Learn · §4958 Compliance
§4958 ComplianceWhat Counts as “Appropriate Comparability Data” for Nonprofit Executive Pay?
The four forms of §4958 comparability data — and what makes a peer set “appropriate” for your organization's budget, sector, and geography rather than merely available.
Under Treasury Regulation §53.4958-6, appropriate comparability data means evidence of what similarly situated organizations — taxable and tax-exempt — pay for functionally comparable positions. The regulation names four acceptable forms: compensation paid by similar organizations for similar roles, the documented availability of similar services in the organization's geographic area, current compensation surveys compiled by independent firms, and actual written offers from similar institutions competing for the same executive. The data must be obtained and relied upon before the board approves the compensation.
The operative word is appropriate — and that is where most comparability files are weak. Data can be genuine, current, and professionally compiled and still be inappropriate for a particular organization if it isn't matched to that organization's circumstances.
What makes comparability data “appropriate” rather than merely available?
Four dimensions of match, in rough order of weight:
Budget size. Organization revenue is the strongest single predictor of nonprofit executive pay. A national median for “Executive Director” blends $500K organizations with $50M ones and describes neither; comparables should bracket your organization's budget.
Functional comparability of the role. Titles mislead — a “Vice President” at one organization is a department head at another. What matters is scope: staff and budget managed, program complexity, fundraising accountability. Comparables should match the job, not the title.
Sector. An arts organization, a health system, and a food bank of identical budget pay differently. Peer sets should reflect mission area (NTEE classification is the standard taxonomy), with judgment about when cross-sector data is legitimately relevant — the regulation explicitly permits taxable-employer data for comparable roles.
Geography. Labor markets are local. Metro-level data matters most for staff roles; for executives, regional cost-of-labor differences still move the range meaningfully.
How many data sources should a board use?
The regulation sets no minimum for organizations over $1 million in gross receipts (below that, three comparables from similar communities suffice). Practitioner consensus, however, is that a defensible file rests on multiple independent sources — commonly three — because each source has different methodology, sample, and lag. A typical strong file pairs a filings-based analysis (IRS Form 990 data, modeled to the organization's profile) with an independent survey and, where relevant, state-association data or an actual written offer.
What about data freshness — and the 990 lag problem?
All Form 990-based data shares one structural weakness: filings become public 18–30 months after the compensation was paid. A board setting fiscal 2027 pay on raw 2024 filings is anchoring two to three years behind the market, which systematically understates reasonable compensation in inflationary periods. The remedies are to use the most recent filings available and to age the data to the decision date using a documented index — the BLS Employment Cost Index is the standard instrument — with the adjustment method recorded in the comparability file. An unadjusted number isn't wrong, but a board should know what year it is actually looking at.
What should the comparability file contain?
Enough that a reviewer three years later can reconstruct the decision: the peer organizations or survey sources used and why they were selected, the data vintage and any aging adjustments, the percentile range and where the proposed compensation falls, total compensation rather than base salary alone, and the date the body reviewed it. This file is one of the three legs of the rebuttable presumption of reasonableness — the others are advance approval by an independent body and contemporaneous minutes.
Frequently asked questions
Can we use for-profit salary data?
Yes — the regulation expressly includes compensation paid by similarly situated taxable organizations for functionally comparable positions. It should supplement, not replace, nonprofit data for roles where the markets genuinely overlap.
Is free data from 990 lookups enough?
Raw lookups of a few organizations' filings can satisfy the letter of the rule for small organizations, but the appropriateness test still applies: cherry-picked or mismatched comparables are the most common audit vulnerability. The selection logic matters as much as the data.
Does a consultant's report count as a survey?
A custom comparability study and an independent published survey are both acceptable forms. What the IRS looks for is independence, relevance to the role and organization, and reliance before the decision.
Where does the percentile target come from?
The data describes the market range; the board chooses the position within it and documents why. Paying at the 75th percentile can be reasonable with a documented rationale (performance, complexity, retention risk); paying at the median without any rationale can still be challenged if the peer set was wrong.
Educational information, not legal or tax advice. CauseComp builds budget-, sector-, and geography-matched comparables sets from 610,000+ Schedule J records — aged to the current year with the methodology documented on every report.
Educational content from CauseComp, a service of RB Consulting Services, LLC. Provides data and documentation to support board deliberations — not legal advice.