Getty Images
In the United States, the debate has been raging

Marc Fleurbaey is the Robert E. Kuenne Professor in Economics and Humanistic Studies at Princeton. He is a former editor of the journal Economics and Philosophy, and much of his research has focused on the evaluation of public policy.

The COVID-19 crisis puts all policymakers in a difficult position. In addition to the great uncertainty about the parameters of the pandemic and about the short-term and long-term consequences of the economic slump, there is an ethical dilemma: how to balance the value of saving human lives against the value of preserving people’s livelihoods. In the United States, the debate has been raging between those who say “the cure cannot be worse than the disease,” and those who say “we are not going to put a dollar figure on human life.”

Whether they like it or not, policymakers must address this difficult trade-off. In theory, separating every person from everyone else for two or three weeks would immediately stop the infection spread at little economic cost. This is totally impractical, but less extreme lockdown measures observed in many countries do reduce the reproduction rate of the pandemic to such low numbers that this, actually, would fully extinguish the pandemic within two or three months. It is also possible, under a more lenient approach, to keep the pandemic under control by a stop-and-go policy of repeated lockdown episodes of a few weeks each until a vaccine is found. The problem is: Are all these measures worse than the health crisis itself?

When Princeton students were sent back home, just before the spring break, I was intrigued by this exceptional situation and built a model on a simple Excel spreadsheet, first for my own curiosity, but also as a possible teaching tool for my Woodrow Wilson School class on the microeconomic evaluation of public policy. Several friends and colleagues have helped me improve it, including my daughter, a physicist. 

The model simulates the pandemic as well as the lockdown and testing policies available, and includes a set of evaluation tools for the comparison of various policy options. You can download it from https://bit.ly/fleurbaey. (Users can change all parameters and assumptions and determine the timing and intensity of contact reduction and testing policies.) In a rough but informative way, the model takes account of inequalities in income and life expectancy across social groups and allows for various assumptions about the distribution of the economic cost and the fatality burden among these groups. Such assumptions relate to policy choices about social protection, income support, and access to health care.

The evaluation tools included in the model belong to two ethical approaches, which are the most common for such assessments. To be clear, yes, these methods all “put a dollar figure on human life,” but they do so in a principled way. The main reason they put a value on human life is that each one of us does it all the time. We all take risks every day, implicitly deciding that life is not worth more than the value of the risk taken. I used to ride my bike to the University even though the risk of a fatal accident, while low, was much higher than the risk I’d face driving my car. I thought the advantages of bike-riding — such as exercise and long-term health benefits — outweighed the greater risk I was taking on. Similarly, the main agencies of the federal government use a key number, the “Value of a Statistical Life,” to assess the benefits of various programs with consequences on population health and compare these benefits to the program costs. Without such a number, they would have no idea whether a safety program is worth the cost.

The ethical calculus based on the value of a statistical life is straightforward. In the current pandemic, the number of deaths avoided by the various measures taken by most governments is very high: about 1 percent of the population in developed countries. This huge number — more than 3 million people in the U.S. — comes in part from the overwhelmed hospitals being unable to receive all patients when the “curve” is not “flattened.” The value of a statistical life in the U.S. is generally taken to be around 150 times the annual per-capita income, or about $10 million. Therefore, the health benefits of avoiding all these deaths are worth about 150 percent of a year of income: more than $30 trillion! Assuming that the serious recession that would have occurred in absence of any government intervention is worsened only by about 10 percent of a year of income under the lockdown policy, the conclusion is obvious. The cure is vastly better than the disease.

An evaluation that gives special priority to the worse-off will favor policies granting all people equal access to health care and guaranteeing a lifeline to economically vulnerable workers and small businesses.

However, there are several issues with this overly simple approach. First, it does not account for the fact that the victims are mostly elderly people who do not lose many years of life. There is another number called the “Value of a Statistical Life-Year,” which can be used to assess the years lost instead of the lives lost. Considering that the average victim loses 10 years of life, and that the value of a statistical life-year is about three times the annual income per capita in the U.S., one obtains an estimate of the health benefits of about 30 percent of a year of income. This is still much greater than the predicted economic cost, but in the same order of magnitude — therefore, one should pay close attention to how the health and economic consequences unfold in the coming months. 

Another ethical problem with the Value of Statistical Life approach is that it does not pay attention to the unfair distribution of background conditions (such as income and life expectancy) among social groups, or to the distribution of the health and economic burdens of the ongoing crisis among them. In the U.S., in particular, disadvantaged communities are paying a heavy price on all fronts. These communities often include workers who are exposed to the virus or are among the first to be laid off or lose their businesses.

This is why I prefer an alternative approach that makes an evaluation of the well-being of the population, taking account of these fairness issues in the distribution of burdens and benefits. This “social well-being” approach is more demanding because it requires more data and more detailed predictions of the consequences of the crisis and the policies for different groups, but it provides more relevant evaluations. In particular, it allows the decision-maker to pick a key ethical parameter: the degree of priority granted to the worse-off. An evaluation that gives special priority to the worse-off will favor policies granting all people equal access to health care and guaranteeing a lifeline to economically vulnerable workers and small businesses. It will give more weight to health benefits if the victims of the virus are among the worse-off, and more weight to economic costs if the burden falls on the most disadvantaged. My colleagues and I are debating whether the elderly victims of the virus are really among the worse-off, since they include many pensioners who do not risk losing their jobs and who have been lucky to live to an old age — the average age of a COVID victim in developed countries is close to 80. That’s similar to the life expectancy of the population, which means that, on average, those who do not die from the virus will not live longer than the average victim of the virus. On the other hand, the average COVID victim has lived during times that were much less economically affluent than recent years. 

At any rate, in the current crisis both approaches — in all their variants — strongly favor an ambitious public-health policy combining strict shutdown, widespread testing, and universal mask wearing to promptly quash the spread of the virus at relatively little economic cost. Not many countries have succeeded in achieving this feat, but some have, including some low-income nations. It requires swift action and works best with clear vision, great transparency, a high level of trust and cooperation of everyone involved, and extensive support measures to help the economically vulnerable to go through the hiatus. This is really hard to achieve because the policy appears very painful in the moment, and if relaxation of the effort comes too early, most of the effort will have been done in vain and the pandemic will restart. When the pandemic remains rampant and one has to examine new measures as cases spike again, the question of the hard trade-off between lives and livelihoods reemerges. The tools offered in this model can be applied again and may help us better understand how to navigate this crisis of historic proportions.