Consequences of animal product consumption (combined model)

(this is mostly old content—I’m just combining this and this into a single post with slight updates)

This spreadsheet estimates the consequences of animal product consumption. I start by discounting 1kg purchased of each animal product by the cumulative elasticity factor. But when other people stop buying an expensive animal product, they may start buying a different one. Therefore, I make rather arbitrary guesses for these effects—how much more chicken will people buy when they give up beef, etc. So for each additional 1kg of beef you buy, I estimate that the additional production will be 0.68kg of beef, 0.05kg of chicken, 0.01kg of salmon and so on.

I use data on lifespan and meat yields, assumptions for intrinsic moral weight relative to humans informed by intuition and neuron counts (pigs = 0.17, cows = 0.15, turkey = 0.07, chicken = 0.06, fish = 0.03), and assumptions for quality of life (combined ratings from Charity Entrepreneurship and from Compassion by the Pound) to then estimate the net welfare lost per kg purchased. The figures are in human-equivalent welfare-days on a −100 to +100 scale, where −100 means none of your needs are met (note, still much better than torture). So when the spreadsheet says 1kg of chicken causes −37.7 welfare (cell AI3), that means that buying 1kg of chicken causes an equivalent amount of animal suffering to the feeling of being fully hungry, miserable, in pain, etc for about 6.5 hours out of a 17 hour waking day, instead of feeling completely neutral during that time.

Then I use data on CO2e emissions per kg of food produced. Note that this doesn’t account for the counterfactual emissions from eating plants—but these are generally much lower than the emissions from animal products. Using both short-run and long-run estimates of the costs of climate change and other environmental costs upon humans, I get a measurement of the human cost of each kg of CO2e.

With some typical calorie/​kg numbers, I get combined figures for the welfare costs of 750 calories which is approximately one meal.

I then use estimates of effective charity impacts to see how much money it would cost to Pareto offset these impacts (i.e.: donate to climate change charities to offset the CO2e, donate to animal charities to offset the animal suffering).

Finally, I look at the amount of additional farm animal suffering that is predicted when people gain income in a set of African countries supported by GiveWell aid programs—the “meat eater problem.”

Note: this model ignores wild animal suffering, leaving it all to be considered unclear. The calculations for wild fish are just about the possible indirect effects on farm animal products.


  • Farmed fish is the worst product, poultry/​eggs/​pork/​beef/​wild fish are similar in the middle, and cheese/​milk are the most benign

  • For someone with access to adequate vegan food, all typical animal products have enough costs to outweigh the direct pleasure and convenience of consumption, with the possible exception of small amounts of dairy

  • Using decently conservative assumptions for charity impacts, the ethical costs of a high-meat diet can be offset with approximately a nickel a day donated to relevant effective charities (or less, if you select a more effective cause area instead of a climate & animals bundled donation)

  • Increasing annual income for a poor person by $1,000/​year has a farm animal welfare cost equivalent to adding several weeks of misery. To me, given the direct noneconomic benefits of efforts such as malaria prevention and deworming, such efforts seem robustly good enough to outweigh the animal costs in the short run. Things like industrialization and GiveDirectly seem to have unclear short run welfare impact, depending on the characteristics of the recipient populations