Partners of different ages are not one pension formula.

A four-partner firm will often open the conversation with a fair-sounding sentence. Everyone should get the same percentage. The sentence is about partnership politics, and it is understandable. It is a poor instruction for a defined benefit or cash balance design. The statute does not experience a 62-year-old partner and a 40-year-old partner as the same planning problem. The older partner is closer to the age at which the benefit limit is measured. The younger partner has decades of interest credits still in front of any retirement date the document assumes. A formula that treats them as interchangeable will either waste the older partner's capacity, overpromise to the younger partner, or fail the tests that compare benefits rather than deposits.

This briefing is about that gap. It is written for small firms, not for a hundred-lawyer institution with a legacy pension and a committee. The questions are who is an owner, how old each owner is, what the firm actually pays, and what the associates and staff must receive before any of the partner credits are defensible. Sterling Pension Group works with firms of that scale from West Hartford. We are a third-party administrator. We coordinate independent Enrolled Actuaries. We are not an actuarial firm, and nothing here is a valuation.

Age is an input, not a courtesy

A traditional defined benefit plan promises a benefit, usually described as an annuity at a retirement age. The contribution is whatever the Enrolled Actuary calculates is needed to fund that promise under the law. All else equal, funding a given dollar of future annuity costs more as the participant gets closer to retirement, because there are fewer years for the trust to earn the assumed return. The section 415(b) dollar limit is also less severely reduced when retirement is near the ages the statute treats as normal. For 2025 that dollar limit is an annual benefit of $280,000, reduced for early payment and for fewer than ten years of participation. The IRS describes those reductions on its page of defined benefit plan benefit limits. The limit is a benefit, not a contribution cap. Two partners can both be inside it and still have very different deductible deposits.

A cash balance plan states the promise as a hypothetical account. Each year the document credits a pay credit, and it credits interest on the account. The account is a record. It is not a separate brokerage account with the partner's name on it, and it is not the same thing as the assets in the trust. The pay credit still has to fit under the benefit limit once it is projected forward with the interest crediting rate to a retirement age. That projection is why age dominates the design. A large pay credit given to a 40-year-old compounds for a long time. The same dollar credit given to a 62-year-old has little time to compound before the limit is applied. The older partner can often receive a much larger credit for that reason alone. The younger partner's maximum is smaller even when the younger partner bills more hours and originates more work.

None of those maxima is a number you should pull from a chart. The 2025 cost-of-living limits fix the benefit ceiling and the compensation cap. They do not fix the contribution. An illustration for a specific partner is a calculation on that partner's age, compensation, the interest rate in the document, the form of payment, and the assets already in the plan. It is illustrative. It is not an IRS cap, and it is not portable to the partner down the hall.

The compensation the firm actually pays

Law firms do not all pay partners the same way, and the plan can only use the compensation the entity type allows. If the firm is a partnership or an LLC taxed as a partnership, a partner's plan compensation is generally earned income from self-employment, after the adjustments the statute requires, not the draw and not the gross origination credit. If the firm is an S corporation or other corporation, shareholder-employees are limited to W-2 wages. Distributions and K-1 allocations are not wages. That point is the whole of the briefing on S corporation compensation, and firms that converted to S status without rethinking payroll walk into it every year.

Compensation taken into account is also capped. For 2025 the cap is $350,000. A partner whose earned income or wages exceed that figure is, for formula purposes, a partner who earns $350,000. The rest of the profit still matters to the household and to the bank. It does not raise the pension credit. Publication 560 is the short IRS explanation of how small-business plans use compensation. It will not resolve a partnership agreement. It will keep the census from using draws, guarantees, and originations as if they were the same word.

Guaranteed payments deserve their own line on the census. They may be part of earned income. They are not automatically the compensation definition written into the document. Bonuses paid to associates are wages. Profit allocated to a partner who did not perform services may not be earned income at all. The bookkeeper who exports "partner income" as one column is not wrong for the financial statements. The export is wrong for the plan until someone splits it.

Same percentage, different promise

A design that credits every partner with the same percentage of pay looks even and often tests badly, or simply fails to do what the older partners hired the plan to do. Cross-testing, which the next sentences describe only in outline, compares the value of benefits. A dollar credited to a younger person is a more valuable retirement benefit than a dollar credited to an older person, because it has longer to grow. The test notices that. So does the 415 limit. Equality of percentages is not equality of benefits, and it is not the maximum the firm is allowed to fund.

Firms usually need a schedule of credits by person or by class, not a single rate. The classes have to reflect real categories the document can state, and they have to pass coverage and nondiscrimination once the staff are included. A schedule that happens to give the name partner a large credit and the new equity partner a token credit will be examined as a benefit comparison, not as a matter of seniority etiquette. Age can support a difference. A difference that is only about who controls the firm, with no relationship to the benefit rules, will not.

Here is the shape of an illustration, not a result and not a limit. A 62-year-old equity partner with compensation at the $350,000 cap may see a cash balance pay credit well into six figures, sometimes several times the credit available to a 40-year-old income partner whose compensation is $220,000. The younger partner's credit might be a fraction of the older partner's even in a design everyone considers generous. Those sentences are not a quote. Change the ages, the rate, the existing account, or the staff census and the figures move. The point of stating them at all is to retire the expectation that four partners will each receive the same deposit.

Associates are not a rounding error in that picture. If the plan covers anyone who is not an owner, the credits for owners have to be defensible against the benefits for everyone else. Staff cost is the practical name for that constraint. A firm that wants large partner credits should expect a real employer contribution for eligible associates and staff, often delivered through a 401(k) and cash balance combination so the testing can compare benefits rather than raw dollars. The gateway and the general test are technical. The business fact is not. People who are not partners will receive money, and that money is part of the cost of the partner credits. A design that ignores them is not a design yet.

Owners who are over 5 percent owners are highly compensated by ownership alone, regardless of pay. Associates can be highly compensated because of pay. For the 2025 plan year the lookback is generally 2024 compensation, measured against the threshold then in effect, which was $155,000. The $160,000 figure in the 2025 COLA release is the threshold that will be applied to 2025 pay when the following year's test is run. Title does not decide the label. A senior associate over the lookback threshold and a junior equity partner are not automatically in the same testing bucket, and they are not automatically in different ones.

What a firm should decide before anyone models a credit

The first decision is whether every equity partner wants a plan at all. A cash balance credit is a promise of the firm, funded by the firm, even when the hypothetical account is communicated as if it were the partner's. Partners who expect to leave, partners who are already drawing down, and partners whose compensation swings with originations should say so before the formula is written. A fixed credit sized to a peak year is a bad promise in a soft year. That problem has its own briefing later this spring. It is enough, now, to say that unanimity is not required by the statute and is often required by the partnership agreement. Read the agreement before the illustration, not after a partner objects.

The second decision is the census of who is not a partner. Counsel, contract attorneys, of-counsel lawyers who are common-law employees, paralegals, and staff all may be employees for coverage purposes. A leased employee or a shared-services arrangement with another entity can pull more people in. Related firms under common control are one employer for these rules even when they keep separate letterhead. A real estate holding company, a consulting LLC, or a spouse's practice can change the test. The page for law firms is the short version of who we sit down with. The long version is the payroll register plus a list of every entity any partner owns.

The third decision is cash. Defined benefit minimum funding is a legal deposit, not a discretionary profit-sharing vote. For a calendar-year plan the minimum for a year is generally due by September 15 of the following year, and a missed minimum carries a 10 percent excise tax reported on Form 5330, with a further tax if the shortfall is not corrected. The deduction often follows the firm's return, including extensions, under section 404(a)(6). Those dates sometimes fall on the same day for a partnership or corporation. They are not the same rule. A firm that cannot write the check in a down year should not adopt the formula that assumes it can.

Elective deferrals, if the firm also sponsors a 401(k), are separate and individual. For 2025 the deferral limit is $23,500, the ordinary catch-up is $7,500, and the catch-up for ages 60 through 63 is $11,250 in place of the ordinary catch-up. Annual additions, other than catch-up, are limited to $70,000. An older partner's cash balance credit does not increase a younger partner's deferral limit. Publishing the IRS figures in a partner meeting prevents the rumor that the pension "uses up" everyone's 401(k).

How the work is divided

The Enrolled Actuary calculates the range: the minimum the plan must receive and the maximum that can be deducted, given the formula and the census. Sterling Pension Group does not sign that certification. We collect the census, keep the document aligned with the formula the partners actually adopted, prepare the annual administration, and coordinate the actuary's work with the firm's CPA. Investment advice is not administration. The trustee or the investment adviser manages the trust. The hypothetical account still grows at the rate in the document, which may differ from what the portfolio earned. That difference is a funding fact the partners should hear before they adopt a rate.

This briefing is not tax, legal, or actuarial advice. Partnership agreements, buy-sell terms, and the question of who is an employee are legal questions. The deduction is a CPA question. The certified contribution range is an actuarial question.

What to do in the next two weeks

Write down each partner's date of birth, ownership percentage, and the compensation the plan will actually be allowed to see. For a corporation, that is projected W-2 wages. For a partnership, that is projected earned income, not the draw. Note anyone who will be 60, 61, 62, or 63 this year, because the 401(k) catch-up, if you have one, is $11,250 rather than $7,500.

List every person on payroll who is not a partner, with date of hire and whether they are full time. List every other entity a partner owns. Do not wait for a complete model to do this. A model built without it will be redone.

Decide, in a partners' meeting that is allowed to be blunt, whether equal percentages are a political requirement or a habit. If they are a requirement, say so, and expect the illustration to be constrained by the youngest partner and by the staff. If they are a habit, ask for credits that differ by age and still pass testing. Either request is legitimate. Only one of them should be implied.

When that page exists, bring it to a first conversation. We will not give you a partner-by-partner maximum in that conversation. We will tell you whether the ages, the entity, and the staff make a cash balance conversation worth the actuary's time.