Originally Posted: July 2, 2020
DALLAS (SMU) – The “back to the city” movement popular with Millennials has resulted in the revitalization of many urban areas, but has also made these city dwellers living in close proximity, frequently sharing rail cars and buses, more susceptible to risk during a pandemic. Population density and dependence on public transportation are just two of the factors that came into play in a new statistical study of how 3,000 U.S. counties have fared during the COVID-19 pandemic.
Produced by SMU economist Klaus Desmet and his UCLA colleague Romain Wacziarg, the working paper issued by the National Bureau of Economic Research confirms some previous assumptions about the spread of COVID-19, but raises intriguing questions, such as: How have counties that more strongly support President Donald Trump managed to elude the worst of the COVID-19 outbreaks? Why are locations with a greater share of highly educated people also seeing higher rates of infection?
As COVID-19 spread quickly in New York in March, many scientists theorized it would just be a matter of time before similar rates of infection impacted U.S. cities and states farther West and South. But Desmet and Wacziarg wanted to delve more deeply, so from March 15 to May 26, 2020 they examined variation in COVID-19 cases and deaths on a daily basis using two approaches:
- comparing counties day by day, providing snapshots of the drivers of disease
- comparing counties at the same stage in terms of days since cases and deaths reached a certain threshold per capita.
The two economists wanted to understand whether spatial variation in disease severity is mostly related to timing — proceeding on the theory that all counties would eventually succumb to the same infection patterns as New York; or whether spatial variation in cases and deaths reflects underlying fundamental differences across locations — measuring the impact of correlates such as population density, modes of transportation, housing arrangements, age, race and health.
The data show that it’s not just about timing: places differ in fundamental ways that make some more susceptible to the disease. One key predictor of how badly a county fares is density. When talking about density, it does not simply refer to how many people we can fit in a square mile, but rather to the likelihood of meeting other people.
For example, use of public transit and living in multi-unit housing worsen the outlook substantially. Similarly, a higher share of the population residing in nursing homes turns out to be more important than underlying health conditions and age per se. And connectivity matters , including density but also proximity to large international airports. Ideal circumstances for economic productivity in normal times make us vulnerable during a pandemic. These findings are consistent with standard epidemiological models of infectious disease: the most important driver of disease spread is the infection rate — the number of individuals an infected person interacts with long enough to pass on the disease – and that increases with connectivity.
But there are other relevant findings. “Locations with a higher share of African Americans fare worse in the incidence of COVID-19 as well as counties with many members of minority groups, such as Hispanics, are disproportionately impacted , as are counties with many poor people,” said SMU’s Desmet, an SMU economics professor who collaborated with UCLA economist Wacziarg. Desmet hypothesized that the higher incidence could be due to underlying health conditions or the types of jobs they performed.
“Counties with a higher percentage of Trump voters do better. This last result has to be interpreted with much caution: it does not mean that voting for Trump causes less disease, but rather that counties that vote more for Trump have other characteristics that lower COVID-19 incidence,” said Desmet. “One obvious such characteristic (of Trump voters) is lower density, but this finding holds up even after taking into account differences in density. Independently of what exactly is driving this result, it can go some way toward explaining the partisan divide on policies of lockdown and reopening.”
As we wait for more effective treatments and a vaccine for COVID-19, what is the appropriate course of action?
- finding ways to reduce the negative effects of our dense way of living.
- protecting nursing homes by turning them into “bubbles”.
- reducing unnecessary travel.
On the policy side, the findings by Desmet and Wacziarg suggest that a once-size-fits-all approach to lockdowns and reopenings is misplaced. The question of whether we should reopen the economy eludes a simple “yes” or “no” answer. It depends on local conditions and specificities. More susceptible places should exercise more caution, while locations with characteristics less vulnerable to the virus are better positioned to reopen sooner and to a greater extent. READ MORE