In Doing Two Things at Once, I wrote about the propensity towards hackneyed charity schemes. People just turn two things they like into a “charitable cause.” There’s really an underlying issue that’s much more systemic.
Jason Crawford, in a discussion of funding models, compares non-profits and for-profits in terms of “metabolisms.” In a for-profit entity, there is a clear feedback loop: Profits (from sales) allow a business to generate returns for investors, who, in turn, loan capital to the company. The for-profit uses that capital, at least in part, to create products for customers. Customers buy those products to create sales for the company.
In a non-profit organization, there is a missing part of the product loop: There is usually no way that the beneficiaries of charity can return feedback to the non-profit and its donors. Obviously there’s been a big push for charities to collect data, but the data is only directionally connected to impact (it isn’t perfectly accurate) and that’s only somewhat connected to contributions (some donors will donate or not donate regardless of what the data shows). It’s nowhere near the tight signal that sales are.
In an ideal world, the amount of contributions from donors to an organization would reflect the value that beneficiaries received from an organization. But we know that isn’t the case, because satisfying beneficiaries is rarely the sole goal of donors. Where investors in for-profits relentlessly pursue monetary returns, non-profit beneficiaries pursue other things like self-worth and status. That’s why people are allowed to pair two random things and call it a charity. Charity need not fit the needs of the beneficiaries, just the interests of donors.
I don’t consider this an Effective Altruist critique either. EA focuses on how we can do the most good. Part of that is by serving the needs of beneficiaries, but EA also raises the question of who the beneficiaries of charitable actions should be. Since EA looks to have the most impact, it would typically prefer to do health and poverty work in developing countries over developed ones, as the marginal dollar is able to do more good in the former. That sort of reasoning isn’t required for my argument here: Even if a charity is trying to help people who aren’t destitute (say, poets in Brooklyn), we should still be better attuned to what those beneficiaries actually want if we are to call our actions charity at all.
There’s a classic essay on this subject, Jeffrey Friedman’s There is No Substitute for Profit and Loss. He focuses on the specific role of profits to push the point I’ve been making even further. It’s not simply that profit signals are a little bit better than an alternative feedback method like doing good data collections. Profit is what tells us whether any proxy—data, metrics, mission statements, management styles—are achieving results. We can’t use secondary metrics to deduce a magic formula which implies profitability, because it is profitability itself that guides the entrepreneur to serve their customers.
There are some new structures we could try to make charity better: We could create an organization that pays non-profit organizations on behalf of them making their beneficiaries happy. Such an organization could try to make donations function more like profits by donating it proportionally to the amount of value a non-profit creates for its beneficiaries. Perhaps such an organization would allow you to choose a group of beneficiaries one wants to help then donate accordingly. But the “accordingly” is the hard part. Do you survey people and ask what they valued? The information is inevitably of a lower quality than their revealed preferences through spending. Indeed, in Friedman’s terms, the profits would be signaling to us the quality of our survey data: Do answers match revealed preferences?
Why not instead just give money directly? Or perhaps start a grant program? In those cases you could still take advantage of the corrective power of profits.
There can be an advantage for charity that isn’t tied to the desires of beneficiaries: Some things do more good than people value them. Insecticide-treated malaria nets are the perfect example. Someone may place little value in a net, perhaps due to their time preference, but using a net would do a lot of good for recipients and nets have positive externalities. So ignoring profits might be good for things with positive externalities, as they would be naturally undersupplied.
At the same time, the median charity scheme is probably creating much less value than that, so direct cash transfers likely do good on most margins.
But this is all cheating. The reason bad nonprofits exist is because people have already shown they don’t care about the wellbeing of beneficiaries. I’m optimistic enough to believe that if my theoretical meta-charity made it clear that this was the place to give to if you really cared about a certain group, some people would donate to it instead. But it still wouldn’t fix the underlying problem. After all, an organization with similar methods already exists: It’s called Givewell + OpenPhil.
In the end, it’s Hansonian. Charity, by and large, isn’t about charity.