This post is part of a symposium on the future of cost-benefit analysis. Read the rest of the symposium here.
In Spring 2020, many economists argued that the COVID-19 economic shutdowns were justified on Cost Benefit Analysis (CBA) grounds. It may initially seem difficult to justify lifesaving measures using a monetary calculus. If anything is beyond price, we might think, it must be the value of human life. This difficulty is surmounted by using the concept of the “value of a statistical life” (VSL). In fact, justifying mortality reduction policies using the CBA and VSL framework is routine in regulatory law.
In the following post, I argue that this framework is a poor one for justifying lifesaving pandemic-mitigation measures. Even if a policy measure does not pass the standard CBA test, it may nonetheless be justified. Conversely, if a policy measure passes the CBA test, there could nonetheless be reasonable arguments against it. This is because CBA reserves the determination of “value” exclusively for individual economic agents in a market, and insists that policy be based on such individual valuations. However, decisions that belong to us socially, as members of a polity, need not be held hostage to the specific normative perspective and theoretical logic of neoclassical welfare economics. Instead, I argue, they should be arrived at in and through democratic institutions and processes.
To make this case, I discuss, first, the theoretical framework for CBA; second, the application of CBA in finding the value of a statistical life (VSL); and lastly, the implications for the use of the CBA and VSL framework in the current pandemic.
CBA and its framework in economic theory
On an intuitive level, requiring a cost-benefit test seems hard to contest. After all, who would argue that a policy is justified if the benefits do not outweigh the costs? But this intuition in favor of CBA rests on an illusion, namely, that there is one value-neutral way of weighing the costs and benefits of a policy or rule. This is not the case.
To see this, we need to look more closely at the logic behind CBA as provided by economic theory. Central to this logic is the neoclassical conception of the individual economic agent. The agent has preferences assumed to be exogenous. That is, they are said to be independent of the set of institutions in which the agent exists, including the institution of the market itself. Furthermore, preferences are interpreted in terms of individual rates of substitution – i.e., the rate at which an individual would substitute one good for another. This means the very process of valuation itself is inextricably linked to exchange; there is no “valuation” that is not valuation in terms of some other good. Coupled with the institutions of prices and income, one can then find the agent’s “willingness to pay” for a good (or “willingness to accept” to part with one).
This notion of the “willingness to pay” (WTP) of the individual economic agent is essential to CBA, because the question of whether the social benefit of a policy is greater than the social cost assumes that social benefit is properly interpreted as nothing other than the sum of each individual’s willingness to pay as revealed in market exchanges.
CBA and the value of a statistical life
But the concept of WTP, resting on exogenous preferences revealed in exchanges, creates an obvious difficulty in the case of non-market goods. There are at least two ways to understand the nature of this difficulty: the first is that it reflects a mere methodological problem, in that no market exists to reveal the underlying preferences. The second is that it may reflect a deeper ontological problem, in that the exogenous preferences simply do not exist. The problem is generally understood in the former sense, while the latter possibility is rarely entertained (with the exception of some heroic dissenting scholars).
The most provocative example of this difficulty is the question of the value of human life. Because the commitment to the picture of agents valuing goods in terms of their exchangeability is so thoroughgoing, it must extend to life itself. The only difficulty, therefore, is the methodological one of locating a market that reveals these valuations. Serendipitously, the economics profession has located markets performing this methodological function, most notably the labor market, which ostensibly reveal workers’ rate of substitution between income and mortality risk. For example, labor market studies might reveal the wage premium necessary to entice a person to become a coalminer. It is in answering questions like these that economists determine the value of a statistical life.
“VSL” and the pandemic
At least two criticisms of this theoretical and methodological structure behind VSL estimates are highlighted by the pandemic. The first is specific to the notion of worker rates of substitution between income and risk. In the pandemic-induced recession of 2020, with a restricted range of job opportunities, and little publicly provided material support, newly-acknowledged “essential” workers had little choice to but to accept increasingly risky jobs with little or no hazard pay. In fact, the worse the bargaining position for workers, the less they will “value” their lives – lest they risk their livelihood.
The second criticism focuses on the endogeneity of the underlying preferences. Even assuming away the question of worker bargaining power, we are still left with the question of whether, as a matter of policy, our social willingness to pay for mortality risk should be defined exclusively in terms of agents’ WTP as revealed by the market. Notably, some labor market research finds that the VSL peaks somewhere midlife and then declines with age. If the problem of valuation for non-market goods (such as life) were the mere methodological one of finding a market for mortality risk, then policy should be dictated by this (or similar) research. But if the problem is ontological, and preferences are endogenous to the institutional context, then markets (as one such institution) cannot be said to simply reveal preferences. Instead, they shape them. There is therefore no theoretical justification for eschewing a social willingness to pay to protect the lives of the elderly—as we might be inclined to in a pandemic—that is arrived at through democratic and deliberative institutions, rather than confining ourselves to WTPs of individual agents ostensibly found in markets.
In short, to insist on using individual WTP as an ontological ground for social decisions is to abandon the possibility of a democratic determination of valuation. The policy space should reclaim such a possibility.