Counting Your Chickens

ABSTRACT

Suppose that, for reasons of animal welfare, it would be better if everyone stopped eating chicken. Does it follow that you should stop eating chicken? Proponents of the ‘inefficacy objection’ argue that, due to the scale and complexity of markets, the expected effects of your chicken purchases are negligible. So the expected effects of eating chicken do not make it wrong.

We argue that this objection does not succeed, in two steps. First, empirical data about chicken production tells us that the expected effects of consuming many chickens are not negligible. Second, this implies that the expected effect of consuming one chicken is ordinarily not negligible. Parity between your purchase and other counterfactual purchases, and uncertainty about others’ consumption behaviour, each tend to pull the expected effect of a single purchase toward the average large scale effect. While some purchases do have negligible expected effects, many do not.

KEYWORDS:

• animal ethics
• decision theory
• vegetarianism
• inefficacy

Acknowledgments

The authors are listed alphabetically. Thanks to Andrew Bacon, John Hawthorne, Christian Tarsney, the Big Decisions working group at USC, the Pascalian Risks seminar at the Global Priorities Institute, and several anonymous referees for very helpful comments and discussion.

Notes

1 A parallel debate has emerged regarding the individual contributions to climate change (Sinnott-Armstrong 2005; Kingston and Sinnott-Armstrong 2018; Broome 2019).

2 For simplicity, we will speak as if people only bought whole chickens; but of course many chickens are instead cut into parts and sold that way, or as ingredients in other foods.

3 Similar appeals are made in other contexts (Parfit 1986; Gibbard 1990; Broome 2019). Nefsky (2019) offers further references and discussion.

4 Even so, the expected effect is larger than you might think: we calculate more than $0.08$ chickens produced per chicken purchased (that is, per 20 sandwiches). (Details omitted for lack of space.) If the harm to the chicken is more than $12$ times greater than the benefit of eating twenty chicken sandwiches, then even in this case the sandwiches lose the contest. But we can bolster Chignell’s artificial example by imagining an even stronger track record: say KCF’s monthly order hasn’t changed for a hundred years, rather than ten.

5 These statistics are from the United States Department of Agriculture (USDA) National Agricultural Statistics Service and the USDA Economic Research Service.

6 Thanks to an anonymous referee for pressing this.

7 We thank a referee for forcefully pressing this.

8 This was the reaction of a referee.

9 McMullen and Halteman (2019) give a related (informal) argument based on uncertainty about production costs.

10 This calculation also depends on a parameter $b$, which we here set to $2$: that is, the expected effect of purchasing one million chickens is at most two million.

11 The uncertainty argument also depends on a subtle independence assumption: namely, that the probability of various dependency hypotheses (‘if $n$ are consumed, $k$ will be produced’) are probabilistically independent of how many chickens are actually consumed. This is not necessarily true. For example:

Newcomb’s Chicken An extremely reliable predictor has analyzed your social media history and predicted whether you will purchase a chicken. If they predict you will, they produce two. If they predict you won’t, they produce one.

Given the information that you won’t purchase a chicken, it is very likely that the predictor predicted that you wouldn’t buy a chicken, and so produced just one. But also, given this information, it is very likely that the predictor would have produced just one chicken even if, improbably, their prediction was wrong and you did purchase one after all. Thus, given this information, the conditional efficacy of your purchase is close to zero. Similar reasoning applies to the case where you do in fact make the purchase.

12 Budolfson draws an analogy to voting, claiming that the probability of casting a decisive vote in a large election is ‘nearly infinitesimal’. We are persuaded by Barnett (2020) that this is not true: in competitive elections, that probability is often greater than $1/N$, where $N$ is the total number of voters.

13 This terminology comes from supervaluationism, but the idea as we apply it here is common to many leading theories of vagueness.

14 As we noted (foonote 11), it also relies on the simplifying assumption that how many chickens are purchased is probabilistically independent of the dependency hypotheses ‘if $n$ chickens were purchased, then $k$ chickens would suffer’. Given these two assumptions, we can derive the formula for $E_k$ by unpacking these definitions: (The derivation is omitted for reasons of space.) $\begin{array}{cc}C\left(n\right)& =\sum _k\mathrm{Pr}\left(n\mathrm{chickens}\mathrm{are}\mathrm{purchased}\text{□}\to k\mathrm{chickens}\mathrm{suffer}\right)\cdot k\\ E_n& =\sum _k\mathrm{Pr}\left(\mathrm{I}\mathrm{purchase}n\mathrm{chickens}\text{□}\to k\mathrm{chickens}\mathrm{suffer}\right)\cdot k\\ & -\sum _k\mathrm{Pr}\left(\mathrm{I}\mathrm{purchase}\mathrm{no}\mathrm{chickens}\text{□}\to k\mathrm{chickens}\mathrm{suffer}\right)\cdot k.\end{array}$

Additional information

Funding

Jeffrey Sanford Russell was supported by a grant from Longview Philanthropy.