Moral public goods

Link post

Au­to­mat­i­cally crossposted

Sup­pose that a king­dom con­tains a mil­lion peas­ants and a thou­sand no­bles, and:

  • Each no­ble makes as much as 10,000 peas­ants put to­gether, such that col­lec­tively the no­bles get 90% of the in­come.

  • Each no­ble cares about as much about them­selves as they do about all peas­ants put to­gether.

  • Each per­son’s welfare is log­a­r­ith­mic in their in­come.

Then it’s si­mul­ta­neously the case that:

  1. Nobles pre­fer to keep money for them­selves rather than donate it to peas­ants—money is worth 10,000x as much to a peas­ant, but a no­ble cares 1,000,000 times less about the peas­ant’s welfare.

  2. Nobles pre­fer a 90% in­come tax that is re­dis­tributed equally—a tax that costs a par­tic­u­lar no­ble $1 gen­er­ates $1000 of value for peas­ants, since all other no­bles will also pay the higher taxes. That makes it a much bet­ter deal for the no­bles (un­til the to­tal in­come of no­bles is roughly equal to the to­tal in­come of peas­ants).

In this situ­a­tion, let’s call re­dis­tri­bu­tion a “moral pub­lic good.” The no­bles are al­tru­is­tic enough that they pre­fer it if ev­ery­one gives to the peas­ants, but it’s still not worth it for any given no­ble to con­tribute any­thing to the col­lec­tive pro­ject.

The rest of the post is about some im­pli­ca­tions of tak­ing moral pub­lic good se­ri­ously.

1. Jus­tify­ing redistribution

This gives a very strong eco­nomic ar­gu­ment for state re­dis­tri­bu­tion: it can eas­ily be the case that ev­ery in­di­vi­d­ual prefers a world with high re­dis­tri­bu­tion to the world with low re­dis­tri­bu­tion, rich and poor al­ike. I think “ev­ery­one prefers this policy” is ba­si­cally the strongest ar­gu­ment you can make on its be­half.

(In fact some peo­ple just don’t care about oth­ers and so not ev­ery­one will benefit. I’d per­son­ally be on board with the purely self­ish peo­ple just not fund­ing re­dis­tri­bu­tion, but un­for­tu­nately you can’t just ask peo­ple if they want to pay more taxes and I’m not go­ing to sweat it that much if the most self­ish peo­ple lose out a lit­tle bit.)

I think this ar­gu­ment sup­ports lev­els of re­dis­tri­bu­tion like 50% (or 30% or 70% or what­ever), rather than lev­els of re­dis­tri­bu­tion like 99% that could nearly level the play­ing field or en­sure that no billion­aires ex­ist. I think this enough to cap­ture the vast ma­jor­ity of the pos­si­ble benefits from re­dis­tri­bu­tion, e.g. they could get most house­holds to >50% of the av­er­age con­sump­tion.

This ar­gu­ment sup­ports both for­eign aid and do­mes­tic re­dis­tri­bu­tion, but the for­eign aid com­po­nent may re­quire in­ter­na­tional co­or­di­na­tion. For ex­am­ple, if ev­ery­one in de­vel­oped coun­tries cared equally about them­selves, their coun­try, and the world, then you might end up with op­ti­mal do­mes­tic poli­cies al­lo­cat­ing 10% of their re­dis­tri­bu­tion abroad (much less in smaller coun­tries who have min­i­mal in­fluence on global poverty, a lit­tle bit more in the US), whereas ev­ery­one would pre­fer a mul­ti­lat­eral com­mit­ment to spend 50% of their re­dis­tri­bu­tion abroad.

2. There are lots of pub­lic goods

I think it makes sense for states to di­rectly fund moral pub­lic goods like ex­is­ten­tial risk miti­ga­tion, ex­plo­ra­tion, ecolog­i­cal preser­va­tion, arts and sci­ences, an­i­mal welfare im­prove­ments, etc. In the past I’ve thought it usu­ally made more sense to just give peo­ple money and let them de­cide how to spend it. (I still think states and philan­thropists should more of­ten give peo­ple cash, I just now think the pre­sump­tion is less strong.)

In fact, I think that at large scales (like a na­tion rather than a town) moral pub­lic goods are prob­a­bly the ma­jor­ity of pub­lic goods. Car­ing slightly more about pub­lic goods slightly changed my per­spec­tive on the state’s role. It also makes me sig­nifi­cantly more ex­cited about mechanisms like quadratic fund­ing for pub­lic goods.

I en­joyed David Fried­man’s The Machin­ery of Free­dom, but it re­peats the com­mon liber­tar­ian line that dona­tions can help the poor just as well as taxes:

If al­most ev­ery­one is in fa­vor of feed­ing the hun­gry, the poli­ti­cian may find it in his in­ter­est to do so. But, un­der those cir­cum­stances, the poli­ti­cian is un­nec­es­sary: some kind soul will give the hun­gry man a meal any­way. If the great ma­jor­ity is against the hun­gry man, some kind soul among the minor­ity still may feed him—the poli­ti­cian will not.

This seems to­tally wrong. The use of co­er­cive force is an ac­tive in­gre­di­ent in the state feed­ing the hun­gry, as it is with other pub­lic good pro­vi­sion. Anar­chists ei­ther need to make some spec­u­la­tive pro­posal to fund pub­lic goods (the cur­rent menu isn’t good!) or else need to ac­cept the pareto in­effi­ciency of un­der­fund­ing moral pub­lic goods like re­dis­tri­bu­tion.

3. Altru­ism is not about consequentialism

Con­se­quen­tial­ism is a re­ally bad model for most peo­ple’s al­tru­is­tic be­hav­ior, and es­pe­cially their com­pro­mises be­tween al­tru­is­tic and self­ish ends. To model some­one as a thor­ough­go­ing con­se­quen­tial­ist, you have two bad op­tions:

  1. They care about them­selves >10 mil­lion times as much as other peo­ple. Donat­ing to al­most any­thing is in in­sane, no way the re­cip­i­ent val­ues the money 10 mil­lion times more than I do.

  2. They care about them­selves <1% as much as ev­ery­one else in the whole world put to­gether. When choos­ing be­tween pos­si­ble wor­lds, they would gladly give up their whole fu­ture in or­der to make ev­ery­one else’s life a lit­tle bet­ter. Their per­sonal prefer­ences are nearly ir­rele­vant when pick­ing poli­cies. If they found them­selves in a very pow­er­ful po­si­tion they would be­come rad­i­cally more al­tru­is­tic.

I think nei­ther of these is a great model. In fact it seems like peo­ple care a lot about them­selves and those around them, but at the same time, they are will­ing to donate small amounts of their in­come.

You could try to frame this as “no one is al­tru­is­tic, it’s just a sham” or “peo­ple are ter­rible at moral­ity.” But I think you’ll un­der­stand most peo­ple’s al­tru­ism bet­ter if you think about it as part of a col­lec­tive ac­tion or pub­lic goods pro­vi­sion prob­lem. Peo­ple want to e.g. see a world free from ex­treme poverty, and they are (some­times) will­ing to chip in a small part of that vi­sion for the same rea­son that they are will­ing to chip in to the lo­cal pub­lic park—even though the ac­tual con­se­quence of their dona­tion is too small for them to care much about it.

On this per­spec­tive, donat­ing to lo­cal char­i­ties is on much more even foot­ing with donat­ing to dis­tant strangers. Both are con­tri­bu­tions to pub­lic goods, just at differ­ent scales and of differ­ent types, and that’s the thing that most unifies the way peo­ple ap­proach and think about them. The con­se­quen­tial­ist anal­y­sis is still rele­vant—helping the poor is only a moral pub­lic good be­cause of the con­se­quences—but it’s not that the lo­cal char­ity is just a con­se­quen­tial­ist er­ror.

In ad­di­tion to mi­s­un­der­stand­ing nor­mal hu­mans, I think con­se­quen­tial­ists some­times make re­lated er­rors in their own judg­ments. If a bunch of util­i­tar­i­ans want to en­joy a nice com­mu­nal space, it’s worth­while for each of them to help fund it even though it nei­ther makes sense on util­i­tar­ian grounds nor for their own self-in­ter­ests. That’s a good norm that can leave ev­ery util­i­tar­ian bet­ter off than if they’d spent the same money self­ishly. I think that a lot of moral in­tu­ition and dis­course is about this kind of co­or­di­na­tion, and if you for­get about that then you will both be con­fused by nor­mal moral dis­course and also fail to solve some real prob­lems that ev­ery­day moral­ity is de­signed to solve.