Cash Balance Plans for High-Income Business Owners: A Tax Strategy Guide

7/8/2026

Our post on defined benefit versus cash balance designs covers the structural comparison and concludes that cash balance is the better answer for most closely held clients.

This post covers the decision that follows: whether to actually do it — how the deduction works, the interaction most analyses omit, the modeling that has to happen first, and how it ends.

How the Deduction Actually Works

The contribution is an employer deduction, taken in the year, and it must be funded by the applicable deadline — generally the due date of the return including extensions — to be deductible for that year. Confirm the current deadline, because the funding date rather than the adoption date is what secures the deduction.

It is a deduction against ordinary income at the owner's marginal rate, plus state tax where applicable. For an owner at a high marginal rate in a high-tax state, the combined effect is substantial and is the entire attraction.

It reduces adjusted gross income, which produces a second-order benefit worth modeling: thresholds and phase-outs the owner had exceeded may come back into range. Depending on the client's situation this can be worth a meaningful amount beyond the headline deduction, and nobody computes it unless asked.

There is no employment tax effect, since the contribution is not compensation.

And the honest framing of the benefit: assets grow tax-deferred and distributions are ordinary income later. So the value is deferral plus rate arbitrage — and the rate arbitrage only exists if the owner's rate in retirement is lower than it is now. A client at the top rate today who will remain at the top rate in retirement receives the deferral benefit and not the arbitrage, which is still worth having and is a smaller number than the marginal-rate calculation suggests.

The Interaction Most Analyses Omit

The technical point that changes the arithmetic for pass-through owners, and it is routinely left out of the pitch.

A retirement plan deduction reduces qualified business income, which reduces the qualified business income deduction for an owner claiming it. So the net benefit of the contribution is less than the marginal rate applied to the contribution — the deduction saves tax at the marginal rate and simultaneously gives back a portion of a deduction the owner was otherwise receiving.

The size of that offset depends on the owner's situation, including whether they are above or below the applicable thresholds and how the limitations apply to their business. For some clients the interaction is modest; for others it materially reduces the plan's after-tax value.

Confirm the current mechanics and model it specifically for the client. A plan recommended on a gross marginal-rate computation, without this interaction, has been oversold — and the client will discover it when the return is prepared.

One further interaction: wages affect both the plan contribution capacity and the qualified business income computation, sometimes in opposite directions, which is the modeling point our post on year-end planning raises for S corporation owners.

Who This Actually Fits

Described in terms the practitioner can assess:

Income consistently and substantially above what a defined contribution plan can absorb. Not one good year — several, with a reasonable expectation of continuation. An owner who is not already maximizing a defined contribution plan should do that first.

Owner age from roughly the mid-forties upward. The shorter the funding horizon to retirement age, the larger the permitted annual contribution, which is why these plans are compelling at fifty-five and marginal at thirty-five.

A favorable demographic — few employees, or employees younger and lower-paid relative to the owner. The staff cost is what makes or breaks the arithmetic, not the owner's capacity.

The ability to commit for several years, because the contribution is a legal obligation rather than a discretionary one and the permanency expectation applies.

A professional practice in many cases, given the demographic and the insurance premium exemption available to certain professional service employers discussed in our design post.

The Modeling to Do Before Adoption

The practitioner's actual work, and the reason to be involved before the plan is sold.

1. The census and the actuarial projection. Complete and accurate compensation, dates of birth, hire dates, and ownership including attribution. The output that matters is the owner's contribution against the required staff cost — and that ratio is the decision.
2. The gateway cost. Where the design pairs the cash balance plan with a cross-tested profit sharing allocation, a minimum allocation gateway applies to non-highly compensated employees. That gateway is the real cost of the structure and it determines whether it works for this workforce.
3. The qualified business income interaction, per above.
4. Multi-year cash flow, including a bad year. The contribution is mandatory. Model the year the business earns half what it expects, and ask whether the client can fund it.
5. The all-in cost, which clients are never quoted fully:

• Actuarial valuation and certification, annually
• Third-party administration
• A plan audit if the participant count crosses the applicable threshold — a real annual cost that a growing employer triggers without noticing
• Insurance premiums where the plan is covered, and the exemption analysis where it may not be
• Investment management
• Legal and document costs, including restatements

6. The comparison against the simpler alternative. A well-designed profit sharing plan alone, with a cross-tested allocation, may deliver a substantial share of the benefit with none of the mandatory funding, the actuary, or the permanency exposure. Model both, because for a client at the margin the simpler answer is frequently better and nobody presents it.
7. Controlled group and affiliated service group aggregation, asked about explicitly, because an owner with several entities may have one combined plan universe rather than one per entity.

The Investment Question Specific to These Plans

The point clients misunderstand most, and it changes their expectations.

The plan credits participants with a defined interest credit — a fixed rate, an index-linked rate, or the plan's actual return. Where the credited rate is fixed or index-linked, the employer bears the gap between the credited rate and actual investment performance.

Which produces a counterintuitive conclusion: a conservative portfolio matched to the crediting rate is usually the right investment policy, because it minimizes funding volatility. Outperformance reduces future required contributions and underperformance increases them, and the asymmetry matters more to a business owner than the upside.

So this is not an aggressive growth vehicle, and a client who expects to earn market returns inside it has misunderstood the structure. Where the client genuinely wants investment risk, an actual-return crediting design transfers that risk to participants and removes the funding volatility — which is a design decision to make at adoption rather than a portfolio decision to make later.

Set the expectation explicitly at adoption, because a client disappointed by conservative returns in year three will question the whole arrangement.

The Employee Side Is a Real Benefit

Worth mentioning because it is the non-tax argument.

A cash balance plan presents as an account with a balance, which participants understand — unlike a traditional pension promise, which employees discount to near zero in perceived value. So the same employer dollars generate materially more retention effect.

For a professional practice competing for staff, that is not a rounding argument.

The Exit

Plan it at adoption, because the exit determines whether the strategy was sound.

Termination requires full funding. A plan cannot simply stop — on termination the accrued benefits must be fully funded, which means a business terminating in a bad year may face a larger contribution than it was making. That is the risk clients least anticipate.

Benefits are then distributed or annuities purchased, with the associated administrative process and cost.

The permanency expectation applies, per our design post: terminating after two or three years having taken large deductions invites a challenge to the plan's qualification and to the deductions taken.

A business sale requires the plan to be addressed — terminated, frozen, or assumed — and the treatment affects the transaction. A buyer will diligence an underfunded plan, and an unfunded obligation is a purchase price adjustment.

And a freeze is the middle option clients do not know exists: benefit accruals can generally be stopped prospectively, with notice, leaving the existing obligation funded but no new accruals. That is frequently the right answer for a business whose circumstances changed, and it preserves the deductions already taken far better than a termination does.

The Risks, Stated Plainly

Mandatory funding. Not discretionary, in any year.

Underfunding consequences, including quarterly contribution requirements and an excise tax.

Interest rate sensitivity, which affects the funding requirement independently of investment performance.

Demographic drift. A workforce that grows or gets younger changes the required staff cost, and a design that worked at adoption can become expensive within a few years. This is why the annual review matters.

Administrative burden, which is permanent.

The permanency exposure if the client's circumstances change early.

Structured coverage is available through the retirement plan administration catalog, the 401(k) Training and Certification Program, HS 326: Planning for Retirement Needs, the Certificate in Integrated Wealth Planning and Advice, the Retirement Tax Guide, the Certificate in S Corp Transactions, and the wealth planning training for accountants and CPAs catalog.

The Annual Review

Not optional, and it is where the accountant's ongoing value sits.

Can the client still fund it, given this year's results and next year's outlook.

Has the demographic changed — new employees, changed compensation, an ownership change.

Is the funded status where it should be, and is the interest crediting rate still matched to the investment policy.

Does a prospective amendment need to be made, and if so it must happen before the accrual rather than after, per the anti-cutback discussion in our design post.

Is the contribution still the best use of the money, against the client's other objectives.

Where This Goes Wrong

• Recommended on a gross marginal-rate computation with no qualified business income interaction modeled
• The contribution presented as discretionary
• The client not maximizing a defined contribution plan first
• The staff cost and gateway not modeled before adoption
• The all-in cost understated, particularly a plan audit triggered by participant count
• The simpler alternative never modeled for a client at the margin
• Controlled group aggregation not asked about
• An aggressive investment policy against a fixed crediting rate, producing funding volatility
• The client expecting market returns inside the plan
• Adopted as a one-year strategy, creating a permanency problem
• Termination attempted in a bad year, when full funding is required
• A freeze never presented as the middle option
• The plan unaddressed in a business sale
• No annual review, so demographic drift makes it expensive unnoticed

The summary for a practitioner: this is the largest deduction available to a closely held business owner and it is a legal obligation rather than a strategy. Model the staff cost, model the qualified business income interaction because it reduces the benefit and nobody else will, model a bad year, quote the all-in cost including a possible plan audit — and tell the client at adoption that a conservative portfolio is the correct investment policy and that a freeze exists if their circumstances change.

Frequently Asked Questions

How does the deduction work?

It is an employer deduction against ordinary income at the owner's marginal rate plus state tax, taken in the year and required to be funded by the applicable deadline — generally the return due date including extensions. It also reduces adjusted gross income, which can restore thresholds and phase-outs the owner had exceeded, a second-order benefit worth modeling.

What interaction do most analyses omit?

That a retirement plan deduction reduces qualified business income, and therefore reduces the qualified business income deduction for a pass-through owner claiming it. The net benefit is consequently less than the marginal rate applied to the contribution, and the size of the offset depends on the client's specific situation. A plan sold on a gross marginal-rate computation has been oversold.

Is the tax benefit permanent?

No. Assets grow tax-deferred and distributions are ordinary income later, so the value is deferral plus rate arbitrage — and the arbitrage exists only if the owner's retirement rate is lower than today's. An owner who will remain at the top rate receives the deferral benefit and not the arbitrage, which is worth having and is smaller than the headline figure.

What should be modeled before adoption?

The census and actuarial projection showing the owner's contribution against the required staff cost; the gateway cost of any cross-tested profit sharing component; the qualified business income interaction; multi-year cash flow including a bad year; the full annual cost including actuarial, administration, insurance, investment, and a plan audit if the participant count crosses the threshold; and a comparison against a well-designed profit sharing plan alone.

What is the right investment policy for a cash balance plan?

Generally conservative and matched to the interest crediting rate, because where the credited rate is fixed or index-linked the employer bears the gap between it and actual performance. These are not aggressive growth vehicles, and a client expecting market returns inside one has misunderstood the structure — if they want investment risk, an actual-return crediting design is the decision to make at adoption.

What happens if the client wants out?

Termination requires full funding of accrued benefits, so a business terminating in a bad year can face a larger contribution than it was making — the risk clients least anticipate. A freeze, stopping future accruals prospectively with notice, is the middle option most clients do not know exists, and it preserves the deductions already taken far better than an early termination does.