S-Corp Salary vs. Distribution: How the Split Works (and the 60/40 Rule Myth)

# S-Corp Salary vs. Distribution: How the Split Works (and the 60/40 Rule Myth)

If you own an S corporation and actually work in it, money reaches you two ways. Some of it arrives as W-2 wages, run through payroll, with taxes withheld. The rest arrives as a distribution: a transfer of company cash to a shareholder, with no payroll tax attached. Same bank account, same person, very different treatment.

Search for how to split the two and you will hit a number within about thirty seconds: 60/40. Sixty percent salary, forty percent distribution. It gets repeated in forum threads, in videos, in the onboarding emails of budget incorporation services, and occasionally by people who should know better. It is not in the Internal Revenue Code. It is not in the Treasury regulations. It is not in an IRS publication, revenue ruling, or audit technique guide. Nobody who quotes it can produce the citation, because there isn't one.

This piece walks the split as it actually works: what has to be paid first, how each side is taxed, where the 60/40 idea came from, and what really sets the number. If you want the method behind the salary figure itself, that lives in a separate walkthrough of how the reasonable salary is determined, and the broader standard sits under S-corp reasonable compensation. Owners will find the mechanics here directly useful. Firms will recognize the file they have to be able to defend.

Is the S-corp 60/40 rule real? >No. There is no 60/40 rule in the Internal Revenue Code, the Treasury regulations, or IRS guidance, and no safe-harbor percentage of any kind for S-corp reasonable compensation. The standard is reasonable compensation for services actually performed, judged on the facts and circumstances of one shareholder-employee at a time. A shareholder-employee who provides more than minor services must be paid reasonable wages, and the IRS can recharacterize distributions as wages when the salary is too low. A 60/40 split can be perfectly defensible for one owner and indefensible for another with identical revenue, because the two owners do different work. The percentage is an output of the analysis, never an input to it. What holds up under review is a salary figure tied to the owner's real duties, hours, and comparable market wages, with the sourcing written down.

Key Takeaways

• No safe-harbor percentage exists - 60/40, 50/50, 70/30, and every other ratio are folklore. The standard is facts and circumstances applied to a specific shareholder-employee.
• Salary is determined first, distributions follow - reasonable compensation is priced off the work performed, not carved out of profit as a percentage.
• The tax difference is the reason this gets policed - wages carry payroll tax, distributions do not, which is exactly why an artificially low salary attracts attention.
• Ratios break the moment profit moves - a fixed percentage swings the salary with a windfall year or a loss year, while the owner's actual duties barely change.
• The IRS publishes factors, not formulas - time and effort devoted, dividend history, what comparable businesses pay, and six more. Not one of them is a percentage.
• Documentation is what survives - a salary figure carrying its source, occupation, geography, and period beats a ratio that cites nothing, every time.

How salary and distributions work in an S-corp

An S corporation does not pay federal income tax on its operating income. Profit passes through to shareholders on a Schedule K-1 and is taxed to them whether or not a single dollar of cash ever leaves the company. That is the first thing to internalize, because it explains why distributions are not a second taxable event in the ordinary case. The income was already taxed on the way through.

What is left is a question of employment tax. And that question is settled before anyone decides what to distribute.

The salary-first requirement

Treasury regulations treat a corporate officer who performs more than minor services and receives remuneration as an employee for employment tax purposes. Not a contractor, not merely an investor, an employee. Which means the compensation paid for those services is wages, and wages get run through payroll.

The sequence matters more than most owners realize. Reasonable compensation for services is set on its own terms, priced against what the work is worth. Distributions are what the company chooses to move out of the business afterward, subject to cash, basis, and the shareholders' agreement. The salary is not a slice of the profit. It is the price of the labor, and it happens to be paid out of the same pot.

Read the order backwards and everything downstream goes wrong. An owner who starts with "profit was 400 thousand, so 60 percent of that is my salary" has picked a number with no relationship to the job. Some years that produces a wage well above anything the market would pay for the role. Some years it produces a wage a court would call unreasonably low. It is unstable in both directions, and its instability has nothing to do with the owner's work changing.

Revenue Ruling 74-44 is the canonical statement of the failure mode: where shareholders took corporate payments styled as dividends in place of reasonable salaries, the IRS recharacterized those amounts as wages subject to employment taxes. Courts have run the same play repeatedly. In the Radtke and Spicer Accounting cases, the shareholder who did the professional work drew no wages at all and took distributions instead, and the amounts were treated as compensation. In Watson, the taxpayer did pay a salary, just a thin one relative to what a comparable professional in that role earned, and a portion of the distributions was reclassified. That last one is the important precedent for most owners: paying some salary is not the test. Paying a reasonable one is.

How each is taxed

Wages carry Social Security and Medicare tax, split between the employee and the corporation, remitted with payroll deposits and reported on the quarterly employment tax returns and the annual W-2. The corporation deducts the gross wages and its share of the payroll tax, which reduces the profit that passes through on the K-1.

Distributions carry no payroll tax. They are also not deductible by the corporation, because they are not an expense. Under the basis and accumulated adjustments account ordering rules, a distribution is generally tax-free to the shareholder to the extent of the previously taxed income and stock basis behind it, with amounts beyond that treated as a dividend to the extent of any accumulated earnings and profits, and as gain thereafter.

Put those two side by side and the arithmetic of the incentive becomes obvious. Every dollar shifted from the wage column to the distribution column is a dollar that escapes the payroll tax layer. That is the entire reason the standard exists and the entire reason examiners look at it. There is one qualitative wrinkle worth naming without pretending it resolves anything: the wage figure also interacts with the qualified business income deduction in ways that can cut against a low salary for some owners, so the "lower is always better" instinct is not even reliably true on its own terms. That interaction depends on income levels and business type, and it is a conversation for your own tax professional.

The consequences when the number is wrong are not limited to writing a check for the payroll tax that should have been paid. Reclassification pulls in interest and can pull in penalties, amended payroll filings, and a knock-on effect on the shareholder's return. The practical cost, and what an examination of this issue actually looks like, is worth understanding before you set a number, starting with what reclassification costs.

The 60/40 rule myth

Rules of thumb survive because they answer a hard question instantly. This one answers it wrongly, and it answers it wrongly in a way that feels authoritative, which is the worst combination available.

Where the 60/40 idea came from

Nobody can point to an origin document, which is itself the tell. The most plausible account is practitioner shorthand. Somebody observed how a set of carefully built salaries happened to land relative to profit, rounded it, and repeated it as a target. An observation about outcomes got promoted into a rule about inputs.

From there it spread the way all convenient numbers spread. Incorporation services put it in onboarding material because it makes the S election feel turnkey. Content marketers repeated it because it is concrete and rankable. Owners repeated it to each other because it turns a judgment call requiring evidence into a calculation requiring a calculator.

It also has cousins. You will see 50/50 quoted with the same confidence, and a "one third salary, one third distribution, one third retained" formulation that appears to be a garbled import from business valuation heuristics that have nothing to do with employment tax. None of them carry any more authority than 60/40, which is to say none.

The honest test is simple. Ask anyone who quotes a ratio for the citation. Not the blog post that repeated it, the primary source: the Code section, the regulation, the ruling, the case. The request ends the conversation every time.

Why the IRS has no such safe harbor

The standard is reasonable compensation for services actually rendered. That phrasing is doing specific work: it anchors to the services, not the profit, and it uses "reasonable," which is a facts-and-circumstances word rather than an arithmetic one.

The IRS has published what it looks at, and it is a list of factors, not a formula. Training and experience. Duties and responsibilities. Time and effort devoted to the business. Dividend history. Payments to non-shareholder employees. Timing and manner of paying bonuses. What comparable businesses pay for similar services. Compensation agreements. The formula the company itself uses for determining compensation. Read that list and notice what is absent: any percentage, any ratio, any threshold below which you are safe.

Consider two owners with identical revenue and identical profit. One runs a consulting practice where she personally delivers every engagement, works full time, and is the entire product. The other owns a small equipment rental operation with a manager running daily operations, and puts in perhaps eight hours a week on strategy and banking relationships. A 60/40 split is probably too low for the first owner and possibly generous for the second. Same numbers on the income statement, opposite conclusions, because the analysis has never been about the income statement.

Ratios also fail in the two years that matter most. A windfall year inflates the salary far past anything the role commands, which is money left on the table and, in a bonus-heavy pattern, an invitation to look at the timing and manner of paying bonuses. A loss year drives the ratio-derived salary toward zero, which is precisely the fact pattern the recharacterization cases were built on. The owner's job did not change in either year. Only the denominator did.

A safe harbor is a legal construct: a specific rule that, if followed, protects you from a specific challenge. Congress writes them, and Treasury implements them. No such construct exists here. Treating a ratio as one gives you the feeling of protection with none of the substance, and the feeling is what stops owners from building the file they would actually want to have.

What actually determines the split

The split is not a decision. It is a result. You determine one number properly, and the other side of the ledger is whatever is left after the cash and capital rules are applied.

Reasonable compensation sets the floor

Start with the work, not the money. An owner-operator typically wears several hats at once: producing the service, selling it, managing people, keeping the books, running the entity. Each of those hats corresponds to an occupation that has an observable market wage. Break the year into those roles, assign realistic hours and a candid proficiency level to each, and price each one against wage data matched on occupation, geography, and period.

Three approaches show up in practice. The cost approach, sometimes called the many-hats method, builds the figure up from the priced roles just described and fits most closely-held operating businesses. The market approach compares the owner's total compensation against what similar businesses pay someone in the equivalent seat. The income approach works backward from what an independent investor would require as a return, and is generally the least applicable to small service companies. Whichever one you use, the choice itself gets documented alongside the reason for it.

The output is a salary. Not a percentage, a salary, expressed in dollars, derived from labor. If you want a rough starting figure before you do the full build, you can estimate a reasonable salary and treat the result as a hypothesis rather than a conclusion.

Two consequences follow that owners tend to resist. First, the reasonable figure does not care whether the company had a good year. A slow year does not make the owner's work less valuable, and pushing the salary to zero to match the P&L is the exact pattern that gets unwound. Second, in a genuinely bad year, the answer is documentation rather than arithmetic. If a real reduction in the owner's hours drove the wage down, that is a fact worth documenting. Cash constraints are a weaker argument when distributions went out in the same year. What is not defensible is a number that moved for no recorded reason.

Distributions follow

Once compensation is set and actually run through payroll, distributions become a cash and capital question rather than a compensation question. How much cash can the business release without starving working capital. How much basis and accumulated adjustments account balance stands behind the distribution. Whether what is moving is really a distribution or a repayment of shareholder loans, which is a different animal with different documentation.

One structural constraint deserves a flag. An S corporation may have only one class of stock, and the governing provisions have to confer identical rights to distribution and liquidation proceeds. Persistently disproportionate distributions among shareholders will not automatically terminate the election, but the pattern invites questions about whether the arrangement really does confer identical rights. In a multi-shareholder S corporation, distributions that track ownership percentages are the clean path, and departures from that should have a reason on the record.

The other thing distributions cannot do is retroactively repair a thin salary. A large December bonus dropped in after the accountant runs the numbers may fix the payroll tax arithmetic, but timing and manner of paying bonuses is one of the factors on the list, and a bonus with no articulated basis reads exactly like what it is. The reasoning behind the compensation figure is worth far more when it was recorded during the year than when it is reconstructed the following spring.

Getting the split defensible

Here is the part that separates an owner who is fine from an owner who merely feels fine: not the number itself, but whether the number can be explained years later by someone who was not in the room.

Evidence over rules of thumb

Picture the question arriving three years from now, in writing, from an examiner. It is never "what percentage did you use." It is "how did you arrive at this figure, and what is it based on." A ratio has no answer to that. It cites nothing, because there is nothing to cite.

A defensible file answers at field level. Every figure ties to a specific source, a specific occupation code, a specific geography, and a specific period. Where the preparer adjusted a raw wage figure because the owner's proficiency, hours, or local market warranted it, the adjustment carries a written rationale, the prior value, and who made it. Nothing moves without a reason on the record.

That standard is harder to meet than it sounds, and not because the analysis is difficult. It is harder because files decay. The spreadsheet is on a laptop that got replaced. The wage data was pulled from a page that has since been revised, so the number no longer reproduces. Worse, plenty of firms and owners buy a wage report from a subscription vendor, and the report is only retrievable while the subscription is current. The retention obligation outlives the subscription. The evidence should too.

A study built to survive that timeline looks like a workpaper, not a printout: structured intake capturing the owner's roles and hours, versioned wage evidence with sourcing attached at field level, a calculation that reproduces from those inputs, and an immutable evidence manifest recording who prepared it, what changed, and who approved it. That last piece is what converts a number into a record.

When to bring in a firm

If you are an owner reading this, the useful next step is not to pick a percentage. It is to assemble the raw material and hand it to someone qualified to conclude on it: a written breakdown of the hats you wear and the hours behind each, your payroll history, comparable job postings or offers you have seen for the roles you fill, and any real change in your involvement this year versus last. Bring that to your CPA, or to a firm that runs TracePrep, and you will get a figure with reasoning attached instead of a number pulled from a ratio.

If you are the firm signing the return, the exposure is yours as much as the client's, and the practical problem is consistency. Preparers each build the analysis a little differently, the reviewer inherits a conclusion they cannot fully trace, and the workpapers scatter. TracePrep runs the workflow as one path: structured intake, versioned wage evidence tied to source, occupation, geography, and period, a deterministic calculation, and a Reviewer who controls the conclusion before anything is delivered. The product supports that professional judgment; it never substitutes for it. Finalized evidence and workpapers stay with the Firm permanently, regardless of subscription status. Year two clones the prior year's facts forward with change analysis only, so it costs a fraction of year one.

Interview. Evidence. Calculate. Review. Sign. That is the whole shape of it, and none of the steps is a percentage.

Every figure traces to a source, at field level, 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.