On Moral Offsetting

And the doing/allowing distinction

I thought Grok was no longer woke, but alas…

Some in the effective altruism community advocate the moral offsetting of animal products - the practice of purchasing animal products but then donating to an effective animal charity. The idea is that by doing this, you can make it so that your overall impact on animals is net-neutral - or even net-positive.

The reaction from many vegans to this is one of deep repugnance and disgust. They analogize it to “offsetting” acts of egregious wrongdoing against humans.

For example, suppose someone purchases child pornography - and as a result, causes one additional instance of child rape.¹ He then proceeds to “offset” the act by donating to an organization that fights against child sexual abuse - let’s suppose that this donation results in two instances of child rape being prevented.

Surely such an act would be morally worse than the act of neither purchasing child pornography nor donating.

While I agree this is the “common sense” verdict, I actually don’t buy this value judgment. How can an action that results in results in more child rape possibly be morally better than an action that results in less?

In my view, those who make the judgment that the act of purchasing child pornography and subsequently donating to an anti-child sexual abuse charity is morally worse than the act of not purchasing child pornography and not donating to an anti-child sexual abuse charity have lost the forest for the trees.

They correctly recognize how awful it is that the offsetter has caused an additional instance of child rape. But they fail to recognize how even more horrendous it is that the non-offsetter has caused two additional instances of child rape.

Anti-offsetters seem to place moral priority on utterly trivial details - like how exactly one caused an instance of child rape to occur - rather than what’s obviously morally paramount - the child rape itself. To conclude that the offsetter’s actions are worse than the non-offsetter’s is to conclude that the particular manner in which one causes child rape is more important than the amount of child rape one causes! But this is completely crazy!

To me, saying that causing two additional child rapes via not donating to an anti-CSA charity is morally better than the act of causing one additional child rape via the act of purchasing child pornography is like saying that it’s morally better to cause two additional child rapes via pressing a blue button than to cause one additional child rape via pressing a red button.

One may resist my description of the person who doesn’t donate to the anti-CSA charity as causing child rape. But we don’t take issue with describing people as causing harm in similar cases:

For example, suppose a mother refuses to feed her child. As a result, the child ends up dying from lack of food. We don’t have any issue with describing the mother as having caused her child’s death. So I see no reason to object to it here.

I once had a hedonist friend of mine analogize caring about things like virtue intrinsically to caring intrinsically about the amount of blue in the world. As a pluralistic consequentialist, I naturally thought of such an analogy as absurd!

Perhaps my comparison to caring intrinsically about the distinction between causing harm via doing vs. allowing to caring intrinsically about causing harm via pressing a red button or a blue button strikes some as absurd in a similar way.

In that case, let me try to motivate the case that the distinction that doing and allowing harm is not a morally relevant one:

First of all, it’s important to disentangle two different things: whether doing vs. allowing harm is itself of moral importance or whether it’s merely correlated with factors of moral importance.

Just like one can think the typical human is more valuable than the typical pig without thinking that species is itself of moral importance, one can think that the typical act of killing is worse than the typical act of letting die without thinking that the distinction between doing vs. allowing harm is itself of intrinsic moral importance.

I think that many cases of doing harm come with bad-making features that instances of allowing harm don’t. For example, suppose someone tortures a turkey in their basement. Now compare this to someone who decides to buy a Spotify Premium subscription instead of donating to an effective animal charity.

I think there’s a level of malice involved in the case of the turkey torturer that simply isn’t present in the case of the Spotify Premium subscriber - even if both result in the same amount of pain.

My differential attribution to malice isn’t reducible to the fact that the turkey torturer does harm whereas the Spotify Premium subscriber allows harm - as I’d say similar things about the turkey torturer vs. someone who purchases a turkey from the grocery store for Thanksgiving dinner. Even though both cause harm - maybe even the same amount of harm - it seems clear that the turkey torturer exhibits a level of malice that the Thanksgiving cook simply doesn’t.

One morally relevant difference is that the person who tortures a turkey in their basement is likely deriving some sort of sadistic pleasure from the act of harming the bird that the people who purchase the Thanksgiving turkey or Spotify Premium subscription aren’t.

Now we can construct cases where this isn’t the case - but there are still morally relevant differences to my mind. The turkey torturer is viscerally aware of the suffering they’re inflicting - whereas the Spotify Premium subscriber or Thanksgiving cook may not even be thinking of the fact that they’re causing suffering at all.

Even if they are, the knowledge of the harm they’re causing is a lot less visceral and more abstract.

Suppose that when the Thanksgiving cook went to buy the turkey, graphic and detailed imagery of the future agony that their decision would result in played in their head. If they still went ahead and purchased the turkey anyway, I think this would reflect way worse than their moral character than the average non-vegan! Ditto for the Spotify Premium subscriber.

I believe that if we get rid of these other morally relevant factors, it seems clear that the distinction between doing and allowing harm is not itself morally relevant:

Suppose someone is going for a morning jog. Unbeknownst to them, if they step on a particular part of the sidewalk, this will result in someone being tortured to death. Now consider the case of a couch potato. They plan to stay in all day watching Netflix. Unbeknownst to them, if they don’t go outside and step on a particular part of the sidewalk, this will result in someone being tortured to death.

Note that the morning jogger does harm - whereas the couch potato allows harm. Yet it seems that the moral statuses of these two actions is the same. It does not seem like the couch potato acted any better than the morning jogger.

Let’s consider two other cases. Suppose that ten children are set to be horrifically tortured - a bystander has the power to avert this simply by blinking his eyes. However he decides not to - because he likes when children are in pain. In the other case, someone is aware that if he blinks his eyes, ten children will suddenly be horrifically tortured. He decides to blink his eyes - because he likes when children are in pain.

Like the previous two cases, I don’t think there is any moral difference between these two cases. The blinker doesn’t act any worse than the non-blinker.

The Moral Character of Individuals vs. The Moral Status of Actions

Now that we’ve disentangled features simply correlated with doing vs. allowing harm and doing vs. allowing harm itself, it’s important to disentangle two other concepts: the moral character of individuals vs. the moral status of actions.

Judging that act X is morally better than act Y need not require judging that the person who performs X is morally better than the person who performs Y.

Let’s go back to the turkey torturer. Suppose that after he gets done torturing the turkey in his basement, he suddenly becomes worried that he’ll be sent to Hell for what he’s just done - unless he donates a sufficient amount of money to an animal charity in order to prevent two turkeys from being tortured. He then proceeds to do this. He does not do this out of any actual concern for animal suffering itself - he only does it because he thinks it’ll prevent him from getting sent to Hell.

I propose that the turkey torturer’s actions are morally better than both the actions of the Thanksgiving cook and the Spotify Premium subscriber. After all, the turkey torturer’s actions result in substantially less severe suffering - and I think the amount of severe suffering produced is the primary determinant of an action’s moral status.

That said, it also seems clear to me that the turkey torturer is a much worse person than both the Thanksgiving cook and the Spotify Premium subscriber. They possess a level of extreme maliciousness that the other two simply do not.

There are other cases where we can see this as well.

Suppose someone is convinced that pressing a button will prevent a squirrel from having their head bashed in with a crowbar. All the evidence available to them suggests that this is the case. Seeking to spare the squirrel, they promptly press the button - however it turns out that pressing this button actually results in a squirrel being attacked with a crowbar.

Likewise, suppose someone is similarly convinced that this pressing this button will prevent a squirrel from having their head bashed in a crowbar. However, they don’t really care about animals - and so refrain from pressing the button. As a result, the squirrel is spared from being attacked with a crowbar.

I think the second action is a morally better one - an action that results in a squirrel not having their head bashed in with a crowbar is way better than one that does result in that happening! - but the second person is a morally worse person.

On Demandingness

One concern one may have with the view I’ve laid out here - that there is no fundamental moral distinction between doing and allowing harm - is that it leads to a morality that is simply too demanding.

One may worry that if failing to prevent an act of torture is just as bad as committing torture, then we’re all moral monsters tantamount to torturers. But as I’ve already shown, judgments of acts and character can come apart. Judging two acts as equally bad does not entail judging the two agents performing those acts as being equally bad.

Still, one may worry that even if we’re better people than torturers, we’re still all constantly performing actions just as bad as those of torturers.² Surely it can’t be that such a large proportion of our everyday actions are so morally atrocious!

While such a conclusion may be unsettling, I remain drawn to it anyway. In a similar vein, it is like veganism. Accepting ethical veganism also means indicting a large proportion of our everyday actions as morally egregious.

It is uncomfortable to think that we are currently witnessing an atrocity far worse in moral severity than the Holocaust - that virtually everyone else participates in and with seemingly no end in sight.

But just because a conclusion is uncomfortable doesn’t mean it isn’t correct.³ And when I reflect on whether there is any moral difference between doing and harming harm, I see no way to escape the conclusion that there isn’t one.

There’s no reason to think that a large proportion of our everyday actions aren’t seriously morally objectionable. And there’s no reason to think that “demandingness” is a mark against a moral view.

Suppose that every time we blinked, it caused an innocent person from a distant galaxy to be teleported into a brazen bull. Humans blink an average of between 14,000 and 19,200 times a day. Surely it would be extremely demanding to insist that we refrain from blinking. Yet surely we still ought to refrain from blinking.

Likewise, it may be demanding to insist that we refrain from purchasing luxury goods and instead donate our excess income to charity. Yet surely we still ought to.

On Hypocrisy Allegations

There’s an inevitable thought that comes to people’s heads when they hear this sort of stuff - well sure, that’s all very well and good Travis, but do you live up to this advice? Do you donate all your money to charity? Do you spend every waking hour working to earn additional money to donate - or do you spend your time writing Substack articles about philosophy and complaining about the color scheme of Bugs Bunny in Season 2 of The Looney Tunes Show?

This is an objection that every consequentialist will inevitably be hit with. You don’t really believe what you say - you don’t act in accord with the purported moral principles you swear by.

There are a few things to say about this. First of all, it’s unclear why someone acting out of accord with a given moral view would necessarily mean they don’t actually hold that moral view. If I say that a given action is morally bad and yet I perform this action anyway, why does that mean I must not actually think that action is morally bad?

Note that we don’t say the same thing about prudential judgments. Just the other day, I ate an entire bag of Sweet Spicy Chili Doritos. At the time, I would’ve acknowledged that this was prudentially bad - that I shouldn’t do it. Yet I did it anyway. Does this mean I didn’t really think that eating that entire bag of Doritos was prudentially bad?

This seems absurd. Yet I’m not sure what’s supposed to make it absurd in this case but not in the moral case.

I have a variety of different goals. Sometimes these goals conflict with each other. When I make moral or prudential judgments, these are judgments about a subset of my goals. To judge something as morally (or prudentially) better or worse is to judge how conducive or detrimental it is to achieving my moral (or prudential) goals.

Since I have conflicting goals, that means unfortunately I can not simultaneously realize all my goals. I have to choose some goals to act in accord with - to the detriment of my other goals.

Oftentimes I’ll find myself motivationally compelled to act in accord with my non-moral goals - to the detriment of my moral goals. But this doesn’t mean I don’t actually have the moral goals I claim to - or that my judgments about what’s most conducive to those goals are false.

Second of all, the mere fact that someone is a hypocrite is not itself an indictment of their position. Let’s go back to the case I mentioned previously - where we discover that every time we blink a person ends up being tortured. Suppose that someone insisted that we should stop blinking - yet this person didn’t completely refrain from blinking themselves. The mere fact that they’d be hypocrite doesn’t mean they’d be wrong that we should stop blinking!

Are Animal Product Offsetters Similar in Character to Child Porn Offsetters?

Even if one accepts the verdict that purchasing child pornography and donating to an anti-CSA charity is ethically better than not purchasing child pornography and not donating (assuming it leads to less overall child victimization), one still may note that we tend to view the child porn offsetter has having a substantially worse moral character than the person who simply doesn’t watch child pornography at all (and doesn’t donate either).

Should we therefore similarly conclude that while purchasing animal products and donating to animal charities is a better action than not purchasing animal products and not donating (assuming it leads to less overall animal suffering), animal product offsetters nevertheless have a substantially worse moral character than vegans (who don’t donate)?

I don’t think so. I think watching footage of children being raped and deriving sexual enjoyment out of it displays callousness that simply enjoying the taste of animal products doesn’t. Animal product purchasers are more detached from the abuse that they fund - whereas child porn purchasers actively watch the abuse they fund and get off on it.

I think those who purchase child pornography are more analogous to those who purchase animal crush videos. If someone purchased animal crush videos - then proceeded to donate to animal charities to offset it - I would think of this person of having similar moral character to the child porn offsetter and much worse moral character than the animal product offsetter.

And if humans were being factory farmed and the sale of their flesh and secretions were as ubiquitous as the sale of animal products is, I would not view those who offset human products as being any worse in moral character than those who offset animal products.

Practical Objections

So far I’ve focused on principled objections to moral offsetting - objections that take for granted that offsetters’ actions result in the same or less amount of suffering as vegans and charge that their behavior is still morally worse anyways.

Yet there is a second kind of objection to offsetting - and one I’m more sympathetic to - that questions whether the amount that non-vegans donate is really sufficient to offset the amount of harm produced through their diet.

Offsetting proponents often cite very optimistic estimates for how easy it is to offset - I’ve seen people claim that you prevent an animal from being factory farmed for just 25¢ and that you can offset an entire year’s worth of non-veganism by donating $5.

I for one am very skeptical of these figures. And some of the claims I’ve seen by offsetting proponents have been quite sloppy. For example, in this article by Jim Davies, he states:

But let’s look at the kind of person who would become vegan. How many animals would they eat if they ate meat? Or, to put it another way, before somebody goes vegan, how many animals did they eat? For most vegans, the answer is probably lower than the population’s average. Some people are vegan (or are better able to maintain veganism) because they don’t like meat as much as others like meat. But let’s assume, just for simplicity, that if a vegan started eating meat, they’d eat what the average American ate, and so, by virtue of their vegan diet, they are saving 105 animals a year. That’s the immediate effect of veganism on animals, though there are other, indirect effects, such as making veganism more socially acceptable to others.

He uses this to come to the conclusion that one can more than offset a year’s worth of non-vegan by simply donating $100. But there’s a big issue with this:

The 105 animals estimate is from Animal Charity Evaluators’ Effects of Diet Choices report. However, this report is about how many animals would be spared if the average person globally went vegan, not how many animals would be spared if the average American went vegan.⁴

ACE’s report actually does make mention of a previous estimate by Harry Sethu about how many animals would be spared from an average American going vegetarian: between 371 and 582!⁵ And this is for the average American going vegetarian, not going vegan!

So in order for an American to offset their diet, they’d have to donate substantially more.

(It’s also additionally worth noting that ACE’s report itself states that their figure is likely an underestimate. Their report also only considers animals that are vertebrates.)

My Own Behavior

Perhaps you want to inquire about my own behavior. Do you, Travis Talks, offset? Or are you a full-blooded vegan?

When it comes to the things I buy⁶, they are overwhelmingly vegan (I’d say over 99%) - though I have offset on a handful of occasions. In most of these cases, it was because the non-vegan version of what I wanted to buy was cheaper and I determined that it would be more conducive to the goal of helping animals to donate the money I saved to charity instead of paying. In two other cases it was because I missed a particular food and decided to get them.

On the handful of occasions where I have purchased animal products, I’ve bought the ones that generate lower amounts of suffering (like cow products) while still straying clear from the ones that generate the most amount of suffering - like chicken, eggs, and farmed fish. My donations tend to be about 40% or more of the price of the animal product purchased - just to be on the safe side.

I used to be stricter than avoiding animal products - lately I’ve become less so due to the considerations I’ve mentioned here. Just recently I bought a new pair of shoes for $60 - I didn’t strictly need a new pair of shoes. My old shoes weren’t in very good condition - but I could’ve gotten away with wearing them for a bit longer. I could’ve bought a cheaper pair of new shoes. But I didn’t. I bought a new pair for $60 (though this is still better than some of my more unvirtuous friends, who spend $100+ on shoes…).

Doing this instead of donating the money certainly causes much more suffering in expectation than spending $2.24 on a pint of dairy ice cream. So I figured it’s somewhat arbitrary for me to continue doing stuff like this while being super strict about avoiding animal products.

It may be suggested that I’m simply trying to rationalize my own behavior here. Perhaps - though the same could be said about the vegans purchasing luxury goods who defend the doing vs. allowing distinction.

We’re All Complicit

I hope the main takeaway from my article is that we’re all complicit - whether we purchase animal products or just luxury goods - in untold amounts of harm to animals. None of us has clean hands.

This shouldn’t be taken as a dig at vegans. I don’t want to come across like the crop death morons who claim that since vegans kill animals too, that means there therefore must be nothing morally objectionable about funding a modern-day Auschwitz.

Obviously being vegan is morally superior to being non-vegan - all else being equal. There’s no doubt about that.

The point is just to have some perspective. There’s a contingent of vegans out there that will go insane over seeing someone make the slightest stray from veganism - I once saw a group of vegans on Twitter freaking out over an otherwise vegan person eating a box of breakfast cereal that contained a small amount of animal-derived vitamin D.

And it’s like come on. Spending $10 on a month of Spotify Premium instead of donating is resulting in way more animal suffering in expectation than just buying a box of Cinnamon Toast Crunch or whatever. At some point it seems like the goal isn’t actually reducing animal suffering but just maintaining personal purity.

So while we’re all complicit in harm to some degree or another, let’s still do what can to alleviate the suffering of animals. For some this means total abstention from animal products. For others this means consuming more animal products - but also donating more to animal charities. And that’s fine.

1. One may object that this hypothetical is disanalogous to purchasing animal products since it doesn’t actually result in additional animals being factory farmed. While it is true that any individual given purchase of animal products is unlikely to result in additional animals being factory farmed, the expected value of the action is still one of additional animal suffering. The expected value of purchasing a chicken, for instance, is one additional chicken being bred into existence. If one is tempted to just ignore the expected value and look at what’s most likely to occur, here’s an intuition pump to dissuade you from this: Suppose that pressing a button has 0.1% chance of sentencing a billion people to be tortured and a 100% chance of giving you a lollipop. Pressing the button will almost certainly result in you benefitting and no else being harmed - yet the expected value of the action is that ten million get tortured. It seems obviously wrong to press the button!

2. Technically I don’t actually think that our typical failure to donate is just as bad as committing a typical act of torture - even though I think failing to prevent torture is just as bad as committing torture. This is because I think that virtuousness and viciousness influence more than just the moral character of people - they influence the moral status of actions themselves. On my view, an action performed with malicious intent is morally worse than an otherwise identical action performed without malicious intent - since malicious intent is itself intrinsically disvaluable.

3. This shouldn’t be taken as an endorsement of moral realism - that there’s some correct or incorrect moral conclusion independent of how we feel about it. I am a moral anti-realist - however the mere fact that a conclusion is uncomfortable doesn’t mean it doesn’t follow from your values. For example, suppose it turns out the only way to prevent one million babies from boiled alive is to ensure one baby is boiled alive. It may be an uncomfortable conclusion to accept that it could morally obligatory to boil a baby alive - however it may nevertheless follow from one’s values. None of this assumes the existence of objective values.

4. Davies even says earlier in his article that the report is about how many animals would be spared if the average person went vegan - which makes it all the more strange that he suddenly starts acting like it’s applicable to the average American.

5. One caveat about this report is that some of the animals included may be non-conscious.

6. I say “things I buy” because I do regularly consume free animal products. I consider this no morally different than consuming free vegan food.