When you last went through a self-checkout at Woolies, were you asked to round up your bill for charity? Perhaps your $49.31 shop became $50.00, with the remaining 69 cents going to a good cause. Individually, these decisions seem almost trivial. Collectively, they can add up quickly. One Woolworths round-up campaign raised $770,000 for Foodbank in a single month, despite every individual donation being less than a dollar (Foodbank, 2026).
Branesh Prakash
ESSA Monash Clayton
[Bran is a Master of Applied Econometrics student at Monash University and an economist with the Reserve Bank of Fiji, currently on study leave as an Australia Awards scholar. Beyond his work and interests in macroeconomics, policy analysis and applied econometrics, he is a supporter of the Effective Altruism movement. He is particularly interested in the economic ideas that underpin its work: scarcity, opportunity cost, marginal impact and the allocation of resources towards where they can have the greatest effect. This article reflects that broader interest in how economic thinking can help make charitable giving more deliberate and effective. He encourages fellow economics enthusiasts to explore Effective Altruism and contribute to discussions around the economics of doing good.]
Disclaimer: The views expressed in this article are those of the author and do not necessarily reflect the views of affiliated organisations.
There is a lot to like about this model. Supermarkets already have the payment infrastructure and millions of customer transactions, allowing them to aggregate tiny donations that few of us would ever transfer independently. In effect, they dramatically reduce the transaction costs of fundraising. Hundreds of thousands of amounts that are almost immaterial to individual shoppers can become a sizeable pool of charitable funding.
But there are really two economic problems here: mobilising charitable dollars and allocating them. A checkout round-up can be remarkably successful at the first. Its success tells us much less about the second. Therefore, an important question is not whether Foodbank—or whichever charity happens to appear on the screen at your local grocery store—does good work. It is whether our next dollar would do more good there than somewhere else for what we are trying to achieve through our giving.
Who gets the chance to ask? How solicitation shapes where we give
We probably do not make most charitable decisions by sitting down with a list of every worthy cause, comparing them and constructing an optimal portfolio. Instead, charities enter our consideration because we encounter them: a friend runs a fundraiser, an appeal arrives after a natural disaster, somebody knocks on the door, or a supermarket puts a question on a screen.
Bekkers and Wiepking’s (2011) review of charitable giving reports that around 85 percent of donation acts in one US survey and 86 percent in a Dutch survey occurred following a solicitation. Before effectiveness is compared, then, there is already a selection mechanism at work: some causes obtain our attention while thousands of others never enter our choice set. Even the form of the request influences whether we choose to donate. Sudbury and Vossler (2022) study the kind of checkout decision Woolworths uses. These requests are typically unexpected, small and made within seconds. They find that round-up requests produce higher donation rates than comparable fixed requests, with a “loose change” effect helping explain the difference. Their framework also highlights decision costs. “Would you like to donate 69 cents?” is a much simpler problem than “Would you like to donate—and, if so, how much?” (Sudbury & Vossler, 2022).
All this highlights that the architecture surrounding a decision matters. Attention, convenience and the way we are asked help determine where charitable dollars flow before the relative impact of different charities has properly entered the calculation. This whole process is part of an allocation mechanism—though not one necessarily designed to direct money toward where it best advances some underlying objective.
It is the next dollar that matters—and it depends what our objective is
Suppose instead that you have already decided to donate $100. Economics gives us a familiar problem: scarce resources have alternative uses. Once the $100 has been committed to charity, giving it to one cause means foregoing what the same $100 could have achieved through another. The relevant question moves from “Is Charity A effective?” to “What additional good will the next dollar given to Charity A achieve compared with the next dollar given elsewhere?”
That word—marginal—matters. A charity can have an excellent track record without necessarily being the best destination for another $100 today. Perhaps its highest-value programs already have sufficient funding, while additional donations would be directed towards lower-priority opportunities. Another organisation may have an unfunded opportunity where an additional dollar produces a much larger benefit. As with other forms of resource allocation, what matters for the next decision is not simply the average return earned in the past, but the return available at the margin. But unlike financial returns, the “return” from charitable giving is not straightforward to define. Before asking which option is more effective, we first need to ask: effective at achieving what objective?
A straightforward answer might be improvements in overall human or societal welfare. But donors may also care about helping locally, supporting a cause connected to their family, assisting a particular community, fairness, environmental protection, among other things. Andreoni’s (1990) theory of “warm-glow” giving makes an important economic point: people can derive utility not only from the total amount of good produced from it, but from their own act of giving.
Berman et al. (2018) show how powerful these preferences can be. Even when effectiveness information was made easy to compare, participants frequently preferred a less effective charity when its cause was personally more appealing. In an experiment where participants were randomly allocated to donate to a charity or invest in a stock, around a third selected the highest-rated charity option, in contrast to half of the respondents who chose the best investment stock. Therefore, people appear to regard charitable choices as legitimate expressions of personal values rather than purely optimisation problems (Berman et al., 2018).
That does not make such choices economically irrational. There need not be one universally correct charitable objective. The point is therefore not to remove personal values from charitable giving, but to make those values explicit rather than allowing whichever solicitation reaches us to determine the allocation. If helping your local community matters to you, that can receive weight in your objective. If reducing extreme deprivation matters more, weight that instead. But there is still a useful question once those preferences are acknowledged: given what I care about, where will my marginal dollar advance it most?
Why we rarely know where the next dollar will do most good
Answering that question is difficult partly because charitable giving has an unusual information problem. When we buy a product, we generally experience its quality ourselves. With charitable giving, somebody else receives most of the benefit. The organisation we donate to knows much more than we do about how our money is spent, and measuring its true impact may require asking a difficult counterfactual: “what happened because this program existed that would not otherwise have happened?” Finding credible answers takes time, and time has an opportunity cost.
Krasteva and Yildirim (2013) build this insight into an economic model of charitable giving. When acquiring information requires time and effort, someone contemplating a smaller contribution has less incentive to investigate before giving. This makes intuitive sense—for example, spending half an hour researching the globally optimal destination for a 69-cent Woolworths round-up would itself be a remarkably inefficient use of resources. But here, the individual and aggregate perspectives diverge. No individual shopper has much reason to conduct due diligence over 69 cents. Yet once hundreds of thousands of these individually sensible decisions are pooled into $770,000, information about where charitable resources have their greatest marginal effect becomes much more valuable. This may embed a two-part allocation. In a checkout campaign, the supermarket has already selected the charity that will receive the round-ups, and the shopper only decides whether their marginal 69 cents should join that pool. The shopper’s choice is whether to allocate that marginal amount to a pre-selected cause rather than retain it—or potentially direct it elsewhere. This also demonstrates our information problem as it is sensible for the individual not to investigate the selected charity for a tiny checkout donation, but when deciding where our broader charitable giving should go, we may still want information about which options best advance our objectives. Search costs may explain only part of the story here as impact information may not enter into people’s donation decisions at all. In Metzger and Günther’s (2019) laboratory experiment, around 57 percent of participants donated, yet only 29 percent chose to purchase additional information beforehand, and demand was lowest for information about the donation’s actual impact: only 22 percent sought it, compared with 38 percent who sought information about the type of recipient. Strikingly, participants who had chosen not to acquire information barely changed their giving even when information was subsequently provided for free. The authors suggest they may not have understood, trusted or placed much value on the information provided. Yet, those who actively sought impact information responded strongly to it, giving more to higher-impact projects and less to lower-impact ones.
We therefore face an aggregation problem where individually small donations rationally attract little research, yet large numbers of low-information participation decisions can collectively channel substantial charitable resources towards causes selected by intermediaries. At that scale, the case for undertaking due diligence is much stronger than it is for any single shopper. For individuals making larger or planned donations, specialised evaluators can perform a similar function—incurring the research cost once and making the resulting information available to many donors. In other words, the supermarket solves the transaction-cost problem by pooling tiny donations; a good evaluator can partly solve the information-cost problem by pooling research.
Lesson: separate how much to give from where to give
Perhaps, then, the lesson from checkout charity is not that we should reject it. In one respect, supermarkets have solved an economic problem exceptionally well: they have made giving cheap and easy. The question is whether we can reduce the information and decision costs of allocating our donations just as effectively.
One simple approach is to separate two decisions that solicitations usually bundle together: How much do I want to give? And where do I want it to go? Instead of reconsidering both questions every time an appeal appears, we might first choose a charitable budget that suits our circumstances. Then, once or twice a year, incur the larger fixed cost of thinking seriously about its allocation: decide what outcomes matter to us, examine evidence, and compare where an additional dollar could have the greatest effect.
GiveWell offers one illustration—not a definitive answer—of what this kind of deliberate allocation can look like (see GiveWell, 2025). Additionally, it does research and provides information that we cannot gather ourselves. GiveWell uses cost-effectiveness estimates as an important input when comparing funding opportunities, while explicitly cautioning that its numbers are “extremely rough” since calculations require uncertain empirical assumptions and philosophical judgments, such as how to compare increased income with preventing a death. It therefore does not recommend treating its estimates as literal rankings carved in stone. That uncertainty, however, is not a reason to abandon comparison. It is a reason to be modest about what comparison can achieve. Nor must deliberate giving eliminate spontaneous generosity. Someone might reserve part of their charitable budget for the causes that unexpectedly move them and allocate another part more deliberately.
There may thus be no single correct destination for everyone’s next $100. Economics cannot tell us how much weight we should place on poverty alleviation, medical research, our local community or a cause close to home. But it can remind us that every $100 has an opportunity cost—and that, given our objective, the right comparison concerns what the next dollar achieves.
Perhaps the best place to start is to decide what “doing good” means to us before somebody else places the choice on our screen.
References
Andreoni, J. (1990). Impure Altruism and Donations to Public Goods: A Theory of Warm-Glow Giving. The Economic Journal, 100(401), 464-477, https://doi.org/10.2307/2234133
Bekkers, R., & Wiepking, P. (2011). A Literature Review of Empirical Studies of Philanthropy: Eight Mechanisms That Drive Charitable Giving. Nonprofit and Voluntary Sector Quarterly, 40(5), 924-973, https://doi.org/10.1177/0899764010380927
Berman, J., Barasch, A., Levine, E., & Small, D. (2018). Impediments to Effective Altruism: The Role of Subjective Preferences in Charitable Giving. Psychological Science, 29(5), 834-844, https://doi.org/10.1177/0956797617747648
Foodbank (2026). Thank you Woolworths and your amazing customers. https://www.foodbank.org.au/woolworths-round-up-success/
GiveWell (2025, May). GiveWell’s Cost-Effectiveness Analyses. https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/cost-effectiveness-models
Krasteva, S., & Yildirim, H. (2013). (Un)Informed charitable giving. Journal of Public Economics, 106, 14-26, https://doi.org/10.1016/j.jpubeco.2013.06.011
Metzger, L., & Günther, I. (2019). Making an impact? The relevance of information on aid effectiveness for charitable giving. A laboratory experiment. Journal of Development Economics, 136, 18-33, https://doi.org/10.1016/j.jdeveco.2018.08.015
Sudbury, A., & Vossler, C. (2022). Checking out checkout charity: A study of point-of-sale donation campaigns. Journal of Economic Behavior and Organization, 204, 252-270, https://doi.org/10.1016/j.jebo.2022.10.026
