A man walked through an empty Brooklyn Bridge Park in New York on Tuesday. PHOTO: MARK LENNIHAN/ASSOCIATED PRESS


Economics vs. Epidemiology: Quantifying the Trade-Off

par Featured articles licensed from The Wall Street Journal | le 16 April 2020




Their estimates show the benefits of social distancing exceed the costs, but they will need to delve deeper


Greg Ip – The Wall Street Journal

The rising economic toll of pandemic-induced shutdowns is fueling suspicions that government leaders are listening too much to epidemiologists and not enough to economists.

That is understandable. Yet if economists were in charge, the policy response probably wouldn’t be much different. While epidemiologists and economists study different problems, their approaches are similar, with a heavy reliance on statistics and an awareness of trade-offs.

Epidemiologists know that extensive social distancing is costly, in terms of forgone wages, damage to physical and mental health, and harm to children’s learning when schools are closed. Economists, similarly, know that even without lockdowns, a pandemic inflicts costs in terms of absenteeism, reduced consumption, lost lives, impaired health and uncertainty

Costs and Benefits

Economists expect GDP to run well below its pre-pandemic baseline for the next three years, while lockdowns appear to have saved thousands of lives.

Sources: WSJ economists’ survey & calculations (GDP); Alessandro Vespignani, Northeastern University (deaths)


Sources: WSJ economists’ survey & calculations (GDP); Alessandro Vespignani, Northeastern University (deaths)

Devising a policy response requires input from both disciplines—epidemiologists to estimate the benefits of social distancing in lives saved, and economists to calculate the costs in forgone jobs and income. So far, their work suggests the benefits exceed the costs, though that conclusion is surrounded by uncertainty. As governors explore when and how to reopen their economies, economists and epidemiologists need to help them devise interventions that both maximize lives saved and minimize the economic cost.

Basic pandemic models start with estimates of how much of a population is susceptible, how many and how quickly other people are infected by each infected person, and what proportion of the infected will recover. Those models can then be enriched with more granular data about geographic and demographic subpopulations.

Alessandro Vespignani, director of the Network Science Institute at Northeastern University, heads a team that has modeled the spread of Covid-19 and numerous previous pandemics. Its model divides the U.S. into segments of 10 square miles, integrates airline traffic and commuting patterns, and then adds restrictions such as school closures as they are implemented.

The benefit of mitigation measures can be derived by comparing current projections of the number of deaths to a baseline that assumes no mitigation measures are adopted. A team at Imperial College in London in mid-March put that baseline at 2.2 million deaths in the U.S., and Mr. Vespignani’s team got similar results. Such a scenario would probably never happen in real life: People would social distance, whether or not governments ordered it. But it is a useful benchmark nonetheless.

The Northeastern team estimates the actual number of U.S. deaths as of April 15 at 34,000 (higher than the number so far reported) compared with 159,000 in the unmitigated scenario—implying that mitigation has saved more than 100,000 lives so far. By month’s end, the model projects 53,000 deaths, compared with 584,000 without mitigation, implying that half a million lives will be saved. There are wide uncertainty bands around all those figures.

What’s the value of those lives saved? Every life is priceless, of course, but people nonetheless implicitly put a price on it—think of life insurance. Federal regulators put the statistical value of a life at around $10 million. Multiply that by Northeastern’s estimates, and the benefits of mitigation measures so far come to $1 trillion. That grows to $5 trillion by month’s end. If you assume 2 million deaths are avoided, the value grows to $20 trillion—about a year’s worth of gross domestic product.

These are rough measures, and the ways of calculating the benefits can vary. In a March paper, Michael Greenstone and Vishan Nigam of the University of Chicago assign a lower value to older people and a higher one for the young. On that basis, they conclude that three months of social distancing could save 1.7 million lives worth $8 trillion.

So much for the benefit. What about the cost? Economists surveyed by The Wall Street Journal forecast annual GDP growth of 0.9% through 2022. That is down from a forecast of 1.9% in January, before social-distancing measures were adopted. I estimate that amounts to a loss of $3 trillion in today’s dollars. So in this accounting sense, the trade-off is worth it. Note that governments can’t recoup that loss simply by lifting restrictions because many people would stay at home out of fear of infection.

Now, the caveats. My estimate of forgone GDP excludes the years after 2022, as well as nonmonetary costs such as less socializing and long-term harm to mental health or learning. And economists may be underestimating how persistent the pandemic’s hit to employment and incomes will be. The most pessimistic forecast in the latest survey of economists yields a three-year loss of $7.5 trillion—more than twice as much as the median scenario.

The death estimates are also uncertain. If fatality rates have been overestimated, then baseline estimates of 2 million deaths may be too high. Mr. Vespignani said it isn’t likely to change much as epidemiologists are reasonably confident about the plausible range of fatality rates. But he said epidemiologists know a lot less about Covid-19 than the flu or Ebola. For example, they don’t know the proportion of asymptomatic infections, which affects estimates of how far and how fast infections are transmitted.

But the biggest gap in our knowledge involves the costs and benefits of individual mitigation measures. Epidemiological models can estimate the aggregate benefit in terms of lives saved, but are much less precise about the benefit of any individual measure.

Filling this gap is crucial now so that governors can evaluate the most effective, least costly ways to reopen the economy, such as wearing masks, isolating vulnerable groups, testing aggressively, tracking and tracing, or a combination.

Mr. Vespignani’s group is now making such estimates for the Boston area, using detailed household data to determine the feasibility and effect of contact tracing and isolation. He hopes this will identify strategies that are less disruptive than total lockdowns.

“The decisions that will be made in the next few weeks must include economists and epidemiologists and public-health people, to find the trade-off between those two viewpoints,” he said.