Demographic Disparities in COVID-19 Disruptions
Summary
An article titled Demographic Disparities in COVID-19 Disruptions: What Has Shaped Them?, published in the Federal Reserve Bank of St. Louis Review, Fourth Quarter 2023, by a lead analyst in the bank's Institute for Economic Equity. It uses the Community Impact Survey conducted by the Federal Reserve System in August 2021 to compare COVID-19 disruption reported by nonprofit organizations serving communities of Color and White communities. The article reports that communities of Color were 20 percentage points more likely to be significantly disrupted. It finds that unequal access to government relief was as important as challenges returning to work, and that together these challenges explain approximately 30 percent of the disruption gap. It closes with regression tables and appendix figures.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
Full text
ARTICLE
Demographic Disparities in COVID-19 Disruptions: What
Has Shaped Them?
Violeta A. Gutkowski
Abstract
This article uses the Community Impact Survey implemented by the Federal Reserve System in 2021 to identify
COVID-19 disruptions on low- to moderate-income communities. I find that communities that were primarily of
Color were more likely to be significantly disrupted by COVID-19 than White communities. I also assess the impor-
tance of certain challenges, such as returning to work or unequal access to government relief, in shaping the observed
demographic differences in disruption.
JEL codes: D04, D63
Federal Reserve Bank of St. Louis Review, Fourth Quarter 2023, 105(4), pp. 261-79.
https://doi.org/10.20955/r.105.261-79
1. INTRODUCTION
In the past years, we have seen waves of COVID-19 infections disrupting communities in the United States.
The pandemic has interfered with lives across the country, and its effects on low- to moderate-income (LMI)
individuals and communities have been significant. Though the peak of distress has passed, people of Color
are still bearing major disturbances. Figure 1 shows that at the peak of COVID-19, distressed communities
were facing serious disruptions regardless of demographic composition. However, 18 months into the pan-
demic, communities of Color were 20 percentage points more likely to be significantly disrupted than White
communities even though the disruptions had already halved.1
This article analyzes a different and unique channel to assess the disruption gap between the two com-
munities—unequal access to government COVID 19 relief—and compares it with other possible explanations,
such as challenges returning to work. Government relief has been very critical to LMI communities (Chalise
and Gutkowski, 2021). It has kept millions out of poverty, helped people access health coverage, and reduced
hardships like the inability to afford food or meet other basic needs (Budget and Policy Priorities, 2021). Nev-
ertheless, equally distributing these funds was challenging. Holtzblatt and Karpman, 2020 show that the receipt
of Economic Impact Payments from the CARES Act occurred more slowly—or not at all—for some groups.
Having no recent history of filing taxes, lacking internet access, and being unbanked were some of the barriers
to receiving government payments. Many were also excluded from unemployment benefits. For example,
people of Color who often work in low-wage or part-time jobs, or as independent contractors, may have been
more likely to be excluded or not eligible for regular unemployment insurance (Janger, Rubin, and Singh,
2020).
1. The distinction between primarily White communities and communities of Color comes from the following survey question: “Does
the entity you represent primarily serve a community of Color?” See Section 2.1.1.
Violeta A. Gutkowski is a lead analyst in the Institute for Economic Equity at the Federal Reserve Bank of St. Louis. The author
thanks William Rodgers and Nishesh Chalise for their suggestions and support.
©2023, Federal Reserve Bank of St. Louis. The views expressed in this article are those of the author(s) and do not necessarily
reflect the views of the Federal Reserve System, the Board of Governors, or the regional Federal Reserve Banks. Articles may be
reprinted, reproduced, published, distributed, displayed, and transmitted in their entirety if copyright notice, author name(s), and
full citation are included. Abstracts, synopses, and other derivative works may be made only with prior written permission of the
Federal Reserve Bank of St. Louis.
261
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Figure 1
General Economic Disruption
Peak of Disruption Current General Disruption
Color: No Color: Yes Color: No Color: Yes
100 100
80 80
60 60
Percent percent
40 40
20 20
0 0
NOTE: The figure shows the percentage of respondents who reported disruption at the peak of distress (left panel) and at the time of
the survey (right panel) for entities serving communities that were primarily of Color and those that were not.*
*Questions on disruption were the following: “Thinking about the period of peak of distress caused by the pandemic, what level of
disruption did COVID-19 have on economic conditions in your community at that time?” “Currently, what level of disruption is
COVID-19 having on economic conditions in your community?”
Furthermore, not all small businesses had access to Paycheck Protection Program (PPP) loans due to lack
of eligibility, burdensome paperwork, and lack of knowledge and understanding of the program. According
to Liu and Parilla, 2020, it took 31 days for small businesses with paid employees in majority-Black zip codes
to receive PPP loans—7 days longer than those in majority-White communities. Lack of a bank relationship
was also a burden. The initial demand for loans significantly exceeded the ability of banks to deliver them,
and they appear to have prioritized firms with which they had a preexisting relationship (Granja et al., 2022;
Li, Strahan, and Zhang, 2020). Thus, larger firms—which tend to have ongoing bank relationships—accessed
PPP funds sooner than smaller firms, on average.
I find that unequal access to government relief plans was as important as challenges in returning to work
in explaining variation in disruption between White communities and communities of Color. Returning to
work has been particularly challenging to many due to risk of exposure, lack of childcare, or lack of public
transportation. Together, these challenges can explain approximately 30 percent of the disruption gap. These
findings highlight the importance of targeting, implementing, and distributing government relief packages
and the relevance of childcare in the US economy.
The difference in the disruption across demographics groups could be explained by a variety of factors. For
example, Maxwell, 2020 suggests that economic hardship, food and housing insecurity, and unequal access to
mental healthcare services could have contributed to explaining differences in the overall disruption caused by
COVID-19. Gemelas, Davison, et al., 2022 show that people of Color experienced larger income losses, which
could have made it harder for them to make ends meet. In addition, workers in communities of Color had
less ability to telecommute, which may have resulted in unequal health risks affecting overall employment and
labor force participation (OECD, 2022).
Differences in access to healthcare and health insurance could also have contributed to the gap. Gangopad-
hyaya, Karpman, and Aarons, 2020 show that between late April and mid-July, more than three million adults
lost employer-sponsored insurance (ESI). These losses were concentrated among Hispanic adults, young adults,
men, and adults who did not attend college. Though the reduction in ESI was partially offset by a rise in public
coverage, groups that faced the largest ESI losses also saw the largest increases in uninsurance.
The next section provides more information about the data from the Community Impact Survey (CIS)
as well as the main variables of interest. Section 3 presents the main analysis and results, followed by some
262
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
robustness exercises. I conclude in Section 4 with some discussion about the findings and further possible areas
of research toward an equitable recovery path.
2. DATA AND ANALYSIS
2.1 Community Impact Survey (CIS)
The data for this analysis come from the CIS, conducted by the Federal Reserve System in collaboration with
eight national partners in August 2021.2 The survey was designed to monitor the impact of the pandemic on
low-to moderate-income (LMI) communities and the organizations that serve them. It provides an insightful
and informative snapshot of how COVID-19 affected people and organizations as of the dates the survey was
administered.
Survey respondents included representatives of nonprofit organizations, government agencies, financial
institutions, and other organizations who serve LMI communities across the US.3 Responses were collected
through a convenience sampling method that relied on contact databases to identify representatives of these
community organizations. These representatives were invited to participate in the survey via emails, newslet-
ters, and social media posts.
The survey was organized into three different sections: respondent profile questions, questions about the
impact of COVID-19 on the entities providing services, and questions about the impact on the people and
communities these entities served. It is worth noting that respondents to this survey were service provider
organizations, and thus all of the answers about the impact on people and communities come from their per-
spective.
This article focuses on the disruption caused by COVID-19 on LMI people and communities, differences
in the impact across demographic groups, and how challenges returning to work or accessing government
funding have shaped the observed differences in disruption. Given the diversity of responding entities, I focus
on nonprofit organizations that completed the full survey (N=2291). In the next section, I describe the main
variables of interest for this work.
2.1.1 Communities of Color
Respondents were asked if they served a community that was primarily of Color (possible responses were yes,
no, or unsure). Throughout the article, I refer to “community of Color” as those who responded with yes (47
percent of responses) to this question and “White” as those who responded with no (48 percent of responses).4,5
Communities that were primarily of Color included neighborhoods where the primary population was Amer-
ican Indian or Native Alaskan (4 percent), Asian (3 percent), Hawaiian or Pacific Islander (2 percent), Hispanic
or Latino (41 percent), non-Hispanic Black or African American (45 percent), and Other (5 percent).
2.1.2 Disruption
My main variable of interest is the COVID-19 level of economic disruption on people and communities. The
survey question asked “Currently, what level of disruption is COVID-19 having on economic conditions in
your community?” (possible responses were significant disruption, some disruption, no disruption, or unsure).
Figure 1 shows the general economic disruption on LMI communities disaggregated by the demographics
served. At the peak of distress, disruption levels were high for all demographic groups, with more than 80
percent of respondents reporting significant disruption. While significant disruption almost halved in 2021 for
all LMI communities, it was 20 percentage points more likely for communities of Color to be significantly
disrupted than White communities.
While economic conditions could be interpreted differently across respondents, they provide a broad overview
of how respondents saw COVID-19 disruptions impacting their communities. To have a better understanding
of COVID-19’s impact on several dimensions, the survey also asked about disruption in different segments of
the economy that are vital for the development of LMI communities. These segments were financial stability
(related to income loss and income instability), small business (including short-/long-term closure, supply chain
disruptions, and reduced demand), access to healthcare (such as access to adequate healthcare, access to health
insurance, and mental health services), services for children (including the availability of childcare and adequate
access to K-12 education), housing stability (involving evictions, back rent, foreclosures, and homelessness), and
basic consumption needs (for example, food, household essentials, and other personal needs).
2. For more information on the survey, please visit https://fedcommunities.org/data/main-street-covid19-survey-2021.
3. There were approximately 3,700 responses from entities who answered most of the survey, including the questions related to being
a direct service provider or not and COVID-19’s level of disruption during the peak of the pandemic and at the time of the survey.
4. Five percent of responses were “unsure,” which are not included in the study.
5. The survey did not explicitly ask whether the respondent served a primarily White community.
263
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Figure 2
Significant Disruption in Segments of the Economy
Financial Stability Disruption Small Business Disruption Health Care Disruption
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Services for Children Disruption Housing Stability Disruption Basic Consumer Needs Disruption
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Color: No Color: Yes
NOTE: The figure shows the percentage of respondents who reported significant disruption in their communities in each segment of
the economy at the time of the survey, by the demographics served with 95 percent confidence interval bars.
There was some variation in responses for these different segments of the economy with significant dis-
ruption ranging between 30 and 60 percent of responses (Appendix Figure A.3). Nevertheless, communities of
Color were more likely to be significantly disrupted than White communities for all segments of the economy
surveyed (Figure 2). Appendix Tables A.1 and A.2 provide a disaggregation of responses by demographic com-
position, type of engagement with community and location, and the distribution of impact for each segment.
2.1.3 Challenges
The next set of relevant variables relate to the barriers or challenges that affected people and communities.6
Respondents were asked to rate the seriousness of the following challenges: 1) returning to work (such as
childcare, public transportation, COVID-19 exposure risk), 2) accessing government funding (lack of eligibility
or capacity for processing applications), 3) relationship with banks to access capital, 4) and applying for funds
(including complex process, burdensome paperwork, internet access).
More than 65 percent of responses indicated that returning to work was a significant challenge and 38
percent claimed that accessing government funding was a severe hurdle. Lack of a bank relationship was a
serious barrier for almost 30 percent of respondents, and more than 45 percent indicated applying for funds was
a severe obstacle. Appendix Table A.3 provides the distribution of challenges among respondents. However, this
distribution was not symmetric across demographic groups: The likelihood of these obstacles being severe was
higher for communities of Color.7 Figure 3 shows there is a statistically significant difference in the likelihood
of communities of Color reporting that these difficulties were acute relative to White communities."
Differences in barriers to returning to work could be related to unequal access to childcare due to high costs,
inflexible schedules, closures, and illnesses due to COVID-19 exposure. Lee and Parolin, 2021 find that Black,
Latino, and Asian families have been exposed to childcare closures to a much greater extent than White families.
Labor market segmentation and differences in occupations could also account for the observed differences in
6. They were asked the following question: “During the COVID-19 pandemic, how challenging were the following for the people and
communities you serve?” Respondents replied with “significant challenge,” “challenge,” “no challenge,” or “unsure.”
7. See Appendix Figure A.4 for responses across White communities and communities of Color.
264
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Figure 3
Significant Challenges by Demographic Group
RTW Sig. Challenge Acc Govt Sig. Challenge
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Bank Relat Sig. Challenge Funds App Sig. Challenge
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Color: No Color: Yes
NOTE: The figures show the percentage of respondents who reported significant challenges for the communities they serve by
demographic group served. Top left: Return-to-work challenges (such as childcare, public transportation, COVID-19 exposure risk). Top
right: Challenges accessing government funding (lack of eligibility or capacity for processing applications). Bottom left: Relationship
with banks to access capital. Bottom right: Applying for funds (including complex process, burdensome paperwork, internet access).
difficulties returning to work. Accessing government funds has been a challenge to many, particular to people
of Color due to lack of eligibility, lack of knowledge or resources to apply to them, and lack of historical bank
relationships, which favored previous connections when expediting funds. Holtzblatt and Karpman, 2020 show
that the receipt of Economic Impact Payments from the CARES Act occurred more slowly—or not at all—for
some groups. Barriers to payment receipt included having no recent history of filing taxes, lacking internet
access, and being unbanked. Autor et al., 2022 show that the PPP—another well-known federal program to
mitigate COVID-19 disruptions—incidence was highly regressive, with about three-quarters of PPP funds
accruing to the top quintile of households.
2.1.4 Control Variables
Within the respondent profile part of the survey, entities were asked whether they served primarily in urban,
suburban, or rural communities and whether or not they were offering services directly to people or small busi-
nesses (direct service providers). Suburban observations account for 13 percent of observations with a similar
demographic composition and outcome than rural areas. I exclude these observations to reduce the amount of
noise introduced in the analysis and to facilitate the interpretation of the coefficient on location. Given that the
geographic location of service as well as direct service ( or not) could have a different demographic composition,
I control for these variables in my analysis (see Appendix Figures A.1 and A.2).
3. RESULTS
In this section, I first provide supportive evidence that communities that were facing major challenges were
more likely to be significantly disrupted at the time of the survey. I then show, in Section 3.2, that the disruption
gap between White communities and communities of Color was approximately 17 percentage points and
that incorporating these challenges reduces the disruption gap by a third. Section 3.2.1 assesses the relative
importance of each obstacle in explaining the disruption gap. The main finding is that all the challenges
265
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Table 1
Challenges and Likelihood of Disruption
General economic significant disruption
(1) (2) (3) (4) (5) (6)
Urban 0.124*** 0.108*** 0.0973*** 0.105*** 0.0997*** 0.0837***
(0.0215) (0.0212) (0.0212) (0.0214) (0.0213) (0.0210)
DSP 0.0474 0.0489* 0.0443 0.0455 0.0375 0.0420
(0.0293) (0.0288) (0.0283) (0.0286) (0.0284) (0.0279)
RTW 0.198*** 0.116***
(0.0209) (0.0224)
Acc Gvt Fund 0.238*** 0.137***
(0.0209) (0.0253)
BK Relat 0.211*** 0.0716**
(0.0238) (0.0280)
Fund Appl 0.212*** 0.0813***
(0.0207) (0.0260)
Constant 0.342*** 0.225*** 0.270*** 0.305*** 0.273*** 0.192***
(0.0302) (0.0315) (0.0294) (0.0296) (0.0297) (0.0306)
Observations 2,291 2,291 2,291 2,291 2,291 2,291
R-squared 0.015 0.051 0.069 0.048 0.059 0.095
NOTE: Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1
have been similarly important in explaining the disruption gap. Section 3.3 provides results for my preferred
specification and contrasts challenges returning to work with overall access to government support during the
pandemic. Last, in Section 3.4 I perform some robustness checks on the impact on a variety of segments of
the economy to further support the main findings as well as to control for median household income in the
environment (Section 3.5).
3.1 Challenges and Likelihood of Disruption
I first provide evidence of the positive correlation between the challenges and the likelihood of being signifi-
cantly disrupted. I use the following regression:
(1) SigDisrup = α + β ∗ Challenge + γ ∗ urban + δ ∗ DSP + υ,
where SigDisrup is a dummy variable that equals one if disruption was significant and equals zero otherwise.
Challenge is a dummy variable that equals one if the respondent indicated significant challenge and equals
zero otherwise. The four challenges analyzed are returning to work (RTW), accessing government funding
(Acc Gvt Fund), bank relationship (BK Relat), and the fund application process (Fund Appl). To control for
differences in disruption due to geographic location, I incorporate Urban, a dummy variable that equals one
if the entity offered services primarily in urban areas. DSP is a dummy that equals one if the responding
entity was a direct service provider (offering services directly to individuals, families, or small business owners).
Controlling for the type of engagement with the community is important since direct service providers could
be less optimistic about disruption and recovery than indirect providers given the nature of their engagement
with the community, and thus the results could be biased (see Appendix Table A.2).
Table 1 shows that the challenges considered are relevant in explaining the variation in the disruption as
well as the location of service. The communities that were more likely to face any of these challenges were
more likely to be significantly disrupted given the positive sign and statistical significance of the coefficients.
The table shows that urban areas were more likely to be severely disrupted. Consistent with earlier findings
by Brooks, Mueller, and Thiede, 2021 and Parker, Horowitz, and Minkin, 2021, this could be due to higher
exposure risk in urban areas than rural areas as well as distinct disruptions in the labor market. Column 2
266
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Table 2
Demographics, Challenges, and COVID-19 Economic Disruption
General economic significant disruption
(1) (2) (3) (4) (5) (6) (7)
Color 0.191*** 0.169*** 0.143*** 0.131*** 0.140*** 0.137*** 0.108***
(0.0205) (0.0219) (0.0219) (0.0220) (0.0223) (0.0220) (0.0220)
RTW 0.178*** 0.108***
(0.0211) (0.0224)
Acc Gvt Fund 0.217*** 0.129***
(0.0212) (0.0252)
BK Relat 0.184*** 0.0594**
(0.0240) (0.0278)
Fund Appl 0.191*** 0.0765***
(0.0209) (0.0259)
Urban 0.0649*** 0.0599*** 0.0542** 0.0587*** 0.0545** 0.0493**
(0.0227) (0.0222) (0.0223) (0.0226) (0.0223) (0.0220)
DSP 0.0333 0.0368 0.0336 0.0340 0.0270 0.0333
(0.0290) (0.0286) (0.0281) (0.0285) (0.0283) (0.0278)
Constant 0.365*** 0.306*** 0.205*** 0.248*** 0.280*** 0.250*** 0.180***
(0.0145) (0.0300) (0.0313) (0.0293) (0.0295) (0.0296) (0.0305)
Observations 2,291 2,291 2,291 2,291 2,291 2,291 2,291
R-squared 0.037 0.040 0.069 0.084 0.064 0.075 0.105
NOTE: Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1
shows that communities that faced serious barriers in returning to work were almost 20 percent more likely
to be heavily distressed at the time of the survey. Lack of regular hours for childcare, unpredictable closures
due to COVID-19 exposure or illnesses, and inflexible schedules all could have contributed to these observed
differences.
Similarly, communities with major issues accessing government funds were also more likely to be seriously
disrupted (Column 3). Federal COVID-19 relief has helped millions of people navigate the crisis and has kept
many out of poverty. However, those who were not able to access these funds could not benefit from the gen-
erous package. In addition, communities that lacked a bank relationship (Column 4), as well as those that faced
burdensome paperwork, lack of internet, and a complex application process (Column 5), were approximately
20 percent more likely to be extremely disrupted by COVID-19. These barriers related to being unbanked,
not being within the network of lending banks, lacking knowledge, or lacking capacity were clear obstacles
in the process of applying and thus accessing funds. Column 6 incorporates all four challenges together. The
coefficients on each challenge fall; however, all four challenges remain positive and significantly different from
zero, indicating their relevance when assessing the disruption.
3.2 Challenges and the Disruption Gap
In this section I focus on the disruption gap between White communities and communities of Color and find
that the explored challenges can explain almost 30 percent of the disruption gap. To this end, I run a similar
regression as before but now include Color—a dummy variable that equals one if the community served was
primarily of Color—as one of my main explanatory variables of interest:
(2) SigDisrup = α + θ ∗ Color + β ∗ Challenge + γ ∗ Urban + δ ∗ DSP + υ.
Table 3 shows the results. The first two columns indicate that communities of Color were 19 percentage
points more likely to experience significant disruption at the time of the survey and almost 17 percentage points
when accounting for differences in location and type of engagement with the community. Columns 3 to 6
incorporate, one by one, the different challenges of interest. In all cases, the coefficients on the challenges
267
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Table 3
Demographics, Challenges, and COVID-19 Economic Disruption
General economic significant disruption
(1) (2) (3) (4) (5) (6) (7)
Color 0.191*** 0.169*** 0.143*** 0.131*** 0.140*** 0.137*** 0.108***
(0.0205) (0.0219) (0.0219) (0.0220) (0.0223) (0.0220) (0.0220)
RTW 0.178*** 0.108***
(0.0211) (0.0224)
Acc Gvt Fund 0.217*** 0.129***
(0.0212) (0.0252)
BK Relat 0.184*** 0.0594**
(0.0240) (0.0278)
Fund Appl 0.191*** 0.0765***
(0.0209) (0.0259)
Urban 0.0649*** 0.0599*** 0.0542** 0.0587*** 0.0545** 0.0493**
(0.0227) (0.0222) (0.0223) (0.0226) (0.0223) (0.0220)
DSP 0.0333 0.0368 0.0336 0.0340 0.0270 0.0333
(0.0290) (0.0286) (0.0281) (0.0285) (0.0283) (0.0278)
Constant 0.365*** 0.306*** 0.205*** 0.248*** 0.280*** 0.250*** 0.180***
(0.0145) (0.0300) (0.0313) (0.0293) (0.0295) (0.0296) (0.0305)
Observations 2,291 2,291 2,291 2,291 2,291 2,291 2,291
R-squared 0.037 0.040 0.069 0.084 0.064 0.075 0.105
NOTE: Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1
are positive and significant, between 18 and 22 percent, indicating that even within communities with similar
demographic compositions, communities that were facing these challenges were more likely to be severely
disrupted. Most importantly, the coefficient on Color decreases between 15 and 22 percent when incorporating
the challenges as explanatory variables of disruption. This is not surprising given that Figure 3 presents evidence
that communities of Color were more likely to face serious hurdles. When all four challenges are considered,
the coefficient on Color falls by approximately 35 percent. In other words, the observed differences in the
severeness of the obstacles faced by the different demographic groups can explain 35 percent of the disruption
gap.
3.2.1 Relative Importance of Challenges
In this section I perform an Oaxaca decomposition to assess the relative importance of each challenge in ex-
plaining the disruption gap between White communities and communities of Color.
Of the 19-percentage-point difference in disruption between communities of Color and White communi-
ties (disruption gap of 19 percentage points), 8 percentage points can be explained by location, direct or indirect
service, and all four challenges—this would account for 43 percent of the disruption gap. Figure 4 shows that
approximately 13 percent of the disruption gap can be explained by unequal access to government relief plans
and almost 9 percent by challenges returning to work. Lack of a bank relationship explains around 5 percent
of the disruption gap, while barriers in applying for funds explains 7.5 percent. All four challenges can explain
approximately 35 percent of the disruption gap, and location differences account for another 8 percent. While
these numbers suggest that unequal access to government funding had a larger role in explaining the disruption
gap, I cannot reject that this barrier was as relevant as the other ones (Figure 4B). In other words, I cannot
claim that this obstacle was more important than the other challenges considered in explaining the disruption
gap.
One could think that unequal access to government funding encompasses having difficulty applying for
funds or lacking bank relationships to access funds since, in the end, communities either received funds late or
not at all. Thus the correlation between these challenges might be reducing the overall importance of funding
268
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Figure 4
Disruption Gap: Explanatory Power of Challenges
A. Disruption gap explained B. Percentage explained by challenges
NOTE: The figure shows the percentage of the disruption gap explained by each variable of interest (panel A) and 95 percent confidence
intervals for each variable of interest (panel B).
Figure 5
Returning to Work versus Unequal Access to Government Funding
t!
A. Disruption gap explained B. Percentage explained by challenges
NOTE: The figure shows the percentage of the disruption gap explained by each variable of interest (panel A) and 95 percent confidence
intervals for each variable of interest (panel B).
relative to return-to-work challenges.8 To address this concern, I test the robustness of my earlier findings by
running equation (2) again but now having Return to Work and Acc Govt Fund as explanatory variables:
(3) SigDisrup = α + θ ∗ Color + β1 ∗ RTW + β2 ∗ Acc.Govt + γ ∗ Urban + δ ∗ DSP + υ.
Table 4 presents the results. Return-to-work barriers on their own explain approximately 15 percent of
the gap (Column 2), while unequal access to government funding explains 22 percent of the gap on its own
(Column 3). Together, they explain 30 percent of the gap (Column 5).
3.3 Returning to Work versus Funding
To test the relative weight of returning to work versus access to government funding, I perform an Oaxaca
decomposition. I find that unequal access to government funding is as important as barriers to returning to
work in shaping the disruption gap. While unequal access to government funding increases its impact on the
gap, on average, so do barriers to returning to work (Figure 5). Their confidence intervals still overlap (Figure
5B), and thus unequal access to government funding remains as relevant as return-to-work barriers in shaping
the disruption gap.
Next, I create a dummy variable, All Funding, that encompasses all disruptions related to funding (lack of
bank relationship, challenges accessing government funding, and difficulty applying for funds). Mainly, this
variable equals one if any of the above mentioned implied a significant barrier and equals zero otherwise. Table
4, Columns 4 and 6 show that this new variable is very similar to Acc Govt Fund. The coefficient on Color barely
8. Responses indicating significant challenge are positively correlated with a coefficient between approximately 0.4 and 0.5.
269
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Table 4
Returning to Work versus Funding
General economic significant disruption
(1) (2) (3) (4) (5) (6)
Color 0.169*** 0.143*** 0.131*** 0.131*** 0.119*** 0.120***
(0.0219) (0.0219) (0.0220) (0.0219) (0.0219) (0.0219)
RTW 0.178*** 0.126*** 0.120***
(0.0211) (0.0220) (0.0223)
Acc Govt Fund 0.217*** 0.180***
(0.0212) (0.0223)
All Funding 0.216*** 0.178***
(0.0207) (0.0221)
Constant 0.306*** 0.205*** 0.248*** 0.214*** 0.187*** 0.163***
(0.0300) (0.0313) (0.0293) (0.0300) (0.0306) (0.0309)
Observations 2,291 2,291 2,291 2,291 2,291 2,291
R-squared 0.040 0.069 0.084 0.085 0.097 0.096
NOTE: Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1
Obs: 2,291; Controls: DSP and Urban
differs from those in Columns 3 and 5. The change in the disruption gap is fairly small when I compare All
Funding to Acc Govt Fund.
Overall, this preferred specification provides supportive evidence that unequal access to government funding
as well as challenges returning to work were important features shaping the observed disruptions in communi-
ties. Together, they can explain almost 30 percent of the disruption gap, and both challenges were as important
in explaining differences in disruption.
3.4 Segments of the Economy Disruption Gap and Challenges
The survey asked not only asked about general economic conditions but also about COVID-19 disruptions on
communities along different segments of the economy—financial stability, small business, access to healthcare,
services for children, housing stability, and basic consumption needs. All these segments of the economy are
vital for vulnerable communities. As mentioned earlier, serious disruption along these areas varied between 30
and 60 percent. Nevertheless, across all segments, it was more likely for communities of Color to be significantly
disrupted. Disruption gaps also varied across segments, with financial stability and services for children having
the largest gaps: 22 and 16 percentage points, respectively.
I next examine the role that challenges played in shaping these observed gaps. Table 5 shows the disruption
gap along the different sectors as well as the percentage of the gap explained by challenges. Challenges and
location explain approximately 40 percent of the disruption differences in general economic disruption, financial
stability, services for children, and basic consumption needs. More interestingly, they can be held accountable
for most of the disruption gap in small business and housing stability, explaining 91 percent and 71 percent of
the disruption gap, respectively. It is worth noting that the disruption gap in small business is small relative to
the rest of the sectors analyzed.
Appendix Figure A.5 compares each one of the challenges to assess their relative importance in each seg-
ment. All four challenges were important and equally relevant in shaping the observed disruption gap for all
six sectors except for services for children, where, reasonably, bank relationships seem to have no explanatory
power on the disruption gap. In other words, the data suggest that challenges returning to work, lack of a bank
relationship, unequal access to government funding, and barriers applying for funds all seem to be important in
shaping the differences in the observed disruption across demographic groups. Together with the differences in
community location, these challenges can explain more than 40 percent (up to 90 percent for certain sectors) of
the observed disruption gap. These findings highlight the importance of the implementation and distribution
270
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Table 5
Segments of the Economy and the Disruption Gap
Disruption gap Percentage of gap explained
General disruption 19.1% 43.2%
Financial stability 21.9% 43.3%
Small business 4.2% 90.8%
Access to healthcare 10.8% 57.8%
Services for children 16.4% 41.5%
Housing stability 10.9% 71.2%
Basic consumer needs 15.2% 42.8%
of COVID-19 government relief packages as well as the relevance of childcare, transportation, and healthcare
in our economy.
3.5 Disruption Gap and Income
In this section I examine whether some of the racial disparities observed in the results stem from income dif-
ferences in LMI communities and find that they do not. This is because income differences do not explain the
variation in the disruption levels or the variation in the severity of the challenges faced. Moreover, there are
income differences between White communities and communities of Color, even when comparing among
LMI communities.
Matching communities’ median household income to this data set offers some challenges as I do not have
precise information on where communities were located and know only the location of the respondents’ orga-
nization headquarters. A report by Faulk et al., 2021 shows that headquarter location is a strong predictor of
the area served. That said, in some cases organizations serve in multiple locations as well as both regionally and
nationwide. Using organizations’ headquarter zip code as a proxy for the community’s location will provide a
noisy measure of income in the communities served, given that some organizations serve in several locations.
Nevertheless, it is only way I can measure income in the community served given the available data.
Appendix Tables A.4 and A.5 show that incorporating communities’ Income—logarithm of median house-
hold income—does not affect the coefficient or significance of Color. In other words, median household income
does not alter the role that demographic composition plays in explaining general economic COVID-19 dis-
ruptions within LMI communities. This finding is robust when having a dummy variable for high income
instead. High Income is a dummy variable that equals one when median household income is above the sample
median and equals zero otherwise. In all cases, incorporating income does not alter the coefficient on Color or
the relevance of the main challenges analyzed in explaining variation in disruption.
While this finding might be somewhat surprising, Appendix Table A.6 shows that variation in income
cannot explain the differences in general economic disruptions, challenges returning to work, or accessing
government funding. A possible explanation for this is that while there might be some variation in median
income within my sample, the survey was already targeted at low-income communities, where job losses and
financial instability have hit the majority of low-income neighborhoods in a similar manner.
Overall, these last two sections provide some robustness to the core of this article supporting the relevance
of unequal access to government funding as well as the challenges of returning to work as important features
shaping the observed disruptions in communities. The importance of these two indicators remains regardless
of accounting for income or the segments of the economy analyzed, all important for the vitality of LMI
communities.
4. CONCLUSION
The federal government responded to the COVID-19 pandemic with a safety net that went a long way in
preventing widespread hardship (Cooney and Shaefer, 2021). The relief measures it provided reduced poverty,
helped people access health coverage, and reduced hardships such as the inability to afford food or meet other
271
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
basic needs. Funds and resources were substantial for LMI communities, with almost two-thirds of respondents
indicating that stimulus checks, unemployment benefits, and rent relief were very critical during COVID-19
(Chalise and Gutkowski, 2021).
However, as this article finds, the ability to access funds has been unequal between communities of Color
and White communities and can explain the observed differences in disruption levels across demographic
groups. Possible reasons for these differentiated outcomes likely rely on historical inequities that cannot be
easily resolved by distributing funds. Special attention and targeting might be necessary for funds to reach
selected recipients, and further research on this targeting and implementation could help address these issues.
I also find that unequal access to government relief was as important as return-to-work challenges in shaping
the observed demographic differences. Together, these challenges can explain approximately 30 percent of the
disruption gap. While much research has been devoted to explaining how barriers to employment have had a
destabilizing effect during COVID-19, this article highlights that the demographic variation in return-to-work
challenges can explain part of the disruption gap between White communities and communities of Color. The
findings highlight the importance of targeting, implementing, and distributing government relief packages as
well as the relevance of childcare in our economy. Further work is needed to allow a more-inclusive labor
market, such as one with accessible and affordable childcare.
272
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
REFERENCES
Autor, David, David Cho, Leland D. Crane, Mita Goldar, Byron Lutz, Joshua Montes, William B. Peterman,
David Ratner, Daniel Villar, and Ahu Yildirmaz. 2022. The $800 billion paycheck protection program:
where did the money go and why did it go there? Journal of Economic Perspectives 36, no. 2 (May): 55–80.
https://doi.org/10.1257/jep.36.2.55. https://www.aeaweb.org/articles?id=10.1257/jep.36.2.55.
Brooks, Matthew M, J Tom Mueller, and Brian C Thiede. 2021. Rural-urban differences in the labor-force
impacts of covid-19 in the united states. Socius 7:23780231211022094.
Budget, Center On, and US Policy Priorities. 2021. Tracking the covid-19 economy’s effects on food, housing,
and employment hardships. Center on Budget and Policy Priorities. United States.[Web Archive] Retrieved from:
https://www. cbpp. org/research/poverty-andinequality/tracking-the-covid-19-economyseffects-on-food-housing-
and.
Chalise, Nishesh, and Violeta A Gutkowski. 2021. Perspectives from main street: the impact of covid-19 on
communities and the entities serving them. Retrieved from Federal Reserve System Fed Communities, https:
//f edcommunities.org/data/main-street-covid19-survey-2021.
Cooney, Patrick, and H Luke Shaefer. 2021. Material hardship and mental health following the covid-19 relief
bill and american rescue plan act. Ann Arbor, Michigan: Poverty Solutions, University of Michigan.
Faulk, Lewis, Mirae Kim, Teresa Derrick-Mills, Elizabeth Boris, Laura Tomasko, Nora Hakizimana, Tianyu
Chen, Minjung Kim, and Layla Nath. 2021. Nonprofit trends and impacts 2021. The Urban Institute.
Gangopadhyaya, Anuj, Michael Karpman, and Joshua Aarons. 2020. As the covid-19 recession extended into
the summer of 2020, more than 3 million adults lost employer-sponsored health insurance coverage and
2 million became uninsured. Washington, DC: Urban Institute.
Gemelas, Jordan, Jenna Davison, et al. 2022. Inequities in employment by race, ethnicity, and sector during
covid-19. Journal of racial and ethnic health disparities 9 (1): 350–355.
Granja, João, Christos Makridis, Constantine Yannelis, and Eric Zwick. 2022. Did the paycheck protection
program hit the target? Journal of financial economics 145 (3): 725–761.
Holtzblatt, Janet, and Michael Karpman. 2020. Who did not get the economic impact payments by mid-to-late
may, and why? Washington, DC: The Urban Institute.
Janger, Emma, Nicole Rubin, and Sejal Singh. 2020. Making unemployment insurance work for working
people. UCLA L. Rev. Discourse 68:102.
Lee, Emma K, and Zachary Parolin. 2021. The care burden during covid-19: a national database of child care
closures in the united states. Socius 7:23780231211032028.
Li, Lei, Philip E Strahan, and Song Zhang. 2020. Banks as lenders of first resort: evidence from the covid-19
crisis. The Review of Corporate Finance Studies 9 (3): 472–500.
Liu, Sifan, and Joseph Parilla. 2020. New data shows small businesses in communities of color had unequal
access to federal covid-19 relief. Brookings Institution Report.
Maxwell, Connor. 2020. The coronavirus crisis is worsening racial inequality. Center for American Progress web-
site: americanprogress. org/issues/race/news/2020/06/10/486095/coronavirus-crisis-worsening-racial-inequality.
Published June 10.
OECD. 2022. The unequal impact of covid-19: a spotlight on frontline workers, migrants and racial/ethnic
minorities, https://doi.org/https://doi.org/https://doi.org/10.1787/f 36e931e- en. https://www.oecd-
ilibrary.org/content/paper/f 36e931e-en.
Parker, Kim, Juliana Menasce Horowitz, and Rachel Minkin. 2021. Americans are less likely than before covid-
19 to want to live in cities, more likely to prefer suburbs.
273
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
APPENDIX 1.
Table A.1
COVID-19’s Level of Disruption by Segments of the Economy
No disruption (%) Some disruption (%) Significant disruption (%) Unsure (%) Observations
General disruption 1.4 51.2 46.4 1.0 2,391
Financial stability 1.7 41.3 54.3 2.8 2,291
Small business 1.1 34.7 60.4 3.7 1598
Access to healthcare 4.6 43.0 40.8 11.6 2,288
Services for children 1.8 29.1 59.1 10.0 2,287
Housing stability 3.4 31.0 55.3 10.3 2,289
Basic consumer needs 5.5 51.9 35.9 6.7 2,288
Table A.2
Disruption by Color, Type of Service, and Location
General economic disruption
No disruption (%) Some disruption (%) Significant disruption (%) Unsure (%) Observations (% of total)
Color: No 2.5 60.5 36.3 0.8 48
Color: Yes 0.4 42.5 55.8 1.2 52
No DSP 2.7 55.9 40.8 0.7 14
DSP 1.2 50.5 47.3 1.1 86
Rural 2.1 59.5 37.9 0.5 35
Urban 1.1 46.8 50.9 1.3 65
Total obs. 31 1,115 1,010 22
NOTE: This table shows the distribution of responses to general economic disruption disaggregated by communities’ demographics,
type of service provided, and location.
Table A.3
Distribution of Challenges among Respondents
Challenges
RTW (%) Acc Gvt Fund (%) BK Relat (%) Fund Appl (%)
Not a challenge 2.2 7.5 14.3 6.4
Challenge 27.8 41.9 29.6 36.5
Significant challenge 65.9 40.1 27.3 46.4
Unsure 4.1 10.6 28.8 10.6
Observations 2,222 2,233 2,019 2,195
274
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Table A.4
Robustness: Controlling for Income
General economic significant disruption
(1) (2) (3) (4) (5) (6)
Color 0.195*** 0.195*** 0.194*** 0.170*** 0.170*** 0.169***
(0.0210) (0.0210) (0.0210) (0.0226) (0.0226) (0.0226)
Income -0.000126 -0.00267
(0.0258) (0.0260)
High Income -0.0112 -0.0143
(0.0210) (0.0210)
Urban 0.0691*** 0.0693*** 0.0704***
(0.0233) (0.0234) (0.0234)
DSP 0.0472 0.0469 0.0458
(0.0301) (0.0302) (0.0302)
Constant 0.363*** 0.364 0.369*** 0.290*** 0.319 0.298***
(0.0148) (0.285) (0.0186) (0.0310) (0.289) (0.0332)
Observations 2,178 2,178 2,178 2,178 2,178 2,178
R-squared 0.038 0.038 0.038 0.043 0.043 0.043
NOTE: Robust standard errors are in parentheses.*** p < 0.01, ** p < 0.05, * p < 0.1
275
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Table A.5
Robustness: Controlling for Income (2)
General economic significant disruption
(1) (2) (3)
Color 0.111*** 0.111*** 0.110***
(0.0227) (0.0227) (0.0227)
Income 0.00203
(0.0248)
High Income -0.00933
(0.0204)
Urban 0.0533** 0.0532** 0.0542**
(0.0227) (0.0227) (0.0227)
DSP 0.0480* 0.0482* 0.0471
(0.0289) (0.0290) (0.0290)
RTW 0.110*** 0.110*** 0.110***
(0.0230) (0.0230) (0.0230)
Acc Govt Fund 0.124*** 0.124*** 0.124***
(0.0258) (0.0258) (0.0258)
BK Relat 0.0527* 0.0528* 0.0521*
(0.0285) (0.0285) (0.0285)
Fund Appl 0.0786*** 0.0785*** 0.0786***
(0.0266) (0.0266) (0.0265)
Constant 0.164*** 0.142 0.169***
(0.0313) (0.275) (0.0334)
Observations 2,178 2,178 2,178
R-squared 0.105 0.105 0.105
NOTE: Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1
Table A.6
Disruption, Challenges, and Income Differences
Color General sig. disrup. RTW sig. chall. Acc Govt sig. chall.
(1) (2) (3) (4) (5) (6) (7) (8)
Income -0.0540** -0.0106 0.00110 -0.0159
(0.0260) (0.0264) (0.0248) (0.0260)
High Income -0.0367* -0.0184 -0.00735 -0.00918
(0.0214) (0.0214) (0.0206) (0.0209)
Constant 1.110*** 0.535*** 0.580** 0.473*** 0.627** 0.643*** 0.565** 0.395***
(0.285) (0.0151) (0.290) (0.0151) (0.273) (0.0145) (0.286) (0.0148)
Observations 2,178 2,178 2,178 2,178 2,178 2,178 2,178 2,178
R-squared 0.002 0.001 0.000 0.000 0.000 0.000 0.000 0.000
NOTE: Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1
276
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Figure A.1
Demographics and Communities Served
A. Location of service B. Type of service providers
Urban Rural DSP Non-DSP
Color: No Color: Yes Color: No Color: Yes
Figure A.2
Disruption by Type of Service Offered and Location
1
Likelihood of Significant Disruption
.2 .4 .6 .8
0
Rural Urban
No DSP DSP
277
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Figure A.3
Disruption in Segments of the Economy
Financial Stability Small Business Healthcare
100 100 100
80 80 80
% % %
60 60 60
40 40 40
20 20 20
0 0 0
Services for Children Housing Stability Basic Consumption Needs
100 100 100
80 80 80
% % %
60 60 60
40 40 40
20 20 20
0 0 0
Significant disruption Some disruption
No disruption Unsure
Figure A.4
Challenges Faced by Demographic Group
Return to Work Accessing Govt Funds
Color: No Color: Yes Color: No Color: Yes
Bank Relationship Fund Application
Color: No Color: Yes Color: No Color: Yes
Significant challenge Challenge
Not a challenge Unsure
278
Gutkowski Federal Reserve Bank of St. Louis REVIEW · Fourth Quarter 2023
Figure A.5
Confidence Intervals of Challenges by Segments of the Economy
A. Financial stability B. Small business
C. Healthcare D. Services for children
E. Basic consumption needs
E. Housing stability
NOTE: The figures show 95 percent confidence intervals from the Oaxaca decomposition for each of the challenges analyzed. There is
one figure for the decomposition in each segment of the economy.
279
File and source
- File
- demographic-disparities-in-covid-19-disruptions-what-has-shaped-them.pdf
- Size
- 1,636,182 bytes
- SHA-256
- d3db09d97b63caf69f0af86f1cd573bfc2b3aa6bcdfcab5b77d1189548af14bb
- Original
- doi.org