This is a contribution to the Moral Economics sequence.

At a given price level, there is a an optimal bundle of goods and services that maximizes a consumer's preference satisfaction. This means that given that you are consuming such an optimal bundle, you cannot increase your preference satisfaction by buying more of one good or service and less of another - e.g. more gym classes and less video games.

Suppose, however, that an external actor subsidizes gym classes (e.g., your parents, the government), but not video games. Then you might increase your preference satisfaction by buying more gym classes and less video games. Indeed, that is the point of such subsidies.

Similarly, the government's optimal bundle of goods and services - which is a function of the voters' preferences - may change if certain goods and services are subsidized. For instance, if someone is willing to top up each extra dollar the government spends on foreign aid with ten extra cents, it might rationally decide to spend more on foreign aid, and less on other policy areas. (Exactly how large this top-up would need to be would depend on the marginal utility that the government gets from more foreign aid spending, and the marginal disutility it gets from less spending on other policy areas.)

No doubt there are people willing to make such top-ups, which gives room for moral trade  - a notion developed by Toby Ord - between them and the government. Like all forms of trade, such trade would increase both parties' preference satisfaction.

In principle, anyone could take part in such a moral traed (e.g. another government). Here I am going to focus on a special kind of moral trade with the government, however, namely ear-marking your tax money for a special policy area (e.g. foreign aid) for a certain price. People could be allowed to ear-mark all or parts of their taxes for a particular policy area of their choice, if they paid for it via a higher tax-rate.

My guess is that a fair number of voters would accept this offer even on terms which are quite favourable for the government. If so, allowing people to ear-mark their taxes for a price should be very advantageous for the government. The fact that it got more tax money would outweigh the fact that some of the tax money was ear-marked.

Also, note that these ear-marked taxes will to some extent cancel each other out. If four tax-payers, who pay an equal amount of tax, ear-mark their taxes to four mutually exclusive and equally large policy areas, the government will get more tax money in total without the distribition of government spending having been changed. In practice, some areas are going to be more popular to fund, however, which means that this will perhaps not be the normal state of affairs.

Another advantage is that people might feel more motivated to pay their taxes if they knew that their taxes were going to something they felt strongly about. That could mean less tax evasion. It could also mean that people started to think of their ear-marked taxes as something more akin to charity, and less like taxes in the classic sense. It would feel like less of a burden.

One possible worry that deserves comment, though, is connected to the notion of counterfactual trust (which Ord mentions) and replacability arguments (which, e.g. 80,000 hours have used a lot). Suppose that lots of people start ear-marking part of their taxes for foreign aid. Then the government might be disposed to reduce the part of the general pot of money - coming from non-ear-marked taxes - that goes to foreign aid. In the end, the ear-marking will have led to very little extra spending.


This constitutes a breach of counterfactual trust - the government doesn't distribute the general pot of money in the way it would have done if there hadn't been any opportunity for moral trade, because it thinks that in this way, it can maximize its moral preference satisfaction. I don't see a moral problem with that (which can be the case with other forms of breaches of counterfactual trust), but there is a potential pragmatic problem. If tax-payers see this happen over and over again, they might be dissuaded from ear-marking taxes for their favourite cause.

One way to get around this problem is to explicitly refrain from such redistribution of the general pot, at least the first few years. Inevitably, I think that there will be some such redistribution in the end, though.

Potentially, it won't make that much of a difference, though. We know that people have problems getting these replacability arguments. Lots of people don't just want more money to foreign aid - they want their money to go to foreign aid. If that's right, issues of replacability and worries of breaches of counterfactual trust are less important.


Another worry is that this would introduce a certain element of unpredictability in the government budget - particularly if these sorts of ear-marked taxes became common. Temporary shortfalls in the general budget could be covered by increased government borrowing, but unexpected extra funds would normally have to be saved until new capacity to use them has been added. That would probably be a minor problem, though, as would the extra admin work this would incur.


One objection is that this would give rich people, who pay large amounts of taxes, an undue influence over government spending. It is indeed true that this system would give them a better chance of satisfying their moral preferences. However, the government's moral preferences would also be better satisfied under this system (since it would only take part in such transactions that increases its moral preference satisfaction). Thus it is not true that the gains that rich people make happen at others' expense. Everyone wins from trade - this is not an exception.

As some readers might have noticed, there is a similarity between this system and charity tax-deductions. The main difference is that under a system of ear-marked taxes, the controls of where the spending went to would be much more rigorous.Today, you can get tax-deductions for donations to all sorts of highly ineffective charities. Under this system, you could only ear-mark your taxes for something that the government has already deemed spending-worthy.

Has this idea been put forward before? Has it even been tried in practice, perhaps? If not, one could test it by letting people ear-mark small parts of their taxes to a small number of areas - say foreign aid, health care, education, etc. Perhaps it would be most interesting to test it in a country where you can't make charity deductions, since there's bound to be some competition between tax-deductible donations to charity and ear-marked taxes.

I'm not sure whether this would be a good idea, though, and am open to counter-arguments.





More posts like this

Sorted by Click to highlight new comments since: Today at 5:01 AM

GiveWell opposes earmarking funds for charities, because the charities end up spending money how they want to anyway.

In fact, they believe that it's deceptive for charities to advertise donations as "buying" a particular item or service.

"My guess is that a fair number of voters would accept this offer even on terms which are quite favourable for the government." I doubt that. Republicans are opposed to anything that even resembles a tax increase. Democrats support tax increases, but the tax increases are almost always on "someone else".

Thanks for the links - that's interesting. I think the government is quite different from charities, though. It would have the capacity to be more transparent, and would also be under stronger pressure to be so.

Regarding left vs right, a libertarian site in Sweden shared my post, and the reactions from that camp is generally positive. Social democrats (who are suspicious of tax-deductions to charity as well) are generally more suspicious.

Regarding "tax", you need not necessarily see it as a higher tax. Indeed, a law ph.d. student friend of mine tells me that these sorts of ear-marked contributions would be seen as "fees" under Swedish law. Also, like I said above it is a generalization of tax-deductible donations to charity (with the exception that the criteria for what you could donate to is more strict), which means that you could also see this as a sort of donations.

I think this would get more practically interesting if one could somehow reduce the replaceability of the funds. e.g. if the government had to spend half of its budget in proportion to earmarked funds, and could only redistribute the rest.

Interesting. Probably been suggested somewhere else but where I don't know. It's related to the idea of tax-deduction for charity.

Yes, I point that out towards the end. This idea is inspired by that, and is intended to take care of the primary objection against tax-deductions for charity: that the money might be spent wastefully. Under this system, you're only allowed to get a "deduction" for causes the government are already spending money on (and thus think are important).

Ah, missed it.

I think the main differences are that it's more costly to the donor to donate to a tax-deductible charity than to earmark their domations. (40% of cost rather than ~5%), and the replaceability issue that makes it unclear whether or not earmarked taxes do any good at all.

Thanks, Ryan. I'll take the opportunity to develop my thoughts.

Let me focus on Sweden since that's the country I know best and give you a bit of context. The previous right-wing government made charitable giving tax-deductible a few years ago, but the new Social Democratic government is now about to suspend this law.

One important argument against tax-deductions to charity is that charities can be very ineffective and/or work on unimportant causes. As I point out in the OP, this system effectively takes care of that counter-argument, since you could only ear-mark could only ear-mark your taxes for something that the government has already deemed spending-worthy.

Indeed, you could see this scheme as a generalization of tax-deductions to charity. When a donor is giving to a tax-deductible charity, the government's share of the total cost is, as you say (I think) equal to the donor's marginal tax rate (which is above 50 % in Sweden for high-income donors). Under this system, government's share of the total cost could be any proportion of the total cost. Typically, it would be higher, though (i.e. the donor's share of the total cost would be lower).

Also, in Sweden, the government is funding various charitable organizations directly, such as the Red Cross. Thus, in principle you could ear-mark taxes to the Red Cross via this scheme. (You could also lobby the government to start funding other charities, of course.) The difference is that this scheme would give you more leverage than standard tax-deductions would (since the donor's share of the total cost would be lower).

I've said that under this system, you could only ear-mark your taxes to forms of spending the government is already making. That's not the only possible criterion, but I think it's a natural one.

That criterion is obviously much stricter than the criteria that apply to charity tax-deductions in most countries. This is related to the fact that under this system, the government's share of the total cost typically is higher. Not only is the government refraining from taxing the money you decide to give to a particular charity or policy area - it also tops up this money with other tax-payers' money. This means that it could legitimately claim to have more of a say where this money is going to. It could enforce stricter criteria.

This system can also be seen as "matching donations" - the government matches your donations with a certain multiple of your donation. Apparently that already exists in, e.g. Canada. If EAs could get governments to match donations to effective charities, that obviously would leverage our contributions substantially.

I wouldn't be comfortable with this unless there was strong evidence showing it would work.

People without children (in the public schooling system) will often complain that they pay taxes for public schooling. However, they still benefit from being in a well educated society. This is just one example of people being bad at making decisions on where tax money should go.

Another valid point already raised is that earmarking donations in charities is generally ineffective, and I can't imagine governments being particularly different.

Again, the terms of the trade is set by the government. The government does not have to accept any transaction which doesn't increase its preference satisfaction. So even if it is the case that individuals are bad at making decisions on where tax money should go, that would not mean that the outcome of this kind of trade would be negative.

The government already allocates money to different policy areas in the budget. That should make it comparatively easy to see to that the ear-marked tax money actually goes to the policy area they were supposed to go to.

Interesting. One way to solve the replaceability problem is to force the government to announce a preliminary budget before the ear-marking "bids" and pledge to treat the bids as differentials with respect to the preliminary budget.

Curated and popular this week
Relevant opportunities