Functional Expense

Every dollar an organisation spends appears twice on the Form 990: once as a natural expense category (salaries, rent, printing) and once as a functional category (programme, management and general, fundraising). The first classification describes what was bought. The second describes why — which part of the organisation's work the expenditure served. That second split is what most readers, and most donors, actually look at.

The three functional categories are defined in the instructions to Form 990 and in the Financial Accounting Standards Board's ASC 958, which governs nonprofit financial statements ↗. Programme service expenses are costs incurred in carrying out the organisation's exempt purpose — the thing it exists to do. Management and general covers oversight, governance, and anything not attributable to a specific programme or fundraising activity. Fundraising covers the cost of soliciting contributions.

Part IX of the Form 990 is a matrix: natural expense types run down the rows, functional categories run across the columns. The organisation fills in every cell. Salaries, for instance, don't sit in a single column — they are allocated across all three functions in proportion to how staff actually spent their time. The same logic applies to rent, utilities, postage, and most other shared costs.

Allocation is judgment, not measurement

Here is where the machinery gets interesting, and where the numbers become genuinely arguable. For any expense that serves more than one function — which is most of them — someone must decide how to split it. The IRS does not prescribe a single method. The Form 990 instructions ↗ state that organisations should allocate costs on a reasonable basis, consistently applied, and should document that basis. What counts as reasonable is left to professional judgment.

Time studies are one common approach for salaries: staff track how their hours divide across programmes and administrative work over a sample period, and the resulting percentages drive the payroll allocation for the year. Some organisations use square footage to allocate occupancy costs. Others use headcount, or direct costs already allocated, as a proxy for shared expenses. None of these methods is inherently wrong, but different methods applied to the same facts can produce materially different splits — and therefore materially different ratios.

That is not a flaw in the system so much as an acknowledgment that a single hour of an executive director's time genuinely serves multiple purposes simultaneously. The obligation is to apply the chosen method consistently, document it (Schedule O is the natural place), and ensure it reflects economic reality rather than a desired outcome.

Auditors review the methodology when an audit is conducted. The IRS can examine it on review. But for the very large number of organisations that are neither audited nor examined in a given year, the allocation is effectively self-reported.

What the ratio does and doesn't tell you

Because the functional split is visible on the 990, it feeds directly into the programme expense ratio — the fraction of total spending attributed to programme services — that aggregators such as Candid and Charity Navigator surface to the public. An organisation that allocates 85 percent of expenses to programme looks more efficient than one at 65 percent, and the pressure to push numbers toward programme is real.

The risk is that the ratio rewards the appearance of efficiency over the substance of it. Organisations that serve genuinely complex populations, or that invest seriously in back-office infrastructure, may carry higher management and general costs for entirely legitimate reasons. In 2013, Charity Navigator, GuideStar, and the BBB Wise Giving Alliance jointly disavowed the overhead ratio as a reliable indicator of organisational effectiveness — a position the overhead myth open letter articulated publicly — but the ratio persists in practice because readers still reach for it.

What the functional split does do reliably is make cost structure visible and comparable over time within a single organisation. A programme ratio that declines sharply year over year, or an allocation methodology that changes without explanation, is a signal worth following up. The Form 990's public-inspection rules mean anyone can pull multiple years of filings and look at the trend; services like Candid make that comparison straightforward.

The number is a starting point, not a verdict. The methodology note behind it — wherever the organisation has chosen to explain it — is where the actual story lives.