Which Workers Will Benefit from AI?
Earlier job transformations prompted by the onset of manufacturing and IT may hold clues for how fast the labor market will adapt to the latest technology revolution.
Which Workers Will Benefit from AI?Containing the spread of the coronavirus has come at a staggering price. Despite unprecedented monetary and fiscal support from governments, the World Bank predicts global economic output will dive 5 percent in 2020 and per-capita income will shrink in the largest percentage of countries since 1870. Now a new wave of COVID-19 cases in the United States is threatening countrywide reopening plans and pressuring policy makers to impose fresh lockdowns—at risk of deepening the downturn.
Yet, research from Chicago Booth’s Austan D. Goolsbee and Chad Syverson suggests that US government restrictions had only a mildly chilling effect on the economy. They argue that commercial activity had already begun to collapse before late March and early April, when many jurisdictions imposed shutdowns. In that case, government lockdowns didn’t crash the economy—fear of the virus did.
Using foot-traffic data at 2.2 million businesses, the researchers compared consumer behavior across similar stores located in the same metro area but across counties where shelter-in-place orders were issued at different times. Those timing differences allowed the researchers to measure the impacts of lockdowns.
While overall consumer traffic plunged by 60 percent during the crisis, official social-distancing restrictions accounted for only one-tenth of the drop, the researchers find. That is, traffic fell just about as much at stores in counties that hadn’t imposed shelter-in-place orders as at stores in counties in the same metros that did impose them. “Our results suggest that legal restrictions per se didn’t have a big effect on people’s choices to stay at home,” says Syverson. “That means the biggest thing driving the virus recession is the virus itself, or at least people’s fear of being infected by it.”
The researchers used data from SafeGraph, which compiles the number of visits to various business addresses using information drawn from nearly 45 million cell-phone users—about 10 percent of all devices in the US. The data set included businesses in 110 industry classifications where consumer visits are a plausible proxy for economic activity, such as restaurants and retail stores. While policy makers often rely on state-level data, the researchers dug down to the county level to capture consumer behavior in nearby metros that imposed restrictions at different times—or, in the case of eight states, none at all.
Just as the shutdowns may have had less of an impact than often stated, removing them may not ignite a roaring recovery, the researchers warn.
The researchers also find that consumers tended to avoid stores that were larger and busier before the pandemic in favor of smaller ones with fewer patrons—further suggesting that fear, not government lockdowns, drove the economic decline. While all businesses lost traffic from January through the trough in mid-April, the researchers estimate that the smallest 20 percent attracted at least 60 percent more traffic than the rest. And counties with more COVID-19 cases exhibited an even greater relative shift toward smaller businesses.
Just as the shutdowns may have had less of an impact than often stated, removing them may not ignite a roaring recovery, the researchers warn. Jurisdictions that repealed restrictions saw only modest rallies in economic activity, they find, arguing that the effect of repealing a shelter-in-place order “is statistically the mirror image of imposing one, and certainly no larger. The point estimates imply economic activity fell 8 percent when governments instituted the orders and rose 5 percent when they repealed them.”
Goolsbee and Syverson concede that their results do not measure the impacts of shelter-in-place orders on industries such as manufacturing, for which foot traffic doesn’t approximate economic activity.
While people’s behavior may have been driven by fear rather than lockdowns, this doesn’t imply that sweeping lockdowns are ineffectual, Syverson notes. People will take precautions to avoid catching the virus, “but only up to the point where their own cost of being careful—that is, annoyance and inconvenience—equals the benefit of their own reduction in the likelihood of catching the disease,” he says. Their calculations won’t fully account for the costs of spreading the disease. Therefore, he sees a sound economic argument for restricting face-to-face interactions during a pandemic, “above and beyond the amount people will restrict themselves.”
Austan D. Goolsbee and Chad Syverson, “Fear, Lockdown, and Diversion: Comparing Drivers of Pandemic Economic Decline 2020,” Working paper, June 2020.
Earlier job transformations prompted by the onset of manufacturing and IT may hold clues for how fast the labor market will adapt to the latest technology revolution.
Which Workers Will Benefit from AI?Research finds China’s digital coupon programs were a cost-effective way to boost spending.
Why China’s Pandemic Stimulus Worked Better Than the US’sResearch suggests frictions in the building industry could be slowing or halting projects.
US Construction Has a Productivity ProblemYour Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.