COVID-19 and Latino Immigrants

Written by José Moya (Professor of History & Director the Forum on Migration, Barnard College; Director, Greater Caribbean Studies Center, ILAS), with research assistance by Monica Trigos (MPA student, SIPA)

Early in 2020, journalists and the lay public would have predicted, if asked, that Coronavirus was going to be particularly lethal among Latino immigrants.

The prediction would have rested on three types of assumptions. The first, on social class, is both common and accurate. Latino immigrants do have lower levels of education and income than both the native-born population and other immigrants.  The second type–about big, multi-generational Latin American and Latino families—is common but not particularly accurate.  Women actually have fewer children on average in eight Latin American countries than in the US, and fertility rates are below the 2.1 replacement rate in fourteen of the twenty-one countries in the region, meaning that their populations are declining rather than growing (Graph 1). In the US, Latinas do have a slightly higher fertility rate than other groups, but still under the replacement level.  Latinos are more likely to live in multi-generational households than non-Latino-Whites but less so than African- and Asian-Americans.[1] The latter have the highest rate but also the highest levels of income and education of any ethno-racial group in the US, undermining the facile assumption that links size of household with low socioeconomic status. The third type of assumption—that because Latinos are less educated and poorer they would have worse health outcomes—is common, logical, usually correct, but wrong in this case.

The predictions of public health experts, thus, would have been the opposite of those of journalists and the lay public: that Coronavirus was going to be least lethal among Latino immigrants. Medical scientists have long been aware of what came to be known as the “Hispanic epidemiologic paradox”: that Latinos, and particularly Latino immigrants, tend to have better health outcomes than the general population, and even than non-Latino-Whites, despite having lower average income and education (thus the “paradox”).[2] Moreover, the health advantage is noticeably strong among those born in Mexico, precisely one of the poorest and least educated nativity groups among Latinos. Part of this is due to being a younger population.  However, the advantage remains after controlling for age. The most recurrent explanation consists of a combination of the “healthy immigrant hypothesis” (that persons with poor health are less likely to emigrate, which should apply even more when one has to walk across deserts as many Mexicans and Central Americans do); and the “salmon hypothesis” (that older and sicker persons tend to return to their place of origin).[3]  Other explanations have included growing up with healthier diets, lower levels of smoking, more active life styles, and stronger family ties.[4]   Another consistent finding demonstrates that the Latino health advantage declines with length of residence in the US and with American birth, which undermines the automatic association of assimilation with betterment.

The predictions of the public health experts would have been spot on and form part of the explanation for another Hispanic paradox: that in the Coronavirus epidemic, Latinos have both the highest infection rates and the lowest mortality rates.   Moreover, data suggest that the healthiest Latinos—the immigrants—suffer the highest rates of infection and that this is the main explanation for Latinos’ overall higher infection rate.  Coronavirus infection rates among Latinos are highest in states like South Dakota, Iowa, and Oregon where immigrants, particularly recently arrived and undocumented, account for a high proportion of the Hispanic population (Table 1).  Conversely, the infection rates are lowest where the Latino population is largest, mostly native-born, and long-established, like Arizona, Texas, California, and particularly New Mexico, where 49% of the population is Hispanic, 83% of these are native-born, and Hispanics account for only 22% of the Coronavirus cases—a striking degree of underrepresentation.  The exception to this pattern is the presence of Puerto Ricans, who are born US citizens and thus not legally immigrants, and whose sociological profile (in terms of family formation, incarceration rates, substance abuse, public health, etc.) often lies closer to that of African Americans than to that of other Latinos.  States where Puerto Ricans account for a large proportion of Hispanics (Delaware 48%, New York 43%, Rhode Island 36%) also have some of the highest rates of contagion among Latinos.

If the main explanation for the high level of Coronavirus among Latinos is the high proportion of immigrants, and particularly undocumented ones (Latinos account for 76% of these), the explanation for the high level of Covid-19 infection among the latter is manifold but not particularly surprising. Latino immigrants are underrepresented in one high-risk group: prisoners–their incarceration rate, even including the large number held just for immigration issues, is less than one-third that of the native-born US population.[5]  Other than this, they are overrepresented in just about every other high-risk group.  Latino immigrants account for 53% of agricultural workers, 44% of those laboring in meatpacking plants, and for high proportions of nursing home and direct care workers, grocery stockers, fast food employees and food deliverers, truck loaders, house cleaners, and other service occupations that make social distancing difficult.  As an Oregon Occupational Safety and Health Administration official put it, “You cannot telecommute to harvest crops.” In a double whammy, Latino immigrants mostly reside in dense urban environments or in sparsely populated rural areas but in crowded work and residential settings—planting and picking fruits and vegetables in industrialized agribusinesses or working in slaughterhouses and meatpacking plants, and living in barracks.  Latino immigrants are also less likely to have health insurance, job security, paid leave, unemployment compensation, and if undocumented, are not eligible for federal Covid-19 related assistance.  As a result, most of them cannot afford to stay at home and have had to work through the pandemic in close proximity to co-workers, customers, and the public.

Yet the “Hispanic epidemiologic paradox” has not stopped working.  Mortality among Latinos (19 per 100,000 population) is less than half that of Blacks (49) and similar to those of Asian-Americans (18) and Whites (17), (Graph 2).  Yet this measure overstates Latinos’ mortality from Covid-19 because it uses as its base the entire population rather than those infected, and Latinos’ rates of infection are significantly higher than for other groups.  If we use data on the mortality of those infected instead, Latinos appear as the outliers with a mortality (40 per 1,000) that is less than half that of Blacks (94), Whites (84), and Asian-Americans (82), (Graph 3).[6]  The “Hispanic epidemiologic paradox” not only continues to work but seems to be doing so on steroids.   Considering that the main explanation for the paradox is the “healthy immigrant hypothesis,” the difference with Asian-Americans is puzzling. After all, the proportion of immigrants in that group (59% in 2017) almost doubles that of Latinos (33%), and the level of social class selectivity in the migration—which supposedly should make them both less likely to get infected and less likely to die if infected–is much higher.  For example, only 8% of the population in India has a college degree but 72% of Indians in the US do so, while 23% of the population in Mexico have a college degree but only 6% of Mexicans in the US do.  Could it be that when it comes to selectivity by physical health, rather than class/education, having to walk through the Sonora desert presents a less porous filter than taking an airplane or overstaying visas?  Indeed, the movement across the Mexican-US border seems to exemplify a migration of the fittest rather than the “worst,” as President Trump once put it.  This makes recurrent references in the media, and even the scholarly literature, to “underlying medical conditions” as an explanation for the large number of Covid-19 cases among Latino immigrants ironic, and a blatant example of the mindless tendency to place all minorities in the same conceptual bag.


[2] The term first appeared in K. S. Markides and J, Coreil, “The Health of Hispanics in the Southwestern United States: An Epidemiologic Paradox,” Public Health Reports, 101:3 (1986):253-65.

[3] A. F. Abraído-Lanza,, “The Latino Mortality Paradox: A Test of the Salmon Bias and Healthy Migrant Hypotheses,” American Journal of Public Health, 89:10 (1999): 1543-48 found that–controlling for age, income, and education– all Latino groups had lower mortality than non-Latino-Whites, including those for whom the “salmon hypothesis” did not apply: Cubans because of their extremely low rates of return, and Puerto Ricans because their deaths are tabulated in US mortality statistics.

[4] Eileen M. Crimmins, et. al., “Hispanic Paradox in Biological Risk Profiles,” American Journal of Public Health, 97:7 (2007):1305-10.


[6] Data from