Qingfang Wang and Wei Kang, Small businesses and government assistance during COVID-19: Evidence from the paycheck protection program in the U.S (2023)

Summary

A 2023 journal article by Qingfang Wang of the University of California Riverside and Wei Kang of the University of North Texas, Small businesses and government assistance during COVID-19: Evidence from the paycheck protection program in the U.S., in EPA: Economy and Space, Vol. 55(8), pages 2147-2165. The abstract says the study examines the neighborhood-level distribution of Paycheck Protection Program loans in inland Southern California, using spatial regressions and in-depth interviews with small businesses, banks, government agencies and nonprofit organizations. It reports that PPP loans roughly reached their small business targets in total number and value, but that communities with higher shares of pandemic-vulnerable businesses or higher socioeconomic vulnerability were less likely to have received loans.

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

Cite this document

“Qingfang Wang and Wei Kang, Small businesses and government assistance during COVID-19: Evidence from the paycheck protection program in the U.S (2023).” Original at digital.library.unt.edu. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/qingfang-wang-and-wei-kang-small-businesses-and-government-assistance-4d68643435/, SHA-256 405c696fe7773e70e07dc6dacfec956d969265fc98386bfa2ff3bd2212528c02.

Full text

1191943
research-article2023
EPN0010.1177/0308518X231166407EPA: Economy and SpaceWang and Kang

Original Article A Economy and Space
EPA: Economy and Space

Small businesses and
2023, Vol. 55(8) 2147–2165
© The Author(s) 2023

government assistance during Article reuse guidelines:

COVID-19: Evidence from the sagepub.com/journals-permissions
https://doi.org/10.1177/0308518X231166407
DOI: 10.1177/0308518X231166407
journals.sagepub.com/home/epn
paycheck protection program
in the U.S.

Qingfang Wang
School of Public Policy, University of California Riverside, Riverside, CA,
USA

Wei Kang
Department of Geography and the Environment, University of North Texas,
Denton, TX, USA

Abstract
The COVID-19 pandemic has caused unprecedented losses for small businesses in cities across
the globe. Policymakers have relied on a wide range of measures to support ﬁrms and sustain busi-
ness continuity. However, signiﬁcant concerns have been expressed about the degree of equity in
the distribution and efﬁciency of government assistance during the pandemic disruption. Drawing
on the case of the Paycheck Protection Program (PPP) and its implementation in inland Southern
California, this study examines the spatial distribution of PPP loans at the neighborhood level.
Based on spatial regressions and in-depth interviews with small businesses, banks, government
agencies, and nonproﬁt organizations, the study ﬁnds that, in terms of their total number and
value, the PPP loans have roughly succeeded in reaching their small business targets. However,
communities with higher shares of pandemic-vulnerable businesses or higher levels of socio-
economic vulnerability are less likely to have received PPP loans. There have also been spatial spill-
over effects of community vulnerability when it comes to receiving PPP loans at the neighborhood
level. The correlation between fewer PPP loans and community vulnerability also reﬂects both
short-term needs and longstanding challenges facing entrepreneurship and business development
in socioeconomically disadvantaged communities. Moreover, small business resilience and commu-
nity resilience are inseparable, and thus government business assistance must be considered in the
context of local communities.

Corresponding author:
Qingfang Wang, School of Public Policy, University of California Riverside, 900 University Ave, Riverside, CA 92521, USA.
Email: Qingfang.Wang@ucr.edu
2148 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

Keywords
COVID-19, PPP, small businesses, community vulnerability

Introduction
The COVID-19 pandemic has triggered unprecedented ﬁnancial losses across the world. As a key
component of urban economic viability, small businesses suffer badly if they lack access to public
ﬁnancial markets or other ways of managng short-term ﬁnancial challenges. Policymakers around
the world have created a wide range of measures, such as wage subsidies, small business grants, and
guaranteed business loan schemes to support ﬁrms and mitigate the negative economic conse-
quences of the pandemic. In the U.S., small businesses received immediate assistance from the
Federal Emergency Management Agency (FEMA), the Small Business Administration (SBA),
and other federal and local government agencies. However, both the general public and academic
researchers have expressed concerns about the degree of equity in the distribution and efﬁciency of
government assistance programs. For instance, these programs may have disproportionately bene-
ﬁtted certain businesses over others, with the latter including extremely small businesses that have
experienced a high level of pandemic disruption, minority- and women-owned businesses, and
those in communities that have traditionally lacked investment (Ofﬁce of Evaluation Science,
2021; Santellano, 2021). Researchers even argue that these programs may have widened pre-
pandemic economic inequalities across racial groups (Ong et al., 2021).
As part of the U.S. Coronavirus Aid, Relief, and Economic Security (CARES) Act, the Paycheck
Protection Program (PPP) offers guaranteed, forgivable loans to provide liquidity to small and mid-
sized businesses to help prevent job losses. The PPP deployed more than $500 billion within just
four months of its passage into law, making it one of the largest ﬁrm-based ﬁscal policy programs in
U.S. history. Our study investigates the spatial distribution of PPP loans by addressing the follow-
ing questions. How have PPP loans been distributed across neighborhoods? What are the neighbor-
hood characteristics associated with PPP loan distribution? What does the experience of the
program suggest about policymaking to assist small businesses with their recovery from pandemic-
like disruptions?
We conducted our study in the Riverside-San Bernardino-Ontario metropolitan area in inland
Southern California, an area with a population of 4.5 million – about 12% of California residents.
Based on spatial analyses of PPP loans across census tracts, and in-depth interviews, we ﬁnd that if
the distribution of these loans was measured by their total number and their aggregate value, they
seemed to have reached their small business targets. However, communities with more vulnerable
businesses and higher socioeconomic vulnerability were less likely to have received the loans. The
results suggest that both short-term needs and longstanding challenges facing entrepreneurship and
business development in disadvantaged communities; moreover, small business resilience and
community resilience were inseparable and should be jointly considered in designing and imple-
menting government assistance programs like the PPP.
This is one of the ﬁrst systematic analyses of the spatial distribution of COVID-related govern-
ment business-aid programs at the community level in U.S. cities. First, research had demonstrated
signiﬁcant changes in health, social, and political urban landscapes under the impact of COVID
(Florida et al., 2021); but knowledge of the experiences of small businesses during these
changes was extremely limited. This study expanded the scholarship on urban resilience to
include the experiences of small businesses in underserved communities, and thus highlighted
the role of public policies in community resilience-building, as well as their intended and unin-
tended consequences.
Wang et
andal.Kang 21493

Second, the interaction of economic, social, racial, urban, and territorial inequalities had been
reinforced by the uneven impact of the pandemic across urban communities along the lines of
race, ethnicity, gender, class, and other aspects of socioeconomic status (Bambra et al., 2021;
Dokhov and Topnikov, 2021). Our study directly examined the relationship between access to gov-
ernment assistance and community vulnerability. In particular, by triangulating the perspectives
from both supply and demand sides, we identiﬁed the challenges facing socioeconomically disad-
vantaged communities.
Third, existing literature suggested a gap in understanding the links between business resilience
and community resilience in responding to emergencies (Xiao and Peacock, 2014; Xiao and Van
Zandt, 2012). Our ﬁndings reﬂected the relationship between economic and social, and between
business and community resilience, and thus they contributed to interdisciplinary scholarship on
resilience, economic and community development, and equitable recovery and growth. The ﬁnd-
ings also had important practical and policymaking implications for organizations and government
agencies seeking to foster small business development and community resilience.

Literature review
The payroll protection program (PPP) under COVID
The PPP was established by the U.S. CARES Act in March 2020. It was designed to help small
businesses, roughly those deﬁned as having 500 or fewer employees. By providing federally guar-
anteed loans to small businesses with no fees and an interest rate of 1%, the PPP was intended to
provide a direct incentive for small businesses to keep their workers on the payroll, to stabilize the
rate of small business closures, and to control the rise in unemployment rates. The loans have a
maturity rate of two years and repayment may be forgiven in certain circumstances. A small busi-
ness may be eligible to apply for up to two draws of PPP loans.
First Draw PPP loans to eligible borrowers qualify for full loan forgiveness if the following
requirements are satisﬁed during the 8- to 24-week period following loan disbursement: employee
and compensation levels are maintained; and the loan is spent on payroll costs (at least 60% of it)
and other eligible expenses. Small businesses that have previously received a First Draw PPP loan
can apply for a Second Draw loan. These loans qualify for full loan forgiveness if the same require-
ments are satisﬁed during the 8- to 24-week period following loan disbursement (SBA, 2021b).
To apply for a PPP loan, small business owners must go through an SBA 7(a) lender or a fed-
erally insured depository institution, a federally insured credit union, or an institution participating
in the Farm Credit System. The total amount approved as of October 23, 2021, was about 11.5
million dollars, and by that date, there had been applications for 7.9 million dollars of forgiveness.
Most of the loans were small, with the average amount being 67,647 dollars (SBA, 2021a).

Existing studies of the PPP under COVID-19
Extant literature has revealed that the spread of the COVID-19 pandemic and its impact on busi-
nesses had been unevenly distributed geographically (Bartik et al., 2020; Kim, 2021; Zhai et al.,
2021). Some studies particularly suggested that the PPP had beneﬁted small businesses that were
able to obtain the loans, enabling them to retain workers and pay for utilities, rent, supplies, insur-
ance, and other costs (Bartik et al., 2020). However, these studies also pointed out that, despite the
unprecedented amount of aid, funds had not reached the businesses that needed assistance the most.
At the same time, many businesses had been reluctant to apply because of distrust and concerns
about administrative complexity and eligibility, insufﬁcient knowledge about program conditions
4
2150 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

and application procedures, bureaucratic hassles, and difﬁculties establishing eligibility (Bartik
et al., 2020; Fairlie and Fossen, 2022).
These issues had particularly limited access to emergency resources for minority and
women-owned businesses (Fairlie and Fossen, 2022; Ofﬁce of Evaluation Science, 2021).
Kickul et al. (2021) found that signiﬁcant differences existed between male- and female-owned
businesses, with male-owned ﬁrms receiving over 80% of the PPP loans made during the
COVID-19 pandemic. Similarly, small business owners of color reported having difﬁculty acces-
sing PPP funds (Gamboa, 2020; Santellano, 2021). According to another nationwide survey, only
12% of Black and Latino small business owners who had applied for funds through the SBA,
mostly PPP loans, reported receiving the amount asked for, while 26% reported receiving only a
fraction of the amount applied for (Unidos US, 2020).
Research also found spatial differences in access to PPP loans. Atkins et al. (2022) found that
Black-White disparities in PPP loans appeared (marginally) greater in areas with fewer bank
branches, because Black business owners might have been less likely to have prior relationships
with SBA-approved lenders. Borawski and Schweitzer (2021) found that PPP loans had had a
broad distribution to small businesses in low- and moderate-income (LMI) communities;
however, they had reached higher-income communities to a greater extent than areas with Black,
Hispanic, and American Indian or Alaska Native majorities. Furthermore, having examined the dis-
tribution of PPP loans across California’s 53 congressional districts, Ong et al. (2021) argued that
disparities in PPP lending had widened pre-pandemic economic inequalities across racial groups.
Overall, existing studies suggested signiﬁcant unevenness in access to government assistance.
However, based on data from the early stages of the pandemic, most of these studies had
focused mainly on descriptions rather than detailed spatial analyses at the community level. As a
result, these studies failed to link their descriptions to a broader understanding of small business
resilience, especially in socioeconomically disadvantaged communities.

Small business resilience and government assistance
Existing literature on small business resilience has highlighted different factors that inﬂuence busi-
ness preparedness for, and recovery from, hazardous events (Bartoloni et al., 2021; Marshall and
Schrank, 2014; Tibay et al., 2018). For example, ﬁrm-level characteristics include business size,
industrial sector, age of business, emergency planning, ﬁnancial situation, occupancy of premises,
and market range (Adekola and Clelland, 2020; Haynes et al., 2019; Orhan, 2016; Webb et al.,
2000). Further, pre-disaster conditions, prior experiences with disasters, leadership and manage-
ment, the core competence of staff, and effective planning all contribute signiﬁcantly to businesses’
post-disaster recovery (Marshall and Schrank, 2014; Nowell et al., 2017; Webb et al., 2000). In add-
ition, social capital and social networks enable households, businesses, and communities to over-
come shocks, and can provide ﬁnancial and non-ﬁnancial capital to build resilience. They often
offer resources in the form of information, technological knowledge, market access, physical
resources, and emotional resources to keep small businesses resilient (Aldrich, 2012; Danes
et al., 2009).
Researchers have also suggested that public policies are of the utmost importance in shielding a
region against a crisis, mitigating a crisis’s worst impacts, and transforming and reorientating a
region toward recovery (Adekola and Clelland, 2020; Torres et al., 2019; Xiao and Peacock,
2014). Governments often provide grants or low-interest loans to assist households, businesses,
and communities to recover from disastrous events. For example, businesses impacted by
Hurricane Katrina, in 2005, received SBA loans of varying amounts, depending on the likely repay-
ment ability of the owners, and the businesses’ locations and structures. All of this suggests that
government assistance has played a positive role in business survival.
Wang et
andal.Kang 21515

However, not every business in need receives government support. Even in a healthy economy,
racial and ethnic minority- and women-owned businesses have greater difﬁculties in obtaining
ﬁnancial capital to launch their businesses and sustain them, in addition to a wide range of other
challenges (Bates and Robb, 2015; Bates et al., 2018; Teixeira, 2001). Disastrous events further
exacerbate conditions for traditionally disadvantaged groups (Cutter et al., 2000; Peacock et al.,
2014; Wang, 2022). For instance, studies on past disasters have identiﬁed that access to government
assistance is not evenly spread. Based on data from more than 120,000 ﬁrms in 60 countries, Cirera
et al. (2021) ﬁnd that government support has been especially limited for the most vulnerable ﬁrms
and countries. There are mismatches between the number of ﬁrms in need and available resources,
between the ﬁrms most urgently in need and the recipients of aid, and between policies reported as
most needed by ﬁrms and policies that ﬁrms are most likely to encounter. In the U.S., research also
ﬁnds that access to government assistance differs according to the level at which ﬁrms operate, busi-
ness owners’ characteristics, industrial sector, location, and other factors. For example, Josephson
and Marshall (2016) ﬁnd that in Mississippi post-Katrina the likelihood of receiving a loan guar-
antee from the SBA increased with the amount of damage, the amount of revenue the business
usually received, the age of the business, whether the owner was white, and how much experience
of the industry the owner had.
Furthermore, as small businesses are embedded in their social and spatial communities, small
business owners who are able to bridge and receive help from community networks are more
likely to be economically resilient, and links with family and friends, community leaders, and institu-
tions can help small businesses thrive after a natural disaster (Lo et al., 2019; Torres et al., 2019). This
indicates the vulnerability of small businesses that are traditionally disadvantaged, including both
ethnic minority- and women-owned businesses and businesses located in geographic areas that
lack infrastructure and resources (Bates et al., 2018; Rubin, 2010). Therefore, we expect that busi-
nesses located in different communities differ in their access to government assistance, and that socio-
economically disadvantaged locations face greater challenges; it is critical to understand the
implementation of this program and the factors that shape small business resilience at different levels.

Data and method
We conducted our study in Riverside-San Bernardino-Ontario, a metropolitan statistical area
(MSA) in inland Southern California. This region ranks among the 15 most populous MSAs in
the U.S. It is socially and demographically diverse, being home to a plurality of Hispanic/Latino
(51%), white (32%), Black (6.9%), and Asian (6.6%) populations. This demographic and racial
diversity is signiﬁcant because disaster disruptions, including COVID-19, disproportionately
impact and penalize poor people and communities of color (Bambra et al., 2021; Dokhov and
Topnikov, 2021). The Census Bureau’s Small Business Pulse Survey (SBPS) indicated that 89%
of the small businesses in this region had reported a moderate to large negative effect from the
COVID-19 pandemic in early May 2020. By April 2021, the percentage was still as high as 67%.
To address the research questions, we investigated how PPP loans were allocated to neighborhoods
and identiﬁed neighborhood-level characteristics that were associated with the number and value of loans
allocated. The descriptions of all the variables and data sources are provided in Supplementary File 1. To
examine the process of implementation at the local level, we conducted in-depth interviews with both
small businesses and small business service providers from different sectors.

Data
PPP loans. Complete microdata for PPP loans is available to the general public from the SBA (SBA,
2021a). The data set covers the entire lifecycle of these loans, starting on April 3, 2020, and ending
6
2152 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

on May 31, 2021. The total of loans received by businesses in the study area was 19,540 (7.568
billion dollars). As shown in Table 1, a total of 90,977 small businesses received First Draw
PPP loans while 31.4% of these small businesses received a Second Draw PPP loan. Of First
Draw loans, 59.12% went to small businesses reporting only one job, and of Second Draw loans
48.55%, went to such businesses. Of First Draw loans, 12.21% went to businesses with only 2–
4 paid employees, 24% went to businesses with only 5–49 paid employees, and 2.68% went to busi-
nesses with at least 50 paid employees. Businesses with 5–49 paid employees that received a PPP
loan were more likely to obtain another one.
The general pattern indicates that the number of jobs reported is positively associated with the
number of PPP loans approved (Pearson’s correlation coefﬁcient = 0.763, p-value = 0). In other
words, among businesses that received a PPP loan, those reporting more jobs were likely to be
approved for larger loans. This positive relationship holds when turning to tract-level statistics.
The number of businesses that received PPP loans and report only one job is negatively related
to the average loan value in the tract (Pearson’s correlation coefﬁcient = −0.688, p-value = 0).
Businesses existing for more than 2 years accounted for 86.43% of First Draw loans, whereas
only 9.65% (8780) of newer businesses (deﬁned as existing less than 2 years) received First
Draw loans; and newer businesses did not receive Second Draw loans. In addition, 32,788 small
businesses, or 36% of total recipients, that received First Draw loans were in low- and
moderate-income (LMI) communities. A similar pattern holds for the Second Draw–out of all
small businesses receiving Second Draw loans, about 34% were in LMI communities.
Since each loan record provides a detailed address, we further geocoded the addresses to census
tracts for aggregation. Out of a total of 822 tracts in the region under study, three did not receive any
PPP loans. For all the census tracts which received PPP loans (Supplementary File 2), the average
number of loans was 146 and the median was 111. The distribution had a wide range [2, 1832] and
was highly skewed to the right, meaning that most tracts received only a small number of loans.
Only three tracts received more than 1000 loans. This pattern holds if we differentiate First from
Second Draw PPP loans (Supplementary File 3).
For 269 tracts, not a single business that received a PPP loan was in an LMI community, while
for another 104 tracts, all the businesses receiving a PPP loan were in LMI communities. An

Table 1. Characteristics of ﬁrst draw and second draw PPP loans.

First draw Second draw

Aggregate number 90,977 28,563
Urban 90,977 (100%) 28,563 (100%)
Jobs reported
1 53,785 (59.12%) 13,867 (48.55%)
2–4 12,924 (14.21%) 4408 (15.43%)
5–49 21,831 (24%) 9575 (33.52%)
50+ 2437 (2.68%) 713 (2.5%)
Business age
Existing or more than 2 years old 77,720 (85.43%) 28,563 (100%)
New Business or 2 years or less 8780 (9.65%) 0
Startup, Loan Funds will Open Business 14 (0.015%) 0
Change of Ownership 8 (0.009%) 0
Unanswered 4455 (4.9%) 0
Low- and moderate-income (LMI) communities
Loans 32,788 (36.04%) 9779 (34.24%)
Loan values $2,163,199,242 (38.63%) $735,701,416 (37.37%)
Wang et
andal.Kang 21537

average tract received Second Draw loans totaling 30.4% of its First Draw loans. Only two tracts
that received loans from the First Draw (two and ﬁve loans) did not receive any from the Second
Draw. Of all the tracts that received Second Draw loans, one tract in unincorporated San Bernardino
County received 16 loans, accounting for 84% of First Draw loans received – the largest percentage.
Tracts receiving a larger number of PPP loans from the First Draw were likely to receive Second
Draw loans (Pearson’s correlation coefﬁcient = 0.958, p-value = 0). This makes sense, as only bor-
rowers that had already received a PPP loan were eligible for a Second Draw PPP loan. The per-
centage of businesses receiving PPP loans in LMI communities is not signiﬁcantly associated
with any other variables (e.g., the aggregate number of loans or average loan values). This suggests
that PPP loans are likely to be equally or randomly distributed across high-income and LMI
communities.

Small businesses. We acquired the distribution of small businesses from the County Business
Pattern (CBP) for 2019, the most recent report available for our data analyses. The 2019 CBP
reported a total of 77,626 business establishments in inland Southern California, which was
fewer than the total number of First Draw PPP loans in this region. Besides the fact that our
CBP data was slightly out of date, since new businesses were constantly being formed, the CBP
series does not include data on self-employed individuals, who accounted for about 56.6% of
PPP loans and represented 81.6% of all business establishments, based on the 2018
Nonemployer Statistics.
The CBP does not provide data at the census tract level. The smallest geographical unit at which
CBP data is released is the ZIP Code area. Out of 170 ZIP Code areas that intersected with the study
area, the 2019 CBP report suggested that 162 had business establishments. We used the U.S.
Department of Housing and Urban Development (HUD) United States Postal Service (USPS)
ZIP Code crosswalk ﬁles to allocate ZIP Code-level data from the CBP to census tracts. These
crosswalk ﬁles contained the ratios of three address types for the census tracts each ZIP Code over-
lapped with: residential, business, and other (Wilson and Din, 2018).
According to the 2019 CBP, the aggregate number of small businesses in this region ranged from
0 to 1772, with a standard deviation of 158. The indicator of business vulnerability in the commu-
nity was not deﬁned for the seven tracts that had no businesses. The rate for the remaining 815 tracts
varied from 0 to 33.3%, with smaller rates largely distributed in the west of the region.

Neighborhood-level characteristics and key measurements. We measured community vulnerability at
the census level from two dimensions. First, an economic sectoral perspective: if a census tract had
a higher percentage of businesses that were vulnerable to the pandemic, we regarded that neighbor-
hood as more vulnerable. Based on the SBPS, we identiﬁed three sectors that had been most
severely impacted by the pandemic: Accommodation and Food Services (72), Educational
Services (61), and Arts, Entertainment, and Recreation (71). These sectors ranked the highest
based on the average of four indices – Overall Sentiment Index, Operational Challenges Index,
Financial Stress Index, and Expected Recovery Index – each of which is calculated and released
by the Census Bureau. The percentage of small businesses in these sectors was used as an indicator
of small business vulnerability in existing studies (e.g., Wang and Kang, 2021).
The second dimension was measured through a residential perspective: a neighborhood or com-
munity was more vulnerable if it had a higher percentage of residents with risk factors that made
them more vulnerable to the pandemic. We adopted the Community Resilience Estimates
(CREs) – a metric developed by the U.S. Census Bureau during the pandemic – for 2019 to
help identify communities where resources and information might have effectively mitigated the
impact of disasters. The 2019 CREs used the 2019 American Community Survey (ACS) and the
Census Bureau’s Population Estimates Program (PEP) to provide estimates at the census tract
8
2154 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

level. Ten Risk Factors (RFs) were considered in constructing each CRE, including the
income-to-poverty ratio (total family income divided by the poverty threshold <130), the
number of single or zero caregiver households, housing overcrowding, communication barriers,
unemployment, disability, lack of health insurance, the number of inhabitants 65 years or older,
lack of access to a vehicle, and lack of access to broadband internet. CRE3 (i.e., the proportion
of individuals with three or more risk factors) was used in our models as an indicator of community
vulnerability. Supplementary File 4 provides the descriptive statistics for key and control independ-
ent variables.
Additionally, the racial and ethnic composition of neighborhoods, as well as the other demo-
graphic, socioeconomic, and housing variables, were sourced from the 2019 ACS 5-year estimates.
These variables were checked using diagnostic tests, model ﬁt comparisons, and multicollinearity
checks (Supplementary File 5).

Spatial analysis
The distribution of PPP loans was examined in both the aggregate number of loans and the aggre-
gate value of loans across census tracts. First Draw and Second Draw loans were modeled separ-
ately as adjustments had been made to the second draw (e.g., a more stringent eligibility
requirement and the expansion of SBA-approved lenders). We further differentiated employer busi-
nesses from self-employed businesses.

Exploratory spatial analysis. Spatial analyses were adopted to examine the spatial distribution of the
PPP loans (aggregate number and value). We used the widely used indicator of global spatial auto-
correlation, Moran’s I, to formally examine whether neighborhoods receiving a higher number/
value of PPP loans were closer to each other. Moran’s I statistic is deﬁned in Equation (1):
N N
N i j wij (yi −
y)(yj − y)
I= N (1)
S0 )2
i=1 (yi − y

Here, yi and yj are the PPP loans/values for neighborhoods i and j, y is the mean of the PPP loans/
values across all neighborhoods, wij is the i, j th element of the row-standardized spatial weight
N
matrix W, and S0 = Ni j wij . The Queen contiguity spatial weight matrix was adopted to
represent the spatial structure of the area’s neighborhoods, meaning that neighborhoods sharing
a common border, or a common vertex were considered neighbors. The permutation-based infer-
ence was utilized to determine the statistical signiﬁcance of the Moran’s I statistic. A
pseudo-p-value was obtained based on 999 permutations.

Spatial regression models. To examine the role of community vulnerability in determining PPP loan
distributions, we started with the standard linear regression model, which is displayed as follows:
ln(y) = β0 + β1 RateVSB + β2 CRE3 + φX + ε (2)
Here, y is our measure of PPP loan allocations to neighborhoods in the overall area, RateVSB is the
indicator of community business vulnerability, CRE3 is the indicator of community household vul-
nerability, X includes a set of control variables that could confound the relationship between the two
community vulnerability indicators and the outcome loan variables, and ε is the random component
assumed to be independently and identically distributed with mean 0 and variance σ 2 (ε ∼ (0, σ 2 )).
Our control variables in X included small business distribution across neighborhoods and
neighborhood-level demographic and socioeconomic characteristics.
Wang et
andal.Kang 21559

We further conducted diagnostic tests to assess potential spatial dependence in our regression
models. The spatial interaction effect was considered in four potential spatial regression models:
the spatial lag model (SAR) incorporated a spatially lagged dependent variable, the spatial error
model (SEM) contained a spatially autoregressive error term, the spatial lag of X (SLX) model con-
tained spatially lagged independent variables, while the spatial Durbin model (SDM) incorporated
both spatially lagged dependent and independent variables.
With 81.6% of all business establishments in this region being nonemployer businesses, it was
critical to look at both employer and nonemployer businesses, but in separate models. Therefore, we
ran four sets of regression models for each of the two draws of PPP loans: the ﬁrst two examined the
aggregate number and the loan values received by all the businesses; the second two examined the
subsample of employer businesses with more than one person working.
We used the hybrid speciﬁcation strategy outlined in the methodology section to search for the
most appropriate spatial regression speciﬁcations. It turned out that the SDM was the most appro-
priate for all cases except for the models for the total numbers of second-draw PPP loans and the
total values of ﬁrst-draw PPP loans. For these two models, a SAR speciﬁcation was adopted.
Supplementary File 6 documents how we chose different models for the current study.
We relied on the direct, indirect, and total effects as proposed by Lesage and Pace (2009) for our
interpretation of the spatial regression models, as these models might include spatial lags of depend-
ent (and independent) variables and thus require special consideration to properly incorporate the
spatial spillover effect in their interpretation. Speciﬁcally, the average direct impact for an inde-
pendent variable represented the average response of the dependent variable to the independent
variable. Because of the inclusion of the spatial lag of the dependent variable in the SDM and
SAR, the average direct impact included the effect of feedback loops where the PPP loan/value
of a community i affected the PPP loan/value of a neighboring community j, and the PPP loan/
value of j affected that of i (also including longer paths and feedback loops).
The average indirect impact was related to the spatial spillover effects. Speciﬁcally, for an inde-
pendent variable, its average indirect impact was interpreted as its impact on a community’s PPP
loan/value because of a unit change in that variable across all other communities. It could also be
interpreted as the impact of a unit change in that variable in a community on the PPP loan/value of
all other communities. Therefore, if the average indirect impact of community vulnerability was
signiﬁcant, a community’s geographical neighbors’ community vulnerability levels mattered to
the PPP loan/value this focal community received.
The average total impact of an activity is the sum of the average direct and indirect impacts. For
an OLS model, the average indirect impact is 0 and the average total impact is equal to the average
direct impact. These impact statistics have been demonstrated to be more suitable for spatial regres-
sion models than the coefﬁcients appearing in our model speciﬁcations, since the latter are not real
partial derivatives. It is likely that the coefﬁcient of the spatially lagged independent variable is
insigniﬁcant, while its spatial spillover effect (indirect impact) is signiﬁcant. For spatial diagnostic
tests and spatial regression model estimation and inference, we used the open-source R package,
spatialreg (Bivand, 2022).

Qualitative method
To examine how PPPs are received in communities, we conducted 114 in-depth interviews with
small businesses (97) and community stakeholders (17) who worked closely with small businesses.
Supplementary ﬁle 7 provides a summary of the composition of the interviewees. In-depth inter-
views with small business owners were conducted to gain a comprehensive understanding of the
coping mechanisms businesses utilized during the current pandemic. Each interview lasted from
40–100 min and was conducted via zoom or phone, in English or Spanish as preferred by the
10
2156 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

participants. The in-depth interviews with community stakeholders focused on their roles and activ-
ities in small business development, especially in underserved communities. Each interview with
stakeholders lasted about 30–50 min and was conducted via zoom or phone. The process followed
the IRB protocol that had been approved by the leading author’s institution. To keep conﬁdentiality,
we use aliases for the participants in this paper.
Using the qualitative software package QSR NVivo, we employed the techniques of narrative
analysis outlined by Strauss and Corbin (1998) and Smith (2000) to analyze the interview tran-
scripts and notes. The results were based on common themes reﬂected by multiple community sta-
keholders, which allowed us to examine the process comprehensively. However, the scope of the
entire project was broader than that of the current paper, in which we are presenting only ﬁndings
that are relevant to the research questions for this study.

Findings
Spatial distribution of PPP loans
Supplementary File 8 shows the spatial distribution of PPP loans at the census tract level. In
general, neighborhoods in southwestern San Bernardino County and western Riverside County
received a larger number of PPP loans than other areas. This applied to both First and Second
Draw PPP loans. The pattern is slightly different if self-employed businesses (deﬁned as reporting
1 worker) are excluded – the PPP loans are a bit more widely dispersed – but, still, census tracts in
southwestern San Bernardino County and western Riverside County received more PPP loans.
Similar patterns apply to the aggregate values of loans. Our global spatial autocorrelation tests indi-
cated a spatial clustering effect for all the dependent variables (Supplementary File 8). For instance,
the estimates of Moran’s I statistic for the number and value of First draw PPP loans across neigh-
borhoods in inland southern California are 0.2696 and 0.1799, respectively, both statistically sig-
niﬁcant at the 5% level based on permutation-based inference (999 random spatial permutations). In
other words, the larger the number or the higher the value of loans received by a census tract, the
larger the number or the higher the value of loans received by its neighboring tracts. This clustering
effect in the dependent variables warranted special treatment with spatial econometric models in the
regression analyses.

Relationship between neighborhood characteristics and distribution of PPP loans
We discussed the direct, indirect, and total effects due to the endogenous and/or exogenous spatial
interaction effects in the methodology section (LeSage and Pace, 2009). Table 2 displays the results
for average direct effects, which represent the average response of the dependent variable to each
independent variable, with the effect of feedback loops incorporated. The average direct effect of
the proportion of small businesses vulnerable to the pandemic (RateVSB) is negative and signiﬁ-
cant for all cases except for the First Draw for employer businesses (both value and aggregate
number of loans). Speciﬁcally, a unit increase in the proportion of small businesses vulnerable
to the pandemic in one community is associated with a 0.98% decrease in the number of First
Draw PPP loans and a 1.46% decrease in the value of First Draw PPP loans in this community,
other conditions being constant. This also applies to Second Draw PPP loans allocated to employer
small businesses in terms of either the value or the aggregate number of loans.
Community household vulnerability (CRE3) also has a negative effect on loan distribution.
Except for Second Draw PPP loans to employer businesses, a higher level of community vulner-
ability predicts a lower number and value of PPP loans received by that community, ceteris
paribus, suggesting that vulnerable communities are less resourceful in obtaining ﬁnancial
Wang et
andal.Kang
Table 2. Average direct effects.

Dependent variable

All ﬁrms Employer ﬁrms

Aggregate number of loans Aggregate value of loans Aggregate number of loans Aggregate value of loans

First draw Second draw First draw Second draw First draw Second draw First draw Second draw

Model SDM SAR SDM SDM SDM SDM SAR SDM
log(NumLoan1) log(NumLoan2) log(ValLoan1) log(ValLoan2) log(NumLoanMore1) log(NumLoanMore2) log(ValLoanMore1) log(ValLoanMore2)
log(SB) 0.3175∗∗∗ 0.3438∗∗∗ 0.5854∗∗∗ 0.636∗∗∗ 0.5329∗∗∗ 0.5564∗∗∗ 0.7596∗∗∗ 0.9746∗∗∗
(0.0125) (0.0136) (0.0236) (0.0279) (0.0145) (0.0175) (0.0285) (0.0505)
RateVSB −0.0098∗∗∗ −0.0146∗∗∗ −0.0086 −0.0233∗∗∗ −0.0165∗∗∗ −0.023∗∗∗ −0.0029 −0.0277∗
(0.0034) (0.0033) (0.0064) (0.0079) (0.0042) (0.005) (0.0074) (0.015)
CRE3 −0.0136∗∗∗ −0.0123∗∗∗ −0.0196∗∗∗ −0.0159∗∗∗ −0.0072∗∗∗ −0.0054∗ −0.0219∗∗∗ −0.0084
(0.002) (0.0022) (0.004) (0.0049) (0.0025) (0.003) (0.0048) (0.0091)
Black 0.0189∗∗∗ 0.0249∗∗∗ −0.0145∗ −0.0142 −0.003 0.0013 −0.027∗∗∗ −0.0522∗∗∗
(0.004) (0.0034) (0.008) (0.0093) (0.0048) (0.0056) (0.0074) (0.0172)
Asian 0.0092∗∗ 0.0075∗∗ −0.0057 −0.0101 0.0022 0.0014 −0.0012 −0.0139
(0.0043) (0.0038) (0.0083) (0.0106) (0.0055) (0.0064) (0.0076) (0.0196)
Hispanic 0.0022 −0.0002 0.005 0.0077∗ −0.0014 −0.0049∗ 0.0047∗ 0.003
(0.0018) (0.0013) (0.0037) (0.0045) (0.0023) (0.0026) (0.0028) (0.0084)
Education 0.0104∗∗∗ 0.0104∗∗∗ 0.0181∗∗∗ 0.0257∗∗∗ 0.008∗∗ 0.0077∗ 0.0216∗∗∗ 0.0264∗∗
(0.003) (0.0027) (0.0059) (0.0073) (0.0035) (0.0041) (0.0055) (0.0124)
Note: ∗∗∗ p < 0.01; ∗∗ p < 0.05; ∗ p < 0.1; inside the parentheses are standard errors.

11
2157
12
2158 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

assistance. Community household vulnerability plays a larger role in determining overall First
Draw loan allocations than community business vulnerability, while that tendency is reversed
for the Second Draw. Speciﬁcally, a unit increase in the level of community household vulnerability
in one community is associated with a 1.36% decrease in the number of First Draw PPP loans and a
1.96% decrease in the value of First Draw PPP loans in this community, other conditions being con-
stant. These are much larger than the 0.98% and 0.86% (insigniﬁcant) decrease respectively asso-
ciated with a unit increase in community business vulnerability. However, the opposite is true for
Second Draw PPP loans: community business vulnerability is more important in helping access
PPP loans, both in terms of the total number and total value.
Spatial spillover effects are summarized as average indirect effects in Table 3. The proportion of
small businesses vulnerable to the pandemic (RateVSB) is negative and statistically signiﬁcant for
the aggregate number of Second Draw PPP loans, pointing to negative spatial spillover effects. That
is, a unit increase in the rate of vulnerable small businesses in a neighborhood is related to an 8.2%
decrease in Second Draw PPP loans received by all other neighborhoods, with geographically
closer neighborhoods impacted to a greater extent, ceteris paribus. In contrast, the positive and stat-
istically signiﬁcant estimated average indirect effect on Second Draw PPP loan to employer small
businesses suggests positive spatial spillovers. A unit increase in the proportion of vulnerable small
businesses in a community is related to a 4.75% increase in the total value of Second Draw PPP
loans received by employer businesses in its neighboring communities, ceteris paribus.
At the same time, community household vulnerability has more of a negative spatial spillover
effect, as indicated by the many negative and statistically signiﬁcant average indirect effects in
all the models except for the First Draw. This means that a more vulnerable residential community
is associated with a lower number and value of PPP loans received by other neighborhoods and
geographically closer neighborhoods were inﬂuenced to a greater extent. The negative spatial spill-
over effect is strongest for second draw PPP loan values: a unit increase in household vulnerability
in a community is associated with a 4.65% decrease in second draw PPP loan values received by
other neighborhoods.
Combining the average indirect and direct effects gives us the average total effect, as shown in
Table 4. The average total effect for an independent variable indicates the change in the dependent
variable if the independent variable undergoes a unit change in all the communities being scruti-
nized, other characteristics being constant. Due to the strong negative spatial spillover effect on
the number of Second Draw PPP loans caused by the proportion of small businesses vulnerable
to the pandemic (RateVSB), the average total effect of RateVSB is strengthened, from −0.0146
(average direct effect) to −0.0228. That is, a unit increase in the proportion of small businesses vul-
nerable to the pandemic in all communities leads to a decrease of about 2.28% in the number of
Second Draw PPP loans made to these neighborhoods, other conditions being constant.
Similarly, the average total effect of RateVSB is negative and signiﬁcant for the number of First
and Second Draw PPP loans allocated to employer businesses. On the other hand, the average
total effect for the level of community household vulnerability (CRE3) is negative and signiﬁcant
in all models. The negative spatial spillover effects worsened the situation of neighborhoods with
underlying vulnerabilities.

How PPP is received at the community level
Previous studies have suggested that government assistance is crucial to recovery from disastrous
events (Haynes et al., 2019; Josephson and Marshall, 2016). Consistently, the qualitative data
suggest that PPP loans directly help retain employees, pay the rent and the cost of PPE, keep the
business running, and provide certainty and a backup plan. In addition to addressing short-term
needs, PPP assistance has indirect and possible long-term effects. For instance, business owners
Wang et
andal.Kang
Table 3. Average indirect effects.

Dependent variable

All ﬁrms Employer ﬁrms

Aggregate number of loans Aggregate value of loans Aggregate number of loans Aggregate value of loans

First draw Second draw First draw Second draw First draw Second draw First draw Second draw

Model SDM SAR SDM SDM SDM SDM SAR SDM
log(NumLoan1) log(NumLoan2) log(ValLoan1) log(ValLoan2) log(NumLoanMore1) log(NumLoanMore2) log(ValLoanMore1) log(ValLoanMore2)
log(SB) 0.0664∗ 0.1919∗∗∗ 0.2107∗∗∗ 0.2223∗∗∗ 0.1167∗∗∗ 0.1619∗∗∗ 0.2104∗∗∗ 0.3353∗∗∗
(0.042) (0.0286) (0.0629) (0.0587) (0.0388) (0.0403) (0.0485) (0.1073)
RateVSB −0.005 −0.0082∗∗∗ 0.0025 0.0061 −0.0018 −0.0022 −0.0008 0.0475∗
(0.0099) (0.0021) (0.0148) (0.0143) (0.0088) (0.0099) (0.0021) (0.0263)
CRE3 −0.0087 −0.0069∗∗∗ −0.0064 −0.0175∗∗ −0.0168∗∗∗ −0.0151∗∗∗ −0.0061∗∗∗ −0.0465∗∗∗
(0.0063) (0.0016) (0.0091) (0.009) (0.0053) (0.0059) (0.002) (0.016)
Black 0.045∗∗∗ 0.0139∗∗∗ 0.0429∗∗∗ 0.0363∗∗∗ 0.0032 −0.0061 −0.0075∗∗∗ 0.0244
(0.0086) (0.0028) (0.0134) (0.0141) (0.008) (0.0092) (0.0026) (0.024)
Asian 0.0027 0.0042∗∗ 0.0189 0.0155 0.0058 0.0137 −0.0003 0.0136
(0.0094) (0.0021) (0.0147) (0.0156) (0.0089) (0.0101) (0.0022) (0.0284)
Hispanic −0.0029 −0.0001 −0.0038 −0.0056 −0.0018 0.001 0.0013 −0.0013
(0.0034) (0.0007) (0.0053) (0.0056) (0.0033) (0.0035) (0.0008) (0.0107)
Education −0.0052 0.0058∗∗∗ −0.007 0.0009 0.0077 0.0144∗∗ 0.006∗∗∗ 0.0217
(0.0065) (0.0017) (0.0098) (0.0104) (0.006) (0.0063) (0.002) (0.0181)
Note: ∗∗∗ p < 0.01; ∗∗ p < 0.05; ∗ p < 0.1; inside the parentheses are standard errors.

13
2159
14
2160

Table 4. Average total effects.

Dependent variable

All ﬁrms Employer ﬁrms

Aggregate number of loans Aggregate value of loans Aggregate number of loans Aggregate value of loans

First draw Second draw First draw Second draw First draw Second draw First draw Second draw

Model SDM SAR SDM SDM SDM SDM SAR SDM
log(NumLoan1) log(NumLoan2) log(ValLoan1) log(ValLoan2) log(NumLoanMore1) log(NumLoanMore2) log(ValLoanMore1) log(ValLoanMore2)
log (SB) 0.384∗∗∗ 0.5358∗∗∗ 0.7961∗∗∗ 0.8582∗∗∗ 0.6496∗∗∗ 0.7182∗∗∗ 0.9701∗∗∗ 1.3099∗∗∗
(0.0446) (0.0333) (0.065) (0.0565) (0.0387) (0.0403) (0.057) (0.1018)
RateVSB −0.0148 −0.0228∗∗∗ −0.0061 −0.0172 −0.0183∗∗ −0.0251∗∗∗ −0.0038 0.0198
(0.01) (0.0052) (0.0143) (0.0129) (0.0083) (0.0089) (0.0095) (0.0234)
CRE3 −0.0223∗∗∗ −0.0192∗∗∗ −0.026∗∗∗ −0.0334∗∗∗ −0.024∗∗∗ −0.0205∗∗∗ −0.028∗∗∗ −0.0549∗∗∗
(0.0067) (0.0036) (0.0092) (0.0083) (0.0053) (0.0059) (0.0063) (0.015)
Black 0.0639∗∗∗ 0.0388∗∗∗ 0.0284∗∗ 0.0221∗∗ 0.0003 −0.0048 −0.0345∗∗∗ −0.0278
(0.0084) (0.0057) (0.0115) (0.0109) (0.0069) (0.0076) (0.0094) (0.0188)
Asian 0.0118 0.0117∗∗ 0.0132 0.0054 0.008 0.0151∗ −0.0016 −0.0003
(0.0087) (0.0058) (0.0127) (0.0118) (0.0073) (0.0081) (0.0098) (0.0203)
Hispanic −0.0007 −0.0003 0.0011 0.0021 −0.0032 −0.0039 0.006∗ 0.0017
(0.0029) (0.002) (0.004) (0.0036) (0.0024) (0.0025) (0.0036) (0.0067)
Education 0.0052 0.0162∗∗∗ 0.0111 0.0266∗∗∗ 0.0157∗∗∗ 0.022∗∗∗ 0.0276∗∗∗ 0.048∗∗∗
(0.0063) (0.0042) (0.009) (0.0079) (0.0054) (0.0056) (0.007) (0.0147)
Note: ∗∗∗ p < 0.01; ∗∗ p < 0.05; ∗ p < 0.1; inside the parentheses are standard errors.

0(0)
EPA:
EPA:Economy
Economyand
andSpace
Space55(8)
Wang et
andal.Kang 15
2161

use the assistance to improve physical infrastructure and equipment, streamline the structure of the
labor force for the long term, and hire website developers for digital branding and marketing. As
Bella explains, “… the help I did already receive honestly was such a huge help because that actu-
ally allowed me to invest more in my marketing strategy. So that was for me what really like pushed
us to the next level of my business … It gave me an opportunity to get my employees raises and then
invest in advertising.”
Supplementary File 9 summarizes the main themes that are related to small businesses’ experi-
ences in access to the PPP loans. However, needs far exceed the available funds; furthermore, not
every business qualiﬁes for these programs. Even business owners who do qualify have difﬁculty
accessing the PPP (Bartik et al. 2020). As such, PPP assistance may not reach those communities in
the most need (Ong et al., 2021; Santellano, 2021). The statistical analyses demonstrate that both
the proportion of pandemic-vulnerable businesses and the vulnerability of residential communities
are negatively related to the number and value of loans received. The qualitative analysis clearly
identiﬁes the challenges facing local communities. With 81.6% of all small business establishments
in this region being nonemployers, many businesses may not qualify for the PPP at all.
Furthermore, qualiﬁcation is only half the battle. Unfamiliarity with how the programs work, con-
fusion about paperwork, and fear of additional debt may deter business owners from seeking help.
The following experiences are representative of the challenges encountered:

“I didn’t know how to ﬁll out the application. I think I ﬁll it out wrong because I did it on my own and it
was a lot of like questions, and I didn’t understand a lot of the stuff.” (Betty)

“… it was still kind of like a process of understanding what it was. How did it work? Um, you know,
what’s at stake too, because, you know, the forgiveness part am I going to be able to repay it? Are we
going to still be in business long enough to be able to repay that? Or is it now going to become, you
know, a personal debt that I’m going to carry on with my business? If I just, if I’m not able to stay
open and close, you know, those were all questions that came through my mind.” (Bella)

The businesses’ experiences are consistent with the comments made by community stake-
holders. For example, one representative from the federal government said that,

“[…] in some instances, it’s just the inability of our federal relief programs of really getting to the hands
of the people that need it the most. I think another challenge is, we don’t have a lever on additional
stimulus packages. […] Businesses have been struggling with receiving additional support from the
federal government. While the state has stepped in to offer some relief for small businesses, it’s still
not enough.” (SBM-4)

Before COVID-19, both businesses and stakeholders in these communities emphasized the need
for training and education in basic procedures such as initiating paperwork for starting a business,
ﬁnancial literacy, certiﬁcation, creating a business plan, and industry-speciﬁc knowledge.
Moreover, negative perceptions of place and opportunity and lack of resources have long frustrated
economic resilience and entrepreneurship. When Covid-19 hit, the pandemic disruption simply exa-
cerbated the challenges at both local and regional levels. As one stakeholder representative com-
mented, “… it’s what COVID did, you know, the crisis is a great, great revealer. It revealed the
exacerbation of the problems that existed way before COVID. […] It’s important for us to
know.” (HK1223, lender) A federal government agency participant echoed this complaint,
“Covid-19 has really laid bare the disparities that exist for businesses owned by people of color,
and women, and other and other minorities.” (SBM-4)
16
2162 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

Discussion
First, does the PPP reach its targets? The short answer is yes. The regression results indicate that the
total number of businesses signiﬁcantly predicts the total number and values of PPP loans received by
census tracts in the inland Southern Californian region. Consistent with existing studies (Bartik et al.,
2020), the qualitative analyses indicate that the PPP is key to coping with the pandemic disruption of
small businesses; however, as both regression results and interview data have shown, businesses in
the most need and/or vulnerable communities may not be able to access PPP assistance.
Second, the mismatch between the receiving of PPP loans and community vulnerability reﬂects
both the short-term needs and the long-term challenges facing entrepreneurship and business devel-
opment in disadvantaged communities (Bates et al., 2018; Fairlie and Robb, 2008; Wang, 2022).
Qualitative analyses suggest that information about, and understanding of, the programs are critical
for accessing government assistance. Successful recipients are bigger employers, have a higher level
of human capital, including ﬁnancial literacy, have good relationships with banks, are well-connected
to business communities and business service sectors, and/or use social media. Stakeholders also
suggest that federal government agencies and big corporations’ work with small businesses is
mainly executed through collaborations with local organizations and small business development
partners, using various databases of certiﬁed small businesses (e.g., minority-, women-, and
veteran-owned businesses). This means that as information, capital, and resources are allocated
and transferred through these networks, those who are not part of them will be excluded. To compli-
cate matters, most government agencies do not have a budget for marketing themselves.
Community-based organizations also face stringent ﬁnancial constraints and language limitations
for non-English speaking communities. Therefore, under COVID-19, it is largely up to the businesses
themselves to reach for support through their social networks. Thus, access to information and
resources is largely determined by businesses’ pre-pandemic connections to different networks.
Third, there is a close relationship between business recovery and community socioeconomic
conditions (Wang, 2022). The signiﬁcant spatial spillover effects that the statistical analyses
reveal indicate that vulnerability at the community level is negatively associated with the
number and total value of PPP loans granted in neighboring communities. In other words, the
implementation of government aid programs reaches far beyond the individual business and neigh-
borhood level. Qualitative analyses also suggest the blurring of boundaries between work and
home, family responsibilities, and childcare during the pandemic. Access to government programs
like the PPP offers more than mere assistance with business operations; it also alleviates business
owners’ personal challenges, which are intertwined with their business careers. The intertwining
relationship between small businesses and their local communities suggests that the design and
implementation of government assistance like PPPs can be improved by including additional
factors at the business owner level and their geographic location characteristics.
The current program was explicitly designed to focus on emergency relief for small businesses,
instead of targeting local residential communities (e.g., individual households). Should the program
be designed to fulﬁll both business and community needs? This is a fundamental issue for policy-
makers to consider before designing such a program. If business assistance is not considered as pro-
viding a framework for the residential community, it may cause unintended consequences,
including increased racial inequality or neighborhood disparities, as argued by existing studies
(e.g., Ong et al. 2021; Santellano, 2021).

Conclusion
The case of inland Southern California suggests that PPP loans have reached small businesses in
general, whether we measure distribution by the total number or the total value of loans, as business
Wang et
andal.Kang 17
2163

counts are strongly associated with both loan counts and total values. However, communities with
more pandemic-vulnerable businesses, or those with higher socioeconomic vulnerabilities, are less
likely to receive PPP loans. This indicates that these loans may fail to reach the communities in most
need. Furthermore, neighborhoods with neighboring communities that are vulnerable also tend to
receive lower numbers and lower values of PPP loans. This suggests the implementation and deliv-
ery process of government aid is closely related to the socioeconomic characteristics of local com-
munities. It calls for policymaking that assists small business recovery from a holistic perspective,
one that considers multiple factors at the business owner, business, and community levels.
Findings from this study have important implications for small businesses, government agen-
cies, and nonproﬁt organizations. Small businesses need to keep themselves as well informed as
possible by proactively researching information on ﬁnancial assistance; maintaining contact with
professionals, banks, business organizations, and local government agencies; and having a social
media presence. Government agencies need to maintain a ﬂow of accurate, timely, and consistent
information and guidance; communicate through social media; reach out to local people, especially
to traditionally hard-to-reach communities; and provide information in multiple languages.
Governments also need to innovate to adapt to new situations; listen to and understand their com-
munities’ needs; streamline government services; and incorporate public input and needs into sens-
ible policies. At the same time, services from nonproﬁt organizations have a more urgent role
during the pandemic disruption. These organizations should advocate for, and promote, their
mission and services in local communities; and stay relevant by suppressing misinformation and
disseminating accurate information. For traditionally underserved communities, inclusive entrepre-
neurship ecosystems are crucial for the region to grow. Providing readily available services to help
with education, technological developments, business strategy, and investment in social and ﬁnan-
cial support will open up the entrepreneurship ecosystem to traditionally underserved communities.

Declaration of conﬂicting interests
The author(s) declared no potential conﬂicts of interest with respect to the research, authorship, and/or publi-
cation of this article.

Funding
The author(s) disclosed receipt of the following ﬁnancial support for the research, authorship and/or publica-
tion of this article: The research for this study was supported by the NSF [2151970].

ORCID iD
Qingfang Wang https://orcid.org/0000-0002-7285-4027

Supplemental material
Supplemental material for this article is available online.

References
Adekola J and Clelland D (2020) Two sides of the same coin: Business resilience and community resilience.
Journal of Contingencies and Crisis Management 28(1): 50–60.
Aldrich DP (2012) Building Resilience: Social Capital in Post-Disaster Recovery. Chicago and London:
University of Chicago Press.
18
2164 EPA:Economy
EPA: Economyand
andSpace
Space55(8)
0(0)

Atkins R, Cook L and Seamans R (2022) Discrimination in lending? Evidence from the Paycheck Protection
Program. Small Bus Econ 58: 843–865. DOI: 10.1007/s11187-021-00533-1.
Bambra C, Norman P and Johnson NPAS (2021) Visualising regional inequalities in the 1918 Spanish ﬂu pan-
demic in England and Wales. Environment and Planning A: Economy and Space 53(4): 607–611.
Bartik AW, Bertrand M, Cullen Z, et al. (2020) The impact of COVID-19 on small business outcomes and
expectations. Proceedings of the National Academy of Sciences 117(30): 17656–17666.
Bartoloni E, Arrighetti A and Landini F (2021) Recession and ﬁrm survival: Is selection based on cleansing or
skill accumulation? Small Business Economics 57(4): 1893–1914.
Bates T and Robb A (2015) Has the community reinvestment act increased loan availability among small busi-
nesses operating in minority neighbourhoods? Urban Studies 52(9): 1702–1721.
Bates T, Bradford WD and Seamans R (2018) Minority entrepreneurship in twenty-ﬁrst century America.
Small Business Economics 50(3): 415–427.
Bivand R (2022) R packages for analyzing spatial data: A comparative case study with areal data.
Geographical Analysis 54(3): 488–518.
Borawski G and Schweitzer ME (2021) How well did PPP loans reach low- and moderate-income communi-
ties? Federal Reserve Bank of Cleveland, Economic Commentary 2021-13. DOI: 10.26509/frbc-ec-
202113.
Cirera X, Cruz M, Davies E, et al. (2021) Policies to support businesses through the COVID-19 shock: A ﬁrm
level perspective. The World Bank Research Observer 36(1): 41–66.
Cutter SL, Mitchell JT and Scott MS (2000) Revealing the vulnerability of people and places: A case study of
Georgetown county, south Carolina. Annals of the Association of American Geographers 90(4): 713–737.
Danes SM, Stafford K, Haynes G, and Amarapurkar SS (2009) Family capital of family ﬁrms: Bridging
human, social, and ﬁnancial capital. Family Business Review, 22(3): 199–215.
Dokhov R and Topnikov M (2021) Everyday mobility as a vulnerability marker: The uneven reaction to cor-
onavirus lockdown in Russia. Environment and Planning A: Economy and Space 53(4): 612–615.
Fairlie R and Fossen FM (2022) Did the paycheck protection program and economic injury disaster loan
program get disbursed to minority communities in the early stages of COVID-19? Small Business
Economics 58(2): 829–842.
Fairlie R and Robb AM (2008) Race and Entrepreneurial Success. Cambridge, MA: The.
Florida R, Rodríguez-Pose A and Storper M (2021) Cities in a post-COVID world. Urban Studies:
00420980211018072.
Gamboa S (2020) Few Hispanic business owners got coronavirus relief loans, Latino survey ﬁnds. Available at:
https://www.nbcnews.com/news/latino/few-hispanic-business-owners-got-coronavirus-relief-loans-latino-groups-
n1192086 (accessed 24 October 2021).
Haynes GW, Danes SM, Schrank HL, et al. (2019) Survival and success of family-owned small businesses
after hurricane katrina: Impact of disaster assistance and adaptive capacity. Journal of Contingencies
and Crisis Management 27(2): 130–144.
Josephson A and Marshall MI (2016) The demand for post-Katrina disaster aid: SBA disaster loans and small
businesses in Mississippi. Journal of Contingencies and Crisis Management 24(4): 264–274.
Kickul J, Grifﬁths MD, Robb CC, et al. (2021) All for one? The Paycheck Protection Program distribution
disparity. Journal of Entrepreneurship and Public Policy. 10(3): 323–335.
Kim D (2021) Visualizing the regional patterns of two crises: The COVID-19 outbreak and decreasing MSME
sales during three different phases of 2020 in Korea. Environment and Planning A: Economy and Space
53(7): 1591–1593.
LeSage J and Pace RK (2009) Introduction to Spatial Econometrics. Boca Raton, FL: Chapman and Hall/CRC.
Lo AY, Chow ASY, Liu S, et al. (2019) Community business resilience: Adaptation practice of micro- and
small enterprises around the Pearl River Estuary. Climatic Change 157(3–4): 565–585.
Marshall MI and Schrank HL (2014) Small business disaster recovery: A research framework. Natural
Hazards 72(2): 597–616.
Nowell B, Bodkin CP and Bayoumi D (2017) Redundancy as a strategy in disaster response systems: A
pathway to resilience or a recipe for disaster? Journal of Contingencies and Crisis Management 25(3):
123–135.
Wang et
andal.Kang 19
2165

Ofﬁce of Evaluation Science (2021) Increasing Access to Small Business Grant and Loan Programs for
Historically Underserved Groups. Available at: https://oes.gsa.gov/assets/abstracts/2006-smallbusinessaccess
report-nov2020.pdf (accessed 23 October 2021).
Ong P, Gonzalez S, Domínguez-Villegas R, et al. (2021) Disparities in the Distribution of Paycheck Protection
Program Funds in California’s Congressional Districts. Available at: https://latino.ucla.edu/wp-content/
uploads/2021/02/LPPI_PPP-Congressional-Analysis-Report_03.03.21.pdf.
Orhan E (2016) Building community resilience: Business preparedness lessons in the case of Adapazarı,
Turkey. Disasters 40(1): 45–64.
Peacock WG, Van Zandt S, Zhang Y, et al. (2014) Inequities in long-term housing recovery after disasters.
Journal of the American Planning Association 80(4): 356–371.
Rubin JS (2010) Venture capital and underserved communities. Urban Affairs Review 45(6): 821–835.
Santellano K (2021) Compounded inequality: How the US paycheck protection program is failing Los Angeles
latino small businesses. Ethnic and Racial Studies 44(5): 794–805.
SBA (2021a) PPP data. Available at: https://www.sba.gov/funding-programs/loans/covid-19-relief-options/
paycheck-protection-program/ppp-data (accessed 23 October 2021).
SBA (2021b) PPP loan forgiveness. Available at: https://www.sba.gov/funding-programs/loans/covid-19-
relief-options/paycheck-protection-program/ppp-loan-forgiveness (accessed 23 October 2021).
Smith CP (2000) Content analysis and narrative analysis. In: Reis HT and Judd CM (eds) Handbook of
Research Methods in Social and Personality Psychology. New York, NY, US: Cambridge University
Press, pp.313–335.
Strauss A and Corbin J (1998) Basics of Qualitative Research Techniques. Thousand Oaks, CA: Sage
Publications.
Teixeira C (2001) Community resources and opportunities in ethnic economies: A case study of Portuguese
and black entrepreneurs in Toronto. Urban Studies 38(11): 2055–2078.
Tibay V, Miller J, Chang-Richards A, et al. (2018) Business resilience: A study of Auckland hospitality sector.
Procedia Engineering 212: 1217–1224.
Torres AP, Marshall MI and Sydnor S (2019) Does social capital pay off? The case of small business resilience
after Hurricane Katrina. Journal of Contingencies and Crisis Management 27(2): 168–181.
Unidos US (2020) Federal Stimulus Survey Findings. Available at: https://theblackresponse.org/wp-content/
uploads/2020/05/COC-UnidosUS-Abbreviated-Deck-F05.13.20.pdf.
Wang Q (2022) Planning for an inclusive entrepreneurial ecosystem: COVID-19 and business resilience in
underserved communities. Journal of the American Planning Association: 1–15. DOI: 10.1080/
01944363.2022.2105740.
Wang Q and Kang W (2021) What are the impacts of COVID-19 on small businesses in the US? Early evi-
dence based on the largest 50 MSAs. Geographical Review 111(4): 528–557.
Webb GR, Tierney KJ and Dahlhamer JM (2000) Businesses and disasters: Empirical patterns and unanswered
questions. Natural Hazards Review 1(2): 83–90.
Wilson R and Din A (2018) Understanding and enhancing the US department of housing and urban develop-
ment’s ZIP code crosswalk ﬁles. Cityscape (Washington, D C ) 20(2): 277–294.
Xiao Y and Peacock WG (2014) Do hazard mitigation and preparedness reduce physical damage to businesses
in disasters? Critical role of business disaster planning. Natural Hazards Review 15(3): 04014007.
Xiao Y and Van Zandt S (2012) Building community resiliency: Spatial links between household and business
post-disaster return. Urban Studies 49(11): 2523–2542.
Zhai W, Liu M, Fu X, et al. (2021) American inequality meets COVID-19: Uneven spread of the disease across
communities. Annals of the American Association of Geographers 111(7): 2023–2043.

File and source

File: wang-2023-small-businesses-and-government-assistance-during-covid-19-evidence-from-the-pay.pdf
Size: 391,570 bytes
SHA-256: 405c696fe7773e70e07dc6dacfec956d969265fc98386bfa2ff3bd2212528c02
Our copy: wang-2023-small-businesses-and-government-assistance-during-covid-19-evidence-from-the-pay.pdf
Original: digital.library.unt.edu