7 lessons from the early days of generative AI

The bottom-line benefits of second chance hiring

Federal spending was responsible for the 2022 spike in inflation

Credit: Rob Dobi

Ideas Made to Matter


The cost of uncoordinated responses to COVID-19


As federal, state, and local governments have reopened businesses and relaxed social distancing and business closure orders, new research documents the heavy consequences of uncoordinated approaches to COVID-19 policies.

A recently published paper by MIT Sloan professors and a team of researchers found COVID-19 policies from state and local governments can significantly impact health and behavior in other communities — both nearby areas, through travel, and geographically distant places, through digital and social connections.

The paper, “Interdependence and the Cost of Uncoordinated Responses to COVID-19," was published at the end of July in the Proceedings of the National Academy of Sciences.

Using an analytical model to gauge the cost of uncoordinated responses to the coronavirus pandemic, the study found that uncoordinated policies can decrease total welfare by 69%.

This is the result of significant geographic and social “spillovers” as people, ideas, and media move across borders, the study found. Some states are taking geography into account, banding together into regional groups to coordinate reopening plans, though coordination has been difficult. Yet social connections, including those made possible by digital technology, are also powerful influences — especially now, as people spend more time online during the pandemic and look to others for social cues about how to behave.

States connected geographically or socially that adopt the same stay-at-home and business closure policies see a 13% reduction in location visits and a 9.1% reduction in the fraction of people leaving home, the study found. And the researchers’ model found that a third of a state’s social and geographic peer states adopting a shelter-in-place policy is as effective at reducing mobility as the state implementing its own shelter-in-place policy.

“We’re all in this together,” Aral, the director of the MIT Initiative on the Digital Economy, said during a recent webinar. “We can’t have regulatory policies that are not in lockstep with each other.”  He said the study has been shared with state and local government officials, in hopes that it inspires greater coordination.

For the study, researchers looked at daily, county-level data on shelter-in-place and business closure policies; anonymized movement data from more than 27 million mobile devices; social network connections among 220 million Facebook users; daily temperature and precipitation data from 62,000 weather stations; and county-level census data. 

Within a given county, the researchers found that shelter-in-place policies led to a 3.2% decrease in the number of devices leaving their homes, and a 6% decrease in the number of locations visited.

But that estimate does not account for geographic spillovers — the effect that policies in nearby areas have on a given area. Accounting for those spillovers, stay-at-home orders were less effective, reducing the number of mobile phones leaving home in a county by 2%. But when half of a county’s geographic abutters also implemented shelter-in-place orders, it further reduced that number by another 1.4%.  Similarly, taking into account geographic spillovers, a shelter-in-place order reduces the number of locations visited in a county by just 4%. When half of a county’s geographic abutters also implemented shelter in place orders, it further reduced the number of locations people visited by 2.3%. 

The researchers also looked at how travel between two adjacent counties was affected when one county implemented a shelter-in-place policy and the other did not. Travel from the implementing county to the non-implementing county increased by 0.55%, while travel from the non-implementing county to the implementing county decreased by 1.2%. When both implemented a policy, travel between the two counties decreased by 0.51% — validating the importance of coordinating geographically connected regions to reduce travel, the researchers said, including from “closed” areas to “open” areas.

When Georgia resumed economic activity before other states in late April, it saw a 13% increase in out-of-state visitors, Aral said.


All states that influence each other adopting the same shelter and closure policies would result in a 13% reduction in locations visited.

Geographic proximity isn’t the only factor. Researchers also found that people are significantly influenced by the behavior of their peers in other states when they are calibrating their behaviors and choices. Some states are especially connected to others through travel and digital or social connections — for example, the study found that Florida’s mobility is most affected by New York’s shelter-in-place policies.

To map these connections, researchers created coordination maps for all 50 states that show the 20 states with which it has the most connections — in other words, which states should consider coordinating responses. For example, Massachusetts was found to be most influenced by California and New Hampshire, while California is most influenced by Texas, New York, Washington, and Florida.

The interdependence estimates were also combined with state populations to find the most influential states whose shelter-in-place policies would lead to the greatest changes in movement around the U.S. The results were highly correlated, but not equal to, a state’s population size, the study found. California was found to be the top potential influencer, followed by Texas, Florida, New York, and Illinois.

Taking into account social and geographic spillovers, all states that influence each other adopting the same policies would result in a 13% reduction in locations visited and a 9.1% reduction in the fraction of devices leaving home — an effect twice as large as geographic proximity alone. “These social spillovers may be even more relevant to the spread of COVID-19 as shelter-in-place orders have increased our reliance on digital connections, creating record-breaking usage of social media and video conferencing to maintain our social ties across geographic distance,” the researchers wrote.

Read more about coronavirus research 

Going forward

States are navigating through various reopening plans, with some states pausing or reversing reopening plans, and more coronavirus-related closures and openings will take place in the future. Divergent policies have been enacted, Aral said — like when Illinois announced that it would extend its stay-at-home policies on April 23, while the next day a neighboring state, Iowa, announced plans to allow many businesses to reopen. More recently, states have issued a patchwork of travel restrictions, with some states requiring visitors to quarantine and others opting not to restrict travel.  

The study found that considering these large spillovers, utility can be up to 69% lower with uncoordinated responses, making efforts to curb the virus significantly less effective. States also might have to compensate for another’s looser restrictions by imposing stricter, more costly policies to achieve their desired targets. But when states coordinate, spillovers help them achieve their targets as they provide “free treatments” as the states positively influence each other.

“We need nationally coordinated efforts against the pandemic,” Aral said.

The paper’s other authors are David Holtz, Michael Zhao, Seth G. Benzell, Cathy Y. Cao, M. Amin Rahimian, Jeremy Yang, Jennifer Allen, Avinash Collis, Alex Moehring, Tara Sowrirajan, Dipayan Ghosh, Yunhao Zhang, Paramveer S. Dhillon, and Christos Nicolaides.

Read "Interdependence and the Cost of Uncoordinated Responses to COVID-19"

For more info Sara Brown Senior News Editor and Writer