How to Build a Defensible Comparability Study

When a nonprofit board sets executive compensation, the IRS doesn’t just ask “is this reasonable?” It asks “can you prove it?” A defensible comparability study is how you prove it. Done well, it protects your organization from intermediate sanctions under IRC §4958 and gives your board confidence that they’re making decisions based on real data rather than gut feelings.

Here’s how to build one that holds up.

1. Define your peer group criteria before looking at compensation

This is where most studies go wrong. If you pick your peer group after seeing the comp data, you’ve introduced selection bias that an IRS agent will flag. The Treasury regulations under 26 CFR 53.4958-6 require that the authorized body “obtain and rely upon appropriate data as to comparability prior to making its determination.” Your criteria should be documented before you pull a single salary figure.

A defensible peer group typically filters on:

• Mission similarity — NTEE classification or a clearly articulated mission description. A mid-size environmental nonprofit shouldn’t benchmark against a large hospital system.
• Budget size — Total expenses within a reasonable range (typically 50%–200% of your organization’s budget). A $2M organization pays differently than a $50M one for the same role.
• Geography — National, statewide, or metro-area comparisons depending on your labor market. A San Francisco nonprofit competes for talent differently than one in rural Mississippi.
• Role comparability — You’re comparing like to like. CEO to CEO, CFO to CFO. Title normalization matters because nonprofits use wildly inconsistent titles.

2. Use reliable, verifiable data sources

The Treasury regulations specifically recognize “compensation levels paid by similarly situated organizations, both taxable and tax-exempt, for functionally comparable positions” as appropriate comparability data (26 CFR 53.4958-6(c)(2)). Form 990 filings are the gold standard because they’re public records filed under penalty of perjury.

Other acceptable sources include compensation surveys from recognized firms (ERI, Mercer, SullivanCotter) and actual written offers from competing employers. But 990 data has a unique advantage: anyone can verify it. Your board, an IRS agent, or a journalist can pull the same filings and confirm your numbers.

What to watch out for

• Stale data — 990 filings run 1–2 years behind. Make sure you’re using the most recent available filings and note the tax years in your documentation.
• Apples-to-oranges compensation — Some studies mix base salary from one source with total comp from another. Form 990 Part VII reports three columns (reportable, related org, and other compensation). Pick one definition and use it consistently.
• Small sample sizes — A peer group of 3 organizations isn’t a real benchmark. The regulation provides a safe harbor for small organizations (under $1M gross receipts) using three comparable orgs, but larger organizations should aim for 15+ peers.

3. Analyze the data honestly

A good comparability study reports the full distribution, not just the number that supports a predetermined conclusion. At minimum, include:

1. Median compensation for the role within your peer group
2. 25th and 75th percentiles to show the reasonable range
3. The number of organizations in the sample
4. Any adjustments you made for geography, budget size, or role scope

If your current compensation falls at the 90th percentile, that’s not necessarily wrong — but the regulation requires that you document “the basis for the determination” when compensation exceeds the comparability range (26 CFR 53.4958-6(c)(3)(ii)). Conversely, if you’re at the 20th percentile and losing talent, the data supports a case for an increase.

4. Document the board’s process

The comparability data is only half the story. Under the rebuttable presumption standard, the IRS also requires that your board (or a designated committee) followed a deliberate process:

• The committee reviewed the comparability data before setting compensation
• Members with conflicts of interest recused themselves from the vote
• The decision was recorded concurrently — before the later of the next board meeting or 60 days after the determination

If you can satisfy all three requirements — independent body, comparability data, concurrent documentation — the IRS cannot challenge your determination based solely on the ground that compensation is unreasonable. They must first produce contrary evidence to rebut your comparability data. That’s a significant legal advantage.

5. Keep it current

A comparability study isn’t a one-time document. Best practice is to refresh the analysis every 2–3 years or whenever there’s a material change in the executive’s role or the organization’s budget. A study from 2018 won’t protect a compensation decision made in 2026.