How to Determine S-Corp Reasonable Compensation

# How to Determine Reasonable Compensation for an S-Corp Shareholder

The question shows up the same way every year. A shareholder-employee ran the business, took distributions, and either paid himself nothing on a W-2 or paid himself a round number somebody picked three years ago. Now the return is close to going out, and the figure on line one of the officer compensation schedule has to be one your Firm is willing to stand behind.

There is no formula that produces that figure. S-corp reasonable compensation is a facts-and-circumstances determination, which means the number is only as good as the reasoning underneath it. A percentage of net income is not reasoning. Neither is last year's figure carried forward because nobody asked.

What follows is the method from the Firm's chair. Decompose the shareholder's role into the work actually performed, price each piece of that work against credible published wage data, then record the judgment calls a Preparer made and why. Subscription wage-report vendors will hand back a computed number quickly. A number is not a defense. The reasoning is what a Reviewer signs, and the reasoning is what an examiner reads years later.

How do you determine reasonable compensation for an S-corp? >Price the work the shareholder-employee actually performs, rather than applying a percentage to profit or distributions. Three approaches carry the analysis. The cost approach, often called many-hats, splits the owner's year into distinct roles, assigns hours and proficiency to each, and prices each role against published wage data. The market approach compares the whole position to pay for similar positions at similar businesses in the same labor market. The income approach reasons from the return the business earns on its assets and capital, and is generally reserved for facts the first two cannot reach. Whichever approach you choose, the figure ties back to duties, hours, and a wage source you can name. Reasonable compensation is a facts-and-circumstances determination, so the documentation is the answer.

Key Takeaways

No safe-harbor percentage exists - The 60/40 and 50/50 splits circulating in owner forums are folklore, not authority, and a figure derived from one has no factual support behind it.
The IRS lists factors, not a formula - Training and experience, duties and responsibilities, time and effort devoted, dividend history, payments to non-shareholder employees, timing and manner of paying bonuses, comparable businesses' pay, compensation agreements, and the formula used to set compensation.
Three approaches carry the analysis - Cost (many-hats), market, and income. The cost approach fits most owner-operated S-corps; the other two answer facts it cannot reach.
Every figure carries four attributes - Source, occupation, geography, and period. A wage number without all four is an assertion, not evidence.
Judgment belongs on the record - Each Preparer adjustment carries a rationale, a prior value, and a timestamp, so the Reviewer evaluates the reasoning instead of rebuilding the analysis.
The Reviewer concludes - Evidence supports professional judgment. It does not substitute for it, and no documentation package guarantees an examination outcome.

What the IRS actually requires

Start with what the statute and the case law actually ask for, because most of the confusion in this area comes from importing a rule that does not exist.

The reasonable-compensation standard

A shareholder-employee who performs services for an S-corporation must be paid reasonable compensation for those services, and that compensation is treated as wages subject to employment taxes. Distributions are not wages. When a shareholder-employee performs meaningful services and takes distributions in place of salary, the IRS may recharacterize those distributions as wages, and courts have done exactly that. The consequence is back payroll tax, plus penalties and interest, assessed against the corporation.

Notice what the standard is not. It is not a floor tied to distributions. It is not a ceiling tied to profit. It is a measure of what the services themselves are worth in the labor market where they were performed. A shareholder who works forty-five hours a week running a construction firm has a compensation figure driven by that work, whether the year was profitable or not. A shareholder who genuinely performs no services has a different analysis, and that analysis is worth documenting too, because the absence of services is itself a fact that needs support.

The standard cuts both ways, which practitioners sometimes forget. Compensation can be unreasonably high as well as unreasonably low. The C-corporation case law that built this doctrine mostly involved the IRS arguing that compensation was excessive and should be recharacterized as a dividend. That body of law is where the factors came from, and it is why the factors are framed as a test of reasonableness rather than a test of adequacy.

Why there is no safe percentage

Owner forums keep circulating a split. Sixty percent salary and forty percent distributions, or fifty-fifty, or a third. None of those come from the Code, the regulations, or a revenue ruling. They are heuristics that spread because they are easy to remember, and they are the single most common weak point in a reasonable-comp position.

Here is the practical problem with a percentage. It derives the compensation figure from the wrong variable. Profit is a function of pricing, capital, market conditions, and the labor of everyone in the business. Compensation for the shareholder's services is a function of what those specific services cost to buy. A firm that had a terrible year does not owe its owner a smaller salary because the market rate for a construction supervisor did not fall. A firm that had a windfall year does not owe a larger one either, unless the owner's own effort produced the windfall.

There is a second problem, and it is the one that hurts under examination. A percentage leaves nothing to show. When an examiner asks how the figure was determined, the answer "sixty percent of net income" invites exactly one follow-up question, and there is no second answer behind it. A duty-by-duty build with wage sources attached invites a different conversation. The examiner may disagree with a proficiency assignment or a geography selection. Disagreement about an input is a far better position than having no inputs at all.

Understanding how the salary and distribution split works mechanically is useful, but it is downstream of this. Determine the compensation figure on its own facts first. The distribution is what is left, not the other way around.

The three approaches in IRS training material

Reasonable compensation analysis borrows its structure from valuation practice, and the same three approaches appear in IRS training material on the subject. They are not interchangeable. Each one answers a different question, and choosing among them is itself a documented decision.

Cost (many-hats) approach

The cost approach treats the owner as a bundle of jobs and prices each one separately. The premise is straightforward. If the shareholder spends part of the year selling, part supervising crews, part doing the books, and part on general management, then the compensation for those services equals what it would cost to hire people to do each piece.

This fits the overwhelming majority of owner-operated S-corps, and it is the approach most reasonable-comp work should default to. It handles the reality that owners of small businesses do not hold one job. It produces a figure that is defensible attribute by attribute rather than in one lump. And it degrades gracefully: if an examiner rejects one role assignment, the rest of the build survives.

It has limits. The cost approach can understate compensation when the shareholder's contribution is genuinely singular, when the value comes from a relationship or a skill that has no clean occupational analog. A rainmaker whose personal book drives most of the revenue is not priced correctly by a sales-manager wage. It can also understate compensation for a shareholder who supervises a large organization, because the wage data for individual functions does not capture the span of control. When either condition shows up, note it and consider a cross-check.

Market approach

The market approach prices the whole position at once. Rather than splitting the year into functions, it asks what businesses of similar size, industry, and region pay someone in the equivalent seat, and it uses that as the comparison.

This works well when the shareholder holds a recognizable executive role and there is credible survey data for that role at that revenue band in that industry. It works poorly when the business is unusual, when the owner's role does not map to a titled position, or when the available survey data is thin enough that the comparison rests on a handful of observations.

The discipline here is comparability. A general manager figure pulled from a national survey and applied to a two-person shop in a rural county is not a comparison, it is a placeholder. Document the size band, the industry classification, the geography, and the survey period you matched on, and note where the match is imperfect. Imperfect matches are normal. Undisclosed imperfect matches are the problem.

Income approach

The income approach reasons from the business's returns. It starts with the return that capital and assets should earn, treats that portion of profit as attributable to the investment rather than to labor, and reads the remainder as the return on the shareholder's services.

It is the least commonly used of the three in ordinary practice, and it should be. It requires assumptions about required rates of return that are themselves contestable, and it reintroduces the profit-linkage problem the cost approach was designed to avoid. Where it earns its place is as a sanity check, or in fact patterns the other two approaches cannot handle: a business whose value comes overwhelmingly from the owner's personal effort, or one where an independent capital-return figure is genuinely available.

Documenting the approach chosen and why

Whichever approach carries the analysis, the choice itself goes in the file. Two sentences is often enough. Name the approach, name the fact pattern that made it fit, and name what you did with the approaches you did not use.

A cross-check strengthens the position considerably. Run the cost approach as the primary build, then compare the result against a market figure for the equivalent titled position. If the two land close, say so. If they diverge, explain the divergence rather than quietly picking the smaller one. An examiner who sees a documented divergence and a stated reason is reading a workpaper. An examiner who sees one number and no alternatives is reading a conclusion.

Step by step: decompose the role

The cost approach is where most of the work happens, so this is the part worth doing carefully. Three passes: list the hats, assign time and proficiency, map each duty to a wage source.

List every hat the owner wears

This starts with an interview, and the interview is the part firms most often shortcut. The shareholder's own description of the year is the primary evidence. It is also the part they have the least practice articulating, because nobody who runs a small business thinks of their week as a set of occupational categories.

Ask about the actual year, not the job title. What did the week look like in the busy season versus the slow one. Who else in the business does each function, and what happens to that function when the owner is out. Which decisions cannot be made without the shareholder. What changed from the prior year: a hire, a location, a product line, a client concentration shift.

Push past the first answer on management. Owners describe themselves as running the business, which is true and useless. Running the business decomposes into recognizable pieces: business development, operations supervision, financial oversight, purchasing, human resources, and direct production or service delivery. Most owner-operators have several distinct roles once you separate them properly. Very small businesses often have more, not fewer, because there is nobody to delegate to.

Record the interview as intake, not as notes. The distinction matters six months later, when a Reviewer wants to know whether the twelve percent allocated to bookkeeping came from the shareholder or from a Preparer's assumption. Structured intake makes that answerable. A memo does not.

Assign time and proficiency

Two variables drive each role's contribution: how much of the working year went into it, and at what level the shareholder performs it.

Time first. Allocate the shareholder's hours across the roles you listed, and make the allocation reconcile to a defensible total. A fifty-hour week for a working owner is ordinary; a claim of eighty for fifty-two weeks needs support. Seasonal businesses should show seasonal allocation, because a flat annual percentage misrepresents a year that had a four-month peak. If the shareholder cannot estimate a role's share within a reasonable band, record the uncertainty rather than smoothing it over.

Proficiency next, and this is where judgment enters properly. The same duty is worth different amounts depending on who is performing it. An owner who handles the books at a bookkeeper's level is not priced at a controller's wage, and an owner with twenty years of licensed trade experience supervising a crew is not priced at an entry percentile. Proficiency drives which point in the wage distribution applies. Anchoring every role at the median is a decision, and it is often the wrong one in both directions.

State the proficiency basis explicitly. Years in the role, credentials held, whether the shareholder trains others in the function, whether the work would need a supervisor if someone else did it. These are the same considerations the IRS factors point at: training and experience, duties and responsibilities, time and effort devoted. Writing them down at the role level is how those factors show up as evidence instead of as a citation.

Map each duty to a wage source

Now price it. Each role gets a wage figure, and each wage figure carries four attributes: the source, the occupation, the geography, and the period.

Published wage surveys from federal statistical agencies are the ordinary starting point, because they are public, methodologically documented, and available at occupation and metropolitan-area detail. Industry compensation surveys supplement them where the occupational categories are too coarse. Whichever you use, the citation has to be specific enough that a reader can retrieve the same figure. A source name alone fails that test. Source plus occupation code plus geography plus survey period passes it.

Geography is the attribute practitioners most often get wrong. Wage data varies substantially between metropolitan areas, and between a metro area and the surrounding non-metro region. Pull the geography where the work is performed, not where the corporation is registered and not the national figure. When a shareholder works across multiple markets, document which one you selected and why.

Period matters for the same reason. Survey data is collected and published on a lag. Using a figure from a survey period that does not align with the tax year is defensible if you say so and explain the adjustment; it is a weakness if it is silent. Name the period on every figure.

The output of this pass is a role-by-role build: duty, hours or percentage of time, proficiency level, wage source with its four attributes, and the resulting dollar contribution. Sum the contributions and you have a compensation figure with a visible derivation. Anyone who wants to argue with it has to argue with a specific input, which is precisely the position you want to be in. If you want to estimate a starting figure before running the full build, treat the result as a directional check on the build rather than as a substitute for it.

Documenting judgment so it survives review

A build like the one above involves a dozen or more judgment calls. Those calls are legitimate. Undocumented, they are the weakest part of the file.

Recording preparer adjustments

Every place a Preparer departs from a mechanical result is an adjustment, and every adjustment needs three things on the record: what the value was before, what it is now, and why it changed.

The common ones are predictable. A wage percentile moved off the median because the shareholder's experience justified it. A geography swapped because the primary work location differs from the business address. A role dropped because the interview established that a manager actually handles it. A time allocation revised after the shareholder reviewed a draft and corrected it. Each of those is defensible. Each of them is also invisible in a finished number.

Rationale text should name the fact that drove the change, not restate the change. "Adjusted to seventy-fifth percentile" says nothing. "Adjusted to seventy-fifth percentile: shareholder holds a master electrician license and supervises four licensed journeymen, per intake" says something a Reviewer can evaluate and an examiner can test.

This is the mechanic TracePrep records structurally. A Preparer adjusts a figure, and the prior value, the rationale, the timestamp, and the identity of the person who made the change go onto the record automatically. Nothing moves without a reason attached. When the Study is finalized, the evidence manifest is immutable, and it stays with the Firm regardless of subscription status. A breakdown of what a full Study contains covers the rest of what belongs in the finished package.

What the Reviewer needs to see

The Reviewer carries the professional risk on this figure. Their job is not to redo the analysis. It is to decide whether the analysis supports the conclusion, and that decision requires a specific set of things visible without excavation.

Four items, at minimum. The approach selected and the reason. The role build with every wage figure traceable to source, occupation, geography, and period. Every Preparer adjustment with its rationale and prior value. And, in year two and after, a change analysis against the prior year that shows what moved and why.

That last item is what turns reasonable compensation from an annual rebuild into an annual update. Prior-year facts clone forward. The Reviewer looks at the deltas: the shareholder took on a new role, a location changed, wage data for the primary occupation moved, hours shifted after a hire. Reviewing changes against a signed prior-year baseline costs a fraction of reviewing a build from scratch.

What the Reviewer should never have to do is ask where a number came from. If that question has to be asked, the file has already failed the test that matters, because the examiner asking it three years from now will not have anyone available to answer.

The conclusion is the Reviewer's. Evidence supports professional judgment; it does not replace it, and no amount of source-tracing guarantees how an examination lands. What source-traced evidence does is put the Firm in a position to answer questions from its own records, on its own timeline, years after the return was signed.

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.