Preparer penalties under Section 6694: what signing a specialty study really costs

A CPA who signs a return carrying a boutique's R&D credit study or cost segregation report has adopted every number in it. Section 6694 makes the signing preparer — not the boutique — liable when a position on the return lacks the required confidence level, and the penalty is measured against the preparer's fee: the greater of $1,000 or half the preparation income for an unreasonable position, the greater of $5,000 or 75% of that income if the conduct was willful or reckless. The statute has a well-marked safe path, but it runs through documented diligence, not through the referral itself.

This guide maps the penalty structure and then turns to the question that matters in a specialty-tax practice: what, concretely, discharges the signing CPA's exposure when someone else built the study.

Who is a "tax return preparer"

Section 7701(a)(36) defines a preparer as any person who prepares for compensation all or a substantial portion of a return or refund claim. Two consequences matter here. First, the definition reaches non-signing preparers: under Treas. Reg. §301.7701-15, a specialty firm whose study constitutes a substantial portion of the return — a six-figure credit usually does — can itself be a preparer subject to Section 6694, whether or not anyone there is a CPA, attorney, or enrolled agent. Boutiques that structure to avoid Circular 230 practitioner status do not thereby escape the preparer-penalty regime. Second, and less comfortingly, the boutique's own exposure does not reduce the signing preparer's: only one person is the signer, and the regulations put primary responsibility for the return's positions there.

The Section 6694(a) standards ladder

Section 6694(a) applies when an understatement of liability is due to an "unreasonable position." Whether a position is unreasonable depends on where it sits on a ladder of confidence standards, and on whether it was disclosed.

Position standards under Section 6694(a):

Substantial authority is an objective weighing of authorities — code, regulations, cases, published rulings — and it is the same standard that protects the taxpayer from the Section 6662 substantial-understatement penalty. We treat the weighing methodology in detail in the substantial authority standard, and the taxpayer-side penalty in accuracy penalties under Section 6662. The practical upshot for credit work: the signing CPA should be able to articulate, before signing, why the study's positions clear substantial authority. "The provider is reputable" is not authority. Authority is Section 41(d), Treas. Reg. §1.41-4, and the case law applying them — and for contested fact patterns (funded research, internal-use software, wage estimation) the case law cuts both ways often enough that the weighing has to be genuine. Our R&D credit case-law map is the survey.

Where a position honestly sits below substantial authority but has reasonable basis, disclosure on Form 8275 converts it into a defensible filing. Preparers underuse this. Disclosure raises exam salience, and clients dislike it — but it is the statutory design for exactly the marginal position, and it beats both silent aggression and declining the engagement.

Section 6694(b): the tier without an exit

Section 6694(b) escalates the penalty — greater of $5,000 or 75% of preparation income — where the understatement is due to a willful attempt to understate liability or to reckless or intentional disregard of rules and regulations. There is no reasonable-cause exception. In the specialty context, the fact patterns that reach recklessness are recognizable: signing a return with a credit study the preparer knows describes activities the client does not perform; ignoring a prior exam that disallowed the same methodology; or continuing to use a provider after learning its studies are templated fiction. A preparer who documents concerns and signs anyway has drafted the government's exhibit.

Section 6695: the mechanical penalties

Alongside 6694 sit the Section 6695 per-failure penalties, indexed for inflation: failure to furnish a copy of the return to the taxpayer (§6695(a)), failure to sign (§6695(b)), failure to include the PTIN (§6695(c)), failure to retain a copy or list (§6695(d)), and failure of the paid-preparer due-diligence requirements of §6695(g) — which by their terms cover the earned income credit, child tax credit, American opportunity credit, and head-of-household status, not business credits, but which signal the direction of travel: Congress is comfortable putting affirmative diligence duties on preparers. These are strict, additive, and embarrassing; a firm-level e-file and signature checklist eliminates them. The IRS's tax professionals hub collects the current penalty amounts and PTIN rules.

Relative severity of the preparer penalty tiers (illustrative $10,000 preparation fee):

§6695 mechanical failure~$60

§6694(a) unreasonable position$5,000

§6694(b) willful/reckless$7,500

Illustrative. Section 6694 amounts are the greater of the flat figure or the percentage of preparation income; Section 6695 amounts are per failure and inflation-indexed (shown at a representative amount).

How a boutique's study creates signing exposure

Walk the typical fact pattern. A client engages a contingent-fee R&D shop; the shop delivers a study claiming $400,000 of qualified research expenses and a credit near $40,000; the CPA transcribes the numbers to Form 6765 and signs. If the study's wage allocations are estimates untethered from time records, or the "business components" fail the Section 41(d) tests, the position may lack substantial authority — and the person who took that position, for Section 6694 purposes, is the signer. The exposure compounds on amended returns: a refund claim is itself a return for preparer-penalty purposes, and the erroneous-refund-claim penalty of Section 6676 sits on the taxpayer's side of the same filing.

The same mechanics apply to cost segregation (a study allocating 45% of an ordinary office building to 5-year property invites the question the Cost Segregation Audit Techniques Guide trains examiners to ask) and to energy studies certifying efficiency levels the building cannot meet.

The reliance defense, and what it actually requires

The regulations are more generous than preparers sometimes fear — and more demanding than boutiques suggest. Under Treas. Reg. §1.6694-1(e), a preparer may generally rely in good faith without verification on information furnished by the taxpayer or by another advisor or preparer. But the same regulation withdraws the protection where reliance is unreasonable: the preparer may not ignore the implications of information furnished or actually known, and must make reasonable inquiries where the information appears incorrect, inconsistent, or incomplete. Treas. Reg. §1.6694-2(e) folds reasonable, good-faith reliance on advice into the reasonable-cause defense — the full regulatory text lives in Title 26 of the eCFR.

Translated into practice, the defense is built from five artifacts:

1. A vetting file on the provider. Evidence the firm chose the specialist with care — sample deliverables reviewed, exam-support terms obtained, references checked. The standing process is the subject of our playbook for vetting specialty providers.
2. A facial-plausibility review of the deliverable. Not a re-performance — a reviewer competent in the return reading the study for internal consistency and consistency with what the firm knows about the client. The CPA who prepares the books knows whether the client "employs twelve engineers in systematic experimentation." Ignoring that knowledge is precisely what §1.6694-1(e) forbids.
3. Tie-outs. Study totals reconciled to the return, wage figures to payroll, cost basis to the fixed-asset ledger. Discrepancies are the "incomplete or inconsistent information" that obligates inquiry.
4. Questions asked and answered, in writing. The inquiry duty, discharged and documented. What goes in this file is covered in workpaper standards for credit claims.
5. A disclosure decision. For positions the review leaves at reasonable-basis strength, a documented Form 8275 conversation with the client — and a documented refusal to sign if the client rejects both disclosure and adjustment.

What diligence does not require

Neutrality demands the other boundary too. The signing preparer is not required to re-perform the engineering, re-interview the client's technical staff, or independently verify every allocation — the regulations say so, and a diligence standard that required duplication would make specialty referral pointless. Nor does every credit study demand the full apparatus: a small, well-documented payroll-offset credit for a client whose research activity the firm observes directly needs proportionate review, not a forensic one. The penalty regime scales with the position's size and shakiness; so should the file.

And sometimes the answer is that no amount of process cures the position. If the review convinces the preparer the study will not clear reasonable basis even with disclosure, Section 6694 has one remaining defense: do not sign. Circular 230 §10.34 points the same direction for practitioners — see Circular 230 explained — and a firm that has never walked away from a study has not tested its own standards.

A closing frame

Preparer penalties are not the IRS's principal weapon — assessment is sporadic, and the dollar amounts are modest next to malpractice exposure. Their real function in a well-run firm is as a design spec: the statute tells you exactly what the government thinks a careful signer does. Meet that spec on every specialty deliverable — vetted source, reviewed study, tied-out numbers, documented inquiry, honest disclosure calls — and the same file that defeats a 6694 assertion defeats the accuracy penalty for the client and the negligence claim from the plaintiff's lawyer. One process, three defenses.