Reasonable Comp Audit Defense: Surviving IRS Reclassification

# Reasonable Comp Audit Defense: Surviving IRS Reclassification

The notice goes to the client first. The client forwards it to you, usually with a one-line email that ends in a question mark. An S-corporation return you signed, two or three filing seasons back, and the examiner has questions about officer compensation.

Then you open the workpaper file. There's a salary figure, and next to it a note that reads something like "per wage survey, adjusted for experience and local market." Nobody remembers which survey. Nobody remembers what "adjusted" meant, or who did the adjusting, or whether the occupation code matched what the shareholder actually did all year. The number might be perfectly defensible. You just can't show that it is.

That gap is the entire exposure. Reasonable compensation is a facts-and-circumstances standard, so an exam almost never turns on a bright-line arithmetic error. It turns on whether the Firm can reconstruct how the figure was reached, from what evidence, for what period. This piece walks the exam from the examiner's side: what draws scrutiny, what reclassification actually costs, what a defensible position contains, and why the salary-first requirement is easier to defend when the supporting evidence was built before anyone asked for it.

What happens if the IRS reclassifies S-corp distributions as wages? >The examiner treats some portion of the shareholder's distributions as compensation for services. That reclassified amount becomes wages, which pulls in employment tax on both the employer and employee sides, amended payroll filings, corrected wage statements, with interest running from the original due dates and, often, penalties. Downstream items that keyed off compensation, including retirement plan contributions and the qualified business income calculation, may also need rework. Whether reclassification happens at all is a judgment about the facts, and the strength of the taxpayer's contemporaneous documentation is a large part of that judgment. Source-traced evidence supports a defense. It does not guarantee an outcome.

Key Takeaways

• Low or zero officer compensation alongside large distributions is visible on the face of the return - the S-corp return reports officer compensation and distributions on separate lines, so the pattern needs no discovery to spot.
• The examiner's real question is method, not magnitude - "how did you arrive at this number, from what data, for what occupation and period" is harder to answer after the fact than most Firms expect.
• Reclassification costs land in categories, not a single assessment - back employment tax, amended payroll returns, penalties, interest, and unbillable staff hours rebuilding an analysis nobody documented.
• Contemporaneous beats reconstructed - evidence assembled at the time of the determination carries weight that a post-notice rebuild does not, and examiners are practiced at telling the difference.
• The exposure sits with the Firm that signed the return - the client feels the assessment; the Firm absorbs the reconstruction, the client conversation, and the professional risk.
• Retention outlives the engagement year - an exam can open years after filing, so the question is whether the Firm still holds the workpapers, not whether someone once ran the analysis.

How reasonable-comp exams start

Reasonable-comp exams rarely begin with a tip or an anomaly buried in the books. They begin with a return that reads a certain way. The information the examiner needs to form a first impression is already on the pages the Firm filed.

What triggers scrutiny

The classic pattern is a profitable S corporation with an active shareholder-employee, a small or blank officer compensation line, and substantial distributions. Nothing about that pattern is illegal on its face. Plenty of returns look that way for legitimate reasons: a shareholder who genuinely performs minimal services, a year with an unusual liquidity event, a business whose profit derives mostly from capital rather than personal effort. But the pattern raises the question, and the question is what starts the correspondence.

Other common signals are structural rather than statistical. Payroll filings that don't reconcile to the compensation reported on the return. A shareholder who is the only person performing billable work at the entity, paired with a compensation figure that would be low for a staff-level employee doing the same job. A comp figure that stayed identical across five years while revenue tripled. A distribution ledger that moves in regular monthly amounts, which reads less like a return on capital and more like a paycheck routed around payroll.

None of this is a formula, and the IRS does not publish a selection threshold that a Firm could design against. That absence matters more than practitioners sometimes admit. There is no percentage split that inoculates a return, no ratio that ends the inquiry. The often-repeated rules of thumb circulating in owner forums have no basis in the statute or the regulations, and an examiner is under no obligation to accept one. What the guidance does supply is a list of factors, among them time and effort devoted to the business, dividend history, and what comparable businesses pay for similar services. The full list sits under S-corp reasonable compensation. Courts have recharacterized distributions as wages when those factors pointed one direction and the return pointed another.

So the trigger is a pattern, and the defense is a record. Those are not symmetrical. The pattern is generated automatically by the act of filing. The record only exists if somebody built it.

What the examiner asks for

The opening request is narrower than most people expect and harder to satisfy than it looks. The examiner wants the basis for the number.

In practice that resolves into a specific set of items. What the shareholder actually did during the year, described as roles and time rather than as a job title. What wage data supported the figure, from which source, covering which occupation, which geography, and which period. What adjustments were applied to that raw data and why. Whether a compensation agreement exists and whether the entity followed it. What non-shareholder employees were paid for comparable work. How distributions were timed and characterized in the entity's own records.

Notice the shape of that list. Almost every item is a documentation question rather than a valuation question. The examiner is not asking the Firm to prove the salary was correct to the dollar, because a facts-and-circumstances standard does not produce a single correct dollar. The examiner is asking whether a reasoned method was applied and can be shown.

A Firm that can produce an interview record, a role decomposition, sourced wage evidence with occupation and geography identified, and a written rationale for each adjustment is having a very different conversation than a Firm producing a spreadsheet tab and a memory. The second Firm may well have picked a defensible number. It is now defending it with recollection, which is the weakest evidence available and gets weaker with every month that passes.

What reclassification costs

When an examiner concludes that some portion of distributions represented compensation for services, the adjustment does not stay contained in one line of one return. It propagates.

Back payroll tax

Reclassified distributions become wages, and wages carry employment tax on both the employer and the employee side. That is the core of the assessment. Around it sits a cluster of corrections that consume real hours: amended employment tax returns for the affected quarters, corrected wage statements for the shareholder, and a matching amendment to the shareholder's individual return where the corrected wage figure changes the outcome.

From there the effects reach items that were computed off compensation in the first place. Retirement plan contributions calculated on a compensation base may be out of tolerance in either direction. The qualified business income deduction interacts with wages paid, so a change in the wage figure can shift a deduction the shareholder already claimed. Any allocation, accrual, or benefit that keyed off officer compensation now keys off a different number.

The categories are what matter here, not a headline figure. Specific rates, wage bases, and thresholds change from year to year and are separately verified before they belong in a workpaper. What is stable is the structure of the exposure: employment tax on the reclassified amount, amended filings across multiple periods and multiple returns, and a downstream recomputation of anything that referenced compensation.

Penalties and interest

Penalties in this area come from more than one direction, and they stack in ways that surprise clients. There is potential exposure tied to the accuracy of the return itself. There is separate exposure tied to employment tax obligations that were not deposited or reported when due, because the wages, in the examiner's view, existed all along. Interest accrues from original due dates rather than from the date of the notice, so a three-year-old return arrives with three years of accrual already attached.

Reasonable cause is the practical relief valve, and reasonable cause is an evidentiary argument. A taxpayer who set compensation through a documented process, using identified wage data, with a written rationale, is making a materially stronger reasonable cause argument than one who set it by feel. Again, that argument depends on a record that existed before the notice.

There is also a cost that never appears on an assessment. The Firm absorbs the reconstruction: pulling old files, chasing wage sources that may no longer be accessible in the same form, interviewing a client about what they did four years ago, drafting a response, and doing all of it against a deadline. Those hours are usually unbillable, and they land in the middle of a filing season rather than politely in August. Then there is the client conversation, which is the part nobody schedules time for. The client's first question is rarely about the tax. It is about who decided on that number.

What a defensible position looks like

A defensible reasonable compensation position is not a better number. It is the same number with a record attached. Two Firms can land on identical figures for the same shareholder and be in completely different positions under exam, and the difference is entirely in what they can put in front of the examiner.

Contemporaneous evidence

Contemporaneous means the evidence was created as part of determining the figure, not as part of defending it. That distinction is not a technicality. Documentation assembled after a notice arrives is doing two jobs at once, reaching a conclusion and justifying one, and examiners read it accordingly.

Contemporaneous evidence for a reasonable comp determination generally includes a record of the shareholder interview and when it happened, a decomposition of the shareholder's roles with the share of time attributable to each, the wage data actually consulted with its source and vintage identified, the adjustments applied with the reasoning stated at the time, and the resulting figure with the person who prepared it and the person who approved it both identified.

Reconstruction fails in specific, predictable ways. Wage survey editions get revised or retired, so the exact data consulted three years ago may not be retrievable in its original form. Memory of a shareholder's role mix drifts toward whatever supports the number already filed. Adjustments that were reasonable in the moment become impossible to justify because the reason has been lost. None of this means a reconstructed position cannot succeed. It means it succeeds against friction that a contemporaneous record does not face.

A traceable methodology

Traceability is the property that lets someone other than the original Preparer follow the number back to its origin. In practice it means every material figure ties to a source, an occupation, a geography, and a period, and it does so at field level rather than as a bibliography stapled to the back.

That granularity is what an exam actually tests. "We used national wage data for accountants" is a citation. "This figure is the median for this occupation code, in this metropolitan area, for this survey period, adjusted upward by this amount because the shareholder also performs the functions described in the second role line, with that rationale recorded on the date the adjustment was made" is traceable. The first can be challenged with a single question. The second answers the question before it is asked.

Method consistency matters alongside traceability. When three Preparers in the same Firm build three reasonable comp analyses three different ways, the examiner only ever sees the one in front of them, but the Firm carries the risk of all three. A repeatable method with a defined evidence standard means the weakest file in the drawer is built to the same evidence standard as the strongest. That is the point of standardizing, and it is worth more than any single well-built analysis.

This is also where the honest limit belongs, stated plainly rather than buried. Source-traced evidence supports a defense. It does not guarantee an IRS outcome, and no methodology, software, or documentation standard can promise one. Reasonable compensation is a judgment about facts, and judgment can go against a well-documented taxpayer. What good evidence does is put the Firm in the argument on its own terms, with a Reviewer who signed a conclusion they could see the basis for, instead of in a reconstruction race against a clock. What a defensible Study contains is worth reading end to end.

Building defense in before the exam

Every property described above has to exist before the notice, which means the work belongs in the engagement, not in the response. The practical question for a Firm is whether its normal reasonable comp process produces a defensible record as a byproduct, or whether producing one requires somebody to remember to do extra work.

The evidence manifest

An evidence manifest is the record of the record. It captures who prepared the analysis, what evidence was pulled and when, what each figure traces to, what changed after the initial build, who made each change and why, and who approved the final conclusion. Once the Study is finalized, that manifest is immutable, which is the property that makes it worth anything under examination. A log that can be edited after the fact answers no questions.

The Reviewer sign-off is the control point that the manifest records. A qualified Reviewer inside the Firm looks at the evidence, exercises professional judgment, and takes responsibility for the conclusion. That is a workflow fact about how a Study gets finished, and it is deliberately not an automation claim. Software can gather versioned wage evidence, tie every figure to its source, flag contradictory inputs, and present the whole picture in a form a Reviewer can actually assess. Software does not make the determination. The Reviewer does, and the Reviewer's professional risk is exactly why the evidence in front of them has to be legible rather than merely present.

Confidence signaling helps here in a way that a single number cannot. When a Study surfaces which fields are well-supported and which ones rest on thinner evidence, the Reviewer spends their attention where the exposure actually is. A study that flags which fields rest on thin evidence is a review. A finished PDF with a bolded salary figure at the bottom is a signature.

Workpapers the Firm still holds years later

Exams arrive on their own schedule. The reasonable comp determination made this filing season may not be questioned for years, which means the operative question is not whether the Firm once did the analysis. It is whether the Firm can still open it.

This is where the retention problem bites, and it is a category problem rather than a vendor-specific one. Subscription wage-report tools deliver a report while the subscription is active. Stop paying, change tools, or lose the seat when a staff member leaves, and the underlying evidence behind reports the Firm already issued can become unreachable. The Firm's document retention duty does not lapse when the subscription does. Neither does the client's expectation that their advisor can answer a question about work the advisor performed.

Firm-owned workpapers invert that. Finalized evidence and the manifest stay with the Firm permanently, independent of subscription status, which means the response to a notice starts with opening a file rather than with an email to a vendor. That permanence is also what makes year two cheap: prior-year facts clone forward, the Preparer works through a change analysis rather than a fresh build, and the resulting file shows the year-over-year movement with reasons attached. A five-year series of determinations, each traceable and each showing what changed and why, is a substantially stronger record than five unrelated snapshots. It reads as a process, which is precisely what the factors ask about.

The Firms that handle reclassification exams well are not the ones with better arguments. They are the ones whose ordinary process already produced the file the examiner asked for. Building that into how the Firm approaches S-corp reasonable compensation is cheaper than building it under a deadline, every time.

Every figure traces to a source, and your Reviewer signs off on the record. Get started.

TracePrep is a software product, not a CPA firm, and does not render tax advice. This article is educational. Reasonable compensation is a facts-and-circumstances determination; consult a qualified tax professional about your specific situation. Source-traced evidence supports audit defense. It does not guarantee an IRS outcome.