When you donate to a charity, your money does not flow straight to the cause. Some of it pays for the people who run the organisation, the computers they work on, the training they receive, the cost of raising the next round of funds. These running costs are called overhead. The charity overhead ratio measures what share of total spending goes to those costs rather than to direct programme delivery — and for decades it has been used as a shorthand for how good a charity is.
The conventional wisdom, pushed by ratings agencies and watchdogs for decades, was straightforward: lower overhead equals a better charity. A charity spending 90 cents in every dollar on programmes is more efficient than one spending 70 cents. It felt logical. It felt measurable.
It turns out to be wrong.
The charities that GiveWell recommends, the world's most rigorous charity effectiveness evaluators, have higher expense ratios than the charities they don't recommend. [13]
Read that again.
The organisations spending more on overheads are the ones doing more good. Not despite the overhead. Often, because of it.
This is the overhead myth: the idea that a charity's administrative costs are a meaningful guide to its quality. It's one of the most persistently damaging misconceptions in philanthropy. It's been comprehensively dismantled by researchers, admitted to by the very watchdogs who helped create it, and lived painfully by thousands of charities starved of the investment they need to do their work properly.
How the myth got so powerful
The overhead ratio (what percentage of a charity's spending goes to programmes versus administration) became popular for one simple reason: it's a number, and numbers feel objective.
Charity rating agencies that emerged in the early 2000s needed something measurable. Overhead ratio was easy to calculate from public financial filings and felt intuitively right. If more money goes directly to programmes, that must be better.
The problem is that "overhead" isn't waste. It includes IT systems, staff training, performance measurement, HR, financial management, and fundraising that funds future work. These are investments. Cut them, and you don't get a more efficient charity. You get a more fragile one.
The starvation cycle
In 2009, Ann Goggins Gregory and Don Howard published a paper in Stanford Social Innovation Review naming what was quietly happening across the sector: the nonprofit starvation cycle. [1]
The mechanism is a closed loop:
Funders expect low overhead
↓
Charities under-invest in infrastructure
and under-report their actual costs
↓
Artificially low numbers become the benchmark
↓
Expectations tighten further
↓
Real investment falls again
The consequences, documented across hundreds of organisations, were visceral. Nonfunctioning computers. Staff without adequate training. In one case, furniture so deteriorated that removal workers refused to move it.
One-third of the charities in the study reported zero fundraising costs, a claim that was mathematically impossible in most cases. The underreporting wasn't dishonesty in any meaningful sense; it was survival. When your funder expects 15% overhead and your real cost structure is 33%, you find ways to make the numbers work on paper, even as the organisation quietly hollows out.
A 2015 study by Jesse Lecy and Elizabeth Searing confirmed the trend empirically: using 25 years of sector data, they found a sustained downward trend in reported overhead across thousands of nonprofits. Not because charities were becoming more efficient, but because the pressure to minimise the number had become overwhelming. [2]
"These groups weren't running lean. They were running on empty."
When the watchdogs admitted they were wrong
In June 2013, the CEOs of three of the most influential charity watchdogs in the world published a joint open letter. [7] The signatories were Charity Navigator, the BBB Wise Giving Alliance, and GuideStar, the three organisations most responsible for entrenching overhead ratio as the go-to donor metric.
They called it The Overhead Myth. They launched a campaign with the same name. And they said this:
"The percent of charity expenses that go to administrative and fundraising costs is a poor measure of a charity's performance. When we focus solely or predominantly on overhead, we can and do starve charities of the freedom they need to best help the people and communities they are trying to serve."
"The people and communities served by charities don't need low overhead. They need high performance."
This was not a minor recalibration. These were organisations that had spent years enforcing overhead standards. The letter was an institutional admission that a metric they had legitimised was causing harm.
Two years later, Darren Walker, President of the Ford Foundation and one of the world's largest philanthropic institutions, went further. In a public letter, he described the sector's behaviour as "a charade": [8]
"At Ford, we have been willing participants in this charade. Our policy of 10 percent overhead on project grants in no way allows for covering the actual costs to administer a project."
The same year, Dan Pallotta's TED talk brought the argument to a global audience. Pallotta described how his events company raised $569 million for AIDS and cancer causes over nine years, then collapsed after sustained media criticism of its ~40% overhead ratio. The successor events raised a fraction of what his organisation had generated. His talk has now been viewed more than 5.4 million times. [12]
"We confused morality with frugality."
Australia's version of the same story
This is not only a US problem.
In 2022, the Centre for Social Impact and Social Ventures Australia published Paying What It Takes, a study of nine Australian not-for-profits that found their true indirect costs averaged 33% of operating budgets. Funding agreements typically covered only 10–20%. [9]
Australian businesses spend twice as much per employee on training, IT, and quality systems as the NFPs in the study did.
The researchers coined a term for what results: "volcano funding", where programme delivery is funded but the organisational core isn't. The edges grow. The centre hollows out.
The overhead myth — visualised
The volcano funding problem
Two organisations. Same programme output. Completely different structural health.
All layers resourced
Programmes funded, core starved
The ACNC, Australia's national charity regulator, is explicit about this. Their official guidance states: [11]
"Because charities are so varied, there is not a one-size-fits-all standard ratio or percentage to measure what might be deemed 'reasonable' spending on administration."
The Paul Ramsay Foundation listened. From July 2022, they moved to a 30% standard for indirect costs across all grants, explicitly citing the starvation cycle research as the reason. One of Australia's most significant philanthropic institutions formally broke with the overhead-minimisation model. [10]
Why comparing ratios across charities often makes no sense
Here's the deeper structural problem: comparing overhead ratios across different types of charity is like comparing the rent-to-revenue ratio of a restaurant and a law firm, then concluding the restaurant is better managed.
An advocacy organisation (one that campaigns for policy change, runs awareness programmes, or publishes research) is almost entirely composed of what traditional accounting calls overhead. The expertise, relationships, and communications capacity are the product. There is no physical service to deliver.
A service delivery organisation (housing support, clinical healthcare, crisis counselling, legal aid) needs those same functions plus significant direct service costs. The ratio will naturally look different, and that difference is structural, not a quality signal.
Looking at some of the charities we've profiled on Cicadar, the range of programme ratios reflects operational models, not rankings:
| Charity | Programme ratio | What it reflects |
|---|---|---|
| Amnesty International Australia | 99.5% | Advocacy model: campaigns, volunteer networks, no physical service delivery |
| GIVIT | 94% | Platform model: connects donors to goods, minimal direct service infrastructure |
| Landcare Australia | 81.7% | Coordinator model: 24 staff enabling 140,000 volunteers, infrastructure is the leverage |
| Fred Hollows Foundation | 75% | International operations: compliance, translation, country-office management across 25+ countries |
| ASRC | 66.9% | Full-service model: legal clinics, housing, medical care, crisis support for asylum seekers |
Amnesty's 99.5% is remarkable. ASRC's 66.9% is also remarkable, and arguably harder to sustain given the complexity and intensity of what they deliver. Neither number tells you which does more good. They tell you about operational structure.
What to actually look at
The good news: there are better questions.
1. Cost per outcome: not cost per dollar spent on admin, but what one dollar actually produces. $40 for an eye surgery. $0.99 for a meal delivered. $95 for a mental health counsellor contact. These figures are harder to find and require real research, which is exactly what Cicadar is built to surface.
2. Transparency: does the organisation publish its financials clearly? Does it explain how it measures its own impact? Does it acknowledge what it doesn't know? Willingness to be examined tells you more than any single ratio.
3. Financial sustainability: how many months of reserves does the organisation hold? Is revenue diversified, or dangerously dependent on one funder? An organisation squeezed by overhead pressure often shows up here first: shrinking reserves, rising turnover, key staff departures.
4. Room for more funding: could this organisation meaningfully do more with an additional $10,000? Or is it at capacity? Effective giving means finding organisations where your donation genuinely enables something that wouldn't otherwise happen.
The bottom line
The overhead ratio tells you how a charity accounts for its costs. It doesn't tell you whether a sick child got the medicine they needed, whether a family found stable housing, or whether a hectare of degraded land was restored.
The three organisations that spent years enforcing overhead standards eventually said so themselves. The research confirmed it. Australia's leading funders are responding. The ACNC has said it plainly.
Now donors can act on it.
Sources
[1] Gregory, A. G. & Howard, D. (2009). "The Nonprofit Starvation Cycle." Stanford Social Innovation Review. ssir.org
[2] Lecy, J. D. & Searing, E. A. M. (2015). "Anatomy of the Nonprofit Starvation Cycle." Nonprofit and Voluntary Sector Quarterly, 44(3), 539–563. doi.org/10.1177/0899764014527175
[3] Gneezy, U., Keenan, E. A. & Gneezy, A. (2014). "Avoiding overhead aversion in charity." Science, 346(6209), 632–635. science.org
[4] Altamimi, H. & Liu, Q. (2022). "The Nonprofit Starvation Cycle: Does Overhead Spending Really Impact Program Outcomes?" Nonprofit and Voluntary Sector Quarterly, 51(6), 1324–1348. doi.org/10.1177/08997640211057404
[5] Meer, J. (2017). "Are overhead costs a good guide for charitable giving?" IZA World of Labor, 329. wol.iza.org
[6] Hung, C., Hager, M. A. & Tian, Y. (2023). "Do Donors Penalize Nonprofits With Higher Non-Program Costs?" Nonprofit and Voluntary Sector Quarterly, 52(6), 1587–1608. doi.org/10.1177/08997640221138260
[7] GuideStar, Charity Navigator & BBB Wise Giving Alliance. (2013). "The Overhead Myth" open letter. guidestar.org
[8] Walker, D. (2015). "How We're Addressing the Overhead Fiction." Ford Foundation. philanthropynewyork.org
[9] Centre for Social Impact & Social Ventures Australia. (2022). Paying What It Takes. assets.csi.edu.au
[10] Paul Ramsay Foundation. (2022). "Commits to Paying What It Takes." paulramsayfoundation.org.au
[11] ACNC. "Charities and administration costs." acnc.gov.au
[12] Pallotta, D. (2013). "The way we think about charity is dead wrong." TED2013. ted.com
[13] GiveWell. (2009). "The worst way to pick a charity." blog.givewell.org
[14] Pro Bono Australia. (2022). "Australian charities locked into funding starvation mode." probonoaustralia.com.au
Programme ratio figures for Cicadar-profiled charities are derived from ACNC Annual Information Statements. They are presented to illustrate operational model variation, not as quality rankings.