Macroeconomic Responses to the Coronavirus Pandemic
Robert J. Barro
The best source of a worst-case scenario for the ongoing coronavirus pandemic likely comes from the Great Influenza Pandemic of 1918-1920. Our research found from data on 48 countries that the flu death rate averaged 2.1 percent, corresponding to 40 million worldwide deaths. However, the death rate was highest in poor countries—including several in sub-Saharan Africa, along with India, Guatemala, and Indonesia. This historical experience suggest that the worst mortality from the current pandemic is yet to come as the disease spreads further into low-income countries, typical with poor public-health systems.
The 1918-1920 pandemic was also associated with recessions in many countries. The typical size of the contraction in GDP was about six percent and lasted for around two years; that is, the lost output was about 12 percent of a year’s GDP.
Currently, the absence of effective medical treatments and vaccines leaves the world without attractive choices for countering the coronavirus pandemic. One possibility, which becomes possible once widespread and rapid testing are available, is targeted quarantining of the infected population. Until then, the main option, now being followed in most countries, is to curb economic activity as a way to reduce interactions and contagion. This policy amounts to a decision to reduce world GDP in the short run by roughly 20 percent. In essence, this is a voluntarily implemented negative supply shock, akin to a sudden loss in productivity in each country.
A decline in GDP by 20 percent, if it lasts for a full year or more, would constitute a rare macroeconomic disaster, analogous to the Great Depression of the 1930s. Historically, these depressions have lasted for 2-4 years. However, the reasonable hope in the current environment is that the sharp cut in the flow of output will last for only a few months, after which the virus will be contained. That is, the objective is for a sharp, V-shaped recovery.
The world’s annual GDP today is around $100 trillion, so a 20 percent cut that persists for a year implies lost output of $20 trillion. This contraction by 20 percent of a year’s GDP is in the ballpark of the economic contractions experienced during the 1918-1920 pandemic (about 12 percent of a year’s GDP). In contrast, a contraction that lasts for only three months implies lost output of only $5 trillion (5 percent of a year’s GDP), and one that lasts for three years implies lost output of $60 trillion (60 percent of a year’s GDP). Thus, the duration of the contraction is a crucial variable. For the moment, I assume that it is a good policy choice to engineer a 20 percent reduction in GDP even if that reduction lasts for a full year—though a less persistent decline seems realistic and obviously more attractive.
What are reasonable monetary and fiscal responses to the fall in GDP? Since we have determined that the cost of reduced GDP is worth bearing as a way to combat the pandemic, it would be inconsistent to follow the usual stimulus policies that work by raising aggregate demand and, thereby, increasing real GDP. For example, if aggressive monetary policy—cutting short-term nominal interest rates and raising central bank purchases of assets—succeeds in raising GDP, we would not regard that as success in the current unusual environment. If we did not want GDP to fall by 20 percent, we could have achieved that goal by lessening the constraints on economic activity in the first place.
The same argument applies to a general fiscal expansion; for example, the U.S. policy of having the federal government give most adults a check for $1200. To the extent that this kind of policy succeeds in raising aggregate demand and, thereby, GDP, we have the same situation as with aggressive monetary policy. That is, we would not value the offsetting rise in GDP.
More sensible are targeted policy responses aimed partly at individuals and partly at businesses. One dimension of this policy is a strengthening of the existing social-safety net. In this context, it makes sense to increase accessibility and benefit levels for programs like unemployment insurance, food stamps, and Medicaid (which finances medical expenses for poor persons). These program expansions, some of which are in the recent U.S. package, are much more targeted to the needy than is the passing out of $1200 checks to everyone.
It also makes sense that the recent U.S. package includes policies aimed at limiting the permanent disappearance of businesses and maintaining links between firms and their workers. Much of this response applies to particularly distressed sectors, such as airlines and other travel-related companies. It is also crucial for central banks, including the U.S. Federal Reserve, to avoid major disruptions of financial markets. These actions are already being pursued vigorously, and they provide confidence that the financial disruptions during the Great Recession of 2008-2009 will not be repeated.
Consider now the underlying policy of engineering a 20 percent decline in GDP that lasts for a year. Two things matter here. By how much does one value the potential reduced mortality and by how much does the policy succeed in lowering mortality? If we use the world death rate of 2.1 percent from the 1918-1920 pandemic and apply this number to the world’s current population of around 7.5 billion, we are talking about possibly saving 150 million lives. If we use a standard number from the research literature of about $1 million for the value of a statistical life, then a saving of 150 million lives is worth $150 trillion, about 1-1/2 years of world GDP. This number greatly exceeds the costs of losing 20 percent of a year’s GDP.
We can get some idea from the Great Influenza Pandemic about how a policy of constraining economic activity will hold back the spread of a pandemic and lead, thereby, to a saving of up to 150 million lives. One point is that Australia constrained its economy by implementing an aggressive maritime quarantine that managed to avoid the Great Influenza Pandemic in 1918. Moreover, this policy did not just postpone the effects of the disease—it ended up achieving a much lower overall flu death rate than those experienced in other countries, even in the Southern Hemisphere. Second, there is direct evidence that non-pharmaceutical public-health interventions held down deaths from the flu in 1918-1919. These interventions applied in three main areas: school closings, cancellation of public gatherings, and isolation/quarantine. The evidence is that these policies worked and more effectively the earlier they were implemented and the longer they were kept in place, with some of the closures applying for as long as ten weeks. Currently, most countries are pursuing these kinds of GDP-suppressing activities even more aggressively than was done in 1918. The results from over 100 years ago suggest that these kinds of interventions are worth the cost.
Robert Barro is Paul M. Warburg Professor of Economics at Harvard University and a visiting scholar at the American Enterprise Institute.
RobertJ. Barro, José F. Ursúa, and Joanna Weng, “The Coronavirus and the GreatInfluenza Pandemic: Lessons from the ‘Spanish Flu’ for the Coronavirus’sPotential Effects on Mortality and Economic Activity,” National Bureau ofEconomic Research, working paper 26866, March 2020.
For asurvey of this literature, see Thomas J. Kniesner and W. Kip Viscusi, “TheValue of a Statistical Life,” forthcoming in Oxford Research Encyclopedia ofEconomics and Finance, 2019.
SeeHoward Markel, Harvey B. Lipman, J. Alexander Navarro, Alexandra Sloan, JosephR. Michalsen, Alexandra Minna Stern, and Martin S. Cetron, “NonpharmaceuticalInterventions Implemented by US Cities During the 1918-1919 Influenza Pandemic,”Journal of the American Medical Association 298 (6): 644-654.
如果购买 尝鲜版 （可以看本报告前两部分，“宏观经济信心”和“宏观经济指标”）， 请扫码支付19元。
如果购买 完整版 （可以看报告5个部分的全部内容）， 请扫码支付29元。