Against the backdrop of the risk of a second wave, staggering losses for business and rapidly increasing government deficits, the wisdom of shutting down large parts of the economy is now being scrutinised, as is the support for affected businesses. New research from Rotterdam School of Management (RSM), however, shows that shutting down business activity to combat COVID-19 is very effective, especially at the peak of the crisis when hospital capacity is under pressure, and even if other distancing and hygiene measures are in place. Governments should adopt a targeted approach with business shutdowns as the effectiveness of business closures declines with the scale at which it is used. The research also shows that decisions on business shutdowns should be made centrally and not locally, since neighbouring regions strongly affect each other.
Dr Dion Bongaerts, Francesco Mazzola and Professor Wolf Wagner studied the effects of business shutdowns on deaths from COVID-19 in Italy, finding that the first shutdown saved the economy €9 billion as it meant the lives of many working and non-working age people were saved.
According to the researchers: “From data on local employment structure, we found that about 17% of the economy was shut down on 11 March.
“Although the shutdown policy was centralised, it affected municipalities differently because of a greater or lesser number of businesses in the sectors subjected to compulsory closure. Comparing municipalities with high rates of business shutdowns to those with low rates allows us to isolate the effect of the shutdown from other nationwide developments.
“The results suggest that business shutdowns are effective in saving human lives: municipalities more exposed to shutdowns experienced subsequently lower mortality rates.”
The virus did not reach every municipality simultaneously, and so graphs of mortality rates show municipalities grouped according to the week that the virus arrived in the local population. These charts plot the average daily COVID-19 death rate per 100,000 people for municipalities above and below the median shutdown exposure of each group.
Graphs reveal that around 10 days after the date of shutdown, mortality rates began to decline – and the decline is more pronounced in municipalities that are more exposed to shutdowns compared with municipalities that are less exposed.
Such analysis enabled the researchers to calculate that the first shutdown saved more than 9,000 lives over a 23 days shutdown time period.
This result generally implies that there is a large societal benefit to be gained from closing businesses during a pandemic. The researchers calculated that the first shutdown could be valued at more than €9 billion saved, assuming that the average life expectancy of those who had died from coronavirus would have been another 13 years (the length of time they would have lived if they hadn’t contracted the disease), and the value of a statistical Life-Year to be approximately €80,000 a year, following existing literature.
The researchers also note that business closures may limit the spread of the virus outside a municipality because there are fewer commuters, as well there being general travel restrictions during the sample period).
They found that shutdowns in centres of business consistently reduced mortality rates in neighbouring municipalities. The beneficial effect was large – in fact, it was comparable to the effect of a municipalities' own shutdown.
The beneficial spillover effects of shutdowns also had a large effect on people who are very unlikely to be in work, such as the elderly.
Finally, the analysis suggests that these benefits are relatively more pronounced in the retail sector, because shutting down this sector contributes most to the decline in mortality rates – and because online shopping, and ordering deliveries from local suppliers, provides an obvious alternative.
Ultimately, the results are consistent with a lower effectiveness of a second shutdown, suggesting large benefits from targeted, rather than widespread, business closures.
