As we approach the two-years mark since the onset of COVID-19, countries around the world continue to struggle with the health and economic effects of the pandemic, many facing their third, and even their fourth wave of infections. However, within each country, not all areas have been affected with the same intensity, with some places considerably more vulnerable than others.
Researchers around the world have tackled these diverse outcomes, finding that much of the variation is explained by the effect of policy interventions, like mobility restrictions or vaccination campaigns. However, there is also a complementary explanation. A large share of the differences can be explained by pre-existing city characteristics. Understanding these vulnerabilities is important, both for fighting the continuing COVID-19 threat and for tackling future epidemics.
Much of the evidence on which local characteristics help explain COVID-19 outcomes is based on data from the US and other high-income countries. The results may be very different in low and middle-income countries, like most of those in Latin America and the Caribbean, due to factors like higher levels of residential crowding, labor informality, and economic inequality. In a recent working paper, I examine the correlates of COVID-19 deaths per capita across Brazilian cities during the first year of the pandemic, and contrast the results with equivalent US estimates. The analysis reveals that some cities characteristics are strongly associated with a higher local impact of COVID-19 in both countries. Other characteristics have a sharply different connection to COVID-19 deaths in Brazil than in the United States.
One key similarity in both countries is that, across time, population density is strongly associated with higher deaths per capita. This can be seen in Figure 1 (top-left), which shows the correlation between density — measured as the population living within 1 km of the average person of the city — and cumulative COVID-19 deaths per capita in Brazil throughout the first year of the pandemic, controlling for other city characteristics. This is likely linked to the fact that density is conducive to a higher number of human interactions, and interactions in turn drive infections.
Figure 1. Correlations of Population Density and Median Income with COVID-19 Deaths and the Tendency to Stay at Home across Brazilian Cities
Higher-Income Cities in Brazil versus the US
But not all the drivers of the local impact of COVID-19 in Brazil resemble those from the US. In the US, cities with higher median income had, controlling for all other factors, fewer COVID-19 deaths per capita during most of the first year of the pandemic. In Brazil, the opposite was true. This seems to be at least partially explained by the fact that, contrary to the US, the population from richer and denser cities in Brazil was relatively less likely to stay at home than the national average (Figure 1, right column). As in much of Latin America and the Caribbean, large shares of the population in Brazilian cities live hand-to-mouth, and don’t have enough savings to support prolonged periods of economic inactivity. Higher-income cities tend to have a higher demand for goods and services and therefore support more local jobs. Since a large share of these jobs cannot be performed from home, this creates more incentives for mobility and human interactions, which in turn can expose individuals to additional infections.
This does not mean that high-income individuals were relatively more mobile during the pandemic in Brazil. International evidence consistently suggests that socioeconomically disadvantaged people were less likely to reduce their mobility, and more likely to continue using public transportation. The results likely reflect a higher mobility of socioeconomically marginalized workers (including domestic workers and informal workers generally), who live in high-median-income cities in Brazil. Moreover, these results are not driven by differences in policy responses across cities: the study finds that, holding other covariates constant, local governments in higher-income Brazilian cities were not more or less likely to implement COVID-19 containment policies.
Socioeconomic Vulnerabilities Loom Large
Not all the city characteristics that correlate with more COVID-19 deaths per capita also correlate with increased mobility. In Brazil, cities with high socioeconomic vulnerabilities were disproportionately affected by the pandemic during the first wave, despite the fact that their population was more likely to stay at home during much of the period of the study. As shown in Figure 2, this was true for cities that had a relatively larger share of their households located in favelas (slums), and those that also had higher levels of residential crowding, as captured by the average number of people per room in the local households. In both cases, the effect on the number of deaths per capita was disproportionately large in the early months of the pandemic, particularly during the first wave.
The study shows that some cities were consistently more vulnerable to COVID-19 during the first wave of the pandemic and that these vulnerabilities are related to long-standing city characteristics. The dynamic is likely to continue throughout the pandemic. The city characteristics that correlate with COVID-19 outcomes can vary from country to country and over time. Some of this has to do with how city characteristics shape the mobility — and thus the exposure to the virus — of the local population. But others are not related to mobility, and reflect other long-standing socio-economic challenges, including deficient housing and work conditions. As the threat of new variants of the virus continues, the results from this research support the case for geographic prioritization of containment policy responses. Emphasizing vaccination and other preventive efforts in more vulnerable cities and neighborhoods may boost the effectiveness of these initiatives.
Figure 2. Correlations of the Share of the Population Living in Favelas and Residential Crowding with COVID-19 Deaths and the Tendency to Stay at Home across Brazilian Cities