The charity ratings problem
Rees Calder · 4 May 2026 · 6 min read
Charity Navigator has rated over 200,000 nonprofits. It is the dominant force in charity evaluation, shaping billions of dollars in giving each year. Donors see four stars and feel confident. Three stars and they hesitate. Two stars and they walk.
Here is the problem: those stars mostly measure financial housekeeping. They do not, in any rigorous sense, measure whether the charity is actually good at helping people.
What the stars actually measure
Charity Navigator's rating system has three components. Financial health (accounting ratios, revenue trends, working capital). Accountability and transparency (board independence, audit practices, donor privacy policies). And since 2020, an "impact and results" score that is, charitably, a work in progress.
The financial metrics reward low overhead and strong accounting controls. The accountability metrics reward good governance paperwork. Neither tells you the thing you actually want to know: does this charity produce results, and how do those results compare to what other charities achieve with the same money?
A charity can score four stars by running an extremely well-administered programme that achieves nothing. A charity can score three stars by running a messy but wildly effective programme with slightly high fundraising costs. The rating system cannot tell these apart.
The overhead obsession
The most damaging legacy of Charity Navigator's early methodology is the overhead myth: the idea that a charity's ratio of programme spending to administrative costs tells you something meaningful about quality.
It does not. Dan Pallotta's 2013 TED talk made this case persuasively, and the evidence has only accumulated since. The Bridgespan Group's 2019 research found that nonprofits investing more in professional development, data systems, and talent (all classified as "overhead") consistently outperformed peers on programme outcomes. The Stanford Social Innovation Review published a letter signed by the heads of GuideStar, BBB Wise Giving, and Charity Navigator itself in 2013, explicitly telling donors to stop using overhead ratios as a primary indicator.
And yet the architecture of the rating system still rewards low overhead. Charities game it: recategorising administrative costs as programme costs, underpaying staff, skipping monitoring and evaluation, all to hit the magic ratio that earns a fourth star.
What effective evaluation actually looks like
GiveWell, by contrast, evaluates roughly 9 top charities at any given time. It asks a fundamentally different question: how much measurable good does this charity produce per dollar donated, compared to the next best alternative?
This means GiveWell examines:
- What is the intervention's evidence base? (Are there randomised controlled trials? How large? How recent?)
- What is the cost per unit of outcome? (Cost per life saved, cost per DALY averted)
- Is the charity room-limited? (Can it productively absorb more funding?)
- What is the counterfactual? (Would this work get funded anyway without your donation?)
The result is a tiny list of charities where GiveWell has high confidence the marginal dollar does exceptional good. The Against Malaria Foundation: roughly $5,500 per life saved. Malaria Consortium's seasonal chemoprevention: roughly $3,000-5,000 per life saved. These numbers come from transparent models with explicit assumptions you can examine and disagree with.
Charity Navigator rates 200,000 organisations. GiveWell rates 9. That ratio tells you something about the depth of analysis required to actually answer the question "does this charity work?"
The middle ground that doesn't exist
You might think there should be an evaluation system that combines Charity Navigator's breadth (rating thousands of charities) with GiveWell's depth (measuring actual outcomes). That system does not exist because it cannot: rigorous cost-effectiveness analysis requires months of staff time per charity, and there are not enough analysts in the world to do that for 200,000 organisations.
The UK has the Charity Commission, which is a regulator, not an evaluator. It checks legal compliance. It does not and cannot assess whether charities are effective. In the US, GuideStar (now Candid) provides raw financial data. BBB Wise Giving Alliance checks governance standards. Impact Matters tried automated cost-effectiveness estimates before being acquired by Charity Navigator in 2020, and the integration has produced limited results.
What this means in practice: for the overwhelming majority of charities, no one has done the work to determine whether they are actually effective. Donors are flying blind, guided by financial ratios that correlate weakly with outcomes.
The behavioural damage
The most insidious effect is on donor behaviour. A 2024 study published in the Journal of Economic Behavior & Organization found that Charity Navigator ratings significantly influence giving patterns. Donors increase giving to higher-rated organisations and decrease giving to lower-rated ones. This sounds like ratings working as intended, except that the ratings measure the wrong things.
The result: capital flows toward well-administered organisations that file good paperwork, and away from messier organisations that might produce better outcomes. A health charity in sub-Saharan Africa running community health worker programmes at $2 per DALY averted but with slightly high transport costs loses donors to a four-star US environmental education nonprofit producing unmeasurable outcomes at vastly higher cost per beneficiary.
What to actually do
If you want the best available evidence on where your money does the most good: Use GiveWell's top charity list. It is narrow by design. That narrowness is the feature, not the bug. The charities on that list have survived thousands of hours of scrutiny.
If you want to evaluate a charity not on GiveWell's list: Ask three questions. First, what is the evidence that this type of intervention works? (Look for systematic reviews or RCTs, not testimonials.) Second, what does the charity report as its cost per outcome? (If they cannot answer this, that is informative.) Third, has anyone independent evaluated this charity's effectiveness? (Founders Pledge, The Life You Can Save, Animal Charity Evaluators each cover different domains.)
If you see four stars on Charity Navigator: Know what you are looking at. The charity files good accounts. It has an independent board. Its fundraising costs are moderate. These are fine things. They are not evidence of impact.
One sentence
Charity Navigator tells you whether a nonprofit is well-run on paper. It does not tell you whether it works. For that, you need evaluators willing to measure outcomes, and there are very few of those.
Sources used: Charity Navigator methodology documentation (charitynavigator.org/about-us/our-methodology, 2025), GiveWell cost-effectiveness analyses (givewell.org, 2024), "Navigating Charity Navigators in the Modern Era" (EA Forum, 2024), Bridgespan Group research on nonprofit overhead and performance (2019), Stanford Social Innovation Review "The Overhead Myth" joint letter (GuideStar, BBB Wise Giving Alliance, Charity Navigator, 2013), Dan Pallotta "The way we think about charity is dead wrong" (TED, 2013), Journal of Economic Behavior & Organization study on charity ratings and donor behaviour (2024, DOI: 10.1016/j.jebo.2024.106918).