![Sample our Geography journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5937/TOC-Subject-Sample-2021-Geography.jpg"})
Abstract
Mitigation policies during the COVID-19 pandemic were justified on benefit-cost-analysis (BCA) grounds, with the value of a statistical life (VSL) measuring the benefit of saving lives. This essay argues that a BCA and VSL framework is not a value-neutral basis for pandemic policy: first because the characterization of labor-market behavior as choices merely regarding money and risk of mortality implicitly endorses those institutions that constrain such choices, as illustrated by pandemic economic conditions; and second because this characterization assumes that individual subjects have exogenous preferences, and accordingly have one “true” valuation, thus ignoring how subjectivity is constituted by capitalist institutions. Such VSL estimates in pandemic scholarship illustrate the search for a single valuation regarding elderly life, whereas the “good” of human life warrants valuation in our capacity as citizens in democratic institutions rather than as market agents. The language of overdetermination highlights the contradictory effects of the mutual constitution of institutions and the individual subject.
Acknowledgments
I would like to thank George DeMartino, who generously shared ideas after a conference presentation of an earlier version and again after reading a more recent version of this essay. I would also like to thank the referees, all of whom made invaluable comments. Additionally, I would like to thank Yahya Madra, who illustrated for me connections between the core ideas of this essay and other relevant scholarship. I am also grateful for the valuable insight provided by Martha McCluskey. Last, I would like to thank Kerry Fast for copyediting the essay.
Notes
1 See, e.g., Executive Order 12866, Federal Register, Volume 58, No. 190 (1993).
2 See, e.g., Mueller v. Reich, 54 F.3d 438, 442 (7th Cir. 1995); Natural Resources Defense Council, Inc. v. EPA, 822 F.2d 104, 111 (D.C. Cir. 1987).
3 Michigan v. E.P.A. 576 U.S. 743 (2015).
4 Of course, a cost can be thought of simply as the loss of a benefit, in which case it is no less salient. While costs are commonly thought in terms of explicit monetary outlays, economists generally hold that they should be thought in terms of opportunity costs. But, to be theoretically consistent, this should include subjective dollar valuations placed on any foregone opportunity, even if not strictly “economic” in character (including social, educational, and cultural opportunities, etc.). Thus, the notion of a “cost” should be no less tied to the ontology of subjective valuation than that of a “benefit” and therefore is no less pertinent to the issues raised in this essay. In practice, however, this is not a move generally embraced. For example, in an essay by Robinson, Sullivan, and Shogren (Citation2020; discussed at greater length below), the “costs” in the varying studies they compared were in terms of lost GDP—even though this is surely not the only way to imagine the costs of mitigation measures in a pandemic.
5 It should be noted, however, that the very notion of government “intervention” in the market has been critiqued on the grounds that the state is itself constitutive of the market (see, e.g., Pasquale et al. Citation2019).
6 See, e.g., Mishan (Citation1973, 11–12) who writes that when we “depart from the ideal economic setting” of the perfectly competitive market, BCA is used in policy to “simulate the effects of an ideal price setting.”
7 If each individual separately would pay $1,000 to avoid a 1 in 10,000 risk of death, then in the aggregate individuals would pay $10,000,000 to avoid one death.
8 The RAND Corporation was set up as a nonprofit in 1948, initially funded mainly by the U.S. Air Force. The name comes from an acronym for “Research and Development” (Hounshell Citation1997).
9 RAND’s role in the history of the question of how to price lives is set out in detail by Banzhaf (Citation2014).
10 Of course, unconsidered by Schelling is whether political institutions could themselves be democratic and whether the very act of engaging in democratically constituted political institutions might change the otherwise purported exogenous preferences to which Schelling wants to defer. This issue is taken up in sections 2.2 and 2.4 below.
11 To be sure, labor markets are not the only source of information used for VSL purposes. For example, economists and policymakers sometimes use “contingent valuation” (CV) studies in which individuals are asked questions about the dollar value they place on mortality risk. However, CV is sometimes criticized on the basis that “talk is cheap” (see, e.g., Boudreaux, Meiners, and Zywicki Citation1999) and therefore only actual behavior in the market, rather than mere survey responses, can reveal true preferences. Hence, while CV and other “stated preference” methods are used, labor-market data forms the primary basis for determining the value of a statistical life (see, e.g., U.S. Consumer Product Safety Commission Citation2023, II.B).
12 To be sure, bargaining power in the labor market depends not only on the unemployment rate but also on other sources of material support, including government aid. In spring 2020, U.S. government assistance increased in various forms, including the extension and supplementation of unemployment insurance, stimulus payments to households, and relaxation of the work requirement for eligible families accessing food assistance (Hunter and McGrath Citation2020; Parrott et al. Citation2020). But unemployment insurance is not available for those who voluntarily quit (whether for fear of catching COVID or otherwise), and the remaining assistance did little to cushion the blow of the pandemic for the poorest households. For example, according to Census Bureau survey research, approximately 24 million households reported insufficient food in May 2020—nearly three times that of the highest response in 2019 (Keith-Jennings, Nchacko, and Llobrera Citation2021).
13 Since Dorman’s (Citation2009) observation, some researchers have departed from the assumption of perfectly competitive markets. The results accord with intuition, finding that an increased risk of unemployment decreases compensating wage differentials for mortality risk (see, e.g., Bender and Mridha Citation2011; Guo and Hammitt Citation2009; Leeahtam, Leurcharusmee, and Jatukannyaprateep Citation2014).
14 This virtue is, however, illusory. It dissolves under further investigation, as discussed in section 2.1.3 below.
15 Several scholars have noted the disconnect between the theoretical underpinnings of the VSL approach and the practice of relying on a population-average VSL rather than VSLs for different groups (or, indeed, individualized VSLs). For example, Greenstone and Nigam (Citation2020, 15) note that although a population average VSL “has a legal basis in that it is US Government policy, it is challenging to justify from economic first principles of individual behavior.” Similarly, Revesz and Livermore (Citation2008, 83) observe that the use of a population average is “unmoored from its economic justification.” Sunstein (Citation2014, 86, 99), for similar reasons, refers to a uniform VSL as “obtuse” and “too crude.”
16 This ethical concern with the effect of income distribution on individual dollar valuations has led some authors to advocate for a “prioritarian” social-welfare function as a guide to policy in the place of BCA (Adler et al. Citation2021). A prioritarian welfare function builds on a utilitarian approach insofar as it allows for intersubjective utility calculations. This, of course, is explicitly verboten in the economic mainstream. It is distinctly “prioritarian” (and not simply utilitarian) because it gives greater weight to those who are materially worse off (Adler Citation2020). However, while this approach arguably remedies ethical shortcomings of the VSL approach based specifically on distributional concerns, it addresses neither (1) the legitimacy of the capitalist labor market, in either its imperfect or “ideal” forms, as the proper basis for determining mortality-risk valuation (whether adjusted for income or not); nor (2) arguments regarding the endogeneity of the economic subject (explored below).
17 The highly politicized nature of the pandemic seems to evince this very antinomy, with vociferous objections to public-health measures often voiced on grounds of interference with individual freedom. See Westermeyer (Citation2021).
18 Arguably, if this reasoning were endorsed, it would parallel the reasoning initially considered by certain RAND-affiliated economists. That is, rather than proposing to the relevant political institutions the “right” dollar valuation for life, such valuations would be decided upon by the institutions themselves.