TDLR: Trade-offs between lives and well-being are something that economists think about and society deals with on a number of fronts. With respect to current social distancing measures and the economic pain they cause, most people’s preferences would probably align with maintaining them for now. We can’t be certain, but evidence from past pandemics and what little we know about COVID-19 suggests that we aren’t facing a trade-off between lives and the economy right now

R1 I was gonna write an R1 on bad takes about public transit, but then COVID-19 happened, and no one wants to ride the train right now. Before I start, I want to say that I am neither a public health expert, epidemiologist, health economist, nor a macroeconomist. Research on this is moving at a breakneck pace, and if y’all have anything to add on these fronts, it would be greatly appreciated.

In this post, I want to address claims that social distancing is “worse than the disease,” that the suffering caused by shutting down wide swathes of the economy is worse than the probable deaths that would be caused by letting COVID-19 run rampant. Anecdotally, I’ve seen this sentiment shared frequently online. There are the protests against various measures, and as time goes on, and the economic pain of social distancing increases, sentiment against social distancing is likely to increase. This sentiment was probably most clearly enunciated by Texas Lt. Governor Dan Patrick when he said that, “People his age should be willing to sacrifice themselves for the sake of their children.” This statement attracted a lot of hand-wringing over the value of human life vs economic activity, but also some support. When polled, a large number of Americans say they fear catching the virus or someone they know catching it, more say they are a worried about the economy. Now, you can still be worried about the economy and think social distancing is important. Indeed, social distancing, seems to be broadly followed, but with recent protests, it’s not a sentiment shared by everyone.

Now, let’s delve into the trade-off that the Lt. Governor is describing. Sure it’s macabre, and maybe we don’t like thinking about it. However, is it a fair statement that we can turn a dial between economic well-being and people dying? Does this trade-off currently exist with respect to social-distancing and coronavirus, and is the trade-off worth making on the current margin?

I will claim that while this is a trade-off worth thinking about, social distancing was, and probably still is, probably necessary to prevent mass death and the economic pain that detractors of it claim to care about. I am not claiming that social distancing will always be necessary or that we as a society won’t have to make trade-offs about risks in the future. More this is to claim that social distancing likely is a first-best solution compared to the counterfactual of letting the virus run rampant right now. There are plenty of intermediate steps we could take, but this post is meant to mostly to counter idiots complaining that this is just like the flu and more importantly, hopefully help provide us a framework for thinking carefully about these trade-offs without spitting out platitudes.

Value of Statistical Life

How much is a human life worth? You might say “priceless”, that no price is too high to save a human life, but think about your own situation. Let’s say I offer you a certain amount of money, but there is a small chance you die instead. Obviously, it will depend on the amount of money offered, the risk, and your personal preferences. Indeed plenty of people risk death for money everyday; many of us drive to work, which has a non-zero chance of death.

Now that’s our own preference. We can choose to put ourselves at risk. Why do policymakers have to butt in? Well, there are various market failures that can make this difficult. There might be information asymmetries, like not knowing how much arsenic is in your food. There are also externalities that affect large numbers of people and that we can’t reach a Coasian bargain with them. For instance a highly infectious disease that is asymptomatic or mild for most people, but highly virile and deadly for others. Eichenbaum, Rebello, and Trabandt have a simple paper that makes the case that allowing the economy to operate as normal is much worse than the shutting things down, but how do we determine this value? This is where something called the “value of statistical life” (VSL) comes in.

VSL can roughly be thought of as the marginal rate of substitution one has between a composite good and one’s risk of mortality. To be clear, this is not the value of a human life. If a kid is trapped down a well, rescuers aren’t gonna ignore her because saving her is expensive. VSL is a bit more detached than that. It isn’t people volunteering to die, just seeing what people are willing to accept in terms of risk if they are compensated for it, but who decides this?

The Office of Management and Budget (OMB) manages them, and different agencies use different estimates based on what they’re regulating. For instance, the Department of Transportation might look at how much people spend on safety features as part of their analysis, and the EPA will look at pollution avoidance measures that people take. These estimates rely on a lot of extrapolation and assumptions about human behavior. They generally come from meta-analyses used through the years. In 2020 dollars, they range from $6.2 million according to Miller (2000) to 13.3 million according to Kochi et. al.. Most estimates that are currently used put the VSL at around $10 million.

For instance, different risks might be more salient than others, and that can impact estimates. There’s also the problem of heterogeneity. Different people might have different preferences that affect their willingness to pay. More to the point, inequality can impact things here. If I’m rich, I might be more willing to spend more to reduce my mortality risk by 10% than a poor person might be willing to spend to reduce their risk by 50%. Is the poor person more risk-loving, or do they just have fewer outside options?

This is worth keeping in mind where as of April 8th 1/3 of hospitalizations from COVID-19 complications were among African Americans. This probably at least part reflects the geography of the outbreak to some degree, but even then it shows large disparities in who is at risk. Who is actually at risk of exposure is hard to pin down. An analysis of which jobs could be done mostly from home shows that less exposed occupations are things like computer work, which tend to pay more than maintenance and cleaning jobs, which are the most exposed.

The upshot of this is that even if we knew the “true” willingness to pay of people to avoid being exposed to the outbreak, that might not fit within fairness criteria that we might care about. Economics doesn’t give the answer on that, but it is worth keeping in mind. VSL has serious limitations, and we shouldn’t put it down as the intrinsic value of human life, but it is a useful starting point in thinking about how the public trades-off risk. However, the idea that society can and does trade-off mortality risk and economic well-being is a sound one.

What about this and COVID-19?

OK, so we have a framework to start with, but what does it say about the costs and benefits of social distancing? Here we’re working with a lot of assumptions. Even public health officials don’t know. 538 has an article from April 2nd that show an incredible range of estimates of the predicted death toll. It’s really hard, and we don’t know a lot about the disease, and these models rely heavily on these assumptions about facts that we just don’t know. That has a major effect on predictions.

We don’t know what the infection rate for COVID-19 is, and estimates are highly sensitive to different estimates of these parameters. For instance, Iceland tested a large percentage of their population, and found that a very large percentage of the population is asymptomatic. This might be good, because ceteris parabis, it means fewer people might die from being infected, but on the other hand, it means that the disease can spread more easily. The above paper examines a set of models called SEIR models, which are kind of like DSGE models, where you need to solve a system of equations and plug in parameters to figure out the likely spread. Another option is using agent-based models, where we model outbreaks as complex systems that can’t be easily captured with a few equations.

The IHME model, that the White House is currently using is a bit different, it updates off the past experience of other cities, particularly in Italy and China, to project death tolls. They’ve shown a serious drop in projected deaths, from 100 to 240 thousand US deaths, even with social distancing to less than 70,000 more recently. Does this mean that we’re all just overreacting?

Well no, for one thing, social distancing measures seem to have reduced disease morbidity. How much is hard to know, and I won’t deign to throw out calculations here for a counterfactual if we had no social distancing. That’s hard to know, and we’ll need to know more about the R0 and mortality rate of the disease. Maybe 500,000 people would die without social distancing interventions, which would be valued at 18% of GDP at a VSL of $7 million, but we just don’t know. Maybe it’s a lot lower, we don’t know, but there are some bits of evidence which suggest that we should probably be risk adverse and that opening up prematurely won’t fix things.

Economic Impact of Pandemics:

There are a lot of factors that can determine how important lowering the infection rate is. For one, our healthcare system only has so much capacity. There are only so many healthcare workers, equipment, and beds for patients. You’ve probably seen infographics about “flattening the curve” because without that, we can only take so much. Indeed, some basic modeling, shows that it wouldn’t take that much to overwhelm our healthcare system, which could rapidly raise the mortality rate.

We can also examine the historical impact of pandemics on the economy. Over the long-run pandemics tend to lower output as they kill people, make survivors less productive, and halt economic output from people isolating, even without government directives.

The closest parallel is probably the 1918 H1N1 flu pandemic, sometimes erroneously called the “Spanish Flu”, which probably originated in either China or the United States. This disease killed at least 50 million people worldwide and at least 675 thousand in the United States. For comparison, World War I, which happened concurrently killed about 10 million worldwide and about 116 thousand Americans. The flu was noteworthy, because while it spread across the entire country, different localities had different responses. Many cities, particularly out West, quickly put aggressive social distancing interventions in place, while other cities were slow. More aggressive cities saw lower mortality rates than less aggressive cities, and less bad secondary waves, which were also a huge feature of the pandemic. Furthermore, it appears that aggressive cities did not see a relative decline in manufacturing employment. Now an important difference with the 1918 flu is that the most vulnerable people were young, prime-age adults, compared to today, where the disease tends to be more virile with older people with other health conditions. Still, as with the equity issue we brought up earlier, comparing between groups is hard, and we can’t just handwave them away.

Now that’s all terrible, but what about the very real economic pain now where millions are unemployed? Surely, we can’t dismiss them, and we can’t. However, there are a couple of things to note here. Firstly, ending social distancing prematurely might not bring the economy back. People can still be nervous about doing certain things like going out to eat. Most Americans report being risk-averse about doing things again when directives allow them too. If they perceive the outbreak as still present, they will still exercise caution, regardless of what policymakers say. Besides, if distancing ends prematurely, this can lead to a 2nd wave, and you have to redo everything.

Secondly, as for the pain and misery of joblessness, that needs to be acknowledged and worked on. The government should absolutely boost UI, stimulus payments, and boosting critical sectors like health care. The Fed is currently keeping interest rates low, being accomodating towards fiscal policy in many unprecedented ways, including lending to small businesses and state and municipal governments. Also while, the pain of joblessness is awful, and suicide rates do tend to correlate with joblessness, there doesn’t seem to be a strong link with overall mortality and economic downturns, and if anything mortality rates go down during downturns as behavior changes. This isn’t to minimize this suffering, but more to counter arguments that more people would die anyway, so there aren’t any gains to distancing in terms of mortality.

So in conclusion, there are serious reasons to be critical of the Lt. Governor’s remarks. While conceptually, they could have some merit, it’s hard to see that is currently the case, as of mid-April 2020. Overall, we just don’t know the counterfactual of not social-distancing. However, there is reason to err on the side of caution. It is very probable that not distancing will result in a magnitude of deaths that outweigh any benefits to the economy. Also, historical experiences show that the economic pain of pandemics are real, and often outweigh the pain of mitigation. In the coming months, this debate will continue. We’ll have to figure out how we navigate this world. There are some plans out there, and well none of them look that cheery. We are going to have to make painful trade-offs, but when it comes to containing this outbreak, the evidence leans towards social distancing for the time being.

Through this all, it’s important to remember that the job of economists is not to maximize the monetary value of traded goods. It is to help people manage trade-offs to maximize their well-being. We are facing painful trade-offs in the largest disruption in this country since World War II. It will require social solidarity, patience, and acceptance of uncertainty.