Profit in a partnership follows whatever the partners wrote down, and when they wrote nothing down, it follows neither capital nor work. The Uniform Partnership Act (1997) gives each partner an equal share of profit and the matching share of loss "regardless" of how much capital each one put in (section 401(b)). A partner who contributed $150,000 against a partner's $10,000 still splits the profit in half unless the agreement says otherwise.

That default is almost never what two partners actually want once the business is running, and it is also almost never written into the agreement as the final answer. What replaces it falls into four recognizable shapes, and the right one depends on a single question: does the split need to reward capital, reward work, or reward both in a defined order. This is general information, not legal advice for a specific partnership; a lawyer in the partners' state should review the final clause.

Four ways partnerships actually split profit

Profit-split formulas compared

Formula How it works Fits best when
Pro rata to capital Profit is split in the same ratio as each partner's capital account balance, recalculated as capital changes Both partners are mostly investors, or the business is capital-intensive and returns should track who funded it
Salary, then split the rest Partners who work in the business are paid a guaranteed payment first (taxed as ordinary income, not a profit share), and whatever is left is split by an agreed ratio One or more partners work full time and others do not, and the working partner's pay should not depend on a good year
Preferred return, then split Capital partners receive a set percentage return on their contributed capital first (a "preferred return"); the rest is split by an agreed ratio, often weighted toward the partner doing the work A capital partner wants a floor on return before anyone shares in the upside, common where one partner is mostly financing the other
Tiered waterfall Profit passes through several tiers in order, each with its own split, until the tiers are exhausted; a late tier usually favors the partner who is not getting paid a salary along the way More than two tiers of economics are negotiated at once, for example a return of capital, then a preferred return, then a split that shifts toward the active partner above a target return

The first three are really one idea in different strengths: pay capital, then pay work, in whatever order the partners agree matters more. The waterfall is the same idea stacked twice, usually because one partner is financing the business and the other is running it, and both want the economics to shift as the business performs better than expected.

A two-tier waterfall, worked through

Splitting $180,000 of annual profit under a two-tier waterfall (hypothetical)

Partner A contributed $200,000 in capital and does not work in the business. Partner B contributed $20,000 and runs it full time. The agreement gives A an 8% preferred return on capital first, then splits everything above that 60% to B and 40% to A, recognizing B's work. This year the business earns $180,000 of profit before any split.

A's preferred return (8% of $200,000)$16,000
Remaining profit after the preferred return ($180,000 minus $16,000)$164,000
B's 60% of the remainder$98,400
A's 40% of the remainder$65,600
A's total for the year (preferred return plus 40% share)$81,600
B's total for the year$98,400

A, who put in ten times the capital and does none of the work, still ends the year with more than $81,000 because the preferred return is paid before anything else is split. B, who put in a tenth of the capital and runs the business, ends with the larger total because the second tier favors the active partner. Change either rate (the 8% preferred return or the 60/40 split) and both totals move, which is why the rates themselves, not just the structure, are the part worth negotiating carefully.

Why special allocations need more than an agreement

A partnership agreement can split profit any way the partners want for tax purposes too, but only within a limit the Internal Revenue Code imposes. Under section 704(b), a partner's distributive share follows the partnership agreement unless the agreement is silent on an item or the allocation does not have "substantial economic effect," in which case the IRS reassigns the item according to each partner's actual interest in the partnership. In practice this means an allocation has to track real economic consequences, not just a tax result the partners prefer: a partner allocated a loss has to actually bear that loss in the capital account, and a partner allocated income has to actually receive the matching economic benefit. A waterfall built around real capital accounts and real preferred returns, like the one above, is the kind of allocation that tends to hold up; one built purely to shift taxable income to whichever partner is in a lower bracket that year is the kind that does not.

The objection worth taking seriously is that most two-partner firms are small enough that the IRS will never look closely at how profit was split. That is probably true in any given year. It stops being a comfortable bet the moment the split itself becomes the dispute, whether that is an audit or a falling-out between the partners, because an allocation that was never economically real is also an allocation that is hard to defend to anyone, including the other partner.

Choosing a formula is only half the job. The other half is deciding whether any of the partners should be paid for the work itself, separately from how the remaining profit gets split, which is covered in how business partners pay themselves. And every formula above assumes both partners' capital accounts are tracked accurately in the first place, including when one partner's work is treated as a form of capital rather than a salary; that distinction is the subject of sweat equity partnerships. The split itself, once chosen, belongs in the agreement alongside the rest of what to include in a partnership agreement, since an undocumented waterfall is no more enforceable than the equal default it replaces.