Behind almost every uncertain or aggressive tax position sits a question the taxpayer cannot answer alone: if the IRS disagrees, how likely am I to win — and what protects me if I lose? A tax opinion is the formal written answer. It is a tax professional’s reasoned analysis of a specific position that applies the governing authorities to the facts and concludes, at a stated level of confidence, how the position should be treated. Done properly, it is far more than comfort — it can be the document that stands between a taxpayer and a penalty.
What a tax opinion actually is
An opinion is not a one-line blessing. A properly constructed opinion has a recognizable anatomy: a statement of the facts it relies on and the assumptions and representations it is built on; the issue or issues presented; an analysis of the authorities on both sides — statutes, regulations, cases, and rulings, weighed for relevance and persuasiveness rather than cherry-picked; the application of that law to the facts; and a conclusion stated at a defined level of confidence. That last element is the part clients fixate on, and the part most often misunderstood.
The confidence ladder
Tax practice uses a ladder of confidence standards, from “frivolous” at the bottom to “will” at the top. Two cautions before the ladder itself. First, the percentages everyone quotes — “substantial authority is about forty percent” — are practitioner rules of thumb, not law. No statute or regulation attaches a number to these standards; the figures are a shared shorthand for conveying odds. Second, the two standards that carry the most legal weight — substantial authority and reasonable basis — are objective. They turn on the actual weight of the authorities as the IRS or a court would assess them, not on how confident the taxpayer or the advisor happened to feel. An opinion can conclude that substantial authority exists; whether it actually exists is decided independently, and the mere existence of an opinion does not bind the Service.
| Level | Common odds | What it means / standard | Where it matters most |
|---|---|---|---|
| Will | ~90%+ | Near-certainty; reserved for positions where the law is clear | Highest comfort; rarely written |
| Should | ~70% | High confidence, well above even odds | Strong comfort and marketed-transaction opinions |
| More likely than not | >50% | Better-than-even chance on the merits | Required for tax shelters & reportable transactions; FIN 48 recognition |
| Substantial authority | ~40% | Objective weight of authorities; Reg. § 1.6662-4(d) | Avoids the substantial-understatement penalty with no disclosure (non-shelter) |
| Reasonable basis | ~25–35% | Reasonably based on one or more authorities; Reg. § 1.6662-3(b)(3) | Avoids negligence; avoids substantial understatement if disclosed |
| Not frivolous | ~10–20% | Not patently improper | Below the line for penalty protection |
| Frivolous | under 10% | Patently improper | Never a proper basis; can draw the § 6702 penalty |
At the top, a “will” opinion expresses near-certainty and is reserved for positions where the law is clear; a “should” opinion signals a high level of confidence comfortably above even odds. More likely than not — a greater-than-fifty-percent chance of prevailing on the merits — is the pivotal rung, because it is the threshold the Code itself demands for the riskiest positions. Below it, substantial authority and reasonable basis are the workhorses of ordinary return positions. Not frivolous and frivolous sit at the bottom and are essentially never the stated basis of a written opinion — a frivolous position is one that is patently improper and can itself draw a penalty under § 6702.
What an opinion is for
Clients commission opinions for several overlapping reasons. Penalty protection is the one most often cited: a well-supported opinion is the principal way to establish the reasonable-cause-and-good-faith defense and to meet the substantive thresholds that switch off the accuracy-related penalty. A substitute for disclosure: a substantial-authority opinion can remove the need to flag a position to the IRS on a disclosure statement. Comfort and diligence: a “comfort” opinion assures the taxpayer that a respected advisor has thought the issue through before capital is committed. Direction and cover for the return preparer, who must meet professional standards of his own before signing. And third-party reliance — lenders, investors, and counterparties in a structured transaction frequently require an opinion before they will participate.
How an opinion delivers penalty protection
The accuracy-related penalty under § 6662 is twenty percent of the underpayment (forty percent for gross valuation misstatements, undisclosed foreign assets, and certain non-economic-substance items). Its most common trigger is a substantial understatement of income tax. For an ordinary, non-shelter position, that penalty is avoided in either of two ways: there is substantial authority for the position, or the position has a reasonable basis and is adequately disclosed. A separate trigger, negligence, is avoided where the position has at least a reasonable basis. This is the machinery a tax opinion is built to engage.
Above those item-level rules sits the overarching defense: the reasonable cause and good faith exception of § 6664(c). Reasonable reliance on a competent professional’s opinion can establish it — but the regulation under § 1.6664-4 is exacting. The determination is made case by case; the most important factor is the extent of the taxpayer’s own effort to assess the correct liability; and reliance on an advisor “does not necessarily” demonstrate good faith. To count, the reliance must be reasonable, the advisor competent and free of a disqualifying conflict, and the opinion based on all the pertinent facts — not on assumptions the taxpayer knows to be untrue.
The bar rises sharply for tax shelters and reportable transactions. There, disclosure buys nothing — it does not reduce the understatement — and the reasonable-cause route of § 6664(d) requires more: adequate disclosure, substantial authority, and the taxpayer’s reasonable belief that the position was more likely than not the correct treatment. This is why “more likely than not” is the standard a serious shelter opinion must reach.
The standards also reach the preparer. Under § 6694 and Reg. § 1.6694-2, a preparer needs substantial authority for an undisclosed non-shelter position, reasonable basis for one that is disclosed, and a reasonable belief that a tax-shelter or reportable-transaction position is more likely than not sustainable. A sound opinion therefore protects the advisor who signs the return as well as the client who files it.
Circular 230: the rules that govern the opinion writer
How an opinion must be written is governed by Circular 230 (31 C.F.R. Part 10), the Treasury rules for practice before the IRS. The most important recent change is one many taxpayers have not caught up to. The old § 10.35 “covered opinion” regime — a maze of prescriptive requirements that spawned the disclaimer at the bottom of every tax-related email — was eliminated in 2014 (T.D. 9668, effective for advice rendered on or after June 12, 2014). In its place is a single, principles-based standard in § 10.37: the practitioner must base written advice on reasonable factual and legal assumptions, exercise reasonable reliance, consider all relevant facts he knows or should know, and relate the law to the facts. Notably, the practitioner may not take the likelihood of audit into account — there is no “audit lottery.” Section 10.34 separately sets the standards for positions taken on a return.
Who may rely on an opinion
An opinion is written for a particular client, and almost every well-drafted opinion says so expressly. A standard reliance and limitations paragraph provides that the opinion is rendered solely for the addressee, may not be relied upon by anyone else, may not be quoted or furnished to third parties without the firm’s consent, speaks only as of its date (with no obligation to update it for later changes in law or fact), and is confined to the specific facts, assumptions, representations, transaction, and taxes it identifies. None of that is boilerplate for its own sake. The opinion-giver owes its professional duty to the client; open-ended reliance would extend that exposure to strangers the advisor never agreed to serve, and the analysis is only as good as the facts and assumptions it was built on — lift it onto a different transaction and it no longer means anything.
Third parties may rely only when the opinion lets them, and that permission is given deliberately. The opinion may name additional permitted relying parties — a lender, an investor, a buyer — or the firm may issue a separate reliance letter extending reliance to a named party on stated conditions. Transactional opinions in financings and securities offerings are written this way as a matter of course, with the offering documents defining exactly who may rely and for what. Even then, a permitted party relies only within the opinion’s four corners: as of its date, on the stated facts and assumptions, and for the stated purpose.
For penalty purposes the sharper question is not who is allowed to rely, but whose reliance actually protects them — and the bar is real. Under Reg. § 1.6664-4(c), reliance on an opinion establishes reasonable cause only if the advice was based on all pertinent facts and a correct understanding of the law, did not rest on unreasonable factual or legal assumptions, and did not unreasonably rely on representations the advisor knew or had reason to know were untrue. The Tax Court applies a practical three-part test: the advisor was competent and had the expertise to justify reliance, the taxpayer supplied complete and accurate information, and the taxpayer actually relied in good faith. An opinion obtained on shaded facts, or filed away unread, protects no one.
The promoter’s opinion: often worth nothing
This warning deserves its own heading, because it is where taxpayers are most often misled. When a transaction is being sold — a shelter, a strategy, a packaged structure — an opinion frequently comes bundled with it, written by the same firm that designed and marketed the deal. For penalty purposes that opinion is usually worth little or nothing, and the law says so from two directions.
The first is statutory and governs reportable transactions. A material advisor — under § 6111(b)(1), broadly anyone who gives material aid, assistance, or advice in organizing, managing, promoting, selling, or implementing a reportable transaction and earns more than a threshold fee for it (roughly $50,000 where the benefits run mainly to individuals, $250,000 otherwise, and as little as $10,000 for a listed transaction) — is exactly the advisor on whom § 6664(d) bars reliance. An opinion from such a promoter, from anyone the promoter pays, or from anyone whose fee turns on the tax benefits surviving, is a disqualified opinion that cannot establish the reasonable belief the statute demands. The same is true of an opinion built on unreasonable assumptions, or one that leans on the taxpayer’s representations without testing them.
The second reason is broader and reaches ordinary positions, not just reportable transactions. Courts have held for decades that reliance on a tax professional is reasonable only where the advisor is competent and independent — free of a conflict of interest. The governing three-part test comes from Neonatology Associates, P.A. v. Commissioner, 115 T.C. 43 (2000), aff’d, 299 F.3d 221 (3d Cir. 2002): the advisor must be a competent professional with sufficient expertise, the taxpayer must give that advisor complete and accurate information, and the taxpayer must actually rely in good faith. The advice must generally come from an independent advisor unburdened by a conflict — not from a promoter of the transaction, as the Sixth Circuit put it in Mortensen v. Commissioner, 440 F.3d 375 (6th Cir. 2006). In 106 Ltd. v. Commissioner, 136 T.C. 67 (2011), the Tax Court refused to let an investor rely on the “more likely than not” opinion of the very lawyer who had structured and pitched his shelter.
The red flags are practical and consistent: the opinion arrives with the product instead of from counsel the taxpayer chose; one firm designed, sold, and blessed the deal; the fee is contingent or scaled to the tax savings; and the identical opinion is handed to every buyer. The Third Circuit captured the intuition in Neonatology when it questioned why a taxpayer should take comfort in the opinion of a professional who is paid to promote the very deal the opinion approves. Almost always, he should not. Real penalty protection comes from an independent advisor the taxpayer engages on its own — working from the taxpayer’s actual facts, with no stake in the conclusion. That is the entire point of obtaining your own opinion rather than accepting the one that came in the sales packet.
None of this means an advisor who helps design a plan cannot opine on it. The disqualification is aimed at promoters — the marketing of a packaged shelter to many buyers — and at advisors with a financial stake in the result, such as a fee contingent on the tax savings or an interest in the transaction itself. The Tax Court has been careful here: recognizing that, read literally, the “promoter” label could swallow ordinary tax planning, it confined the label to marketed shelters offered to numerous parties. A client’s own retained counsel — engaged to structure legitimate planning that is not a reportable transaction, advising on the client’s actual facts, and paid a normal professional fee rather than a share of the tax benefit — is not a disqualified promoter. That advisor is precisely the competent, independent professional whose opinion the reliance defense is meant to credit. The line is the advisor’s stake: a normal fee for tailored advice is fine; a fee that rises and falls with the tax result, or an interest in the deal, is the problem.
What an opinion does not do
An opinion is a reasoned judgment, not a guarantee, and three limits are worth stating plainly. It does not bind the IRS or a court: because substantial authority and reasonable basis are objective, the opinion’s conclusion can be wrong and the underlying standard simply not met. Its penalty protection depends on genuine reliance: an opinion procured on facts the taxpayer knows to be false, or from an advisor with a stake in the deal, will not establish good faith. And it offers no defense at all against the strict-liability penalty for transactions lacking economic substance under § 7701(o) — the reasonable-cause exception does not apply to those underpayments, so no opinion, however strong, can avert the penalty (§ 6662(b)(6), (i)).
The practical takeaway
A tax opinion is most valuable before the return is filed, when the position can still be shaped to reach the confidence level the situation requires — substantial authority to file without disclosure, reasonable basis paired with a Form 8275, or “more likely than not” for anything that smells like a shelter. The right standard is a function of the stakes, the strength of the authorities, and whether disclosure is an option. Getting that judgment right, and documenting it properly, is where experienced tax counsel earns its keep.