Slightly Against The "Other People's Money" Argument Against Aid
...
In the comments to last year’s USAID post, Fabian said:
While i am happy for the existence of charity organisations, i don't get why people instead of giving to charity are so eager to force their co-citizens to give. If one charity org is not worth getting your personal money, find another one which is. But don't use the tax machine to forcefully extract money for charity. There are purposes where you need the tax machine, preventing freerider induced tragedy of the commons.
But for charity? There are no freeriders. If you neither give nor receive, you are just neutral. The receivers are not meant to give anyways.
This is a good question. I’m more sympathetic to this argument than I am to the usual strategy of blatantly lying about the efficacy of USAID; I’m a sucker for virtuous libertarianism when applied consistently.
But I also want to gently push back against this exact explanation as a causal story for what’s happening when people support foreign aid.
The “Other People’s Money” Argument
IIUC, the argument is that people who would not donate to charity themselves find it more congenial to vote to tax other people and give their money to charity.
A simple problem with this argument is that actually, each voter’s money will also be taxed. So for example, if there’s a vote on whether to tax everyone an extra $100 and spend the money on foreign aid, then voting in favor of the law costs you $100, the same as if you donated the money yourself voluntarily.
There are two simple ways to rescue the argument (we’ll discuss complicated ways later):
First, you could argue that supporters are using the government as a force multiplier. That is, suppose that 51% of people support spending $100 of their own money on foreign aid. If, instead of donating personally, they vote for a law that taxes everyone $100, they can make their “donation” go twice as far by “matching” it with $100 checks from the 49% of unwilling voters. This doesn’t have quite the same oomph as the accusation of “spending other people’s money because you don’t want to sacrifice your own”, but at least it sort of makes rational sense from a public choice theory perspective.
Second, you could argue that supporters are disproportionately poor people who pay low taxes, and who suffer no personal downside in forcing the rich to donate.
I don’t think either of these hold up. Depending on what program you’re talking about and how you ask, between 60% and 90% of Americans support some of the more popular foreign aid programs. And although I can’t find straight income data, more educated people are more likely to support them, so probably supporters are on average wealthier than opponents. Let’s say that the votes-in-favor-of-foreign-aid coalition controls 80% of national wealth. It can’t be worth their time to pass foreign aid laws (rather than donate directly) just in order to seize an extra 20% from the opponents. The average coalition member probably loses more in lack of flexibility (e.g. about where to direct their donations) than they gain from the extra 20% that seizing opponents’ money gives them!
But this should be mysterious: if the pro-aid coalition is so strong, why don’t they just donate on their own? We can imagine an extreme scenario where 100% of Americans supported taxing everyone $100 and giving it to Doctors Without Borders - isn’t it obviously simpler for each American to donate the $100 themselves without the law?
Yet this whole discussion is predicated on a shared assumption - which I agree with - that without government intervention, most people wouldn’t donate voluntarily, and the amount of money spent on foreign aid would go way down.
This post is about solving this mystery.
The Virtue Signaling Argument
It might seem that a vote on foreign aid offers the following costs and benefits:
VOTE YES: You feel like a good person, but lose $100 in extra taxes
VOTE NO: You feel like a bad person, but at least you keep the $100
But this only applies if you’re the single deciding vote pushing the law from 49.999% to 50.001%, which you never are. In real life, it’s more like:
VOTE YES: You feel like a good person, and your vote doesn’t change your taxes.
VOTE NO: You feel like a bad person, and your vote doesn’t change your taxes.
Since your taxes will be the same either way, voting yes strictly dominates. This is the virtue signaling model of voting, and it implies that even if no single voter supports raising taxes for foreign aid, the electorate might nevertheless vote to raise taxes for foreign aid!
(although most people don’t tell others their votes, maybe this is about self-signaling)
One potential counterargument: votes on these topics tend to reflect polls about voters’ true beliefs pretty well. But you could argue that poll respondents face the same incentive to virtue signal (polls also don’t change your taxes). Answers to poll questions do change based on changing realities (eg people are less likely to support foreign aid if they’re told the budget is tight), but, uh, maybe people are also trying to signal fiscal discipline, and, uh, somehow keep all of these signals straight so that it exactly matches what you would expect if they were voting and responding based on true beliefs.
A stronger counterargument: the Virtue Signaling Argument implies that all “raise taxes slightly to do a nice thing” laws should succeed. But in fact, many of these laws fail. There are dozens of state and local measures like this every election, and they usually follow the pattern you would expect - blue states vote in favor, red states vote against, more likely to pass if the taxes are low and the nice thing is popular.
You could still explain this with more signaling epicycles - the red staters would rather signal fiscal discipline, except on really popular causes where they’d rather signal support - but now the signaling theory has gotten so complicated that it’s almost impossible to distinguish from honestly held beliefs even in principle.
The Insomnia Argument
Fabian’s critique above gestures at the free rider model of fair taxation. In this model, certain taxes (for example, to fund the police) are fair, because there’s no market solution to the same problem that avoids free rider effects. We can’t simply ask each citizen to make their own decision about whether or not to contribute to the police, because police presence lowers crime for everybody. It would be too tempting to defect - that is, to refuse to contribute - trusting that everyone else’s contributions will be enough to maintain adequate police funding level and keep crime low for you. But if everyone does this, the police don’t get funded at all.
Is charity like this? Suppose that we think of charity as purchasing some psychological good - for example, maybe people sleep better knowing that the poor are being helped. Then we can think of non-donation as free-riding on this psychological good - if you donate to charity and solve the African famine, then I can sleep easy instead of fretting about all the poor people starving in Africa.
The obvious counterargument is that many people don’t care. They’re not free riders; they just sleep fine whether poor people are helped or not.
But we usually ignore this when thinking about government. Let’s take an example from the other side of the aisle and imagine a plan to privatize ICE: the government sets laws for what it can do (e.g. who can vs. can’t be deported), but it’s funded entirely by voluntary donations. What goes wrong?
Free-rider theorists would say that there are some benefits to fewer immigrants (e.g. lower crime, less job-stealing) that defectors would enjoy even if they didn’t donate. But many people don’t believe in these benefits. And many other people, whether they believe or not, are unaffected (for example, they live in crime-free gated communities and have un-stealable-jobs like founder/CEO). These people aren’t free riders in the minarchist sense. They just don’t benefit from the policy.
In order to justify immigration enforcement via taxation rather than voluntary donation, you have to argue that the benefit of coordinating the people who do want fewer immigrants is greater than the unfair cost imposed on the people who don’t want that. But then you could make this same argument about charity.
I’m not sure how to think about this one, or how to avoid having it justify almost anything, so I’m presenting it for your assessment but otherwise moving on.
The Bundling Argument
A variation on the above:
Suppose there’s a famine that will kill 50,000 people, and the only way to solve it is through a $5 million project - for example, buying a giant cargo ship to transport food. If there are 50,000 altruists with $100 each, this is a coordination problem, where you need to figure out some way to incentivize all of them to do their part. For example:
They might worry that if they donated, other altruists would free-ride off them.
They might worry that the cause will collect more than $5 million, and the extra would be useless.
They might worry that the cause would collect less than $5 million, and the problem wouldn’t be solved, but nobody would return their money to them.
You could solve some of these with coordination platforms like Kickstarter, and there are clever/complicated solutions to the others, but it’s still not trivial, and lots of projects get sunk by considerations like this.
But this is rarely how real charity works. More likely, your $100 can save one life, and your marginal utility over total lives saved is pretty constant. Making the famine 1 part in 50,000 less bad is 1/50,000th as good as ending it entirely. If the coalition of altruists only reached 49,999 people and saved everyone except for one person, this is still a pretty good outcome, barely any worse than if they’d succeeded completely. So there is no coordination problem here, and no need to call in the government’s coordination-problem-solving ability.
The Bundling Argument asks: what if, psychologically, this isn’t true? What if “your $100 plus equal donations from the rest of our coalition of 49,999 other altruists can completely solve this famine” sounds more appealing then “if you donate $100, you can save one person from a famine which will still kill thousands of others”?
Most people engage with the sorts of distant suffering that require charity primarily through the news. And through the news, a famine that kills 49,999 people sounds the same as one that kills 50,000; from a psychological perspective, your donation did nothing. What you really want is to read the headline “FAMINE SOLVED, GRATEFUL WORLD GIVES THANKS TO CHARITABLE HEROES”. Therefore, people might be more willing to vote for a law that takes $100 from everyone and solves the problem, compared to how willing they would be to donate themselves - even if they aren’t getting any extra leverage from other people’s money.
The Transaction Costs Argument
Economists have invented an instrument called the assurance contract which solves free rider problems without government force.
A leader (the “entrepreneur”) proposes a contract stating that all signatories will donate money, which comes into effect only after a certain number of people sign it. For example, it might say “Everyone will donate 5% of their income to the Pentagon to provide America with a military, and this contract will activate once every American has signed it.”
Making every American sign it is a tough offer - aren’t some people commie traitors who would prefer to see us invaded by our enemies? - but we can lower the threshold to something like 90% with only slight loss of efficiency. Won’t there still be free riders hoping that somebody else signs first and they get to be in the 10% who miss out? There’s a more advanced version called the dominant assurance contract which goes some of the way to solving that problem.
(ACX grantee Spartacus.app helps people create their own assurance contracts for collective action situations)
So: right now, in the real world, do you support replacing military taxation with an assurance contract?
I don’t. Even though the math checks out and the incentives are aligned, I expect it would fail to get 90% of Americans, for many reasons:
Some people would never hear about it, no matter how well-advertised it was.
Some people would hear about it, but constantly put off signing it out of laziness or anxiety.
Some people live paycheck to paycheck and can’t pay 5% of their yearly income.
Some people would misunderstand the situation and come up with incorrect theories of how signing would hurt them, or how they could profit from refusing.
Some people would be so upset at the idea of paying 5% of their income that they would psych themselves into thinking that maybe being invaded by foreigners wasn’t so bad.
The average person nods along to insane statements like “if Elon Musk distributed his fortune evenly, every American would get ten million dollars” and probably doesn’t have the reasoning skills to think about coordination problems clearly.
Along with these reasons, it seems like people don’t donate money even when they care a lot about something. How many people care a lot about wokeness, either pro or con? How many have donated significant amounts of money to organizations promoting or opposing it? Why? Is the answer just free rider problems? Are they just virtue signaling when they talk about wokeness, and they don’t really care?
Forget about reaching 90% - would even half of Americans sign this contract and follow through? Or would we announce the end of coercive military taxation to great fanfare, and then immediately be invaded by Canada and turned into the 11th province?
Contra the economists, I’m not sure that we fund the military through coercive taxation only to avoid free rider problems. I think we fund it through taxation to avoid the same kinds of transaction cost issues that would sink the assurance contract. Since charity suffers these same transaction costs, the same arguments may apply.
The Multiple Preferences Argument
Everyone has multiple conflicting sets of preferences that change based on how they’re being elicited. These go by many names: Near Mode vs. Far Mode, superego vs. id, “my best self” vs. “my regular self”.
Many people say phones are terrible and destroying society and that their life would be much better without a phone and that they wish they could quit their phone. Then they spend all their time on their phone.
I know I should be more social, so when people invite me to a party, I sometimes say yes. Then, when the time for the party comes around, I realize it would involve leaving my house and search for any possible way to get out of it.
Some economists have an irritating tendency to call whichever one ends up being elicited by the immediate environment “the revealed preference” and everything else some kind of fake signaling; I consider this unwise - should we really say that an alcoholic’s “real” preference is to drink too much, then pay $10,000 for rehab as a fake signal to his friends to let him claim he doesn’t want to do it, then go back to drinking because that’s what he really wants, then do another $10,000 rehab stint but this time actually quit for good because his preference coincidentally changed in the meantime? I prefer George Ainslie’s economist-friendly explanation of genuinely time-inconsistent preferences, each of which is able to enlist certain parts of the planning process as allies to its cause.
Preferences around charity display this kind of time-inconsistency. For example, I used to think I “should” donate to charity, but basically never did it. Then I took the Giving What We Can pledge, which forced me to donate a specific amount at a specific time; even though it felt slightly aversive (“aaaah! I’m losing money!”), I did it to satisfy the pledge, and 99.999% of the time (ie every moment except the exact second I clicked on the donate button) I’m happier with my choices. This mirrors how some people feel like they “should” quit alcohol, but don’t do it until they take some kind of sobriety pledge with an organization like Alcoholics Anonymous - and then are happier with their lives at every moment except the exact second that they really want a drink but can’t have one.
On this model, a vote about foreign aid - which involves a tax bill that won’t come due for months, and gets laundered through a government budget so complicated that you never feel the pain from any particular line-item - enlists a different, longer-term preference than a decision whether or not to donate at this exact second (and unless someone has taken the GWWC pledge or an equivalent, the personal decision to donate is just the decision to donate at this exact moment, integrated over every moment of the lifespan). It’s no mystery that people might make different choices on these two decisions, any more than that someone might both use heroin and be willing to vote in favor of a resolution to make heroin vanish from the universe forever.
If this is right, the question becomes: what’s the role of government in these sorts of intrapersonal conflicts? Is it fair for one person’s time-inconsistent preference to enlist government help in its war against that same person’s other preferences, even when it inconveniences third parties (eg the people who don’t want to donate to foreign aid at all?)
There’s no rule saying that government must always support people’s shortest-term and least-willpower-requiring preference; for example, laws against gambling give extra support to the long-term preferences at the expense of the impulsive ones. On the other hand, we don’t want to make a habit of this, or we’ll end up with Prohibition, or bans on porn, or some other dangerously totalitarian attempt to mandate virtue.
My true opinion on this is that I would be satisfied with a law that gives foreign aid by default, but allows any citizen who’s interested to check a box on their tax forms saying something like:
I request to cancel my participation in foreign aid this year and receive an $X tax refund. I understand this will result in Y amount of preventable death and suffering.
…as long as there are similar boxes for everything else the government does that can’t be justified under simple minarchist logic. I predict that only 10% - 40% of voters - the same number who say they disapprove of foreign aid in polls - would check this he box, and I’d be satisfied with this outcome.
In practice, probably this is too weird, and sets a “bad” precedent (from the government’s perspective), and so will never happen. But I still think it’s useful as a thought experiment. Do you agree that, if all foreign aid were cancelled outright, only a small fraction of people would donate on their own? And do you agree that, if the tax form included a box like this, only a small fraction of people would check it?
If so, then voter support for foreign aid can’t be a simple matter of “spending other people’s money”. Something more complicated must be going on.