Who is hardest hit by a pandemic? Racial disparities in COVID-19 hardship in the U.S.

ABSTRACT

The COVID-19 pandemic disrupted Americans’ lives and livelihoods from a wide range of socioeconomic and public health perspectives, some racial/ethnic groups more profoundly than others. Beyond the statistics of confirmed cases and deaths, however, we do not have sufficient evidence to explain the racial disparity in pandemic hardships. By using the Household Pulse Survey, a nationally representative data timely deployed by the U.S. Census Bureau, this paper measures and examines five dimensions of pandemic hardships – food insufficiency, employment income loss, housing instability, health problem, and school closure – and their racial disparities. This paper finds a pervasive racial disparity across those individual hardships as well as compounded types of hardships (e.g. income loss with food insufficiency at the same time), with commonly higher odds of experiencing hardship for non-Hispanic blacks and Hispanics than non-Hispanic whites. The effects of small business closures and reopening policy are unequal across racial/ethnic groups, implying that non-Hispanic blacks and Hispanics may be left behind the benefits from state reopening and socioeconomic recovery. These findings can shed new light on the role of demographic and socioeconomic characteristics as well as statewide contextual variables in explaining racial disparities in the pandemic hardships and suggest new possibilities for COVID-19 research.

Highlights

1. The COVID-19 pandemic disrupted Americans’ lives from a wide range of socioeconomic and public health perspectives

2. Risk of experiencing pandemic hardships – food insufficiency, employment income loss, housing instability, health problem, and school closure – is higher for non-Hispanic blacks and Hispanics than non-Hispanic whites

3. Pervasive racial disparities exist across compounded types of hardships, including income loss with food insufficiency at the same time

4. The effects of small business closures and reopening policy are unequal across racial/ethnic groups

5. Non-Hispanic blacks and Hispanics may be left behind the socioeconomic recovery from the COVID-19 pandemic

KEYWORDS:

• Coronavirus (COVID-19)
• pandemic hardship
• racial disparities
• small business closure
• state reopening policy
• Household Pulse Survey

Acknowledgement

The author benefited greatly from insightful comments by the editor and reviewers, particularly when academic conferences and scholarly discussions are limited due to the COVID-19 pandemic.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 U.S. trends in COVID-19 are available at leading federal agencies and institutes, including but not limited to CDC COVID Data Tracker (CDC, 2020), U.S. Census Bureau COVID-19 Data Hub (U.S. Census Bureau, 2020), Johns Hopkins Coronavirus Resource Center (Johns Hopkins, 2020), Brookings COVID-19 Updates (Brookings, 2020), and the New York Times database (New York Times, 2020). A curated view into the early-stage COVID-19 research is also available from the U.S. National Bureau of Economic Research (NBER, 2020) and the Social Science Research Network (SSRN, 2020).

2 These New York Times’ analyses match an independent and open database titled COVID-19 Racial Data Tracker (COVID Tracking Project, 2020).

3 As noted by Manville et al. (2020), the structure of the Household Pulse Survey data is somewhat unusual because it allows a respondent to answer the survey for a maximum of three consecutive weeks. This means some samples of the data are repeated respondents – for either two or three consecutive weeks – while other samples are single-time respondents. Manville et al. (2020) points that ‘there is no single obviously-correct approach’ for this data structure and finds very similar estimation results from alternative models in their study on housing instability during the COVID-19 pandemic.

4 Phase 1 of HPS ranges from April 23 to July 21 (12 weeks in total) and phase 2 ranges from August 19 to October 26 (9 weeks). Note that week 1 of phase 1 and all weeks of phase 2 cover two weeks but the survey calls these survey periods ‘weeks’ for continuity with the other survey weeks.

5 Examples of near real-time COVID-19 data include monthly Current Population Survey (CPS), Bureau of Labor Statistics (BLS)’s employment data, U.S. Census Bureau’s Small Business Pulse Survey, and smartphone-based mobility data.

6 Universe of the Household Pulse Survey (HPS) is adult population aged 18 and over, and thus population aged 17 and under are not considered.

7 As robustness-check, temporal lags of one-week and two-week were applied to the state reopening policy, which turned the reopening variable insignificant. This means that the incidence of food insufficiency reduces soon after the execution of state reopening policy.

8 Among a wide range of COVID-19 policy areas including lockdown (Aum, Lee, & Shin, 2020) and tracking technology (Sonn, Kang, & Choi, 2020), this paper focuses on pandemic policies for basic necessities.