I love cost-benefit analysis.
But for the coronavirus pandemic, cost-benefit analysis and I are going to have to see a marriage therapist. We might be headed for a divorce.
Consider the following questions: What kinds of restrictions should states be imposing on work, play and freedom of movement? When should they open up for business? How open should they be, exactly, and exactly when?
To answer such questions, governors, mayors and President Donald Trump seem to be engaging in a kind of intuitive cost-benefit analysis as they struggle to balance the value of increased economic activity against the threat to public health.
Regulators and executive-branch policymakers try to be more rigorous in their analysis of costs and benefits. They ask: How do you calculate the benefits of restrictions, and what’s the right measure of costs? They try to come up with reliable numbers.
The goal is to impose restrictions when (and only when) the benefits exceed the costs — and to adopt an approach that has the highest net benefits, that is, benefits minus costs.
You might not think that’s the loveliest way to proceed, but the basic thinking is simple: Official decisions should have the best possible consequences for people. Looking at costs and benefits is the best available way of figuring out what decisions will have the best consequences.
Whenever public health is at risk, there’s a pervasive challenge. To make cost-benefit analysis work, officials need a method to convert human lives into some dollar figure. No reasonable official would refuse to spend $100 to save one life, and no reasonable official would be willing to spend $100 billion to save one life.
So what’s the right amount? Within the U.S. government, the standard value for a human life, is now about $10 million. It is called the “value of a statistical life.” (When I served as administrator of the Office of Information and Regulatory Affairs under President Barack Obama, I saw frequent use of that value, or VSL, as it is called.)
It’s important to say that we’re not really talking about the value of life in some moral or philosophical sense. The real topic is the value of mortality risks. The $10 million figure comes from evidence suggesting that workers will demand about $100 to face a mortality risk of 1 in 100,000. We are talking about the value of eliminating low-probability risks, not the value of saving lives.
Regulators build on that evidence in thinking about how to value what are usually such low-probability risks — say, from highway crashes, from dirty air, from workplace accidents. (True, some of those risks are not low-probability, and so they raise different kinds of questions.) Whether we approve of what they are doing or not (and I do, mostly), at least we can understand their rationale.
Turn to Covid-19 in this light. (Beware: It’s going to get dark pretty soon.)
To get traction on the problem, we have to figure out the magnitude of the mortality risk. Just for illustration, let’s suppose that if the whole nation opens up tomorrow, dropping all stay-at-home orders and other restrictions on crowding and economic life, 1.1 million Americans will die from the disease. (That number is just for illustration, but something like it has been floated.)
That’s a risk of about 1 in 300. How do we value that?
If people demand $100 to face a death risk of 1 in 100,000, would we say that they would demand $333,333.33 to face a death risk of 1 in 300? That sounds crazy, and it is. No evidence supports that number. And if that $333,333.33 number is crazy, then we have no basis for valuing a life at $10 million in the context of the coronavirus.
There’s another problem, which is that different populations face very different death risks from the coronavirus, and also face different costs from restrictions on economic activity.
Some evidence suggests that the death rate of those infected with the coronavirus spikes at age 70 (killing 8.6 percent of them), and is far lower for infected people in their 40s (killing just 0.3 percent).
With respect to restrictions, should we therefore be treating younger people differently from older people? In doing cost-benefit analysis, should we focus on “life years” saved, rather than lives saved? If so, the benefits of lockdowns will start to look a lot lower, because a lot of the lives we will be saving involve people who do not have so many life-years left.
One of the most careful analyses of the costs and benefits of policies relating to the pandemic insists on the importance of treating the young differently from the old. A research team led by the economist Daron Acemoglu of the Massachusetts Institute of Technology concluded that there would be major advantages to a policy that calls for a long and strict lockdown for older people, and much more flexibility for lower-risk groups.
They argue that a very strict lockdown, limited to people over the age of 65, would achieve most of the mortality gains of a broader lockdown, and that it would allow the rest of the population back into the economy much more quickly.
Acemoglu and his colleagues make a powerful argument in favor of a targeted lockdown, imposing the most aggressive controls on the most vulnerable. Any such approach would obviously raise questions of equity (let alone politics). But even if we could answer those questions, cost-benefit puzzles would remain.
For the most vulnerable (perhaps people over 65, perhaps people over 70, perhaps people with pre-existing medical conditions), what kind of lockdown would be needed, and how long? For lower-risk groups, what kind of flexibility, and how much? How, exactly, do we calculate the monetary value of reductions in mortality? Not to mention morbidity?
Cost-benefit analysis, I love you, but we really need to see that therapist.-Bloomberg