Epistemic status: The welfare calculations are based on assumptions whose limitations I discuss throughout. The monetary levels below should be read as order-of-magnitude. The conclusion that there is a tradeoff between health, environmental impact, and animal welfare is robust.

In this post, I argue that a tax on meat consumption, with a lesser tax rate on non-meat animal products, is a necessary and ideal policy for addressing the harms of factory farming. Meat consumption entails a significant tradeoff between environmental and health concerns, which have primarily been the leading motivators of a meat tax, and animal welfare concerns. The former issues point to a higher tax rate on meat from large animals, especially beef. The latter concern points to a higher rate on smaller animals.

A pound of beef costs roughly one nine-hundredth of a cow's life, while a pound of poultry costs about a fifth of a chicken's. Shifting a pound of demand from beef to chicken yields significant reductions in emissions and health risk, but at a cost of more than 100 times as many animals slaughtered, with poorer quality of life. For this reason, animal advocacy organizations such as Faunalytics are wary of meat taxes that target only environmental externalities and are disproportionately focused on red meat.

To illustrate the tradeoff, I focus on welfare and greenhouse gas emissions from poultry and beef consumption. While the ideal Pigouvian tax rate of meat is most likely far higher, I suggest a rate of $1.00 per pound for all kinds of meat, and $0.50 per pound for non-meat animal products, as an appropriate value.

I argue for the principle of a flat meat tax, but there is much remaining work to bring the policy to a reality. More research and consensus is needed on how to monetize animal welfare for the purposes of setting a tax, and much work is needed to build the political will for a tax. Nevertheless, it is my hope to make the case that a meat tax is the only credible pathway toward putting a major dent into farmed animal cruelty.

Claude Opus was heavily used in the course of researching this post. All writing is done directly by humans, though Claude helped in the editing process and formatting references. All sources are human-verified.

Scope

For the sake of a tractable analysis that nevertheless highlights the main point, I focus on two products–poultry and beef–in the United States, and I focus on two major external costs that serve as the justification for a tax: animal welfare and greenhouse gas emissions. Other environmental impacts, such as land use and water consumption, as well as the health implications of diets, are considerable, but I will not attempt to monetize these.

A further scope restriction is that I focus on poultry and beef only and not on eggs or milk. Welfare considerations in turn are restricted to broiler chickens and beef cattle, even though some poultry is produced from layer chickens and some beef from dairy cows. This restriction lowers the per-pound assessment of the welfare cost of meat. I furthermore do not consider coproducts, such as leather, and this restriction raises our assessment. My calculations also omit consideration of broiler breeders and cattle breeding herds, an additional underestimate of welfare impacts.

Animal Welfare

For both chicken and beef, for the sake of having a comparable model, I will apply the framework of Rethink Priorities’ Moral Weight Project. Based on the MWP, I adopt four assumptions: utilitarianism, hedonism, valence symmetry, and unitarianism. Utilitarianism is the framework that assesses moral value as the sum of (expected) utility over all sentient beings. Hedonism holds that an animal’s state is determined strictly by positively and negatively valenced experiences, and I acknowledge that it may be regarded as incomplete from an animal rights perspective. Valence symmetry holds that positively and negatively valenced states of equal magnitude should be treated with equal and opposite weights. This raises the question of how to think about the possibility of a farmed animal with an overall positive welfare, which is addressed below. Unitarianism counts equal amounts of welfare equally, regardless of the species affected. This assumption is necessary to generate monetary values.

The key elements of the framework are as follows.

For each type of animal, I estimate the number of hours at four levels of pain over its life (excruciating, disabling, hurtful, annoying).

There is a multiplier for each level of pain, from which the four values above can be aggregated into a single pain level.

The framework calls for the use of two weights for each species. The first is called the welfare range, which captures how bad the animal’s experience of its worst state is compared to that of a human. The second weight is sentience probability, which is the probability that the animal should be regarded as sentient for the purposes of moral calculus. I aggregate these two parameters into a single value known as the probability-of-sentience-adjusted welfare range.

Once an animal’s suffering is expressed in comparison with that of a human, the value is monetized by anchoring to a human quality-adjusted life-year.

The total welfare cost is then assessed by multiplying the above by the numbers of animals slaughtered each year and their average lifespan. To assess the cost on a per-pound of meat basis, the above is divided by the number of pounds of meat sold annually in the United States.

Welfare Levels

Welfare levels for broiler chickens are readily available from the Welfare Footprint Institute, and I use those values. For beef cows, I am not aware of such numbers having been computed, and so I estimate them based on the most prevalent welfare issues that afflict beef cattle.

Chickens

For broiler chickens, I take the numbers from the Welfare Footprint Institute. Those numbers include 30.2 seconds of excruciating pain per chicken, 50.27 hours of disabling pain, 333.60 hours of hurtful pain, and 324.67 hours of annoying pain. This compares to an average lifespan of 47 days, and so the majority of the broiler’s life is in a state of pain.

Cows

To my knowledge, similar numbers have not been computed for beef cattle, and so I will build an estimate of the major issues in cattle welfare. For this analysis, I made a lot of best guesses and also considered only the most severe pain issues that beef cattle face.

The actual slaughter is left out on the grounds that I imagine it to be a very fast process and thus one that would not contribute significantly to these calculations.

Thermal Stress

Schuck-Paim et al. (2026) estimate the amount of time that cows suffer from heat in five countries in South America. Their figures are conveniently divided into the four pain levels described above. In Figure 3, the study summarizes the pain levels from heat in four thermal load zones. The United States is on the temperate side relative to the countries that are the basis for this study, though there is plenty of ranching in the hot states in the Southwest. I am taking the “High Annual Thermal Load”, which is the third most severe thermal load out of four assessed in the study, as representative of U.S. cattle farming. That scenario estimates 61 heat stress days per year, and on a heat stress day, cows experience an average of 0.7 hours of disabling pain, 3.8 hours of hurtful pain, and 2.6 hours of annoying pain. Note that the subspecies Bos indicus is more commonly raised in tropical climates, and Bos taurus is raised in more temperate climates, and these calculations do not account for the different heat tolerances of these subspecies (see Scheffler 2022).

Schuck-Paim et al. (2026) only study heat stress and not cold stress, but cold is a serious problem too in the American climate. For instance, hundreds of thousands or millions of cattle died in the particularly harsh winters of 1885-86 and 1886-87, an event that is known as the Big Die-Up. Cox et al. (2016) conduct a study on ranching in Belgium and find roughly equal mortality from heat and cold among dairy cows from 2006 to 2009. Belgium has a relatively cold climate compared to the U.S., though as we saw, the U.S. can have harsh winters too in the Plains states. Nevertheless, I estimate cold-induced pain as being a quarter of heat-induced pain among cattle among all the pain levels in the United States to account for the warmer climate in the U.S.

Painful Operations
For this analysis, I will focus on three painful experiences that farmed cattle commonly go through: castration, dehorning, and branding.

Most male farmed cows go through castration. The reasons are to improve meat quality, to prevent breeding except among designated bulls, and to reduce aggression among the herd. The latter two reasons have their own welfare rationales, but be that the case, castration is a painful procedure, especially if done at an older age. Pain is based on a study by Meléndez et al. (2025), which is motivated by efforts to improve animal welfare. The duration of pain from the operation varies, but based on that study, I’ll estimate seven days of pain, which is 140 hours when considering that cows are typically awake for 20 hours per day. I’ll break that down into 0.1 hours (6 minutes) of excruciating pain, 2.9 hours of disabling pain, and 137 hours of hurtful pain. For the purposes of calculation, it is assumed that 50% of cows undergo castration.

Dehorning (removal of horns) or disbudding (removal of buds that grow horns) is a common procedure on cows which is done to reduce the risk of injury to handlers or other animals, save space in transport, and reduce the occurrence of damaged carcasses. Cozzi et al. (2015) estimate that 47% of beef cows in Europe undergo dehorning. Lacking good information specific to dehorning/disbudding, I will use the same pain profile as for castration.

An estimated 44.8% of farmed cattle go through hot iron branding, which remains the most common method of ID for cattle. Identification is important to the industry for tracking disease, establishing ownership, and for other purposes. Less painful methods of identification are an ongoing area of research. I will use mostly the same pain profile for branding as for castration, but the window of hurtful pain is extended to four weeks, in accordance with an Oklahoma State University article, which gives the time as 2-8 weeks.

Disease

For this section, I will focus on two common ailments that beef cattle face.

Bovine respiratory disease (BRD) refers to a complex of infectious respiratory illnesses that affect cattle. Like the common cold in humans, BRD is not caused by a single pathogen. According to the U.S. Department of Agriculture, 14% of cattle suffer from BRD at some point. The Merck Veterinary Manual states that a BRD infection typically lasts for five days (100 waking hours), though potentially longer in the case of a pneumonia infection. For these calculations, I will assume 100 hours of infection, which are divided into 60 hours of disabling pain and 40 hours of hurtful pain. The 60/40 numbers are determined by best guess.

Lameness, which includes foot rot, joint infections, and injuries, is the other most prevalent disease. According to Davis-Unger et al. (2019), 13.9% of farmed cattle experience a disease, of which 32.3% are lameness (note that this source also assesses BRD and disagrees somewhat with the source in the previous paragraph about its prevalence, so I don’t have super high confidence in the precision of these numbers). Biggs et al. (2016) specify that improvement should occur within 3-4 days of treatment. Based on that, a one week (140 waking hours) illness seems reasonable, though there is a long tail for the cases of deeper infection. I’ll calculate with the same 60/40 Disabling / Hurtful split as above.

Total

Adding up everything above, I find 0.14 hours of excruciating pain, 110 hours of disabling pain, 898 hours of hurtful pain, and 347 hours of annoying pain. Note that I am assuming no pain while the animal is asleep.

Aggregating to a Single Value

To aggregate these pain levels into a single cost, I use estimates from Schuck-Paim, Alonso, and Hamilton (2024) and take the geometric means of their high and low values. I am using a value of 1.0 for excruciating pain, 0.01 for disabling pain, about 0.0004 for hurtful pain, and 3 * 10^(-5) for annoying pain. These numbers mean that excruciating pain for an animal is equivalent to the worst state for a human, with all other pain levels far less.

A broiler chicken’s typical lifespan in the US is 47 days, and this yields a welfare level of around 0.00059. A typical beef cow’s lifespan is 18-24 months, and I use 21 months for these calculations. Putting it all together, I find a welfare level of around 0.00011. I should note that Schuck-Paim, Alonso, and Hamilton (2024) caution against aggregating the four pain levels in this manner, as short bursts of intense agony and long periods of chronic pain may be incomparable, but this aggregation is necessary to produce a single cost.

Since my welfare calculations are anchored to a human quality-adjusted life year, it is implicitly assumed that the worst state for a human is as bad as the average living state is good. I strongly suspect that this assumption greatly underestimates the harm of excruciating pain, and this bias could be corrected for using a value higher than 1.0. However, I am not aware of any justification in the literature for doing so. If the value for excruciating pain were to be increased, I suspect that the other values should be increases as well. This would have the effect of increasing my estimated welfare cost, perhaps by an order of magnitude.

Welfare Range and Probability of Consciousness

The welfare range captures the intensity of an animal’s experience, relative to a human. A human has a welfare range of 1.0, while a nonsentient entity has a welfare range of 0.0. Sentience probability is the probability that the animal is sentient in the sense of possessing moral weight. Again, a human would have a value of 1.0–certainty of moral weight–while an inanimate object would have a sentience probability of 0.0.

I adopt values from Duffy et al. (2024), which combines welfare range and sentience probability into a single parameter known as p(sentience)-adjusted welfare ranges. The values most relevant for this analysis are as follows. Note that neither Duffy et al. (2024) nor the Moral Weight Project explicitly report these parameters for cattle, and so I use the values for pigs, as the most closely related large-brained mammal, as a stand-in.

The main factor accounting for the wide range between the 5th and 95th percentiles is that the model of Duffy et al. (2024) is a composite of three models. One is an equality model, which holds that all sentient beings should be given equal weight. One is a neurophysical model, which is based on metrics related to neuron count. One is an additive model, which is based on the number of traits of consciousness, such as number sense and the ability to imagine the past and future, that the animal possesses. The three models yield very different welfare ranges, and those differences are reflected in the trimodal distribution of the final result for most species.

Putting the Numbers Together

A human quality-adjusted life year is valued at $591,000, based on the central estimate here. Their low and high estimates are, respectively, $276,000 and $899,000 (constant 2023 dollars), though my calculations will stick with the central value.

A USDA fact sheet finds that 9.16 billion broilers were produced in 2023. Since their average lifespan is 47 days, that works out to 1.2 billion broiler chicken life-years for the calendar year 2023. The USDA also estimates 32.2 million head of cattle were slaughtered in 2024, which for a 21 month average lifespan works out to 56 million life-years for the calendar year 2024.

Total chicken consumption in the US in 2023 was 46 billion pounds, while beef consumption was 28.7 billion pounds. My calculations account for US domestic consumption and slaughter and thus are not corrected for imports or exports.

For chicken, I find that the total welfare cost is $12.4 billion to $369 billion (50th percentile: $199 billion), or $0.27 to $8.02 per pound (50th percentile: $4.32). The high value is well above the retail price.

For beef, I find that the total welfare cost is $198 million to $3.6 billion (50th percentile: $1.9 billion), or about 0.7¢ to 12.4¢ per pound (50th percentile: 6.5¢).

My values are highly sensitive to how we choose to weight the most severe excruciating pain relative to human existence. Some alternative values are as follows.

The main reason that the welfare cost with chicken is so much higher than for beef is that a single cow produces over 100 times more meat than a single chicken. A secondary reason is that beef cows generally live more pleasant lives, by my estimation methodology, than do broilers. Partially offsetting these considerations is that a cow is assigned more moral weight than a chicken per unit time, and also that a cow lives longer, which gives it time to experience more misery overall.

Environmental Costs

There are several major environmental impacts associated with meat production, and for the purposes of this analysis, I will focus on climate change.

My approach to estimating the external costs of meat consumption from climate change  is straightforward. First I estimate the social cost of carbon, which is the damage done to society from the emission of a ton of carbon dioxide, using off-the-shelf estimates. I then apply the social cost of carbon to estimated emission intensities for meat.

Based on an analysis for Resources for the Future, a reasonable, rough estimate of the social cost of carbon is $100 to $200 per ton of carbon dioxide. The most important factor that affects the SCC is the discount rate, or the extent to which one devalue the future relative to the present. With a lower discount rate, I place more value on damages from emissions in the future and thus estimate a higher SCC. The metareview of Tol (2025) finds that estimates of the SCC have increased in recent years and the central estimate is $190 to $245 per ton, but he notes that this value might be inflated by publication bias.

To estimate the emissions intensity of food, I draw upon Poore and Nemecek (2018) via Our World in Data. They estimate emissions intensities of 99.48 kilograms of CO₂ (equivalent) per kilogram of beef and 9.87 kg CO₂ per kg poultry. These numbers translate into greenhouse gas costs of $4.51 to $9.02 per pound of beef. For poultry, the values are 45¢ to 90¢ per pound.

Beef shows higher impacts than chicken on other environmental metrics that I will not attempt to monetize here. By weight, beef requires more than three times as much water as chicken (Mekonnen and Hoekstra 2011). Poore and Nemecek (2018) also estimate, via Our World in Data, land use for types of meat and find that beef from a beef herd requires more than 20 times the land as poultry for a given amount of protein. Pierer et al. (2014) find that, for a given amount of protein, beef is responsible for twice the nitrogen runoff as poultry.

Other Costs and Benefits

There are several other costs and benefits associated with meat consumption, and here I will briefly touch on two of them: health and consumer welfare. It is questionable whether these costs and benefits should be reflected in a meat tax, since at least some of them are internal to the consumer and thus already reflected in the price.

Health

Springmann et al. (2018) propose a tax on red meat and processed meat on the grounds that these products are harmful to human health. The optimum tax rate that they estimate varies widely by country, and the world average is about 4% for unprocessed red meat and 25% for processed meat, though the optimum rate for the US specifically is likely higher due to higher than average meat consumption and high health care costs. They estimate that the optimum tax would prevent 222,000 premature deaths worldwide per year and save $41 billion of health care costs.

It is well-known that red meat (e.g. beef) consumption has a more negative effect on human health than white meat (e.g. chicken) consumption. Zhang et al. (2025) discuss how red meat increases the risk of colorectal cancer, cardiovascular disease, and type 2 diabetes due to higher iron content and saturated fat. By contrast, they find health benefits to moderate intakes of white meat.

A meat tax that is motivated by human health would, as Springmann et al. (2018) propose, have higher rates for red meat than for white meat, thus shifting consumption toward white meat such as seafood and poultry. This would in fact benefit human health and mitigate environmental problems such as climate change, water consumption, land use, and nitrogen runoff. However, it would worsen the welfare of farmed animals. At any rate, much of the health cost of red meat is borne by the consumer and should, at least in theory, already be incorporated into the consumer’s decisions. Only the portion of health care costs that are supplied by third parties such as governments should be regarded as genuinely external costs.

Consumer Welfare

The biggest cost to a meat tax comes in the form of consumer welfare. Consumers choose to buy meat, and since they would buy less meat in the face of a tax than they otherwise would–that is the point of a tax after all–this decision should be regarded as a loss of consumer welfare.

Broeks et al. (2020) analyze a meat tax, paired with a subsidy for fruits and vegetables, in the Netherlands. Their basis for the tax is health and environmental benefits, and they find that the tax would increase social welfare overall. This is to be expected, as under the theory of Pigouvian taxation, internalizing the external costs of some form of consumption should be welfare-enhancing. Consumer preference, though real, is not a good argument against a tax in the presence of external costs.

Nevertheless, I do suspect that static models, such as that of Broeks et al. (2020), overstate the consumer welfare cost of a tax. The presence of a durable tax will serve as a market signal that will accelerate the development of suitable alternative proteins, thereby lessening the adjustment cost for consumers.

A Note on Positive Animal Welfare

It is theoretically possible that, perhaps with measures to improve animal welfare, animal agriculture will have net positive welfare for that species. Under Rethink Priorities’ model, especially the assumption of valence symmetry, it would seem to follow that farming should be subsidized in this case. I suspect that most readers would find this to be a surprising and undesired conclusion.

This argument is a form of Derek Parfit’s Repugnant Conclusion. Under totalist population ethics, which asserts that moral value is defined as the sum of welfare of all sentient beings, bringing new life in the world with net positive welfare is a morally good thing. That would apply to farmed animals as well.

There are several arguments against subsidizing the farming of happy animals within this framework. First, as Bruers (2026) points out, this case for a meat tax is not based on any particular population ethics theory, and in particular one does not need to accept totalism in order to accept this welfare argument for meat taxation. Second, given how much uncertainty there is in my calculations,  it would be very difficult to establish with confidence that a particular farmed animal has net positive welfare, and the uncertainty may be inherent based on philosophical considerations. Therefore, the precautionary principle may suggest that it would be better to avoid subsidizing meat.

An Argument Against a Purely Innovation-Centric Approach

Much advocacy for farmed animal welfare is centered on developing alternative proteins, such as cultured meat or plant-based substitutes, that serve as a drop-in replacement for animal meat. To be clear, I strongly support this approach, but I do not believe that development in alternative proteins alone will be sufficient to significantly reduce the quantity of factory farming.

For Rethink Priorities, Peacock (2023) investigates the PTC—Price, Taste, Convenience—Hypothesis, which says that if an alternative protein matches or beats conventional meat products on price, taste, and convenience, then consumers will adopt the alternative on a large scale. The paper finds the hypothesis to be false and that social and psychological factors also play a major role in dietary decisions. One especially interesting piece of evidence in that paper is a field experiment described by Malan et al. (2022), in which Impossible plant-based meat was added as an option to the burrito station at the dining halls of the University of California - Los Angeles, along with veggies and steak. Here, the Impossible option is posited to satisfy PTC.

The study finds that, before the intervention, the cafeteria served 86% steak burritos and 15% veggie burritos (percentages don’t add to 100% due to rounding). After the interventions, the figures were 67% steak, 7% veggie, and 26% Impossible. However, Peacock (2023) argues that it is necessary to look across the system, such as students going to other cafeterias to get their steak fix. He argues that the actual reduction of beef entree selection was only 0.3%. I do think that Peacock (2023) downplays the significance of the alternative protein option, but it is a good illustration of the limitations of the PTC hypothesis.

The Rebound Effect

My skepticism about the ability of alternative proteins to displace animal meat is most informed by the experience in other sectors. There is a widespread phenomenon in resource economics known as the rebound effect, which states that either efficiency or alternative supply will not decrease consumption of the primary as much as expected, or perhaps not even at all.

The rebound effect states that when the efficiency of a resource use improves, the actual reduction of the use of that resource is less than what would be expected from straight application of the efficiency. For example, suppose a new engine is invented that improves the efficiency of cars by 5% and changes nothing else. If all behavior remains the same, we would expect a 5% fuel savings among owners of these cars. However, due to the savings, owners of the new cars might drive more or drive faster, which is less fuel efficient, and so the actual fuel savings might be, say, 3%. Then we say that the rebound is 40%, since 40% of the expected savings went into increased consumption. I recommend Saunders (2009) for a good introduction to the topic.

Estimates of the magnitude of the rebound effect vary widely, and they depend greatly on the nature of the energy efficiency. Sorrell, Dimitropoulos, and Sommerville (2009) estimate that the direct rebound effect for household services in the Organisation for Economic Co-operation and Development is generally less than 30%. However, this is only the direct rebound effect, which refers to the rebound in which improved energy efficiency results in increased consumption of the same products. There is also indirect rebound. In the above example, suppose that consumers with more efficient cars also spend some of their newfound wealth on other products, such as airline tickets, so that actual fuel savings is 1%, taking into account direct and indirect rebounds. Now the rebound is 80% rather than 40%.

To make matters even more complicated, there are additionally economy-wide rebounds. The improved technology and fuel savings will foster economic growth, and there is a well-known correlation between energy consumption and GDP. Again, suppose in our example that the more efficient engines cause another 2% of the car’s erstwhile fuel consumption to be spent by other people and in other sectors of the economy as a result of economic growth. Now actual fuel consumption increased by 1%, rather than there being savings, and the rebound is 120%. When rebound exceeds 100%, we call this situation backfire or Jevons Paradox, an idea developed by Jevons (1865) in the context of energy efficiency in steam engines in Britain. Kulmer and Seebauer (2019) estimate economy-wide rebounds of 65% for household energy efficiency improvements but caution wide uncertainty around those numbers.

Rebound effects are also observed with new supply, in which case they are often called induced or latent demand. Following work such as Oliver, Moreno-Cruz, and Gillingham (2025), I use the term “rebound effect” to refer to increased consumption that results from alternative supply, in addition to pure efficiency. “Incomplete displacement” is another term for this phenomenon. York (2012) finds that a joule of non-fossil fuel energy, when added to the energy supply, displaces less than a quarter of a joule of fossil energy. Equivalently, the displacement efficiency of non-fossil energy is at most 25%. This would be analogous to saying that the rebound for non-fossil energy is at least 75%. He further finds that for electricity specifically, the displacement efficiency of non-fossil energy is less than 10%. Qiu, Kahn, and Xing (2019) find that residential solar panel adoption in Phoenix, Arizona has a direct rebound of 18%, which is probably much less than the true rebound when indirect and economy-wide effects are considered.

More recent work has been a bit more optimistic about displacement. Karlilar Pata and Balcilar (2025) find, examining panel data of OECD countries from 2000 to 2020, that 1.15 joules of renewable energy displace a joule of fossil energy, a displacement efficiency of a much higher 87%.

Rebounds are observed outside of energy as well. Ewers et al. (2009) examine the land-sparing potential of improved agricultural yields; the idea is that when yields improve, or farmers get more crop out of an area of land, less land is needed to feed the world, leaving more for nature. However, the paper finds that a portion of the expected savings is used by switching to lower-yield non-staple crops.

Li and Zhao (2018) find that backfire, or a rebound greater than 100%, occurs with the introduction of Low Energy Precise Application irrigation in the presence of high water rights allocation in Kansas. Loch and Adamson (2015) also find rebound effects for water efficiency, this time in the Murray-Darling Basin in Australia.

Pfaff and Sartorius (2015) find rebounds for raw material efficiency that average a modest 3.8%. Lifset and Eckelman (2013) find rebounds for efficiency in the use of primary metals. Zink and Geyer (2017) find rebound effects with material efficiency specifically from recycling and remanufacturing.

Perhaps the most relevant example for alternative proteins is our experience with aquaculture. People of my age grew up with the idea that, when we reach old age, many kinds of fish would no longer be commercially available due to overfishing.

In part due to fears about overfishing, commercial aquaculture ramped up production rapidly. The crossover point occurred in 2013, where more seafood came from aquaculture than from wild catch. But despite the rapid increase in aquaculture production, capture fisheries have been flat for 30 years, rather than declining. Now, this doesn’t prove that aquaculture had little effect on wild catch; it could be the case that wild catch would have continued to grow past the 1990s if aquaculture never took off commercially. However, Longo et al. (2019) confirm what the data appears to show: eight of nine models they constructed to examine panel data of seafood production by country over time did not find a statistically significant impact of aquaculture production on wild catch.

The situation regarding seafood does differ in one important respect from terrestrial animal meat. The stagnation in wild catch is driven by biophysical limits, whereas there is no obvious limit on the horizon to the amount of animal agriculture that can occur. However, the fact that wild catch fishing has not retreated from this limit, despite 30 years of rapid aquaculture growth, is a troubling sign for the ability of alternative proteins to displace animal agriculture.

Although alternative proteins are still an emerging market, some studies cited above provide some evidence that the rebound effect applies to them too. The field experiment of Malan et al. (2022) provides evidence of how plant-based meats such as Impossible displace other vegetarian options along with animal meat. Malila et al. (2024) do the same for insect farming.

A Tax Is More Effective

A meat tax, unlike the commercial availability of alternative proteins, raises the price seen directly by consumers and affects consumption decisions in ways that can be estimated with price elasticities. Luke, Tonsor, and Schroeder (2026), for instance, estimate price elasticities of beef, chicken, and pork. Depending on the type of meat and the model, they find that a 1% increase in price decreases consumption by 0.3% to 2%, albeit with some substitutions with other kinds of meat.

Furthermore, Lucas (2024) shows that Pigouvian taxes can alter preferences, not merely change behavior in the face of static preferences, in ways that a price rise for other reasons will not. A meat tax would send a powerful signal of social pressure against meat consumption, sending a signal that transcends the direct price effect.

Design of a Tax

Meat consumption entails tradeoffs between animal welfare concerns and other environmental impacts, such as greenhouse gas emissions. Furthermore, these quantities, especially welfare, are very difficult to monetize with any precision and in a way that will gain widespread agreement. Therefore, I propose a flat tax rate on all forms of meat, perhaps $1.00 per pound, and a lesser tax rate on non-meat animal products, such as eggs and cheese, of $0.50 per pound. Such a design does not claim to be a Pigouvian tax that sets the rate at a value equal to externalized costs. The rate is, I believe, lower than most reasonable estimates of the external welfare and environmental costs of meat and animal products, but high enough to have a substantial effect on consumer behavior and provide a market signal for accelerated development of alternative proteins.

Note that, although the costs are highly uncertain, the intervals for the external costs for chicken and for beef overlap despite being dominated by very different components.

For ease of administration, I propose that the tax be levied at the farmgate. For imported products, I propose the tax be assessed based on an estimate of the meat and animal product content of the product, with a deduction for any meat tax that might have been paid to another jurisdiction, similar to how the European Carbon Border Adjustment Mechanism works.

Conclusion

It is clear that the technology currently exists to enjoy a healthy—perhaps healthier if done right, according to Key, Papier, and Tong (2021)—and tasty vegan diet at a reasonable price. However, only 4% of Americans identify as vegetarian and 1% as vegan, according to a Gallup poll, and so finding good options, especially when out and about, is still challenging. There are also some negative stereotypes. These are likely more formidable challenges to decreasing meat consumption than the technological development of alternative proteins.

Traditionally, meat taxation and efforts to reduce meat consumption have focused on red meat, such as beef, for its negative impacts on human health and disproportionate environmental impact. However, when examining the problem from an animal welfare perspective, the focus should be on meat from smaller animals such as chickens. To balance these concerns, and in recognition of the enormous uncertainty that goes into the calculations, I propose a flat tax on meat with a lesser flat tax on non-meat animal products.

I expect that a meat tax will face formidable political obstacles, and building the political will for a meat tax is a multi-year, if not a multi-generational, project. However, it is, in my view, the only way in which the evils of factory farming will be reduced in a major way.

References

1) Norman, T. The Downside Of Meat Taxes. Faunalytics. January 2023.

2) Welfare Footprint Institute. Broilers. Welfare Footprint Institute.

3) Fischer, B. An Introduction to the Moral Weight Project. Rethink Priorities. 2022.

4) Schuck-Paim, C., Alonso, W.J., Freitas, A.P., Oliveira, C.P., Fonseca, V.F.C., Borges, T.D. The Welfare Impact of Heat Stress in South American Beef Cattle and the Cost-Effectiveness of Shade Provision. Animals 16(2), 231. January 2026.

5) Scheffler, T.L. Connecting Heat Tolerance and Tenderness in Bos indicus Influenced Cattle. Animals 12(3), 220. January 2022.

6) Wikipedia contributors. Big Die-Up. Wikipedia.

7) Cox, B., Gasparrini, A., Catry, B., Delcloo, A., Bijnens, E., Vangronsveld, J., Nawrot, T.S. Mortality Related to Cold and Heat. What Do We Learn from Dairy Cattle? Environmental Research 149, pp. 231-238. August 2016.

8) Kansas State University Beef Cattle Institute. Castration. Kansas State University. September 2024.

9) Meléndez, D., Marti, S., Gellatly, D., Moya, D., Schwartzkopf-Genswein, K.S. Can We Mitigate Pain Associated with Castration in Beef Cattle at Early (<6 Months of Age) Ages? A Review from a Canadian Perspective. Canadian Journal of Animal Science 105, pp. 1-31. July 2025.

10) Amaral-Phillips, D. A. Are Your Cows Getting Enough Sleep? University of Kentucky Cooperative Extension Service. 2024.

11) Wikipedia contributors. Livestock Dehorning. Wikipedia.

12) Cozzi, G., Gottardo, F., Brscic, M., Contiero, B., Irrgang, N., Knierim, U., Pentelescu, O., Windig, J.J., Mirabito, L., Kling Eveillard, F., Dockes, A.C., Veissier, I., Velarde, A., Fuentes, C., Dalmau, A., Winckler, C. Dehorning of Cattle in the EU Member States: A Quantitative Survey of the Current Practices. Livestock Science 179, pp. 4-11. September 2015.

13) United States Department of Agriculture, Animal and Plant Health Inspection Service. Cattle Identification Practices on U.S. Beef Cow-Calf Operations. National Animal Health Monitoring System, Beef 2007-08.

14) Congressional Research Service. Animal Identification and Traceability: Overview and Issues. CRS Report RS22653.

15) American Veterinary Medical Association. Welfare Implications of Hot-Iron Branding. American Veterinary Medical Association Literature Review. April 2011.

16) Oklahoma State University. Branding Cattle: Animal Welfare Considerations, Legal Implications and Alternatives. Oklahoma State University News. 2020.

17) Wikipedia contributors. Bovine Respiratory Disease. Wikipedia.

18) Snowder, G.D., Van Vleck, L.D., Cundiff, L.V., Bennett, G.L. Bovine Respiratory Disease in Feedlot Cattle: Environmental, Genetic, and Economic Factors. Journal of Animal Science 84(8), pp. 1999-2008. August 2006.

19) Merck Veterinary Manual. Viral Infections Associated with Bovine Respiratory Disease Complex in Cattle. Updated November 2024.

20) Davis-Unger, J., Schwartzkopf-Genswein, K.S.G., Pajor, E.A., Hendrick, S., Marti, S., Dorin, C., Orsel, K. Prevalence and Lameness-Associated Risk Factors in Alberta Feedlot Cattle. Translational Animal Science 3(2), pp. 595-606. April 2019.

21) Oklahoma State University Extension. Foot Rot in Cattle. Oklahoma State University Extension Fact Sheet. October 2016.

22) Schuck-Paim, C., Alonso, W.J., Hamilton, C. Short Agony or Long Ache: Comparing Sources of Suffering That Differ in Intensity and Duration. Effective Altruism Forum. February 2024.

23) North Dakota Farm Bureau. How Long Do Cows Live? North Dakota Farm Bureau. September 2023.

24) Schuck-Paim, C., Alonso, W.J., Hamilton, C. Short Agony or Long Ache? Welfare Footprint Institute. February 2024.

25) Duffy, L., Fischer, B., Barrett, M., Gaffney, L., Lavery, J.M., Miller, R., Schiestl, M., Schnell, A., Shriver, A., Timmerman, T., Trevarthen, A. Some Tentative Welfare Range Estimates. In: Fischer, B. (ed.), Weighing Animal Welfare: Comparing Well-Being Across Species. New York: Oxford University Press. October 2024.

26) United States Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Guidelines for Regulatory Impact Analysis: Standard Values. U.S. Department of Health and Human Services. January 2024.

27) United States Department of Agriculture, National Agricultural Statistics Service. Poultry - Production and Value 2023 Summary. April 2024.

28) United States Department of Agriculture, Economic Research Service. Cattle and Beef: Statistics and Information. U.S. Department of Agriculture.

29) United States Department of Agriculture, Economic Research Service. Poultry and Eggs: Sector at a Glance. U.S. Department of Agriculture.

30) United States Department of Agriculture, Economic Research Service. Meat Price Spreads. U.S. Department of Agriculture.

31) Rennert, K., Kingdom, C., Prest, B. C. Social Cost of Carbon 101. Resources for the Future. Updated March 2025.

32) Tol, R.S.J. Trends and Biases in the Social Cost of Carbon. Annals of the New York Academy of Sciences 1548(1), pp. 248-259. June 2025.

33) Poore, J., Nemecek, T. Reducing Food's Environmental Impacts Through Producers and Consumers. Science 360(6392), pp. 987-992. June 2018.

34) Ritchie, H., Rosado, P., Roser, M. Food Emissions Across the Supply Chain. Our World in Data.

35) Mekonnen, M.M., Hoekstra, A.Y. The Green, Blue and Grey Water Footprint of Farm Animals and Animal Products. Value of Water Research Report Series No. 48. Delft: UNESCO-IHE Institute for Water Education. December 2010.

36) Ritchie, H., Rosado, P., Roser, M. Land Use. Our World in Data.

37) Pierer, M., Winiwarter, W., Leach, A.M., Galloway, J.N. The Nitrogen Footprint of Food Products and General Consumption Patterns in Austria. Food Policy 49, pp. 128-136. December 2014.

38) Springmann, M., Mason-D'Croz, D., Robinson, S., Wiebe, K., Godfray, H.C.J., Rayner, M., Scarborough, P. Health-Motivated Taxes on Red and Processed Meat: A Modelling Study on Optimal Tax Levels and Associated Health Impacts. PLOS ONE 13(11), e0204139. November 2018.

39) Zhang, H., Gan, M., Wu, Q., Dan, H., Chen, L., Niu, L., Zhao, Y., Wang, Y., Zhu, L, Shen, L. Red meat vs. white meat: A comparative analysis of histological characteristics, nutritional profiles, and health implications. Food Chemistry: X 31. 2025.

40) Broeks, M.J., Biesbroek, S., Over, E.A.B., van Gils, P.F., Toxopeus, I., Beukers, M.H., Temme, E.H.M. A Social Cost-Benefit Analysis of Meat Taxation and a Fruit and Vegetables Subsidy for a Healthy and Sustainable Food Consumption in the Netherlands. BMC Public Health 20, 643. May 2020.

41) Bruers, S. Determining an Optimal Welfare Levy. International Review of Resource Economics 20(1), pp. 1-24. June 2026.

42) Peacock, J. Price, Taste, and Convenience: Competitive Plant-Based Meat Would Not Currently Replace Meat. Rethink Priorities. 2023.

43) Malan, H., Bartolotto, C., Wilcots, C., Angelis, P., Ferrone, A., Wible, C., Westbrook, E., Fabris, E., Wang, M.C., Slusser, W., Jay, J.A. Increasing the Selection of Low-Carbon-Footprint Entrées through the Addition of New Menu Items and a Social Marketing Campaign in University Dining. Journal of the Association for Consumer Research 7(4), pp. 461-470. 2022.

44) Saunders, H.D. Theoretical Foundations of the Rebound Effect. In: Evans, J., Hunt, L.C. (eds.), International Handbook on the Economics of Energy, Chapter 8. Cheltenham: Edward Elgar Publishing. 2009.

45) Sorrell, S., Dimitropoulos, J., Sommerville, M. Empirical Estimates of the Direct Rebound Effect: A Review. Energy Policy 37(4), pp. 1356-1371. April 2009.

46) Ritchie, H., Rosado, P., Roser, M. Energy Use per Capita vs. GDP per Capita. Our World in Data.

47) Jevons, W.S. The Coal Question: An Inquiry Concerning the Progress of the Nation, and the Probable Exhaustion of Our Coal-Mines. London: Macmillan and Co. 1865.

48) Kulmer, V., Seebauer, S. How Robust Are Estimates of the Rebound Effect of Energy Efficiency Improvements? A Sensitivity Analysis of Consumer Heterogeneity and Elasticities. Energy Policy 132, pp. 1-14. September 2019.

49) Oliver, M.E., Moreno-Cruz, J., Gillingham, K.T. Microeconomics of the Solar Rebound Under Net Metering. Journal of the Association of Environmental and Resource Economists 12(5). September 2025.

50) York, R. Do Alternative Energy Sources Displace Fossil Fuels? Nature Climate Change 2, pp. 441-443. March 2012.

51) Qiu, Y., Kahn, M.E., Xing, B. Quantifying the Rebound Effects of Residential Solar Panel Adoption. Journal of Environmental Economics and Management 96, pp. 310-341. July 2019.

52) Karlilar Pata, S., Balcilar, M. Decarbonizing energy: Evaluating fossil fuel displacement by renewables in OECD countries. Environmental Science and Pollution Research. 2025.

53) Ewers, R.M., Scharlemann, J.P.W., Balmford, A., Green, R.E. Do Increases in Agricultural Yield Spare Land for Nature? Global Change Biology 15(7), pp. 1716-1726. July 2009.

54) Li, J., Zhao, J. Rebound Effects of New Irrigation Technologies: The Role of Water Rights. American Journal of Agricultural Economics 100(3), pp. 786-808. April 2018.

55) Loch, A., Adamson, D. Drought and the Rebound Effect: A Murray-Darling Basin Example. Natural Hazards 79(3), pp. 1429-1449. April 2015.

56) Pfaff, M., Sartorius, C. Economy-Wide Rebound Effects for Non-Energetic Raw Materials. Ecological Economics 118, pp. 132-139. October 2015.

57) Lifset, R., Eckelman, M. Material Efficiency in a Multi-Material World. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 371(1986), 20120002. March 2013.

58) Zink, T., Geyer, R. Circular Economy Rebound. Journal of Industrial Ecology 21(3), pp. 593-602. June 2017.

59) Longo, S.B., Clark, B., York, R., Jorgenson, A.K. Aquaculture and the Displacement of Fisheries Captures. Conservation Biology 33(4), pp. 832-841. August 2019.

60) National Oceanic and Atmospheric Administration Fisheries. Ending Overfishing Through Annual Catch Limits. NOAA Fisheries.

61) Malila, Y., Owolabi, I.O., Chotanaphuti, T., Sakdibhornssup, N., Elliott, C.T., Visessanguan, W., Karoonuthaisiri, N., Petchkongkaew, A. Current challenges of alternative proteins as future foods. npj Science of Food 8: 53. 2024.

62) Luke, J.R., Tonsor, G.T., Schroeder, T.C. U.S. Meat Demand Elasticity Estimates: Using Publicly Available Data Versus Scanner Data. Agricultural and Resource Economics Review 55(1). 2026.

63) Lucas, G.M. Shaping Preferences with Pigouvian Taxes. Texas A&M University School of Law Faculty Scholarship. 2024.

64) European Commission, Directorate-General for Taxation and Customs Union. Carbon Border Adjustment Mechanism. European Commission.

65) Key, T.J., Papier, K., Tong, T.Y.N. Plant-Based Diets and Long-Term Health: Findings from the EPIC-Oxford Study. Proceedings of the Nutrition Society 81(2), pp. 190-198. May 2022.

66) Gallup. In U.S., 4% Identify as Vegetarian, 1% as Vegan. Gallup. August 2023.