequate disclosure. It just gets you paperwork.

Disclosure isn't a universal shield, though, and treating it as one is a mistake. It won't cure negligence, substantial or gross valuation misstatements, tax shelter items, or understatements tied to undisclosed foreign financial assets. For transactions lacking economic substance under IRC § 7701(o), disclosure knocks the penalty down from 40% to the baseline 20% strict-liability rate, but it doesn't erase the penalty outright.

Where a position sits in that reasonable-basis range, roughly 20 to 25% likelihood, and doesn't fall into one of those excluded categories, disclosing at filing is close to a no-brainer. The cost is a line on the return. The cost of skipping it, if the position gets challenged and fails, is the full accuracy-related penalty. Once a challenge does show up, whether that disclosure statement exists is one of the first things an examiner or appeals officer checks. At that point it's part of the live argument. It's part of the live argument.

Identifying the nature of the challenge before deciding how to respond

Not every notice is the same animal, and responding to the wrong kind of dispute with the wrong tool wastes time the taxpayer doesn't have. Four categories tend to cover most of what these notices involve.

Substantiation disputes are the simplest: the position itself is sound, but the records backing it up are missing or thin. The fix is document production, nothing more exotic. Methodology disputes are different, they turn on whether the examiner's approach to computing or categorizing an item is flawed, and the response there is a technical rebuttal rather than a stack of receipts. Legal interpretation disputes go deeper still, the parties simply disagree about what the law requires, and that kind of disagreement often needs a formal protest, a trip to Appeals, or litigation to resolve. Penalty disputes sit in their own lane: the underlying tax adjustment might be accepted outright, but the penalty attached to it is contested on its own terms, usually through disclosure status or a reasonable-cause argument.

Math and clerical error notices are a distinct, earlier-stage category. The taxpayer gets 60 days to notify the IRS of disagreement and send correcting records before the adjustment turns into an actual assessment, per the Taxpayer Bill of Rights.

The audit format shapes the response environment too. Correspondence audits run entirely by mail and tend to have a narrow scope, but under current processing delays it's easy to let a deadline slip without noticing. Office audits happen at an IRS location with a defined scope and higher stakes than correspondence. Field audits are the broadest and most exposed: the examiner reviews books at the taxpayer's place of business, and that kind of access tends to raise issues nobody planned to discuss.

Taxpayers keep certain rights through all of this regardless of format. They can be represented by an attorney, CPA, or enrolled agent, and in most cases they don't have to show up personally at all. Audit scope is tied to specific issues and specific tax years, and it can't expand without justification.

Getting the classification right matters because the cost curve is steep. A substantiation dispute resolved at examination costs little beyond pulling together documents. A legal interpretation dispute that ends up in Tax Court costs orders of magnitude more, and the window to settle it cheaply at an earlier stage closes fast. With the IRS Independent Office of Appeals having lost more than a quarter of its staff, the agency is more likely to skip Appeals altogether and go straight to a statutory notice of deficiency. Practitioners should classify the dispute quickly and stop assuming the full administrative ladder will still be standing when they need it.

Managing the examination stage sets the ceiling for every stage that follows

The burden of substantiating deductions and credits sits with the taxpayer, full stop, and contemporaneous records are the whole foundation of that burden. A sloppy or over-eager initial response, one that hands over more than the examiner asked for, is one of the easiest ways to widen an audit that didn't need to widen.

Scope control matters here more than most practitioners give it credit for. The audit is tied to defined issues and defined tax years, and there's no obligation to volunteer anything outside that scope. Doing so anyway is probably the single most common way a narrow audit turns into a broad one.

There's a harder edge to this stage too: failing to raise an issue at examination can waive it for good. Issues that don't get developed here typically can't be introduced for the first time at Appeals or in Tax Court later. Given current processing delays, using the IRS Document Upload Tool for examination responses is worth doing wherever it's an option, since it gives confirmation of receipt and avoids the kind of mail delay that automated systems sometimes misread as no response at all.

Representation carries real weight at this stage. A taxpayer can send an attorney, CPA, or enrolled agent instead of appearing personally in most cases, which removes the risk of a client saying something damaging in casual conversation with an examiner.

What comes out of examination is a set of proposed adjustments, not an assessment. The taxpayer still has full rights to respond, provide more documentation, and disagree before anything escalates further. The record built during examination is the same record Appeals and Tax Court will work from later, which is why treating this stage as routine paperwork, rather than the opening act of a legal proceeding, tends to be the most expensive mistake in the whole process.

The IRS Independent Office of Appeals: what it offers, how to access it, and what it no longer reliably provides

In principle, Appeals gives a taxpayer an independent look at the examination's conclusions, run by personnel separate from the examining division and governed by ex parte communication rules meant to keep that independence real. The office handles a wide range of matters: audit adjustments, penalties, collection actions like liens and levies, wage garnishments, passport revocation, payroll tax disputes, innocent spouse relief, and rejected offers in compromise.

Once a notice proposing an adjustment arrives, practitioners generally have 30 days to file a formal protest or appeal request. Missing that window narrows the options considerably, though it doesn't necessarily close every administrative door. Collection matters run on a separate track called Collection Due Process, or CDP, hearings, which the IRS has to offer before levying or immediately after filing a federal tax lien. CDP has its own procedural rules, and an inadequately drafted CDP request can cost a taxpayer the judicial review that would otherwise follow.

The office isn't what it was. The Independent Office of Appeals lost more than a quarter of its staff, and the National Taxpayer Advocate's fiscal year 2026 objectives report warns that compliance-driven performance pressure risks turning Appeals into something closer to an extension of examination than an independent check on it. Practitioner commentary points to a bypass trend already underway: the IRS issuing statutory notices of deficiency directly, more often, pushing taxpayers straight into Tax Court without ever getting an administrative resolution chance. This is a documented operational shift, not speculation.

Congress has responded, at least partially. H.R. 6506, the Taxpayer Due Process Enhancement Act, passed the House on May 19, 2026. It would expand Tax Court jurisdiction to review the underlying tax liability in CDP hearings when properly disputed, bar the IRS from applying overpayments against a disputed liability while a CDP case is pending, and suspend the refund statute of limitations during those proceedings. As of May 30, 2026, the Senate hadn't taken it up, the bill was referred to the Senate Finance Committee on May 20, 2026, and no vote had been scheduled.

Appeals remains the cheapest route to resolution when it's available. But given the staffing losses and the bypass tendency, practitioners are better off building litigation-quality documentation from the start, even in cases they fully expect to settle administratively.

Tax Court and other judicial venues: when litigation becomes the necessary path

A statutory notice of deficiency, the "90-day letter," is what opens the door to Tax Court. The taxpayer has 90 days from the notice date to file a petition, or 150 days if the notice went to an address outside the United States, per the Taxpayer Bill of Rights. Miss that window, and the pre-payment judicial option is gone.

That pre-payment feature is what makes Tax Court distinctive. It's the only judicial forum where a taxpayer can contest an IRS assessment before paying the disputed tax. Every other route, district court or the Court of Federal Claims, requires paying first and suing for a refund afterward.

The Supreme Court narrowed Tax Court's reach in CDP cases with Commissioner v. Zuch in June 2025, eliminating jurisdiction in CDP proceedings once the IRS drops its levy pursuit. H.R. 6506 is Congress's direct answer to that ruling, but as of late May 2026 it hasn't become law, so practitioners handling CDP matters have to work within the narrowed jurisdiction Zuch left behind. Where Tax Court isn't available or isn't the better strategic fit, another federal forum may be the route to take. District Court offers a jury trial, which can matter in fact-heavy disputes, but it requires paying the tax first. The Court of Federal Claims also requires prepayment, though it sometimes carries more favorable precedent on particular issues.

The deadlines here carry no forgiveness. An untimely response to a statutory notice of deficiency ends the Tax Court option outright, and automated IRS systems don't extend deadlines just because the agency's own processing ran slow.

One newer risk deserves a flag: AI-related sanctions in litigation hit $145,000 in the first quarter of 2026 alone, tied to fabricated citations. That's a direct warning for anyone using AI drafting tools in controversy work. Circular 230 § 10.22 requires verifying anything AI-generated before it goes in front of a court, and § 10.36 requires firms to have compliance procedures in place, procedures that OPR Alert 2026-19 reads as including staff training on AI tools and documentation that the training happened.

Choosing a forum isn't a default setting. It's a decision that turns on the case's facts, the circuit's precedent, and whether a jury or a Tax Court judge is more likely to see things the taxpayer's way, and it's a decision that should get made with litigation counsel well before the petition deadline, not in the days right before it.

How the procedural map changes when penalties, rather than the underlying tax, are the central dispute

Accuracy-related penalties occur more than any other kind. They attach to substantial understatements, negligence, and disregard of rules, and the penalty itself runs at 20% of the underpayment tied to the violation.

The main defense against most of these is reasonable cause and good faith, and it requires showing that the taxpayer made a genuine effort to comply given the facts and circumstances at hand, not just that the position was defensible on paper. That defense stands or falls on the same groundwork laid out earlier in this piece: what standard the position met when filed, whether it got disclosed, and what the practitioner documented along the way. When the underlying tax adjustment isn't really in dispute but the penalty is, the fight shifts almost entirely onto that reasonable-cause ground, and the record built at examination becomes the deciding factor in whether it holds.