Economist Sergio Rebelo has spent the past 2 weeks holed up in his Chicago home, working feverishly to crack the economics of the coronavirus.
Armed with a hybrid model that combines how viruses spread with how people work and consume, the Northwestern University researcher is one of a number of macroeconomists now trying to shed light on the balance between the economic impact of locking down major parts of the economy and the economic damage wrought by the disease itself. “When you think about the optimal policy, you really want to see the effect between the economy and epidemiology,” Rebelo says.
It is much more than an academic exercise.
Much of the world economy has shuddered to a halt. In the United States alone, a record 3.3 million people filed for unemployment benefits in late March. President Donald Trump once mused about lifting pandemic restrictions by mid-April to prevent more economic damage, but ultimately settled on extending federal advice to maintain physical distancing through the end of April. Trump reversed course after epidemiologists warned that a return to normal behavior could spark an explosion of COVID-19 infections, killing as many as 2 million Americans. With strong interventions, models reviewed by the White House suggest, deaths could be reduced to 100,000 or perhaps fewer.
Trump’s decision, however, has only put off the question of when, exactly, cities and states should begin to ease up on distancing orders. “If you keep the shutdown going for 2 months more than we need to, that’s just an unbelievably costly mistake. … If we lift the shutdown 2 months too soon, that would be an unbelievably costly mistake,” says James Stock of Harvard University, who is working with public health experts to develop models weighing the economic trade-offs of different containment strategies. (Stock served on the federal Council of Economic Advisers under former President Barack Obama.)
This is new territory for macroeconomists more familiar with gauging how interest rates might influence employment. Even health economists have little experience modeling a pandemic so threatening and disruptive, says Beate Sander, a health economist at the University of Toronto who worked on a 2009 study of the costs of different interventions to treat an influenza pandemic. School closures were the most extreme scenario in Sander’s earlier study. “We couldn’t imagine that this would be something we would be forced to resort to because we would be so unprepared.”
But Rebelo had previously adapted models of epidemics to simulate how changing attitudes spread through a society, contributing to booms and busts in the housing market. When the novel coronavirus appeared in the United States, he started to think about how to put the model to work on a real virus.
The cost of shuttering large parts of the economy is relatively easy for Rebelo and his collaborators, Northwestern University economist Martin Eichenbaum and Mathias Trabandt of the Free University of Berlin, to translate into money, the currency of economics. On the economic side, their model calculates how the disease and government policies would influence how much people work and buy.
But the dollars and cents of a virus are less intuitive. Rebelo uses a modified version of what’s known as an SIR model, an acronym for categories of people: susceptible, infected, and recovered. It simulates how a disease moves through a population based on how infectious and lethal it is, and how much contact people have with each other. To put a price on the results, Rebelo takes the number of predicted deaths and calculates an economic estimate of the value of the lost lives. The approach is similar to the price that the U.S. Environmental Protection Agency used to gauge the costs and benefits of environmental regulations: $9.5 million per life.
His initial modeling efforts showed that even a yearlong lockdown makes economic sense, to allow time for a vaccine to be developed. The pause would shrink the economy by approximately 22%—a cost of $4.2 trillion. By comparison, the model shows that without containment measures, the economy would contract by about 7% over that year—but as many as 500,000 additional lives would be lost, which translates into a loss of roughly $6.1 trillion.
Andrew Atkeson, an economist at the University of California, Los Angeles, agrees that the economics point strongly toward strict measures. If the epidemic is allowed to grow unhindered, he predicts the economy will grind to a halt anyway as people see an explosion of infections and stop going out. “Either you shut off the economy now and have people staying at home, or you let this thing rip and you have people staying at home scared,” he says.
Rebelo cautions that his model is simplistic. But he and colleagues are now working to create more sophisticated scenarios. They hope to include the size of different age groups in the United States to account for differences in how deadly the disease is for different ages, and modified “smart” lockdowns that allow more economic activity. In some scenarios, people who recovered from the disease might return to work if they are shown to be immune. “That’s maybe where all these economies are going because it’s going to be very hard to shut down the economy for a very extended period of time,” he says.
Economists are also weighing subtler interactions between health and economics, including the possibility that the economic shock itself will add to the body count. Public health experts broadly agree that more suicides happen in recessions. Scientists found an additional 4750 suicides in the United States over 3 years attributable to the Great Recession of 2008. Trump pointed to a potential increase in suicides as a reason for loosening restrictions.
Yet economic downturns have typically translated into a net drop in deaths, says Christopher Ruhm, an economist at the University of Virginia who has studied the phenomenon. Although suicides can rise, decreased economic activity can save lives partly because it reduces traffic accidents and air pollution, he says.
There are notable exceptions. Death rates rose in Russia following the collapse of the Soviet Union, because the economic downturn was part of a broader social collapse, Ruhm says. In the case of the coronavirus pandemic, Ruhm says it’s too soon to know, but “my guess would be purely from the economic aspect there would be some modest decline in mortality.”
The modelers still lack basic data. The most critical is a better estimate of how deadly the disease is. In his modeling, which supports long, strict lockdowns, Rebelo used statistics from South Korea, which has some of the most comprehensive testing, to estimate that 0.5% of all infected people die. But if it turns out a lot of people get infected and have few symptoms, the economically sensible approach might be to let the infection spread and accept that there will be some death toll, Stock says. “The policies are extremely different depending upon these parameters that we don’t know.”