The share of a nonprofit’s budget classified as overhead tells you almost nothing about whether the organization does its job well. It is an accounting ratio built from categories the organization assigns itself, it rewards understating the cost of running an institution, and it has no established relationship to whether anyone’s life improved. Donors use it because it is a single number and single numbers feel like judgment. It is a poor one.
What the ratio actually measures
Nonprofits report functional expenses on Form 990 in three buckets: program services, management and general, and fundraising. The overhead ratio is the second and third buckets divided by the total. The Form 990 filed with the IRS is where the figure comes from, and it is a public document for most exempt organizations.
Read that definition again and the problem is visible. Management and general covers accounting, legal, human resources, insurance, audit, technology, and the executive function. Fundraising covers the cost of raising the money that pays for everything else. Neither category describes waste. Both describe the machinery required for an organization to exist and keep existing.
The number is softer than it looks
Expense allocation involves genuine judgment. A program director who spends part of the week supervising staff and part of it delivering services has to be split between categories. A database serves both program tracking and donor management. A regional office hosts both. Two organizations doing identical work can report meaningfully different ratios based on defensible but different allocation choices.
That is not fraud. It is the ordinary result of forcing continuous activity into three discrete boxes once a year. It does mean that comparing the ratios of two organizations to two decimal places is a false precision exercise.
The comparison across causes is worse
An organization that ships commodities has a structurally low ratio, because the commodities themselves count as program expense and dwarf everything else. An organization that does research, litigation, or public education has a structurally high one, because its program is people, and people come with the payroll, benefits, and office costs that partly land in management and general. Ranking the two against each other measures the cost structure of the work, not the quality of it.
The starvation cycle
Sector researchers have documented a self reinforcing loop for two decades. Donors demand low overhead. Organizations respond by underinvesting in accounting systems, technology, evaluation, and staff development, and by reporting their costs as favorably as the categories allow. Weak infrastructure produces weak measurement and weak execution. The organization then has less evidence of impact to show, which increases the pressure to compete on the overhead number instead.
The pathological version is an organization that cannot tell you whether its program worked, because it never funded the evaluation that would have found out, because evaluation is management and general.
What the sector itself said
In 2013 the leaders of BBB Wise Giving Alliance, GuideStar, and Charity Navigator, three of the largest charity evaluation bodies in the country, published an open letter to donors arguing that the overhead ratio alone is a poor measure of performance and should not be used as the primary basis for giving decisions. The three organizations most responsible for popularizing the metric went on record against using it that way.
What to look at instead
None of the alternatives is a single number, which is the honest part of this.
Confirm the organization is what it says it is
Start with the IRS Tax Exempt Organization Search. It shows whether an organization is currently recognized as exempt, its deductibility status code, and whether its exempt status was automatically revoked for failure to file. That check takes a minute and screens out a category of problem the overhead ratio never touches.
Read Part III of the Form 990
Part III asks the organization to describe its program service accomplishments and report expenses against each program. This is where you find out what the organization actually did with the money rather than which bucket the money sat in. Vague, unquantified descriptions repeated verbatim year over year are informative. So are specific ones.
Look for outcome reporting, not activity reporting
Activity reporting counts inputs and outputs: meals served, reports published, people trained. Outcome reporting asks whether anything changed as a result. Most organizations do the first. The ones doing the second are usually explicit about their methodology and, more tellingly, about what did not work.
Check governance and financial health
An independent board majority. An audit if the organization is large enough to warrant one. Enough operating reserve to survive a bad quarter, which a very low overhead ratio often indicates the organization does not have. Executive compensation disclosed in the 990 and reasonable for the size and location of the organization.
Check whether the work matches the problem
This is the least quantifiable and probably the most important. Does the organization’s theory of the problem make sense? Does its program follow from that theory? Organizations that publish their reasoning are easier to evaluate than ones that publish only their results, and it is worth reading how a group defines the problem before deciding whether its solution is credible. Fight For A Living Wage, for instance, frames the issue as affordability across housing, health care, childcare, food, transportation, and education rather than as the wage floor alone, which is a specific and checkable claim about causes. Roundups that group organizations by the problem they work on, like this write-up of groups working on economic hardship, are more useful as a starting list than as a ranking.
A workable framework
Confirm exempt status. Read Part III of the most recent Form 990. Look for evidence the organization measures whether its work does anything. Check that governance and reserves are sane. Then, and only then, glance at the expense ratios, and treat an extremely low one as a question rather than a virtue.
An organization spending 8 percent on overhead is either unusually efficient or unusually underbuilt, and the ratio cannot tell you which. Every other item on the list can.