Trends in Florida’s 2020 COVID-19 Experience: Race, Ethnicity, and Nativity

by Patrick Bernet, PhD


With Florida’s COVID-19 infection rates still raging in early 2021, this second offering from a three-part miniseries explores the influence of race, ethnicity, and nativity.  Each has been linked to outcomes at the individual level, but as community level characteristics, the effects can differ dramatically. (The first entry in this trilogy focuses on the impact of population age structure: Trends in Florida’s 2020 COVID-19 Experience – JPHMP Direct).

Race

Florida counties with higher proportions of Black residents generally experience higher 2020 COVID-19 infection rates, as evidenced in the upward trend in county data points in Figure 1. For each one percent increase in Black population share, there are approximately 100 more infections per 100,000. The increase is not uniform; infections increase rapidly at lower Black population proportions, then slow down beyond a 12-15% Black population share.

Figure 1. COVID-19 infection rates by population proportion Black 

Source: Author’s analysis of COVID-19 case data for 2020 from the Florida Department of Health, using American Community Survey population demographics to compute rates per 100,000 and population proportion Black. The horizontal axis represents the proportion of the population Black. The vertical axis represents total (all age) county infection rates. To facilitate viewing, several points were cropped out: Gadsden County with 55% Black & 8,549 infections per 100,000 and Lafayette County with 13% Black and 16,696 infections per 100,000. 

In any ecological study, it cannot be determined if it is a Republican or Democrat or Black or White individual who is more likely infected, only that the county as a unit has higher overall infection rate. Similarly, voter preferences are just part of the underlying societal dynamics revealing in the table in Figure 1. With voter support averaging 65%, these more-Republican voting counties had much lower proportions of Blacks (9%) and Hispanics (13%), roughly half the level of the remaining counties. Further, these 45 counties were home to just 13% of all Black Floridians, concentrating the remaining 87% in just 22 less-Republican (blue) counties. None of this infers that voter choice is a root cause of the circumstances or disease outcomes; it does, however, serve as a reliable indicator of much more complex underlying societal chasms.

County political preferences add a distinct layer to this relationship. Representing two-thirds of the state, the 45 counties with highest Republican voting preference in the 2020 presidential contest (red dots) are consistently higher than the 22 less-Republican leaning counties (lowest tercile, in blue). The size of the gap is easiest to understand as the vertical gap between more- and less-Republican leaning counties.  For instance, for counties with Black population proportions of 20% or more, there are 2,000 to 3,000 more infections per 100,000 among Republican leaning counties.

Hispanic 

The relationship between Hispanic population proportions and county COVID-19 infection rates in Florida is best viewed in two parts. Among the counties with Hispanic proportions below 15%, there is no relationship between ethnic population shares and COVID-19 infection rates. But among the counties with higher proportions, infection rates start low and increase with Hispanic population share.

The counties with Hispanic proportions below 15% form a vertical column of no distinct shape or slope (each county is represented by dots Figure 2). This implies there is no significant difference in COVID-19 infection rates between counties that are 5% Hispanic or 10% Hispanic. What is curious is the complete reset at 15% Hispanic where infection rates start well below the state’s 6,360 per 100,000 average, then gradually rise with Hispanic population proportions. It is as if the relative size of a counties’ Hispanic community makes absolutely no difference to infection rates until a threshold level.
Figure 2. COVID-19 infection rates by population proportion Hispanic

Source: Author’s analysis of COVID-19 case data for 2020 from the Florida Department of Health, using American Community Survey population demographics to compute rates per 100,000 and population proportion Hispanic. The horizontal axis represents the Hispanic proportion of the population. The vertical axis represents total (all age) county infection rates. To facilitate viewing, one point was cropped out: Lafayette County with 14% Hispanic and 16,696 infections per 100,000. 

The rise in infection rates with higher Hispanic proportions above that 15% threshold does not appear associated specifically with the Cuban community. The counties with higher shares of residents born in Cuba are distributed throughout the vertical range of infection rates (green dots in Figure 2 representing highest tercile ranked by Cuban nativity). That said, Miami is the extreme point on the right: more than half Hispanic, with many hailing from Cuba and infection rates almost twice state averages.

Nativity

Given the high proportion of Hispanics living in Florida, it would not be surprising if a foreign birth perspective also yielded a sharp V-shaped relationship between 2020 COVID-19 infection rates with population proportions. However, infection rates fall gradually with higher proportions of foreign-born residents, then increase gradually above 15-20% (see Figure 3). To identify potential causes for this shape, and having just looked at Cuban nativity, counties are shaded based on proportions of residents hailing from Europe or Asia (represented by blue dots in Figure 3-A). Most such counties had lower-than-average infection rates, which nonetheless mimic the overall decline and incline pattern. Though the inflection point is less sharp from this nativity perspective than seen in ethnicity, the same 15-percent level serves as the pivot. A similar pattern appears when nativity proportions are limited to Cuba or Latin America, meaning the position of the inflection point cannot be attributed to any one region of birth.

Figure 3. COVID-19 infection rates by population proportion foreign-born

A. Infections by % foreign-born with focus on European and Asian nativity

B. Infections by % foreign-born with focus diversity of foreign-birth sources

Source: Author’s analysis of COVID-19 case data for 2020 from the Florida Department of Health, using American Community Survey population demographics to compute rates per 100,000 and population proportion of foreign-born residents hailing from Europe or Asia (panel A), or concentration of foreign-birth source countries and regions (panel B). The horizontal axis in all panels represents the foreign-born proportion of the population. The vertical axis in all panels represents total (all age) county infection rates. Panel B concentrations are based on the Herfindahl-Hirschman Index showing the relative concentration of foreign-birth sources among a county’s residents. To facilitate viewing, one point was cropped out: Lafayette County with 5% foreign-born and 16,696 infections per 100,000.

If the specific country of birth does not seem to explain the upturn, concentration by county of birth may. The Herfindahl-Hirschman Index indicates which counties have their pool of foreign-born residents dominated by just one or two nations or regions of origin (yellow dots in Figure 3-B), and which have a more balanced draw (multicolored cubes). Those counties drawing foreign-born residents from a small number of places form the upward sweeping arm. The increasing infections, then, may be more a function the lack of nation of origin diversity in a county’s foreign-born residents rather than the specific country or region of birth. Paradoxically, the higher the proportion of foreign-born residents in a county, the fewer homelands they are drawn from. If such concentration is associated with nativity-based enclaves, this has significant implications in controlling pandemic spread. Enclaves can insulate residents from outgroup infection surges courtesy of limited interactions with “the outside world.” But once an infection gets in, the tight ingroup social networks can be an accelerant. The latter appears to be the dominant influence. 

Conclusion

The human and financial costs of pandemics are most effectively and efficiently minimized if infections can be avoided in the first place. Even this short article demonstrates that bumper sticker solutions are too myopic; it is not just race, nor ethnicity, nor foreign birth. It’s much deeper. Pandemics travel in waves and, like tsunamis, the damages done are not foretold on the surface but are defined by what lies beneath.

RECOMMENDED:


Patrick Bernet, PhD

Patrick Bernet, PhD, is an Associate Professor at Florida Atlantic University. Dr. Bernet researches pregnancy outcomes, focusing on disparities, program cost effectiveness, and the impact of bias on health from the perspectives of behavioral economics and financial analysis. He also studies the demographics of the COVID-19 pandemic. In addition to Dr. Bernet’s research, he also serves on maternal and child health committees in both Broward and Palm Beach counties and is a volunteer Medicare insurance counselor.

This site uses Akismet to reduce spam. Learn how your comment data is processed.