Building a Compensation Peer Group for a Nonprofit

Every defensible executive pay decision rests on a peer group — the specific set of organizations whose reported pay answers the question "what do like enterprises pay for like services?" The phrase comes from the reasonableness standard itself (Treas. Reg. §53.4958-4(b)), and it makes peer-group construction a legal exercise before it is an analytical one: the group is the board's claim about which market it competes in. This page covers what a defensible group requires. It does not hand you one — and the last section is honest about why.

The three matching axes

1. Size — the axis that moves pay most

Organization budget is the strongest single predictor of nonprofit executive pay in the filings — our variance analysis across 732,319 filings found budget size explains roughly four to six times what either geography or mission does. A peer group whose members are five times your budget is not a peer group; it is an aspiration with citations. Match on budget first, and state the band you matched to.

2. Mission — the market for the skills

Sector matters because it defines the labor market: a hospital system CFO and a food-bank CFO carry the same title into different markets. Mission-matching does not require identical NTEE codes — it requires that a stranger reading the list would accept that these organizations do comparable work with comparable complexity.

3. Market — where you actually hire

Geography is the axis boards debate most and the filings reward least. It still belongs in the definition — a board should be able to say whether it prices against its state, its region, or a national market for a specialized role — but its weight should reflect what the data shows: it is the smallest of the three dials.

What the regulation actually requires

The rebuttable presumption's data prong asks for "appropriate data as to comparability" (Treas. Reg. §53.4958-6(c)(2)) — compensation paid by similarly situated organizations, both taxable and tax-exempt, for functionally comparable positions. Two things follow. First, the comparables must be positionally comparable — matched on what the person does, not what the title says. Second, "appropriate" scales with the organization: the regulation gives small organizations — annual gross receipts under $1 million — a safe harbor at three comparables (Treas. Reg. §53.4958-6(c)(2)(ii)). Larger organizations get no such number. "Appropriate" is judged against their own circumstances, which is why a cohort deep enough that no single filer's unusual year can move the answer is the safer construction.

What makes a peer group fall apart

• Thin cells. A group of four where one member had a CEO transition year is not a benchmark; it is noise wearing a median's clothes. This is why CauseComp publishes nothing below 20 filings — the same discipline applies to a hand-built group.
• Unstated definitions. If the minutes cannot say what qualified an organization for inclusion, the group cannot be defended later — and later is when it matters.
• Post-hoc selection. A group assembled after the pay conversation, or trimmed when the first answer disappointed, is advocacy. Fix the definition before anyone sees a number, and record the definition, not just the members.
• Title matching. "Executive Director" spans organizations of every size; Schedule J's components are comparable, titles often are not.

The honest part: this is slow

Assembling a genuinely matched cohort from public filings is slow work — finding candidate organizations, confirming size and mission from their returns, extracting the right compensation column for the right year, and documenting all of it. It is slow enough that most boards don't do it. That sentence is true, and it is also the reason services exist: CauseComp's Executive tool does this construction against the full filing corpus and names the members (plans and pricing are public), and our research on comparability in practice measures what happens when boards skip the work — the shortcut misses the similarly-situated figure by 38% in the typical case. Whether a board builds its group by hand or buys the construction, the requirements on this page are the same.

A peer group answers "what do they pay?" The prior question — what market are we choosing to pay against, and why? — is a compensation philosophy, and boards that settle it first find the peer-group debate largely settles itself.

Documented comparability data like the above supports — but does not by itself establish — a board's rebuttable presumption of reasonableness under IRC §4958 (intermediate sanctions). The presumption also requires advance approval by an independent board body and contemporaneous documentation. CauseComp provides comparability data for informational purposes and does not provide legal or tax advice.