Abstract

[AI Generated summary of the evaluation.] Bruers (2023) proposes a novel survey-based approach to quantifying the external welfare cost of farmed animals by eliciting individuals’ willingness to pay or accept experiencing an animal’s life in a hypothetical dream scenario. In contrast with Kuruc and McFadden (2023) and Espinosa (2024), who rely on parametric assumptions regarding utility or on animal welfare scoring frameworks, Bruers’ method avoids explicit functional forms and interspecies welfare comparisons, though it risks higher anthropocentrism and uncertainty. Survey results from Flemish-speaking Belgians suggest that most perceive intensively farmed animal lives as not worth living and would demand compensation to endure them, implying a recognized negative net welfare for conventionally farmed animals. Despite the method’s innovative setup and potential policy relevance, the author highlights substantial limitations—hypothetical bias, large outliers, participants’ limited knowledge of rearing conditions, and the unusual nature of imagining another species’ life—that render the exact monetary estimates unreliable. Nonetheless, the paper underscores the challenges of valuing nonhuman welfare and enriches the ongoing discussion about practical, comprehensive ways to incorporate animal welfare costs into policy analyses. [End of AI-generated content]1

Summary Measures

We asked evaluators to give some overall assessments, in addition to ratings across a range of criteria. See the evaluation summary “metrics” . This evaluator declined to give these measures, although they report on some of these below descriptively.2

Written report

Overall assessment

This paper investigates a method to value the welfare externalities imposed on farmed animals in monetary terms. At the time the paper was written (2022), the field of animal welfare economics had only started investigating this issue, and only two other methods had been proposed. On one side, Kuruc and McFadden (cited as a working paper, ultimately published in 2023)[1] proposed estimating the externalities of human diets on animal welfare by adapting a DICE model. The authors assumed a CRRA utility function for humans and calibrated their social welfare function such that a human life below the poverty line (adjusted for purchasing power) would decrease social welfare. Monetized animal welfare was computed by assuming the same utility function for animals and the same contribution to social welfare and by arbitrarily setting the utility of animals on the CRRA utility function.

In contrast, Espinosa (cited as a working paper, ultimately published in 2024)[2] proposed a distinct parametric welfare function for animals using the Five Freedom framework. The Five Freedom Fulfilment Index consists of a welfare score based on the intensity of infringement of an animal’s freedoms that can be interpreted as a QALY-like utility function. For each of the Five Freedoms, the author defined possible levels of impairment (ranging from ‘no violation’ to ‘very severe violations’) with associated ‘violation points’ (from 0 to 4). The welfare score is defined as a linear transformation of the total number of violation points, rescaled such that the maximum score of 1 corresponds to a life without freedom violation and score of 0 corresponds to a life that is neither worthy nor unworthy of being lived. Monetized animal welfare was then computed by weighting the QALY-like utility scores of an animal by a utility potential (based on the relative number of cortical neurons compared to humans) and by multiplying it by the social monetary valuation of a human QALY (as humans have by definition a utility potential of 1).

Bruers (2023)[3] emerged as a third concomitant proposal to economically value animal welfare. The paper starts from the idea that animals, while having preferences, cannot express willingness-to-pay (WTP) or willingness-to-accept (WTA) related to changes that could affect their welfare status. Policymakers and social planners thus lack appropriate measures to estimate the welfare impact of a policy on animals (measures that we could use in benefit-cost analyses for instance). To address this problem, the author proposes to elicit humans’ WTP/WTA to experience an animal’s life (focusing on farmed animals). He argues that these outcomes reflect the external animal welfare costs, i.e., the welfare burden imposed on animals, which is distinct from the altruistic concerns that humans experience when they impact animals. In this setup, a positive WTP to avoid experiencing an animal’s life would reflect the belief that a farmed animal’s life is not worth living.

The strategy proposed by Bruers (2023)[3] significantly differs from Kuruc and McFadden (2023)[1] and Espinosa (2024)[2], with pros and cons. On the one side, Bruers (2023)[3] does not need to make parametric assumptions about the utility functions of humans or animals. In contrast, Kuruc and McFadden (2023)[1] assumed a CRRA utility function for both humans and animals. Espinosa (2024) also remained agnostic about the shape of the utility function for humans, but his original version of the Five Freedom Fulfilment Index imposed significant structure (linearity assumption). In addition, Bruers (2023)[3] does not need to make any assumption about the highly complex question of interspecies welfare comparison (i.e., how to compare utils across species). Kuruc and McFadden (2023)[1] remain silent on the issue (i.e., they only consider a ‘political’ weight parameter about the importance to socially give to animal welfare), while Espinosa (2024)[2] weighs utility points by utility/welfare potentials, building on the idea of Spears and Budolfson (2019)[4]. On the other side, the weaker structure imposed by Bruers (2023)[3] comes at the cost of larger risks of anthropocentrism and, thus, larger uncertainty about the estimates. Lay people might [also] lack the appropriate knowledge about the welfare of the animals and current rearing conditions in their country.

The main differences between the three approaches are summarized in the table below:

From a theoretical perspective, the author derives three types of WTP/WTA. First, he expresses the WTP for meat consumption as the amount of money (or, more precisely, a composite good) that would make him indifferent between his current situation with meat consumption and an alternative situation where he would stop eating meat but would receive this amount of money. Second, the author defines the marginal WTP for animal welfare (altruistic component) as the maximum amount of money an agent is willing to spend to increase the welfare of an animal by one unit [one utility point]. Third, the external cost of the welfare burden imposed on animals is defined as the change in an agent’s endowment that would make him indifferent between (i) experiencing only his life with the reduced/augmented endowment and (ii) experiencing both his life (with the original endowment) and the animal’s life.

To estimate these compensating variations, the author developed a survey administered in June 2022 to a sample (N=500) of the Flemish-speaking Belgians aged 18 to 65. The sample was demographically representative for  age and gender. However, due to the exclusion criteria, the final sample shrunk to N=301 respondents. In this hypothetical survey, the respondent is asked to imagine taking a pill that would make him sleep and live the life of an animal. In this dream, he would experience the entire life of the animal, the same way the animal does (e.g., similar sensory capacities, similar cognition). When waking up, he would not remember this experience. The alternative consists of taking a pill that would induce a dreamless sleep of equal duration.

The author estimates the external cost imposed on animal welfare by eliciting the WTP/WTA for the two pills. He first asks which pill the respondent would prefer taking and then elicits either a WTP for the preferred pill or a WTA for the other pills (WTP or WTA condition, random assignment with equal probability at the respondent level). The author then reports absolute WTP/WTA for pills under various scenarios (conventional farming, intensive farming, label-certified farming, neutral life, and happiest possible life) and relative WTP/WTA (absolute WTP/WTA minus the WTP/WTA for the animal’s neutral life). The latter aims to control for the (dis)utility of experiencing an animal’s life, whatever the welfare state at stake. Next, the author attempts to approximate the WTP for marginal animal welfare improvement by eliciting the participants’ WTP to replace one kilogram of conventional meat with one kilogram of cultured meat (for three types of meat -chicken, pig, cow-, displayed in random order). The author also elicits expectations about the WTP for cultured meat of other Belgian citizens as an attempt to mitigate possible social desirability biases. 

The author finds that most of the participants give a positive WTA for conventionally farmed animals, indicating that they would need to be compensated to experience a farmed animal’s life and, thus, that the animal’s life is seen as not worth living. This result holds for chickens, pigs, and cows. Consistently, people are more and more willing to accept experiences of animal lives when welfare is larger (chicken with better life or maximum welfare scenarios). The author finds positive WTP for cultured meat, which he interprets as positive altruistic preferences towards animals, but detects no statistically significant correlations between one’s WTP for cultured meat and one’s WTA/WTP for experiencing the life of the animal of the associated cultured product.

The author recognizes several important limits with his work.3 He underlines the issues associated with data attrition due to the exclusion criteria. Indeed, about two-fifths of the sample were excluded from the analysis either because they reported inconsistent answers (e.g., larger WTP to avoid the life of the happiest animal than to avoid the lower-welfare lives), reported unrealistic WTA (i.e., more than one million Euros), or reported low confidence levels in their answers. Regarding the latter, the author underlines the large sensitivity in the results depending on whether less-confident respondents are included or excluded from the analysis. He also mentions that respondents might misperceive or misestimate the welfare status of currently farmed animals in their country. He is also concerned about the presence of large outliers in the WTPs, which might originate from hypothetical bias issues. He then concludes that this survey approach might represent a difficult tool to estimate the external costs imposed on animal welfare.

Claims, strength, and characterization of evidence

Bruers (2023)[3] proposes a neat and clever design to discuss the issue of the quantification of animal welfare. At the time of the writing, the community was (and still is) looking at the most appropriate way to express animal welfare in monetary terms. Bruers’ approach avoids some of the drawbacks of concomitant alternative methods. In Kuruc and McFadden (2023)[1] the relative position of the animals on the welfare utility function was somewhat arbitrary. While Espinosa (2024)[2] assessed animal welfare using an animal welfare function, the parametric assumptions are also important limits. Bruers (2023)[3] does not suffer from these issues: By estimating the WTP/WTA of humans to live a farmed animal’s life, the author avoids needing to impose any parametric assumptions on either the humans’ or the animals’ utility functions.

This paper significantly contributes to the discussion on animal welfare valuation by showing some limits of anthropocentric valuation including his own approach.4 The author recognizes that his method is not reliable to obtain accurate estimates of the welfare of animals. However, it makes a great contribution by underlining the numerous issues that make such approaches unreliable  (e.g., the perception of the animal’s capacity to experience the world, the perception of the animal’s farming conditions, hypothetical biases, outliers, and exclusion criteria).5

While the numbers themselves must be interpreted with caution, the paper discusses several elements that will be valuable for future research. First, it shows that most participants would like to avoid experiencing the life of currently farmed animals. This suggests that people may generally acknowledge that most of these lives are not worth living and even worth avoiding. The author concludes that his evidence aligns well with previous findings in the literature (e.g., Espinosa and Treich 2021)[5]. While the point estimates might be unreliable for policy purposes, the results indicate that there could be a social consensus on the fact that reducing intensive farming would be beneficial. Second, from a more practical side, the author underlines how difficult it is to make a definitive conclusion about what society or the representative citizen thinks about this. Indeed, the author finds large differences between the mean and median estimates of WTA/WTP to experience an animal’s life, which reflects the large heterogeneity in views on the topic (i.e., some people have very negative perceptions of the animals’ welfare),  and, possibly, the substantial heterogeneity in the comprehension of the survey (e.g., familiarity with the concepts involved or with the quantification/monetization tasks).

One major source of caution concerns the interpretation of the results related to the marginal altruistic WTP. This quantity is approximated through the WTP for the replacement of one kilogram of conventional meat supply with one kilogram of cultured meat. However, it is unclear whether the cultured-meat WTP effectively reflects altruistic preferences. For instance, consumers could expect the quality to differ between the two products, they could be reluctant to try new food (food neophobia), they could hold naturalistic views about food (i.e., a reluctance against food tech), etc. At the altruistic level, individuals could also think that cultured meat would prevent the birth of some animals, which, even if their net lifetime welfare is negative, would not be socially desirable.6 

Methods: Justification, reasonableness, validity, robustness

Within the limits that the author himself recognizes, and with the exception of the cultured meat questions, the methods are valid, reasonable, and well-justified. From a theoretical perspective, the measure of the human’s perception of the welfare of the animal is well defined: it corresponds to the change in private endowment that makes the human indifferent between living and not living the animal’s life on top of his own life. From an empirical perspective, the author estimates it by considering several challenges, including (i) that some participants might have positive or negative perceptions of the welfare of the animal (i.e., the author uses a first-stage question to elicit the preferred pill), (ii) WTA/WTP framing might alter the results, (iii) outliers could lead to extremely high average WTP/WTA, (iv) some participants might have very low confidence in this unusual exercise, and (v) there might be some intrinsic (dis)utility in living an animal’s life.

The work also contains numerous limitations, and most of them are recognized by the author himself. The main (and most standard) limitation is the hypothetical nature of the survey: participants do not face real incentives, which might lead to underestimating or overestimating the welfare burden of our diet imposed on animals. One of the most intrinsic limits concerns the difficulty for lay people (and humans in general) to project themselves into the lives of animals of other species. One can be indeed ignorant about the needs, the perception of the world, the cognition, etc. of other species, such that estimating their welfare is a daunting task. In addition, humans can also be ignorant about the rearing conditions in “conventional” farms, making the estimates even more uncertain. Ultimately, the layers of complexity (unusual survey, ignorance about other species, uncertainty about rearing conditions) add up and make the point estimates hard to rely on.

Advancing knowledge and practice

The paper brings valuable insight into the broader research question of how to get monetary valuations of animal welfare using the tools and approaches of economics.

Logic and communication

The paper is well-written and easy to understand both for academics and less-specialized readers.

Open, collaborative, replicable science

I did not find an online repository for data and code.7

Relevance to global priorities, usefulness for practitioners

Because the author recognizes that the method might not be reliable for empirical estimation, the usefulness might be limited for practitioners (other than to prevent them from replicating such approaches).

The paper has a high relevance to global priorities given the prevalence of animal factory farming nowadays.

References

[1] Kuruc, K., & McFadden, J. (2023). Monetizing the externalities of animal agriculture: insights from an inclusive welfare function. Social Choice and Welfare. https://doi.org/10.1007/s00355-023-01451-9

[2] Espinosa, R. (2024). Animals and social welfare. Social Choice and Welfare, 62(3), 465–504. https://doi.org/10.1007/s00355-023-01495-x

[3] Bruers, S. (2023). The animal welfare cost of meat: evidence from a survey of hypothetical scenarios among Belgian consumers. Journal of Environmental Economics and Policy, 12(3), 324–341. https://doi.org/10.1080/21606544.2022.2138980

[4] Budolfson, M., & Spears, D. (2019). Quantifying Animal Well-Being and Overcoming the Challenge of Interspecies Comparisons. In The Routledge Handbook of Animal Ethics. Routledge.

[5] Espinosa, R., & Treich, N. (2021). Animal welfare: antispeciesism, veganism and a “life worth living.” Social Choice and Welfare, 56(3), 531–548. https://doi.org/10.1007/s00355-020-01287-7