How NOT to pick a charity

by Adriano Mannino and Tobias Pulver

Imagine two charities that both fight a particular disease in a poor country. Donating to charity_1 means that 95% of your money goes directly to the cause (buying and distributing medicine), while only 5% is spent on salaries and other expenses. Donating to charity_2 means that 70% of your money goes straight to the cause, while 30% is spent on salaries and other expenses.

Which charity do you pick? Easy decision?

A closer look

Now imagine that before committing yourself to donating, you decide to look more into how each charity is allocating its funds:

Charity_1 relies on many volunteers to keep expenses low. The volunteers are very motivated to work for the good cause, but they lack experience in management, health economics and other relevant domains. The charity chose to not invest an resources into tracking and evaluating their work. In case their interventions have any long-term negative side-effects, they likely wouldn’t find out.

Charity_2 has hired a team of experts, people who could for instance be earning high salaries in business management. The salaries the employees receive are lower than what they could earn elsewhere, but they are still competitive enough to attract talented people with the required expertise. Charity_2 also has people working on impact evaluation, tracking the efficiency and even long-term cost-effectiveness of their interventions.

You also find out that an independent scientific charity evaluator has looked into both organizations and concluded that the interventions (which here includes the entire set of activities the charities engage in) carried out by charity_2 are likely about 80 times more effective at curing the disease than the interventions of charity_1 – i.e. for the same input money, you get an 80-fold greater output. This difference is due to significantly better logistics and efficiency benefits from large-scale distributions, comparatively higher general reliability of processes due to more professional staff, as well as from charity_2 having researched and implemented alternative and better methods to fight the disease in a more cost-effective way (e.g., prevention instead of cure). Charity_2 is reaping the benefits of having improved its operations several times based on the findings of its own impact evaluations.

Now, which charity do you donate to?

Busting the overhead myth

This example was hypothetical, but the general scenario is very plausible. Charity evaluator GiveWell has concluded that charities differ in effectiveness by orders of magnitude. And yet, instead of consulting the results of such impact-per-dollar evaluations, the most pressing question for most donors remains how much “overhead” a charity has¹,²: The less money spent on salaries and other administrative expenses, the better the charity.

This is worrying because this measure is completely irrelevant for the effectiveness of a charity; focusing on it is nonsensical and sometimes even counterproductive. As we have seen in the above example, what constitutes “overhead” of a charity is an essential ingredient to the effective functioning of any charity, and it often is what makes good charities particularly effective. There is no clear-cut difference between “money spent directly on the cause” and “overhead”. Good charities are no different from good businesses: Everything a good charities spends money on is there to help it achieve its goal: carrying out its charitable intervention as well and as often as possible. Imagine a hospital congratulating itself on the fact that 90% of its budget go “directly to the patients” – or indeed a standard for-profit boasting that “90% of our budget go directly into building cars, only 10% of our expenses go into planning, designing and advertizing them”. It’s an obvious way to fail – and it’s still accepted as a hallmark of excellence when people think about charity.

Once people pause to think about the actual (ir)relevance of overhead instead of blindly following the rule that it ought to be as low as possible, it quickly becomes obvious that this approach is misguided. In the end, what matters is how much good the charity accomplishes per dollar, not what percentage of the dollar goes “directly” into an intervention that may or may not be effective. If the case against overhead being relevant is so clear, why is it that it is still the most popular criterion for most donors?

In the excellent TED talk titled “The way we think about charity is dead wrong”, Dan Pallotta argues that the obsession with low overhead is a manifestation of a more general problem: People have a mindset that treats the non-profit sector fundamentally different from the for-profit sector. There is an expectation that charity work ought to be direct with no detours and that people need to work for it “out of a genuine motivation to help”, not because of good salaries. This is all fine, of course, but is it reason enough to demonize charities that want to offer competitive salaries in order to attract more skilled people?

Pallotta has the following to say on this: “We have a visceral reaction to the idea that anyone would make very much money helping other people. Interesting that we don’t have a visceral reaction to the notion that people would make a lot of money not helping other people!”

The incentives our attitudes create are very irrational – shouldn’t people get bigger rewards the more good they do? Treating charities completely differently from for-profits is bad, because if there’s one thing for-profit business is good at, it’s optimizing for a goal – profit. Charities could benefit a lot from a business-like mindset by applying it to their goal: helping others, reducing unnecessary suffering.

One contributor to overhead aversion is the fear of being exploited as a donor. This fear is strong: When we imagine that our donations don’t do any help and instead go to the pockets of mean-spirited scammers, we rightly feel very bad and upset. And in this regard, looking at a charity’s overhead ratio is even a decent heuristic: If the charity spends 99% of its funds directly on some intervention in Africa, then, even if the intervention is not helping at all, at least the people behind the charity aren’t trying to steal my money and have a laugh about it. If they were, they would budget higher salaries and fake expenses.

If the reason people focus solely on the overhead ratio when picking charities is fear of scammers, then it appears as though they care infinitely more about charities not being totally fake than they care about charities being positively competent and effective at helping people. Charities’ cost-effectiveness differs by orders of magnitude. If the non-fake charity you pick only reduces 5% as much suffering as the top-rated charities, this disparity is very significant, and the sub-optimal outcome is something that’s at least as important to avoid as falling victim to a total scam: The difference between a scam and the sub-optimal charity you pick is the magnitude of the 5% – while the difference between your sub-optimal charity and the best charities is the magnitude of the 95% disparity.

If you care about helping people effectively, avoiding scams is included in the considerations, but so is much more. Therefore, you should ignore a charity’s overhead ratio and instead consult information on charities’ output : input ratio (i.e., cost-effectiveness) as calculated by data-driven evaluations.


[1] Gneezy, Uri, Elizabeth A. Keenan, and Ayelet Gneezy. “Avoiding overhead aversion in charity.” Science 346.6209 (2014): 632-635.
[2] Caviola, Lucius, et al. “The evaluability bias in charitable giving: Saving administration costs or saving lives?” Judgment and decision making 9.4 (2014): 303.