How to Read Form 990 Schedule J

The nonprofit sector's proxy statement

Public companies disclose executive pay in a DEF 14A. Tax-exempt organizations disclose it in the Form 990, and the detailed breakdown sits in Schedule J. The two instruments serve a similar purpose and are built on completely different foundations, which is why figures pulled from one cannot be dropped beside figures from the other without adjustment.

Form 990 is a public document. Anyone can retrieve it — through the IRS Tax Exempt Organization Search, or through the several nonprofit data services that mirror the filings. There is no subscription between you and the underlying data, exactly as there is none between you and EDGAR.

This guide covers what triggers Schedule J, how its columns are defined, where it disagrees with Part VII, and what has to be normalized before two organizations can be compared.

Part VII first, then Schedule J

The 990 discloses compensation in two places, and reading them in the wrong order is the most common mistake.

Part VII, Section A lists officers, directors, trustees, key employees, and the highest-compensated employees, with reported compensation from the organization and from related organizations. It is the roster. For most board members it will read $0, because nonprofit board service is typically uncompensated.

Schedule J, Part II takes the individuals whose compensation crosses the reporting threshold and breaks their pay into components. It is the detail. Schedule J is generally triggered when an individual's reportable compensation exceeds $150,000, along with several other conditions the form sets out in Part IV.

Read Part VII to establish who is in scope and what the organization reported in total. Then read Schedule J to see what that total is made of. Going straight to Schedule J means you can miss individuals reported in Part VII who never reach the schedule.

The Schedule J columns

Schedule J Part II breaks compensation into columns that do not map one-to-one onto corporate categories:

• (B)(i) Base compensation. Salary. The closest analogue to the Salary column in a Summary Compensation Table.
• (B)(ii) Bonus and incentive compensation. Performance-related cash.
• (B)(iii) Other reportable compensation. Taxable amounts that are neither base nor bonus — severance, payouts of accrued leave, taxable fringe benefits, and similar items. This column is frequently where a one-time event hides.
• (C) Retirement and other deferred compensation. Employer contributions to retirement plans and deferred arrangements.
• (D) Nontaxable benefits. Health coverage and other benefits excluded from taxable income.
• (E) Total. The sum of columns (B) through (D).
• (F) Amounts previously reported on a prior year's 990 — the column that prevents double-counting deferred amounts across years.

Column (F) is the one most often ignored and the one most likely to distort a comparison. Deferred compensation reported as earned in one year and paid in a later year can appear twice across two filings if you total naively.

Why Part VII and Schedule J disagree

The two sections frequently show different numbers for the same person, and the difference is usually not an error.

Part VII reports compensation on the calendar-year basis of the Form W-2, while the organization's own fiscal year may not be a calendar year. Schedule J reports on the organization's fiscal year. For an organization with a June year-end, that difference alone will produce two legitimate but non-identical figures.

Related-organization compensation is the second source of divergence. An executive paid partly by a parent, an affiliated foundation, or a supporting organization has compensation split across entities, each with its own filing. Reading one filing gives you part of the picture.

The practical rule is to reconcile the two before using either. If Part VII and Schedule J do not tie, find out why — the answer is nearly always fiscal-year basis, a related organization, or a column (F) deferral, and identifying which one is a prerequisite to any comparison.

What has to be normalized before you compare

Two organizations' Schedule J totals are not comparable as filed. At minimum, four adjustments are needed.

Fiscal year alignment. Compare tax years, not filing dates. A 990 filed in 2026 may report the 2024 tax year, and organizations with different year-ends are reporting different periods under the same label.

One-time events. A severance payment or an accrued-leave payout in column (B)(iii) can double an executive's apparent total in the year it lands. Whether you include it depends on the comparison you are making, but you have to see it first.

Related-organization aggregation. If an executive is paid across affiliated entities, the comparable figure is the aggregate, which means pulling more than one filing.

Benefit load. Column (D) nontaxable benefits vary widely between organizations. Comparing base salary alone understates the difference; comparing totals without noting the benefit split obscures it.

None of these are exotic. They are the routine work of making two filings say the same kind of thing, and skipping them is how a nonprofit pay comparison becomes indefensible.

Where the 990 is stronger than the proxy, and where it is weaker

Form 990 has one significant advantage over a DEF 14A: it covers organizations that have no equity at all, so total compensation is genuinely cash and benefits rather than a figure dominated by grant-date equity valuations that may never be realized. Nonprofit totals are closer to what was actually received.

It is weaker in timeliness. Form 990 filings are often available considerably later than a proxy statement for the equivalent period, so the most recent nonprofit data you can obtain generally lags the most recent public-company data.

And it shares one limitation with every disclosure regime discussed on this site: it covers only filers. Private-company pay is disclosed nowhere, and no reading technique recovers a figure that was never filed.

Reading Part VII properly first

Because Schedule J gets the attention, Part VII is often skimmed — and Part VII is where the scope of the disclosure is actually established.

Part VII Section A lists each person by name and position, with a set of checkboxes indicating role: individual trustee or director, institutional trustee, officer, key employee, highest compensated employee, or former. Those boxes matter. A "former" designation attached to a large figure usually signals a severance or deferred payout rather than ongoing compensation, and reading it as a current salary badly distorts a peer set.

The section then reports three figures per person: reportable compensation from the organization, reportable compensation from related organizations, and estimated other compensation from the organization and related organizations. The middle column is the one to watch. A modest figure in the first column beside a substantial figure in the second means the organization you are reading is not the primary payer, and the filing in front of you is not the whole picture.

Average hours per week is also reported, including hours for related organizations. For an executive splitting time across a system, that field is what lets you judge whether a compensation figure reflects a full-time role.

Common misreadings

Four errors account for most bad nonprofit compensation comparisons.

• Treating the filing year as the data year. A 990 carries a tax year that is often two years behind the date you retrieved it. Comparing a recently filed return against an older one without checking both tax years compares different periods.
• Missing the related-organization split. Health systems, universities, and large foundations frequently pay executives across affiliated entities. One filing gives one slice.
• Double-counting deferrals. Column (F) exists precisely to flag amounts already reported in a prior year. Summing across years without netting it inflates the total.
• Reading a severance year as a salary year. A departure payment sitting in other reportable compensation can double an apparent total, and the person may not have served the full year.

Each of these is visible on the face of the filing. None of them is detectable in a spreadsheet of totals someone else assembled, which is the argument for pulling the filings yourself.

Where Schedule J adds narrative

Schedule J is not only a table. Part I asks a series of questions about compensation practices — whether the organization provided first-class travel, a housing allowance, a discretionary spending account, or similar benefits, and whether compensation was set using a review of comparability data and contemporaneous documentation.

Part III then provides space for narrative explanation, and organizations use it to explain unusual amounts. If a total looks anomalous, Part III is frequently where the reason is stated plainly — a retention payment, a contract buyout, a one-time settlement of a deferred arrangement.

Reading Part III before drawing a conclusion about an outlier is the nonprofit equivalent of reading proxy footnotes, and it resolves a large share of figures that look inexplicable from the table alone.

A working sequence

Retrieve the filing. Read Part VII to establish the roster and the reported totals. Move to Schedule J Part II for the component breakdown. Check column (F) for previously reported deferrals. Reconcile against Part VII and account for any difference. Record the organization's EIN and the tax year beside every figure you take.

That last step is the one that converts a spreadsheet of numbers into something you can defend. A nonprofit compensation figure without its EIN and tax year is an assertion; with them, it is evidence anyone can check.

Once the figures are reconciled, the Nonprofit ED Total-Compensation Normalization Workbook takes the Schedule J columns you have recorded, reconciles them against Part VII, and turns each one into a single comparable total positioned against the peers you enter. If you would rather start from a finished benchmark, the Nonprofit Executive Director Salary Benchmark Report sets out percentile bands by mission area and budget band, with an EIN and tax year cited behind every band.