The (behavioural) economics of ineffective giving
A look into some of the psychological quirks that dampen our ability to give effectively.
Summary
Charitable giving should be way better than it is. Hundreds of billions of dollars are donated to charity every year, yet most giving is ineffective. This is a problem because it implies that a lot of unnecessary suffering is occurring. For example, effective charities working on global health issues affecting people in low-income countries can be 1,000 times more cost-effective than giving the equivalent amount of cash to the average American. However, some charities with multi-million dollar budgets choose to focus on the latter.
Understanding cost-effectiveness is important because lives are on the line. Donors appear to care about cost-effectiveness when asked, but charitable giving is often directed towards ineffective charities nonetheless. This blog post explores the (behavioural) economics of charitable giving to understand why donors' stated preferences and actions sometimes diverge from one another.
Key topics discussed in this blog include:
Cost-effectiveness: The importance of cost-effectiveness — getting the most bang for your charitable buck — is highlighted using a real-life example.
Bounded rationality: Bounded rationality can explain why donors make suboptimal decisions when it comes to selecting charities.
Evaluability bias: Evaluability bias can lead donors to focus on charities that are easier to evaluate, rather than those that are the most effective.
Bounded compassion: Bounded compassion can sometimes lead donors to feel indifferent between charities that differ greatly in terms of their impact. This is because the emotional limits of donors may not be fully attuned to the relatively large differences in cost-effectiveness between different charities.
Search costs: Search costs can make it difficult for donors to find the most effective charities to support.
Relational switching costs: Relational switching costs can make it hard for donors to switch to more effective charities, even if doing so will do more good in the world.
Preferences for warm-glow giving: Preferences for warm-glow giving can lead donors to focus on charities that make them feel good, rather than those that are the most effective.
Lack of creative destruction: A lack of creative destruction can make it difficult for new, effective charities to gain traction and compete with established charities.
Lack of incentives for charities to do cost-effectiveness analyses: Charities lack the incentives to do cost-effectiveness analyses, thereby making it difficult for donors to find the most effective charities to support.
Introduction
Hot take: charitable giving could be so much better.
Hundreds of billions of dollars are donated to charity every year, yet most giving is ineffective. This is a huge problem: any gap between our current world and an ideal world where charitable resources are allocated based on how much good they can do implies that a lot of needless suffering is happening.
This gap can be enormous in practice. For example, highly effective charities working to improve global health in low-income countries are around 1,000 times more cost-effective than giving the equivalent amount of cash to the average American, yet some charities with multi-million dollar budgets decide to focus precisely on the latter.
With billions being thrown around, these kinds of gaps really matter: donating to ineffective charities can result in needless deaths that stem from a failure to prioritise scarce resources. So long as we live in a world where charities working to prevent kids from dying have lower budgets than the San Francisco Opera House, we ought to strive to do better.
Understanding cost-effectiveness, and why it is so important
First, let me build intuition on the importance of charitable cost-effectiveness with a real life example.
Suppose a donor is deciding between two charities: Sightsavers, a charity that offers corrective surgery for cataracts, and Guide Dogs for the Blind (GDB), a charity that offers sightseeing dogs that can improve the quality of life of individuals with limited vision. GiveWell, a charity evaluator, estimates that between $110 and $1,000 donated to a charity like Sightsavers can counterfactually reverse a severe case of visual impairment. Conversely, financial statements from 2021 seem to indicate that GDB provided a total of 173 clients with guide dog training that year at a per-individual cost of just over $200,000.
(I know, I can’t believe it either.)
Taking the conservative end of the estimates for cataract surgery — and making the rather silly assumption that the benefit gained from not going blind is equal to the benefit of receiving a sightseeing dog — shows that Sightsavers is, at a minimum, 200 times more cost-effective than GDB. If GDB’s budget was instead donated to Sightsavers, we could reasonably expect at least 30,000 people to not lose their vision who otherwise would have.
You don’t need to be a consequentialist, nor a utilitarian, to think that it’s better to have 30,000 people keep their eyesight rather than provide 173 people with guide dogs. Beyond basic intuition that many different moral frameworks can subscribe to, survey data indicates that donors at least claim to care about cost-effectiveness. Participants overwhelmingly stated that, when donating to charity, they want to do the most good possible for the greatest number of people.
To recap: cost-effectiveness is important, as lives are on the line. Donors seem to care about cost-effectiveness, at least to some degree. Despite this, charitable giving is often directed towards ineffective charities. What follows is the question of why.
The rest of this article will attempt to understand at least part of this thorny problem by exploring charitable giving from a behavioural perspective. My hope is that doing so will help us understand why donors’ stated preferences and actions sometimes radically diverge from one another.
Give more, or give more effectively?
Existing research on the social science of charitable giving has mostly focused on exploring how to make donors give more. But there is good reason to think that encouraging donors to give more effectively can increase the amount of good done by far more than simply increasing the amount donated.
Let’s use a hypothetical example to illustrate this point.
Suppose a donor named John is deciding between NetsForGood, which provides anti-malarial bed nets to people living in extreme poverty, and RetroHealth, which provides antiretroviral therapy to people infected with HIV in sub-Saharan Africa.
NetsForGood has a cost-effectiveness of £100 per quality-adjusted life year (QALY) whereas RetroHealth has a cost-effectiveness of £1,000 per QALY. Suppose neither of these QALY estimates are immediately clear to John, and that he views both charities as potentially good choices.
John gives £50 a month to charity. If John allocates his entire charitable giving budget to RetroHealth, it would take him twenty months to create a single QALY. However, if he instead gives it all to NetsForGood, it would take him just two months.
Because of the stark differences in cost-effectiveness between the two charities John is considering, it follows that his choice of where to give could be far more important than his choice of how much to give.
Numbers from the real world: why giving more effectively > giving more
Let’s further ground this with some empirical evidence. One field experiment aiming to understand how to encourage donors to give more showed that donors who received a postcard emphasising the benefits of donating gave 20 percent more than those in a control condition. A 20 percent lift in the amount John gives to charity would result in an additional 0.12 QALYs per year if he donates everything to RetroHealth. Yet if a similar experimental intervention convinced John to allocate even just a quarter of his original charitable giving budget to NetsForGood, it would result in an additional 1.35 QALYs per year.
What gives? Bounded rationality and search costs, actually
One possible explanation for why people give ineffectively is that donors are boundedly rational, meaning they have limited knowledge and face cognitive limitations that often result in decision-making that satisfices rather than maximises. Think of decision-making that is “good enough” versus “the absolute best”. A study by Caviola et al. explores this by describing a broad class of “epistemic obstacles to [effective giving]”, implying that donors give ineffectively partially due to a lack of knowledge about cost-effectiveness.
Moreover, with millions of charities to choose from — and no comparable unit of impact through which donors can easily compare charities — donors often face high search costs (or as Castillo et al. describe it, “nuisance costs”) when deciding where to give. The combination of these search costs and donors’ bounded rationality can help explain why ineffective charities remain funded.
Charity evaluators for reducing search costs
High search costs have led to the emergence of online charity evaluators, such Charity Navigator, which allow donors to quickly compare charities based on key metrics such as financial performance and transparency. However, Charity Navigator’s methodology has led to perverse incentives for charities while simultaneously failing to educate donors about where their money can actually do the most good.
One study found some preliminary evidence that some charities understate administrative expenses (or misreport them as program expenses) and overstate revenues to game Charity Navigator’s rating system. This system, which emphasises easily quantifiable metrics (such as the overhead ratio, which we will explore later in greater detail), can mislead donors into a false sense of effectiveness.
Scope insensitivity: why we’re bad at making sense of big numbers
Bounded rationality does not just cover a lack of knowledge. The other part includes cognitive limitations that result in suboptimal decision-making. Here, there are two I’ll highlight in particular: scope insensitivity and evaluability bias.
A famous study by Desvousges et al. demonstrated scope insensitivity by asking three different groups of survey respondents how much they would spend in order to save a group of 2,000, 20,000, or 200,000 birds from drowning in an oil pool. Respondents reported an average willingness to pay of $80, $78, and $88 respectively for each group of birds.
Counterintuitively, this implies that respondents do not significantly increase the amount they’re willing to spend as the number of drowning birds increases. Scope insensitive donors might therefore be inclined to give irrespective of the size of the problem at hand, leading to relatively unimportant charitable causes receiving inappropriately high levels of funding.
Our compassion has limits
Closely related to scope sensitivity is the idea of bounded compassion, which describes how our empathy for others can be finite. The mismatch between our emotional limits and the relatively unbounded differences in cost-effectiveness differences between charities can mean donors feel indifferent between charities that dramatically differ in impact.
As Eliezer Yudkowsky argues, the human brain is incapable of properly accounting for differences in negative events that differ by orders of magnitude. Therefore, a prospective risk increasing by a factor of ten does not correspondingly multiply our motivation to prevent it from happening. Put plainly, the amount of emotional pull we feel does not always appropriately intensify as the number of people we can help with our giving increases.
While valuing market goods as having diminishing marginal utility often makes sense — one tends to get a lot more value from the first mattress they have compared to the seventh — valuing lives the same way can be disastrous.
Suppose donors’ decision making is influenced by experience in other markets where the best options are not significantly more cost-effective than average options. If this is the case, donors might mistakenly assume that the market for charitable giving is similarly efficient, and therefore avoid the effort of searching for charities based on effectiveness.
There’s some evidence that donors vastly underestimate differences in cost-effectiveness between charities. A survey of the general population demonstrated that donors think the most effective charities working to help the global poor are 1.5 to 2 times more effective than average ones, whereas experts estimated the best charities are actually closer to 100 times more effective.
Evaluability bias: or, why we love easily digestible numbers
Another cognitive bias contributing to ineffective giving is evaluability bias, which describes when we judge something’s importance based on how easy it is to evaluate. Evaluability bias can partially explain why donors’ behaviour might differ from their stated preferences. There is evidence that donors use a charity’s overhead ratio (its administrative expenses relative to its expenditure on recipients) as a proxy for effectiveness, according to a study by Caviola et al. (2014). This is clearly a problem: if I operated a charity that gave direct cash transfers to billionaires, and ran it on my own dime, it would have very little overhead costs while being wholly ineffective at actually doing any good in the world.
Data from DonorsChoose.org, an online platform that connects donors with teachers who put forth requests for funding for classroom projects, demonstrates that donors are price sensitive to overhead costs. A 10 percent increase in the price of giving — the amount a donor must donate to increase a charity’s output by one dollar — was found to reduce the probability that a project gets funded by 3.5 percentage points. This indicates that the “price” of giving (think of this as somewhat analogous to overhead costs) can substantially influence donors’ willingness to give. The tendency to evaluate charities based on more easily measurable (rather than intrinsically important) factors means that donors might mistakenly assume that they are giving effectively when giving to charities with low overhead costs.
Donors face other psychological hurdles
The market for charitable giving might further be distorted on the demand side due to relational switching costs, which describe the psychological costs felt when switching from a preferred to a less-preferred choice. If donors feel a particularly strong emotional bond to their favourite charity, they might show greater hesitation with switching to a more effective charity, as doing so implies that previous giving has been ineffective.
Donors do not just care about effectiveness: a model of warm-glow giving
Beyond donors’ bounded rationality, and the costly search they must undergo when evaluating charities, it seems crucial to understand if donors care about factors other than cost-effectiveness.
Intuitively, it seems they do. For example, there is evidence from Andreoni’s (1990) model of warm glow giving that donors gain utility from the very act of giving, independent of how effective the charity is. Many donors view giving as a personal choice, meaning donations are not morally required to be guided by measures such as cost-effectiveness. Incorporating these preferences can further explain why donors do not give solely to the most cost-effective charities available. Donors often focus on charities that provide them with emotional satisfaction rather than those that are the most cost-effective.
These preferences can sometimes manifest in problematic ways. For donors in high-income countries, there could also be a relative trade-off between effectiveness and emotional pull. And since donors tend to favour charities they’re personally connected to, charities that focus on issues that do not affect people in rich countries could therefore be ignored by those with the greatest access to resources.
Why ineffective charities manage to stay in the market
So far, I’ve mostly explored donor behaviour. But donors are just one part of the charitable giving ecosystem. As we’ll see, charitable giving is inefficient because there is not a proper mechanism for pushing out ineffective charities. In for-profit markets, poor-performing firms are pushed out by competition. However, in the charitable sector, there is no equivalently powerful mechanism for ensuring that resources are allocated to the most effective charities.
Creative destruction
Creative destruction describes how more efficient and innovative firms push incumbent firms out of for-profit markets. However, this phenomenon does not seem to apply to the charitable giving market in the same way it applies to for-profit markets.
There are few (if any) for-profit markets where the best firms can provide goods or services that are orders of magnitude better than average ones for the same cost to consumers. If there were, market pressures would force out the worst-performing firms — nobody would buy a Nokia if an iPhone is 1,000 times better and costs the same. Yet while poor-performing for-profit firms face pressure to leave markets, ineffective charities can manage to stick around by soliciting donations based on things unrelated to how effective they are at doing good in the world.
Take for example Scotscare, a non-profit that provides support for Scottish people living in London (hardly a marginalised group in 2022) that has managed to stay in operation for hundreds of years. Charities like Scotscare compete for donors’ finite attention, at the expense of other charities that are tackling much larger societal issues.
There is some competition between charities, but not on what matters
Despite the prevalence of ineffective charities, it would be inaccurate to claim that charities do not compete at all. Like for-profit firms, charities compete on characteristics that enable them to acquire donations because donors must make trade-offs about which charities to support.
Crucially, these competitive forces do not necessarily mean that charities compete against each other to more effectively deliver goods and services to those in need. And since formal cost-effective analyses require expertise to conduct are expensive and often time-consuming, charities might be more inclined to spend resources on marketing campaigns or fundraisers rather than rigorously evaluating the cost-effectiveness of the programmes they offer.
Conclusion
The market for charitable giving is highly inefficient. Since the welfare of millions of people are on the line, this matters tremendously.
Part of the solution will require an understanding of the behavioural forces contributing to this unfortunate state of affairs. My aim here was not to be comprehensive, but to highlight a few factors that might play outsized roles in ineffective giving.
Such factors include, but are not limited to:
Difficulty understanding cost-effectiveness — Donors may not fully understand the cost-effectiveness of different charities, and may be influenced by biases such as scope insensitivity and evaluability bias.
The costs of giving — Search costs and relational switching costs can make it difficult for donors to find and switch to more effective charities.
Bad incentives for charities — A lack of incentives for charities to do cost-effectiveness analyses and a lack of creative destruction can also contribute to the problem.
Preferences for warm-glow giving — The tendency of donors to focus on charities that make them feel good, rather than those that are the most effective can be a problem because it can lead to charitable resources being allocated locally, rather than globally. Since charitable giving is mostly done by people in high-income countries, this is a huge concern.
I hope that one day the charitable giving ecosystem won’t be so dysfunctional. But changing these pesky artefacts of human psychology might make that a lofty dream. Nonetheless, the stakes are too high to give up without trying.