How reducing documentation burdens impacts equity in access to small business COVID-19 relief funding
Summary
An accepted manuscript of the research article "How Reducing Documentation Burdens Impacts Equity in Access to Small Business COVID-19 Relief Funding" by Elizabeth Bell, Heather Barry Kappes and Miles Williams, published in Public Administration Review in 2023. Working with a local government in the U.S., the authors analyze outcomes for more than 8,500 businesses in a COVID-19 small business relief fund before and after the fund reduced its documentation requirements. Using an interrupted time series design, they report increased application success for underserved small businesses, but find that other applicants benefitted more from the change. The authors conclude that well-intentioned interventions can perpetuate disparity without targeted supports for marginalized communities.
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
Bell, Elizabeth, Kappes, Heather Barry & Williams, Miles (2023) How reducing
documentation burdens impacts equity in access to small business COVID-19 relief
funding. Public Administration Review, https://doi.org/10.1111/puar.13722
https://researchonline.lse.ac.uk/id/eprint/120047/
Version: Accepted Version
This document is the author’s accepted version of the journal article. There may be
differences between this version and the published version.
Items deposited in LSE Research Online are protected by copyright, with all rights
reserved unless indicated otherwise. They may be downloaded and/or printed for
private study, or other acts as permitted by national copyright laws.
LSE Research Online is the repository for research produced by the London School
of Economics and Political Science. For more information, please refer to our Policies
page or contact lseresearchonline@lse.ac.uk
Public Administration Review
How Reducing Documentation Burdens Impacts Equity in Access to Small Business
COVID-19 Relief Funding
--Manuscript Draft--
Manuscript Number: PAR-D-22-00251R2
Full Title: How Reducing Documentation Burdens Impacts Equity in Access to Small Business
COVID-19 Relief Funding
Short Title: How Reducing Documentation Burdens Impacts Equity
Article Type: Research Article
Keywords: Administrative Burden; social equity; program access
Corresponding Author: Elizabeth Bell
The University of Texas at Austin
Austin, Texas UNITED STATES
Corresponding Author Secondary
Information:
Corresponding Author's Institution: The University of Texas at Austin
Corresponding Author's Secondary
Institution:
First Author: Elizabeth Bell
First Author Secondary Information:
Order of Authors: Elizabeth Bell
Heather Barry Kappes
Miles Williams
Order of Authors Secondary Information:
Manuscript Region of Origin: UNITED STATES
Abstract: Scholars have unveiled how and why administrative burdens perpetuate inequity in
access to government benefits. However, less is known about the tools public
managers can use to reduce burdens and disparities in program access. We partner
with a local government in the U.S. to investigate whether a reduction in documentation
requirements increased equity by promoting access for “underserved” small business
owners (i.e. racial/ethnic minorities, women, people with disabilities, and veterans).
Specifically, we analyze outcomes for more than 8,500 businesses in a COVID-19
small business relief fund before and after a reduction in documentation requirements.
Utilizing an interrupted time series design, we find increased application success for
underserved small businesses, but other applicants benefitted more from the program
change, calling into question whether the policy change advanced equity. Our findings
suggest that even well-intentioned interventions can perpetuate disparity if targeted
supports for marginalized communities are not implemented.
Response to Reviewers:
Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Manuscript
1 of 61
How Reducing Documentation Burdens Impacts Equity in Access to Small
Business COVID-19 Relief Funding
Elizabeth Bell
Lyndon B. Johnson School of Public Affairs
The University of Texas at Austin
2315 Red River St, Austin, TX 78712
Elizabeth.bell@austin.utexas.edu
Heather Barry Kappes
London School of Economics and Political Science
Houghton St, London WC2A 2AE, United Kingdom
h.kappes@lse.ac.uk
Miles Williams
Denison University
100 W College St, Granville, OH 43023
williamsmd@denison.edu
Key Words: administrative burden; COVID-19; public administration; local discretion
Acknowledgements: We thank our city partners who were kind enough to share data to build
evidence on equity in small business relief funding, and the staff in other cities who took the time
to talk about their programs and applicant challenges. We thank Ryan Moore, Jasper Cooper,
Mattie Toma, David Schwegman, Rebecca Johnson, Kelly Bidwell, and so many others on the
Office of Evaluation Sciences team for helpful feedback in the development of this project.
2 of 61
Bios
Elizabeth Bell is an Assistant Professor in the Lyndon B. Johnson School of Public Affairs at
The University of Texas at Austin. Her research seeks to enhance the ability of government to
meet equity-enhancing policy goals. Her research examines the equity implications of citizens'
experiences of administrative burdens, bureaucrats' uses of discretionary power, and the politics
of policy design.
Heather Barry Kappes is Associate Professor (Education) in the Department of Management at
the London School of Economics and Political Science. She takes a behavioural science
approach to researching and teaching marketing, trying to understand what influences people as
they pursue their goals. She applied these ideas to government programs as a Fellow at the
Office of Evaluation Sciences, United States General Services Administration.
Miles D. Williams is a Visiting Assistant Professor for Data for Political Research at Denison
University. His research encompasses issues in political economy, public policy, and causal
inference. He is particularly interested in understanding the role of politics in shaping the
allocation of economic assistance, both international and domestic.
Abstract
Scholars have unveiled how and why administrative burdens perpetuate inequity in access to
government benefits. However, less is known about the tools public managers can use to reduce
burdens and disparities in program access. We partner with a local government in the U.S. to
investigate whether a reduction in documentation requirements increased equity by promoting
access for “underserved” small business owners (i.e. racial/ethnic minorities, women, people
with disabilities, and veterans). Specifically, we analyze outcomes for more than 8,500
businesses in a COVID-19 small business relief fund before and after a reduction in
documentation requirements. Utilizing an interrupted time series design, we find increased
application success for underserved small businesses, but other applicants benefitted more from
the program change, calling into question whether the policy change advanced equity. Our
findings suggest that even well-intentioned interventions can perpetuate disparity if targeted
supports for marginalized communities are not implemented.
3 of 61
Evidence for Practice:
1) Documentation requirements for federal funding can impose administrative burdens that
disproportionately prevent marginalized communities from accessing public programs.
2) Local governments can leverage alternative funding streams to reduce the number of
documentation requirements.
3) Reducing documentation requirements increased access to COVID-19 relief funds for
small businesses owned by individuals from historically underserved communities.
4) Small businesses owners from non-underserved communities benefitted more from the
reduction in documentation requirements than business owners from historically
underserved communities.
4 of 61
For decades, public administration scholars have sought to establish social equity as a
foundational anchor for good governance (Blessett et al. 2019; Frederickson 1980; 1990; Gooden
2015; 2014; Johnson and Svara 2011; Svara and Brunet 2005). However, inequities still abound
across policy areas and have been exacerbated by the COVID-19 pandemic (Wright and Merritt
2020; Alon et al. 2020; Baker 2020). We draw on the National Association for Public
Administration’s definition of social equity to inform our work, which states that social equity is
“the fair, just and equitable management of all institutions serving the public directly or by
contract, and the fair and equitable distribution of public services, and implementation of public
policy, and the commitment to promote fairness, justice, and equity in the formation of public
policy” (Svara and Brunet 2005 p. 256; Johnson and Svara 2011). A key mechanism by which
the administrative state has perpetuated and exacerbated inequities is the imposition of
administrative burdens on marginalized communities seeking access to government support;
these burdens range from a lack of information or clarity on eligibility, complex and demanding
paperwork, and psychological burdens like stigma, frustration, and loss of autonomy (Herd and
Moynihan 2018; Christensen et al. 2020; Keiser and Miller 2020; Chudnovsky and Peeters 2020;
Peeters and Campos 2021). These state-imposed administrative burdens reduce equity in
program access, hinder program effectiveness, and diminish the political efficacy that forms the
foundation of a democratic society (Herd and Moynihan 2018; Christensen et al. 2020; Soss,
Fording, and Schram 2011; Heinrich 2016; Barnes 2020; Baekgaard et al. 2021). Given the
charge of public administration scholars to treat social equity as a central pillar of good
governance (Blessett et al. 2019; Frederickson 2010; Wright and Merritt 2020), better
understanding the causes, consequences, and solutions for addressing administrative burdens in
government programs is of paramount importance to public administration research, theory, and
5 of 61
practice (Biden 2021; Office of Management and Budget 2021). In fact, in countries around the
world, scholars and policymakers have begun conceptualizing, measuring, and attempting to
reduce administrative burdens on the public (e.g., SCM Network 2004; Arendsen et al. 2014;
Herd and Moynihan 2018; Nielsen et al. 2017). Concerns about equity were heightened by the
COVID-19 pandemic, as racial disparities in illness and death became apparent (Shah, Sachdeva,
and Dodiuk-Gad 2020).
While extant literature has significantly developed our ability to diagnose the causes and
consequences of administrative burden (Peeters 2019; Chudnovsky and Peeters 2020; Masood
and Nisar 2021; Heinrich 2016), less attention has been paid to the solutions that policymakers
can use to alleviate administrative burden and enhance social equity. Prior literature has
identified two dominant approaches to reducing administrative burden. First, at the front-lines of
government, when administrative capacity is not too constrained and there is sufficient
discretionary power, street-level bureaucrats and third-party organizations can cut red tape on
behalf of their clients (Bell and Smith 2021; Watkins-Hayes 2011; Wiley and Berry 2018;
Heinrich 2016; Heinrich et al. Forthcoming). However, street-level bureaucrats act based on
cognitive biases that leave room for discrimination in the allocation of benefits and sanctions,
especially when capacity is constrained (Soss, Fording, and Schram 2011; Andersen and Guul
2019; Olsen, Kyhse-Andersen, and Moynihan 2020; Guul, Pedersen, and Petersen 2021; Guul,
Villadsen, and Wulff 2019). Therefore, this reduction in administrative burden may be
inequitably applied in cases where implicit bias prevails (Andersen and Guul 2019; Assouline,
Gilad, and Ben-Nun Bloom 2021). Second, moving away from the street-level to public
managers, scholars have examined whether structural changes to program requirements (i.e.,
enrollment mechanisms), and behaviorally informed communications interventions (i.e., nudges)
6 of 61
(Linos, Quan, and Kirkman 2020; Linos and Riesch 2020; Pepin, O’Leary, and Oberlee 2021;
Lopoo, Heflin, and Boskovski 2020; Herd et al. 2013; Hattke, Hensel, and Kalucza 2020;
Baekgaard et al. 2021; Larsson 2021) reduce administrative burden and improve program access.
Some studies find that nudges improve program take-up (Lopoo, Heflin, and Boskovski 2020;
Linos et al. 2022; Marx and Turner 2019; DellaVigna and Linos, 2021), but others find this
approach insufficient for addressing deep-rooted structural inequities (Linos et al. 2022; Weimer
2020; de Ridder, Kroese, and van Gestel 2021), and many studies do not examine for whom take-
up improves (Frederickson 1990; Guy and McCandless 2012).
In this paper, we investigate how reductions in documentation requirements influence
equity rather than overall program take-up for a unique target population that has yet to be
explored in the literature—small business owners. Prior work focuses on relatively homogenous
disadvantaged communities in means-tested programs, whereas our work leverages variation
across advantaged and disadvantaged business owners. To advance the theoretical foundations of
the administrative burden literature, we utilize recent conceptualizations of social equity in
public administration to form expectations for what an “equity-enhancing” policy would look
like in practice. In doing so, we depart from prior literature by examining whether the policy
change closed the equity gaps in access and outcomes rather than just providing a subgroup
analysis, which more effectively accounts for historical and present-day bias in the
administration of public programs. Specifically, we measure whether the burden reduction effort
disproportionately benefited underserved small businesses applicants, which would close the pre-
existing equity gap in the proportion of applicants from underserved communities.
We partnered with a local government, which wished to remain anonymous, in the U.S.
during the COVID-19 pandemic in 2020 to evaluate a structural policy change that reduced
7 of 61
documentation requirements for a city small business relief fund. We chose to study these
questions in the context of the U.S. because the Biden administration established social equity as
a key goal for all bureaucratic agencies, and also encouraged agencies to document and reduce
administrative burdens in public programs (Biden 2021). As part of an executive order, the Biden
administration established that “affirmatively advancing equity, civil rights, racial justice, and
equal opportunity is the responsibility of the whole of our government. Because advancing
equity requires a systematic approach to embedding fairness in decision-making processes,
executive departments and agencies (agencies) must recognize and work to redress inequities in
their policies and programs that serve as barriers to equal opportunity” (Biden 2021). This
executive order added to a context in which heightened concerns about racial inequities of
COVID-19 were becoming apparent (Shah, Sachdeva, and Dodiuk-Gad 2020), and where the
murder of George Floyd produced a sense of urgency around addressing racism (Meikle and
Morris 2022; Nguyen et al. 2021). Although similar issues were raised across the world,
government officials in the U.S. were particularly interested in partnering with researchers to
document the ways in which they were adhering to calls to action from the White House.
We leverage an interrupted time series approach to examine the impacts of reducing
documentation requirements on equity in access to small business COVID-19 relief funding,
which we conceptualize as improving outcomes for underservedi communities (minority-owned,
women-owned, disabled-owned, and veteran-owned small businessesii) more than non-
underserved communities (White, Male, non-Disabled, non-Veterans). After all, if the policy
change helped non-underserved communities more than underserved communities, this would
serve to perpetuate inequity in the absence of other targeted supports for underserved business
owners. We hypothesized that reductions in documentation requirements would increase the
8 of 61
likelihood that businesses owned by members of historically underserved groups gain access to
critical financial support to help weather the COVID-19 pandemic. Results of the interrupted
time series analysis support this hypothesis, showing that the reduction in administrative burdens
led to substantial increases in the likelihood that underserved small businesses submitted
complete applications, went under review, and were funded. However, we also find that at the
time of the policy change, the reduction in documentation requirements helped non-underserved
small businesses more than underserved small businesses. This differential effectiveness calls
into question whether the policy change was “equity-enhancing.” In the Discussion section, we
propose some explanations for why this policy solution may have helped non-underserved small
businesses more. We conclude by proposing research which could identify scenarios in which
reductions in administrative burden may enhance social equity.
Our findings contribute to the growing literature on solutions to administrative burden in
two ways. First, we provide some of the only evidence examining whether reducing compliance
demands increases equity in outcomes. While other studies have examined inequities (either
disparities or disproportions) in process, access, and quality caused by administrative burden, the
studies on unburdening the public focus on overall program take-up rather than closing the
disparities in outcomes across marginalized and non-marginalized clients. We contribute to this
literature by drawing from the rich body of work on social equity in public administration—
which recognizes the historical and present-day injustices in administration—to inform both our
conceptualization and measurement of the impact of burden reduction on disparities in outcomes.
Second, in response to the growing body of literature showing the limitations of nudges in
producing improvements in longer term outcomes of interest, we examine a heavier-handed,
structural policy solution that impacted all applicants unlike more targeted policy changes. In
9 of 61
this way, we provide evidence to public managers interested in reducing administrative burdens
on how reducing documentation requirements may impact the success of underserved applicants.
Administrative Burden and Social Equity
Administrative burdens impose multiple costs for citizens attempting to access
government benefits, including learning, compliance, redemption, and psychological costs
(Moynihan, Herd, and Harvey 2015; Barnes 2020). When applying for public programs, clients
must spend substantial time and effort learning about program eligibility (learning costs),
compiling complex paperwork and attending required meetings with government officials
(compliance costs), redeeming benefits with third party vendors (redemption costs), and
overcoming negative emotional responses such as stress, fear, stigma, shame, and loss of
autonomy (psychological costs) (Hattke, Hensel, and Kalucza 2020; Baekgaard et al. 2021;
Daigneault and Macé 2019; Reijnders, Schalk, and Steen 2018). Other conceptualizations of
administrative burden on businesses rather than individuals have measured the price, time, and
quantity of additional “administrative activities businesses conduct because regulation requires
it" (SCM Network 2004). This standard cost model, developed by the OECD enables a
generalizable measure of the informational obligations and administrative activities of
regulations, but neglects learning and psychological costs, which we find are also present for
businesses applying for public programs (SCM Network 2004). Regardless of the measurement,
experiences of costly encounters with government undermine program take-up (Heinrich 2016;
Heinrich 2018; Nisar 2017; Brodkin and Majmundar 2010; Homonoff and Somerville 2020;
Gray et al. 2021; Carey, Malbon, and Blackwell 2021), especially when there are no supportive
street-level bureaucrats or third-party organizations to take on the burden of applying (Wiley and
Berry 2018; Bell and Smith 2021; Heinrich et al. Forthcoming).
10 of 61
To make matters worse, administrative burdens are not equally distributed; they
undermine program access for marginalized communities the most (Herd and Moynihan 2018;
Ray, Herd, and Moynihan 2022; Blume 2022; Masood and Nisar 2021). This creates a catch-22
wherein those individuals who need the most help from government are often the least able to
access critical support from street-level bureaucrats (Bell and Smith 2021) and gain access to
public benefits (Christensen et al. 2020; Masood and Nisar 2021; Döring 2021; Hattke, Hensel,
and Kalucza 2020). Underlying causes of the unequal distribution of burdens may be explained
by both citizen (i.e., disparities in human/social capital) and state factors (Christensen et al.,
2020; Chudnovsky & Peeters, 2021) but regardless of the cause administrative burdens have
serious consequences for policy outcomes. These inequities in the distribution of administrative
burden have undermined outcomes for LGBTQA+ communities in Pakistan (Nisar 2017),
impoverished adolescents in South Africa (Heinrich and Brill 2015), racially minoritized
communities in the U.S. (Soss, Fording, and Schram 2011; Watkins-Hayes 2009; Ray, Herd, and
Moynihan 2022), people with less administrative literacy/capital (Masood and Nisar 2021;
Döring 2021), and people struggling with physical or mental health issues (Christensen et al.
2020; Bell et al. 2021). For example, when administrative burdens are increased—like when
social security offices in the U.S. that provide application assistance close—take-up of disability
benefits is significantly reduced, especially for those with moderately severe conditions and low
education levels (Deshpande and Li 2019). Even when a tool such digitalization and automation
are deployed in efforts to make child benefit decisions automatic and less burdensome in
Norway, scholars find that inequities still persist for lower income families who are
disproportionately forced to apply manually (Larsson 2021). Moreover, recent scholarship
demonstrates that on top of administrative burdens not being equally distributed, the assistance
11 of 61
necessary to overcome burdens is also unequally distributed (Bell and Smith 2021; Heinrich et
al. 2022). In impoverished communities, there is less support and capacity to enhance access to
programs that promote the health and educational opportunities of children (Heinrich et al.
2022).
Tools for Unburdening the Public
While these inequities in the effects of administrative burdens are well documented, less
is known about the policy tools that public managers and street-level bureaucrats can leverage to
reduce burdens for marginalized communities and enhance social equity. Instead of focusing on
tools to advance social equity, recent scholarship has focused on tools to increase overall take-up
without regard to specific marginalized communities. There are two distinct approaches to the
reduction in administrative burden in the literature on policy solutions: 1) structural changes to
program design and enrollment mechanisms, and 2) behaviorally informed communications
interventions, or nudges, that influence citizen behavior and resilience to burden rather than the
structure of government programs.iii In the first approach, which examines structural changes to
program rules governing the documentation and verification processes in applications, scholars
have found that reducing compliance demands can reduce psychological costs and shifting
burden onto the state can increase take-up (Herd et al. 2013; Baekgaard et al. 2021; Fox, Stazyk,
and Feng 2020; Gray et al. 2021). Specifically, in Denmark, decreasing compliance demands for
unemployment insurance applicants during COVID-19 reduced psychological costs, in the form
of stress and loss of autonomy (Baekgaard et al. 2021). In Wisconsin, the BadgerCare program
increased enrollment in Medicaid by implementing auto-enrollment, using marketing to
encourage people to apply, and putting the burden of verification on the state (Herd et al. 2013).
Moving beyond just Wisconsin as a case, scholars have found that states that eased the
12 of 61
administrative burden in Medicaid applications following the Affordable Care Act saw the
largest increases in Medicaid uptake, even after adjusting for changes in eligibility requirements
(Fox, Stazyk, and Feng 2020). Therefore, reducing administrative burden in the applications
processes for public programs translates to increases in overall take-up, but whether take-up
increases for the most marginalized communities is still up for debate. In fact, none of these
studies investigating the impacts of structural changes examine equity in outcomesiv, which is a
key oversight given the substantial evidence demonstrating that increases in administrative
burden impact marginalized communities the most. We build on this work by investigating
whether structural policy changes that reduce compliance demands can enhance equity in citizen
outcomes.
Rather than focusing on structural changes that reduce burdens, other scholars have taken
a different approach, highlighting the impacts of nudges—typically in the form of behaviorally
informed communications interventions that shift choice architecture—on experiences of
administrative burden, compliance with regulations, and program take-up (Thaler and Sunstein
2009). In the U.S., nudges such as targeted text messages, personalized phone calls, and
personalized letters have increased take-up of programs like Supplemental Nutrition Assistance
Program (SNAP) and Earned Income Tax Credit (EITC) (Lopoo, Heflin, and Boskovski 2020;
Bhargava and Manoli 2015) and enhanced compliance with municipal housing codes (Linos,
Quan, and Kirkman 2020). Overall, nudges have been shown to produce behavior change in
target populations (Mertens et al. 2022), but whether those behavioral changes are enough to
impact program take-up and overcome structural barriers is still unclear (Linos, Reddy, and
Rothstein 2022; Linos et al. 2022). Recent attempts to implement nudges in the Temporary
Assistance for Needy Families (TANF) program (Pepin, O’Leary, and Oberlee 2021) and a
13 of 61
replication of the EITC nudge experiment failed to produce the same positive effects on overall
take-up (Linos et al. 2022). Ultimately, in some cases nudges may serve to increase take-up, but
whether nudges can increase outcomes for the most marginalized communities who often need
more personalized, hands-on support to overcome administrative burdens is still unclear
(Weimer 2020; de Ridder, Kroese, and van Gestel 2021). This makes our investigation into
structural changes even more important, as it presents a departure from the emphasis on nudges
as a policy solution to administrative burden.v
Taken together, while the literature on administrative burdens has employed an equity
lens to provide insight into the disparities caused by burdens, the next area for development
concerns how to alleviate those inequities. In this new area of scholarship, researchers should
strive to not only theorize how to increase overall take-up of public programs but how to “reduce
(and ultimately eliminate) disparity, marginalization, and discrimination while increasing social
and political inclusion” (Blessett, Fudge, and Gaynor 2017, p. 11). While there is some
agreement on general principles and moral imperative of social equity in public administration
(Guy and McCandless 2012), there are multiple components to this concept (Allbright et al.
2018) and multiple theoretical perspectives on what it means to advance social equity in public
administration. In terms of the components, scholars have argued that public administrators
should follow the following commitments to social equity: “(1) procedural fairness, meaning due
process, equal protection and equal rights; (2) equity in the availability of service and benefits;
(3) equity in the process of providing services and benefits; (4) an equal level of outcomes for all
groups; and (5) a guarantee of a place at the table to express views on policy choices and service
delivery” (Guy and McCandless 2012, S12). Moving to the theoretical perspectives, a libertarian
view of equity emphasizes equal rules for everyone and an equal distribution of inputs (also
14 of 61
referred to as procedural justice by Stivers et al. (2023)) which echoes the Weberian “neutrality”
that more recent scholarship has called into question (Portillo, Humphrey, and Bearfield 2022).
On the other hand, a transformative view of equity and social justice takes an antithetical
approach that emphasizes challenging racism and classism through structural changes that take
into account the intersectionality of marginalized identities (Allbright et al. 2018; Stivers et al.
2023; Trochmann et al. 2022). In this paper, we employ a transformative social equity lens in the
context of administrative burden that acknowledges the inequities administrative burdens and the
pandemic create for marginalized communities (Harper-Anderson, Albanese, and Gooden 2023).
In turn, we argue that policy solutions reducing burdens must address the disparities caused by
government policies that impose disproportionate burdens on marginalized communities. This
lens highlights a previously overlooked flaw in the literature on solutions to administrative
burden. Namely, we argue that to understand whether a policy change aimed at reducing burdens
is “equity-enhancing,” we must understand who benefits from the policy change in question
(Frederickson 2010; Berry-James et al. 2021; Guy and McCandless 2012). Thus far,
administrative burden literature has yet to thoroughly address what combination of policy
solutions reducing burdens would serve to benefit marginalized communities. Therefore, existing
policy solutions identified in the literature could be benefitting relatively advantaged
communities more than those facing intersectional structural disadvantage (i.e. helping White
low-income applicants, but not low-income people of color), serving to perpetuate existing
disparities rather than diminish them. For instance, if a reduction in documentation requirements
primarily helps those with high levels of administrative capital/literacy or those that have more
financial resources, the policy could be increasing take-up but perpetuating existing inequity. To
create a policy tool that reduces burden and is “equity-enhancing”—that diminishes or eliminates
15 of 61
disparities in access and outcomes—we argue that the policy change would have to not only
impact take-up among relatively advantaged groups but also for the most severely disadvantaged
who face intersectional systems of exclusion and marginalization (Wright and Merritt 2020;
Blessett et al. 2019; Berry-James et al. 2020). This could necessitate a combination of both
structural changes to the program that impact all applicants, and targeted communications
interventions and support programs that help marginalized communities navigate the application
process. In the remaining sections, we present our specific case—access to small business
COVID-19 relief funds—and how administrative burdens caused striking disparities in funding
receipt. We also present the policy solution we evaluate, which sought to increase social equity
by reducing compliance demands. Then, we evaluate whether the policy closed equity gaps, or
whether disparity prevailed.
Administrative Burden and Equity in Access to Small Business COVID-19
Relief Funds
Historical discrimination has created inequities in small business ownership and growth
in the U.S. For instance, in 2016, women were 51% of the U.S. population but 33% of the small
business owners; racial and ethnic minorities were 40% of the U.S. population but 20% of the
business owners (Liu and Parilla 2020). The disproportionate underrepresentation of underserved
communities in the small business community is likely due to multiple reasons including
process: discrimination among loan officers reviewing loan applications and fairness: a lack of
engagement with underserved communities in traditional economic development offices in local
and state governments, which result in the disparities we see in access and outcomes (Perry
2020; Guy and McCandless 2012). Businesses owned by Black and Latino Americans have less
working capital than White-owned businesses, and women-owned businesses have less working
16 of 61
capital than men-owned businesses (Robb, 2013). Indeed, compared to White business owners,
Black and Hispanic business owners are less likely to start their business with a loan from a bank
or financial institution, more likely to start business with less than $5,000, more likely to report
unmet credit needs, twice as likely to not receive the full amount of funding requested from the
bank, and more likely not to apply for loans due to rejection fears (Robb 2018; Fairlie and Robb
2010). In addition to the street-level discrimination at the micro-level, there are also meso-level
factors such as intergenerational wealth differences that drive disparities as underserved
businesses face disproportionate barriers in credit verification and the provision of collateral for
loans to start businesses (Perry 2020). In sum, inequities stem not only from active bias and
discrimination, but also from systems that reproduce and exacerbate disparities between groups.
Emerging evidence suggests that these underserved communities, which were already
underrepresented in small business ownership and growth opportunities, were particularly
vulnerable during the COVID-19 pandemic beginning in 2020 (Harper-Anderson, Albanese, and
Gooden 2023). Compared to the 22% decrease in active business ownership across all U.S.
businesses early in the pandemic, declines were larger for Black (41%) and Latinx (32%)
owners, and slightly larger for female (25%) owners (Fairlie, 2020). By the end of March 2020,
Black-owned businesses had cash balances 26% lower than the year before, compared to a 12%
decrease for all firms (Farrell, Wheat, and Mac 2020). Although we are not aware of data on
pandemic impacts specific to disabled or veteran-owned small business owners, veterans with
physical and mental health challengesvi and people with disabilities were vulnerable during
COVID-19 and faced disproportionate challenges in the workforce (National Council on
Disability 2021). In sum, the early indications of disparate negative impacts for underserved
business owners, on top of existing disproportionate barriers for these businesses, gave rise to
17 of 61
pressing questions about how to target recovery efforts in a way that reduced rather than widened
disparities.
Study Setting
To study the impacts of burden reduction efforts on equity in program access, we focus
on a policy change that reduced the number of required documents applicants had to provide to
access one city’s small business COVID-19 relief fund. This work was initiated as part of a
collaboration between the Office of Evaluation Sciences and the Small Business Administration
to build evidence on measures to advance equity as part of the effort to implement the Executive
Order on Advancing Racial Equity and Support for Underserved Communities Through the
Federal Government (Biden 2021). This policy established advancing equity as a key priority
and required that “executive departments and agencies (agencies) must recognize and work to
redress inequities in their policies and programs that serve as barriers to equal opportunity”
(Biden 2021). In this way, the pursuit of social equity was legally required for all federal
agencies in the U.S., which is a unique element of the policy context that may not translate to
other settings. As part of these efforts in the U.S., federal agencies were required to identify
methods to assess equity, conduct equity assessments of key programs, allocate resources to
advance fairness, promote the equitable delivery of government benefits, engage with members
of underserved communities, and establish an Equitable Data Working Group.
The Executive Order on racial equity contributed to a context in which cities were also
guided by federal, state, and local legislation on the use of COVID-19 funding. For instance, one
state legislature allocated money to support hospitality-industry businesses, specifying priority to
businesses with a 50% revenue reduction. This requirement meant that counties disbursing the
18 of 61
funds had to require applicant businesses to submit Profit and Loss statements, even though local
officials knew many businesses struggled to provide these (Office of Evaluation Sciences 2021).
This city we focus on created an Emergency Microloan program in March 2020 to
distribute $11 million to small businesses located in the city shortly after the city was put under
lockdown because of COVID-19. The program was implemented by the city’s Economic
Development Department (EDD) and built on an existing microloan framework but had lower
interest rates (either 0% for a term of six months to one year, or 3% to 5% for a term of up to five
years) and a lower loan limit ($5000 - $20,000) than the standard Microloans offered by the
EDD (which range from $5000 - $50,000 at 7-9% interest over 1-5 years). Funds were
distributed on a first-come, first-served basis, meaning that program staff processed applications
in the order they were received, until funding was exhausted. In December 2020, the City
Council approved the conversion of loans to grants if businesses complied with the provisions of
the loan agreement (i.e. spending the funds in accordance with The Coronavirus Aid, Relief, and
Economic Security Act or, CARES Act requirements). These requirements included not
relocating the business outside the city and submitting a certificate of compliance with all
provisions.
To be eligible for the program, businesses had to have 100 or fewer employees, be
negatively impacted by the COVID-19 pandemic, and try to continue or re-establish business
operations after lockdown subsided. Businesses needed a tax registration certificate with the city
and had to demonstrate sufficient credit and cash flow, and if the credit was insufficient the
applicant needed a co-signer. For eligible businesses to receive an emergency microloan, they
had to submit an online application to the EDD, along with supporting documentation
(“submitted” stage). The application was reviewed by program staff, who conducted a
19 of 61
preliminary review to highlight potential ineligibilities or aspects where further documentation
was required. Once program staff deemed the application complete and likely eligible for
funding, it joined a pool of “under review” applications, which were processed roughly in order
of receipt, then undergoing further review by loan officers and with funding awarded until the
pool was depleted (“funding awarded” stage).
If applicants progress through the application stages with equal probability, then an intervention
that increases the likelihood of submission should have a persistent effect on outcomes at later
stages, although the effect may be smaller as the outcomes become progressively less likely.
However, given that progression through these application stages remained somewhat opaque even
after repeated conversations with program staff it is worth examining effects at each stage
empirically. We return to this point below.
Diagnosing Key Barriers in the Process of Applying for COVID-19 Small Business Relief
To better understand the challenges small businesses faced in accessing relief funds in the
U.S., we spoke with around 15 people in a variety of roles across the country including
community nonprofits, small business development centers, chambers of commerce, and city
program directors. These informal interviews highlighted barriers in two categories: behavioral
and structural. For more details on this data collection effort, see Appendix A.
The most notable behavioral barriers were: 1) lack of awareness 2) confusion and
uncertainty and 3) distrust of local governments. Many small business owners from underserved
communities were unlikely to know about the programs; those who did learn about the programs
were frequently confused about how to fill out required financial documents. For example, Profit
& Loss statements and Balance sheets were required by many grant and loan programs; these
documents were particularly burdensome for low-volume cash businesses without a staff
20 of 61
accountant, such as those with sole proprietors, to generate. These businesses are
disproportionately owned by those from racially minoritized groups and socioeconomically
disadvantaged individuals (Morel, Al Elew, and Harris 2021). This confusion and uncertainty, on
top of the time and resources it took to compile and submit documentation, appeared to be
particularly burdensome for racial and ethnic minorities and women, who were hit harder by the
COVID-19 pandemic (i.e. sick friends and family, demands for child care, job losses, and other
caring responsibilities during the COVID-19 pandemic) (Alon et al. 2020; Sevilla and Smith
2020). Interviewees also described uncertainty regarding how many businesses would be funded
and the likelihood that a potential applicant would benefit. This uncertainty was compounded by
a lack of trust in government due to historical and contemporary discrimination in local
government policies and practices. Many underserved small business owners believed that
application hurdles were indicative of the city trying to exclude people “like them” from getting
help. For it to seem worthwhile to put in the time and effort to provide application
documentation, small business owners had to believe that the government would fund someone
“like them”.
The most important structural barriers that were mentioned in the interviews were: 1)
disparities in digital access, 2) disparities in access to capital for collateral, and 3) disparities in
technical business knowledge. Putting together a profit and loss statement or a balance sheet
requires some level of technical business knowledge, which is usually provided by an
accountant. According to our interviews, U.S. small business owners in underserved
communities were less likely to have digital access and access to technical business knowledge
and support. Moreover, small business owners from underserved communities were less likely to
have access to working capital that may be required for collateral for loans. In light of these
21 of 61
behavioral and structural barriers faced by small businesses in underserved communities,
identifying policy solutions that reduce administrative burden in applications for small business
funding is a key priority for the advancement of equity in the small business community after the
COVID-19 pandemic.
Policy Background and Theoretical Hypotheses
The emergency microloan program in the city we studied initially required a long list of
documentation, as shown in table 1 below. These requirements were driven in part by the strict
verification required for the use of Community Development Block Grant (CDBG) funding,
which is a federal program administered by the U.S. Department of Housing and Urban
Development that supports infrastructure, economic development, and housing services. After
realizing that the extensive documentation requirements shut out underserved small business
owners who were disproportionately more likely to need support and less likely to have the
resources needed to overcome administrative burdens, city officials made substantial changes.
Specifically, the city officials decided not to utilize CDBG funding and to instead leverage
sources of funding within the city budget that did not require stringent documentation. As a
result of this change in the source of funding, the mayor announced on Wednesday May 21, 2020
(in the middle of program implementation) that the EDD relaxed documentation requirements for
the emergency microloans. No other major change was made to the evaluation of applicants or
the publicization of the program, and all eligibility requirements were the same for all
businesses.
As seen in the table below, after May 21st the documentation requirements were
significantly less burdensome. By making the profit and loss statement optional, for instance, the
city removed the most significant documentation barrier that we identified in our interviews. We
22 of 61
leverage this major program change to build evidence on the impact of documentation burdens
on equity in access to small business COVID-19 relief funding.
[Insert table 1 here]
With less documentation required, we predict that underserved business owners, who
were already facing unprecedented challenges in the COVID-19 pandemic, would likely have
faced less confusion and paperwork hurdles that led many to give up on the application process.
Therefore, reducing the documentation requirements should increase the likelihood that
underserved businesses make it through each stage of the application process.
H1: Reducing documentation requirements will increase the likelihood that underserved
businesses (a) that initiate an application go on to submit an application; (b) advance to
“under review;” and (c) are awarded relief funding.vii
In Hypothesis 2, we propose two possible scenarios: H2a) a reduction in documentation
burdens may have a bigger impact on underserved business owners, and could narrow disparities
in application and award rates between underserved and non-underserved businesses and H2b) a
reduction in documentation burdens widens disparities because non-underserved businesses are
better able to take advantage of the program after the policy change. In H2a, we test whether the
policy change accomplished the policy goals and could be considered “equity-enhancing” as we
conceptualize in the theoretical framework above. It could be the case that reducing
documentation requirements will be particularly influential for underserved businesses, relative
to non-underserved businesses because of the disproportionate challenges that underserved
businesses reported in filling out the application requirements prior to the program change.
23 of 61
Therefore, once there are less documentation requirements, disparities in social and human
capital could become less important and underserved small businesses that previously did not
apply for the program may be significantly more likely to apply for and access the program
benefits.
H2a: Reducing documentation requirements will have a larger impact on underserved small
businesses relative to non-underserved businesses in the likelihood (a) that they submit an
application; (b) advance to “under review;” and (c) are awarded relief funding.
However, we also recognize that historically underserved businesses could face several barriers
prior to the documentation phase such as a lack of trust or not being plugged into the network
that the city department reaches out to, which could put them at a disadvantage and result in the
policy aiding non-underserved businesses more (H2b). Moreover, underserved businesses may
have disproportionately lower administrative and human capital, which may make navigating
even a more limited number of documentation requirements insurmountable, especially when
compared to non-underserved businesses who do not face structural disadvantages (Masood and
Nisar 2021; Bell et al. 2023; Chudnovsky and Peeters 2021; Barnes 2020; Heinrich et al. 2022).
H2b: Reducing documentation requirements will have a larger impact on non-underserved
small businesses relative to underserved businesses in the likelihood (a) that they submit an
application; (b) advance to “under review;” and (c) are awarded relief funding.
Methods
We test our theoretical hypotheses using data from the city emergency microloan/grant
program. Program administrators provided us with this data (in spreadsheet form). Our raw data
contained all the businesses that initiated applications (opened the online application)viii with
multiple rows for each small business that depicted the status of the application as well as the date
24 of 61
that the application status changed.ix For the interrupted time series analysis, we follow our pre-
analysis plan and transform the raw data into a dataset that includes a row for each day in the
application period for underserved and non-underserved businesses (2 rows per day) with columns
that capture the number of applications reaching each stage of the application process. For the
businesses that initiate or submit multiple applications, we take the time stamp of the first
application initiation but fill in any missing variables using the most recent version of the
application to reduce missingness.x This ensures that we are not double counting single businesses
in our outcome variables, which would introduce bias in the estimation of treatment effects and
standard errors.
For the underserved status variables there was substantial missingness, most of which is
driven by applications that were initiated but never submitted.xi We address this missingness in our
robustness checks section, where we empirically test whether imputation significantly changes the
results. For the business applicants who have missing information on control variables, we do not
impute any of the other information— in constructing the analytical dataset, we simply take the
average on each day for the businesses whose information is not missing. We summarize the
measurement of each key variable in table 2 below. Our outcome measures are proportions (e.g.,
the proportion advancing to the next stage), in line with the social equity literature’s focus on
disproportions as indicators of inequities (Guy and McCandless 2012).
[Insert Table 2 here]
We imported data from the American Community Survey on neighborhood characteristics
including the percentage of the block group that is considered low or moderate income, and the
25 of 61
percentage of the Census block group that is Black or Hispanic. This data allows us to compare
the businesses in the pre- and post-change time-period to enhance our understanding of whether
the applicants are systematically different in ways that could bias the estimation of treatment
effects in the interrupted time series model. We also pulled data from the County Health Dashboard
on COVID-19 daily cases and deaths to account for this potentially confounding time-varying
factor.
Descriptive Statistics
In table 3, we present the descriptive statistics for the business-level and analytical day-
level dataset before and after the policy change. The most important takeaway from table 3 is the
change in the composition of businesses in the pre- and post-program change periods. For
instance, the average number of employees at an applicant business as well as the loan amount
requested decreased after the program change, which likely indicates that there were more very
small businesses and potentially even sole proprietorships that were initiating applications after
the documentation requirements were reduced. These very small businesses and sole
proprietorships face the largest challenges compiling documentation because they often do not
have accountants on staff or do not have the revenue to support hiring additional staff to help
with the application process. The community characteristics of businesses also changed between
the pre- and post-program change period; businesses initiating in the post-program change period
were located in areas with higher Black, Hispanic, and low-income residents. This suggests that
the program reached a broader set of small businesses in the city, in communities of color and
lower income areas that may have initially not known about the program or did not think that the
program was worth applying to. With regard to the outcome measures, there were less
applications that went under review but around the same proportion of businesses applying made
26 of 61
it to the submitted and funded phase. Finally, another factor changing over time was the COVID-
19 cases and deaths, which markedly increased in the post-program change period.
[Insert table 3 and 4 here]
In table 4, we present the mean outcome measures, business characteristics, and
community characteristics before and after the program change separately for underserved and
non-underserved small business applicants. Before the program change, there was a significant
disparity in the application success of underserved relative to non-underserved small business
owners. In the post-program change period, this disparity disappears or shrinks for the outcome
measures. However, these are only descriptive statistics that fail to capture the localized effect of
the program change immediately before and after its implementation. While illuminating, they
do not on their own count as statistical evidence for or against our hypotheses regarding the
impact of program change.
As a supplement to these descriptive statistics tables, we visualize the number of
businesses from underserved and non-underserved groups at each of the four stages of the
application process, in the pre- versus post-May 21st periods in figure 1. One of the takeaways
from these graphs is that it appears non-underserved businesses were quicker in submitting
applications and given that the program administrators used a first-come first-served
disbursement policy, this likely contributed to inequity before the program change. The second
takeaway is that there is a significant jump in the number of submitted applications immediately
after the program change, providing some face validity to our hypotheses.
27 of 61
[Insert Figure 1]
Statistical Models and Hypothesis Tests
In the absence of random assignment, we employ an interrupted time series design to
obtain plausibly causal estimates that reduce the potential for omitted variable bias. Specifically,
we leverage the discontinuity in time created by the program change to estimate the effect of
reducing documentation requirements on the number of underserved small business applicants
successfully completing an application and acquiring funding prior immediately prior to and
after the implementation of the program change. By leveraging variation over time, we provide a
model that is unaffected by typical time invariant confounding variables, as these characteristics
are taken into account when modelling the underlying long-term trend. However, time-varying
confounders, such as seasonality or a COVID-19 outbreak, may introduce bias in the results. For
this reason, we include a vector of observable characteristics that are unaffected by the treatment
that vary over time to better isolate the treatment effect of reducing documentation burdens. We
estimate the following models for each key outcome in a Generalized Linear Regression model:
𝑌𝑖𝑡 = 𝛽0 + 𝛽1𝐴𝑓𝑡𝑒𝑟𝑖𝑡 + 𝛽2𝑇𝑖𝑡 + 𝛽3𝑁𝑜𝑛𝑈𝑛𝑑𝑒𝑟𝑠𝑒𝑟𝑣𝑒𝑑𝑖 + 𝛽4(𝐴𝑓𝑡𝑒𝑟𝑖𝑡 ∗ 𝑇𝑖𝑡 ) + 𝛽5(𝑁𝑜𝑛𝑈𝑛𝑑𝑒𝑟𝑠𝑒𝑟𝑣𝑒𝑑𝑖 ∗ 𝑇𝑖𝑡 )
+ 𝛽6(𝐴𝑓𝑡𝑒𝑟𝑖𝑡 ∗ 𝑁𝑜𝑛𝑈𝑛𝑑𝑒𝑟𝑠𝑒𝑟𝑣𝑒𝑑𝑖 ) + 𝛽7(𝐴𝑓𝑡𝑒𝑟𝑖𝑡 ∗ 𝑁𝑜𝑛𝑈𝑛𝑑𝑒𝑟𝑠𝑒𝑟𝑣𝑒𝑑𝑖 ∗ 𝑇𝑖𝑡 ) + 𝛽8𝑋𝑖𝑡 + 𝜀𝑖𝑡
● where i is an indicator for whether the unit of observation is for underserved or non-underserved
businesses and t represents each day in the application period;
● 𝑌𝑖𝑡 is an indicator of our three outcomes of interest on each day for non-underserved and
underserved business applicants;
● 𝐴𝑓𝑡𝑒𝑟𝑖𝑡 is a dichotomous indicator of whether the day is before or after May 21st 2020 (date the
program changed)
● 𝑇𝑖𝑡 is the time elapsed (i.e. days) since the start of the application period;
● 𝑁𝑜𝑛𝑈𝑛𝑑𝑒𝑟𝑠𝑒𝑟𝑣𝑒𝑑𝑖 is a dichotomous indicator for whether the row is capturing the outcome for
non-underserved businesses (1) on each day or underserved businesses (0);
● 𝑋𝑖𝑡 is a vector of observable characteristics observed over time, including day of the week fixed
effects to capture seasonality, and COVID-19 deaths and cases in the county on each day;
● 𝜀𝑖𝑡 is a robust idiosyncratic error term, clustered by day.
28 of 61
In the equation above,
● β0 represents the baseline level at T = 0 when every other quantity is also set equal to zero,
● β1 is the level change following the program change for underserved businesses,
● β2 is the change in application and funding rates associated with a time unit increase (which
represents the underlying pre-intervention trend),
● β3 indicates the difference in the outcome for non-underserved businesses compared to
underserved businesses,
● β4 indicates the slope change in the number of business applicants making it to each stage of the
application process following the intervention (for underserved businesses),
● β5 indicates how the time trend for non-underserved businesses differs compared to the time trend
for underserved businesses,
● β6 indicates the level change in the outcome of interest for non-underserved businesses, relative
to underserved businesses after the program change, and
● β7 indicates how the slope change in the outcome of interest after the program change differs for
non-underserved businesses, compared to the slope change for underserved businesses
For Hypothesis 1, our coefficient of interest is β1 which captures the level change in the outcome
for underserved businesses right after the program change. H1 will be supported by a positive
and significant β1. For Hypothesis 2, our main coefficient of interest is β6 which captures the
level change in the outcome for non-underserved businesses, relative to underserved businesses
right after the program change. H2a will be supported by a negative and significant β6, and H2b
will be supported by a positive and significant β6. In our pre-registered analysis plan, we
specified the level change as the main test of H2 and H1 rather than the slope change as the
localized effects are more likely to be attributable to the program change and not other
unobservable phenomena. To correct for multiple tests, we use the Holm-Bonferroni methodxii.
We also estimate a secondary analysis for Women-owned and Minority-owned businesses, as
these are the only subgroups for which we have sufficient statistical power. These heterogeneous
treatment effects help us identify the type of businesses that may have been most affected by the
program change.
Results
Our main analysis, presented in table 5, provides support for our first hypothesis. For
29 of 61
underserved businesses, the proportion of initiated applications that were submitted increased by
26.1 percentage points right after the program changed (p < 0.001, 95% CI [19.64, 32.56]).xiii
However, we do not find support for our second hypothesis. The reduction in documentation
requirements was not associated with a larger increase in completed applications for underserved
businesses, relative to non-underserved businesses. Instead, we find that right after the program
changed, the proportion of non-underserved businesses completing applications increased by an
additional 15.7 percentage points (p = 0.049, 95% CI [7.74, 23.66]), relative to underserved small
businesses.
For our secondary outcomes, we find the proportion of underserved businesses with
applications under review increased by 8.4 percentage points right after the program changed (p =
0.032, 95% CI [4.50, 12.26]) and the proportion of underserved businesses that were funded
increased by 3.8 percentage points right after the program change (p = 0.017, 95% CI [2.23, 5.41]).
Nevertheless, non-underserved businesses still saw larger increases in downstream outcomes.
Right after the program changed, the share of non-underserved businesses that went under-review
increased by an additional 24.9 percentage points (p < 0.001, 95% CI [19.33, 30.47]) and the share
that were funded increased by an additional 10.3 percentage points (p < 0.001, 95% CI [7.82,
12.78]) on top of the increases experienced by underserved businesses. Therefore, across all
outcomes of interest, we find support for Hypothesis 1 but do not find support for Hypothesis 2a.
[Insert table 5]
Subgroup Analysis
30 of 61
In line with our pre-analysis plan and focus on outcomes for specific marginalized
communities, we also investigate whether there were particularly large impacts on women-
owned businesses or minority-owned businesses in Tables 6-7. For minority-owned businesses,
there was a large and statistically significant impact on the likelihood of submitting applications
and being funded, but no statistically significant impact on the likelihood of going under review
(though the coefficient is substantively large and positive). Moreover, in line with the main
findings, minority-owned businesses did not benefit more from the program change than non-
underserved businesses on any of the key outcomes.
[Insert table 6]
For women-owned businesses, the overall trends that we present in the main analysis hold. We
find support for Hypothesis 1, with women-owned businesses benefitting from the program
change, but we do not find that they benefitted more than non-underserved businesses.
[Insert table 7]
Discussion
Administrative burdens contribute to persistent disparities in program access, which
threaten the foundational commitment of public administrators to advance social equity (Guy and
McCandless 2012; Johnson and Svara 2011; Bell et al. 2023; Herd and Moynihan 2018). Efforts
to reduce administrative burden have been found to improve program take-up, but extant
31 of 61
literature has yet to thoroughly explore for whom take-up improves. We conceptualize an equity-
enhancing intervention as that which benefits marginalized communities and diminishes
disparities, which we test empirically in the context of small business COVID-19 relief funds. In
our analysis, we leverage a policy change that reduced the documentation requirements for the
small business COVID-19 relief program in the middle of program implementation to estimate
an interrupted time series analysis. Our findings have multiple implications for future research.
First, we find that reducing documentation requirements increased access to small
business COVID-19 relief funding for historically underserved business owners, but that the
increase in the success of applications was larger for non-underserved businesses after the policy
change. This finding suggests that a reduction in administrative burdens can enhance outcomes
for marginalized communities, while at the same time perpetuating existing disparities if the set
of policy tools do not provide targeted supports to those at the largest disadvantage. These
findings are in line with the research on administrative burden, which shows that those with
greater administrative capital are better able to capitalize on public benefits and overcome
burdensome application processes (Masood and Nisar 2021). Even with the reduction in
documentation requirements, applicants still faced substantial compliance demands, which
ultimately may have been the catalyst for the prevailing disparity in application success.
Moreover, our work also contributes to the literature that has examined whether structural
changes that reduce compliance costs can increase citizen outcomes. While prior research has
observed increases in access among Medicaid beneficiaries (Herd et al. 2013), SNAP recipients
(Lopoo, Heflin, and Boskovski 2020), many have not incorporated analyses that take into
account multiple axes of marginalization. Therefore, while these changes may be increasing
access to public programs among low-income clients, they may not translate to increases in
32 of 61
access for racially minoritized low-income clients or clients with low levels of administrative or
human capital, or those that have had negative experiences with government in the past that fuel
mistrust (Bell et al. 2023; Döring 2021; Masood and Nisar 2021; Christensen et al. 2020).
It is possible that closing equity gaps would have required more intensive support to
underserved applicants throughout the application process, or communication targeted
specifically to underserved communities. It is also possible that using a first-come first-served
disbursement model made reducing documentation less equity enhancing than what otherwise
would have been possible in a lottery or points system that considered and weighted applicant
disadvantage (Johnson et al., 2021). Our findings have important implications for theory and
practice on administrative burden, suggesting that even well-intentioned interventions can
perpetuate disparity if there are not targeted supports for marginalized communities put in place.
Second, the nature of this policy change is of theoretical interest to scholars investigating
citizen-state interactions. A key source of administrative burden comes from federal
accountability pressures and regimes that impose documentation requirements on local or state
governments in exchange for federal funding. Specifically, we study a setting in which local
administrators changed the funding stream for the small business COVID-19 relief program from
the Federal Community Development Block Grant (CDBG) funding to local government funds
which offered more flexibility in documentation requirements. By utilizing their discretionary
power to change the funding streams, administrators were also able to reduce the number of
required documents that small business owners had to submit to apply for small business relief
funds. A key cause of onerous encounters with the state in this context was federal
documentation mandates on funding streams, and an overlooked policy solution could be to use
local discretion to counteract these onerous top-down policies with local innovation. In the
33 of 61
future, public managers could consider applying for a waiver from federal documentation
requirements so that they can assess the extent to which documentation reduces equity and the
extent to which these changes would impact fraud/program integrity.
Finally, the reduction in documentation requirements reduced barriers for all types of
businesses, and regardless of the inability to alleviate pre-existing disparities, it may be
worthwhile to pursue these burden reduction efforts to improve the public’s experience of
government. Moreover, it is worth noting that the intervention appeared to reach a different set
of small businesses that were applying for smaller loan amounts after the reduction in
documentation requirements, which may indicate a level of disadvantage, regardless of
underserved status. In this way, there may be other axes of marginalization that we were unable
to capture, like socioeconomic disadvantage, where we might have seen more equity enhancing
effects. Ultimately, we conclude that this burden reduction effort was unable to
disproportionately enhance access among underserved small businesses; however, all small
businesses in the city benefitted from the reduction in burdens which is a notable achievement
that likely enhanced clients experience of government during a time of crisis.
Limitations and Future Directions
One limitation of this study is our inability to say why progression through the
application pipeline differed before and after the documentation requirement change. Reduced
application burdens may have coincided with increased awareness of the program, may have led
different businesses to apply, and/or may have made the same sort of businesses more likely to
progress through the application stages. For example, we observe an influx of smaller businesses
asking for lower loan amounts, suggesting that the businesses in the post-program change time
period are more disadvantaged relative to those in the pre-program change period. This
34 of 61
compositional change could downwardly-bias the treatment effect, making our estimates a
conservative proxy for the true treatment effect. On one hand, policymakers may not be overly
concerned about which of these possibilities accounts for the differences we observed—they may
simply want to know how such a change is likely to impact equity in outcomes. However, future
research may be able to disentangle these potential mechanisms.
In considering the generalizability of these results, it is worth noting that applicant
businesses are not necessarily a random sample of eligible businesses. Experts we spoke to
before conducting this study reminded us that business owners’ beliefs about the motives of
program administrators (e.g., were they trustworthy, was the program designed “for someone
like me”) shape their applications. In a place where mistrust is high, reduced administrative
burdens may not achieve effects like those we observed. However, this need not detract from the
theoretical importance of the effects identified in this study. As Mook (1983, 380) notes, when
the goal is “not making generalizations, but testing them,” generalizability takes on a different
meaning. Our goal is not to generalize to a larger population, but to test whether a generalized
theory when applied to a set of subjects holds up.
In our data, differences between underserved and non-underserved businesses were larger
at the submission stage than the funding stage. Diminishing effect sizes associated with
progression through the pipeline may just reflect greater distance from the intervention.
However, it may also be due to motivation among staff at this program to support underserved
business owners in their community. A similar intervention in a program whose staff and local
officials are not so motivated may be less effective.
Conclusion
35 of 61
As public administration scholars and practitioners strive for social equity in citizen-state
interactions, future researchers should continue to take up the challenge of imagining and
evaluating what it would take to diminish disparity in access to government programs. To find
the appropriate set of policy tools for public managers to craft “equity-enhancing” interventions,
future research should examine how various efforts to reduce administrative burden impact
underserved and marginalized communities. Our study is one step toward this goal, and future
researchers should build on our work by evaluating whether a policy change like the one we
study—in which documentation burdens are reduced for all applicants—will perpetuate disparity
only in the absence of other targeted support for marginalized communities seeking access to
public programs.
36 of 61
References
Allbright, Taylor, Julie Marsh, Michelle Hall, Laura Tobben, Lawrence Picus, and Magaly Lavadenz.
2018. “Conceptualizing Equity in the Implementation of California Education Finance Reform.”
Education Faculty Articles and Research, December. https://doi.org/10.1086/701247.
Alon, Titan, Matthias Doepke, Jane Olmstead-Rumsey, and Michèle Tertilt. 2020. “The Impact of
COVID-19 on Gender Equality.” Working Paper 26947. Working Paper Series. National Bureau
of Economic Research. https://doi.org/10.3386/w26947.
Andersen, Simon Calmar, and Thorbjørn Sejr Guul. 2019. “Reducing Minority Discrimination at the
Front Line—Combined Survey and Field Experimental Evidence.” Journal of Public
Administration Research and Theory 29 (3): 429–44. https://doi.org/10.1093/jopart/muy083.
Arendsen, Rex, Oscar Peters, Marc ter Hedde, and Jan van Dijk. 2014. “Does E-Government Reduce the
Administrative Burden of Businesses? An Assessment of Business-to-Government Systems
Usage in the Netherlands.” Government Information Quarterly 31 (1): 160–69.
https://doi.org/10.1016/j.giq.2013.09.002.
Assouline, Michaela, Sharon Gilad, and Pazit Ben-Nun Bloom. 2021. “Discrimination of Minority
Welfare Claimants in the Real World: The Effect of Implicit Prejudice.” Journal of Public
Administration Research and Theory, May, muab016. https://doi.org/10.1093/jopart/muab016.
Baekgaard, Martin, Kim Sass Mikkelsen, Jonas Krogh Madsen, and Julian Christensen. 2021. “Reducing
Compliance Demands in Government Benefit Programs Improves the Psychological Well-Being
of Target Group Members.” Journal of Public Administration Research and Theory, no. muab011
(April). https://doi.org/10.1093/jopart/muab011.
Baker, Dominique. 2020. “How COVID-19 Exacerbates Existing Inequities in Our Financial Aid System
– Third Way.” 2020. https://www.thirdway.org/report/how-covid-19-exacerbates-existing-
inequities-in-our-financial-aid-system.
37 of 61
Barnes, Carolyn Y. 2020. “‘It Takes a While to Get Used to’: The Costs of Redeeming Public Benefits.”
Journal of Public Administration Research and Theory, no. muaa042 (November).
https://doi.org/10.1093/jopart/muaa042.
Bell, Elizabeth, Julian Christensen, Pamela Herd, and Donald Moynihan. 2021. “Human Capital and
Citizen-State Interactions: The Effects of Health on Administrative Burdens and Program
Access.” In Administrative Burden Workshop.
———. 2023. “Health in Citizen-State Interactions: How Physical and Mental Health Problems Shape
Experiences of Administrative Burden and Reduce Take-Up.” Public Administration Review 83
(2): 385–400. https://doi.org/10.1111/puar.13568.
Bell, Elizabeth, and Kylie Smith. 2021. “Working Within a System of Administrative Burden: How
Street-Level Bureaucrats’ Role Perceptions Shape Access to the Promise of Higher Education.”
Administration & Society.
Berry-James, RaJade M., Brandi Blessett, Rachel Emas, Sean McCandless, Ashley E. Nickels, Kristen
Norman-Major, and Parisa Vinzant. 2021. “Stepping up to the Plate: Making Social Equity a
Priority in Public Administration’s Troubled Times.” Journal of Public Affairs Education 27 (1):
5–15. https://doi.org/10.1080/15236803.2020.1820289.
Bhargava, Saurabh, and Dayanand Manoli. 2015. “Psychological Frictions and the Incomplete Take-Up
of Social Benefits: Evidence from an IRS Field Experiment.” American Economic Review 105
(11): 3489–3529. https://doi.org/10.1257/aer.20121493.
Biden, Joseph. 2021. “Executive Order On Advancing Racial Equity and Support for Underserved
Communities Through the Federal Government.” The White House. January 21, 2021.
https://www.whitehouse.gov/briefing-room/presidential-actions/2021/01/20/executive-order-
advancing-racial-equity-and-support-for-underserved-communities-through-the-federal-
government/.
38 of 61
Blessett, Brandi, Jennifer Dodge, Beverly Edmond, Holly T Goerdel, Susan T Gooden, Andrea M
Headley, Norma M Riccucci, and Brian N Williams. 2019. “Social Equity in Public
Administration: A Call to Action.” Perspectives on Public Management and Governance 2 (4):
283–99. https://doi.org/10.1093/ppmgov/gvz016.
Blessett, Brandi, M Fudge, and Tia Sherèe Gaynor. 2017. “Moving from Theory to Practice: An
Evaluative Assessment of Social Equity Approaches.” Standing Panel on Social Equity in
Governance. Center for Accountability and Performance and National Academy for Public
Administration.
Blume, Grant H. 2022. “‘As Expected’: Theoretical Implications for Racialized Administrative Power as
the Status Quo.” Journal of Public Administration Research and Theory, March, muac013.
https://doi.org/10.1093/jopart/muac013.
Brodkin, Evelyn Z., and Malay Majmundar. 2010. “Administrative Exclusion: Organizations and the
Hidden Costs of Welfare Claiming.” Journal of Public Administration Research and Theory 20
(4): 827–48. https://doi.org/10.1093/jopart/mup046.
Burnam, M. Audrey, Lisa S. Meredith, Terri Tanielian, and Lisa H. Jaycox. 2009. “Mental Health Care
For Iraq And Afghanistan War Veterans.” Health Affairs 28 (3): 771–82.
https://doi.org/10.1377/hlthaff.28.3.771.
Carey, Gemma, Eleanor Malbon, and James Blackwell. 2021. “Administering Inequality? The National
Disability Insurance Scheme and Administrative Burdens on Individuals.” Australian Journal of
Public Administration n/a (n/a). https://doi.org/10.1111/1467-8500.12508.
Christensen, Julian, Lene Aarøe, Martin Baekgaard, Pamela Herd, and Donald P. Moynihan. 2020.
“Human Capital and Administrative Burden: The Role of Cognitive Resources in Citizen-State
Interactions.” Public Administration Review 80 (1): 127–36. https://doi.org/10.1111/puar.13134.
39 of 61
Chudnovsky, Mariana, and Rik Peeters. 2020. “The Unequal Distribution of Administrative Burden: A
Framework and an Illustrative Case Study for Understanding Variation in People’s Experience of
Burdens.” Social Policy & Administration n/a (n/a). https://doi.org/10.1111/spol.12639.
———. 2021. “The Unequal Distribution of Administrative Burden: A Framework and an Illustrative
Case Study for Understanding Variation in People’s Experience of Burdens.” Social Policy &
Administration 55 (4): 527–42. https://doi.org/10.1111/spol.12639.
Daigneault, Pierre-Marc, and Christian Macé. 2019. “Program Awareness, Administrative Burden, and
Non-Take-Up of Québec’s Supplement to the Work Premium.” International Journal of Public
Administration 0 (0): 1–13. https://doi.org/10.1080/01900692.2019.1636397.
DellaVigna, Stefano, and Elizabeth Linos. n.d. “RCTs to Scale: Comprehensive Evidence from Two
Nudge Units,” 65.
Deshpande, Manasi, and Yue Li. 2019. “Who Is Screened Out? Application Costs and the Targeting of
Disability Programs.” American Economic Journal: Economic Policy 11 (4): 213–48.
https://doi.org/10.1257/pol.20180076.
Döring, Matthias. 2021. “How-to Bureaucracy: A Concept of Citizens’ Administrative Literacy.”
Administration & Society 53 (8): 1155–77. https://doi.org/10.1177/0095399721995460.
Fairlie, Robert, and Alicia Robb. 2010. “Disparities in Capital Access between Minority and Non-
Minority Businesses.” U.S. Department of Commerce: Minority Business Development Agency.
https://archive.mbda.gov/page/executive-summary-disparities-capital-access-between-minority-
and-non-minority-businesses.html.
Farrell, Diana, Chris Wheat, and Chi Mac. 2020. “Small Business Owner Race, Liquidity, and Survival.”
J.P. Morgan Chase. https://www.jpmorganchase.com/institute/research/small-business/report-
small-business-owner-race-liquidity-survival.
40 of 61
Fox, Ashley M., Edmund C. Stazyk, and Wenhui Feng. 2020. “Administrative Easing: Rule Reduction
and Medicaid Enrollment.” Public Administration Review 80 (1): 104–17.
https://doi.org/10.1111/puar.13131.
Frederickson, H. George. 1980. New Public Administration. Tuscaloosa: University of Alabama Press.
———. 1990. “Public Administration and Social Equity.” Public Administration Review 50 (2): 228–37.
https://doi.org/10.2307/976870.
———. 2010. Social Equity and Public Administration: Origins, Developments, and Applications.
Armonk, N.Y.: M.E. Sharpe, Inc.
Gooden, Susan T. 2014. Race and Social Equity: A Nervous Area of Government. 1st edition. Armonk,
New York: Routledge.
———. 2015. “‘PAR’’s Social Equity Footprint.” Public Administration Review 75 (3): 372–81.
Gray, Colin, Adam Leive, Elena Prager, Kelsey B. Pukelis, and Mary Zaki. 2021. “Employed in a SNAP?
The Impact of Work Requirements on Program Participation and Labor Supply.” Working Paper
28877. Working Paper Series. National Bureau of Economic Research.
https://doi.org/10.3386/w28877.
Guul, Thorbjørn Sejr, Mogens Jin Pedersen, and Niels Bjørn Grund Petersen. 2021. “Creaming among
Caseworkers: Effects of Client Competence and Client Motivation on Caseworkers’ Willingness
to Help.” Public Administration Review 81 (1): 12–22. https://doi.org/10.1111/puar.13297.
Guul, Thorbjørn Sejr, Anders R. Villadsen, and Jesper N. Wulff. 2019. “Does Good Performance Reduce
Bad Behavior? Antecedents of Ethnic Employment Discrimination in Public Organizations.”
Public Administration Review 79 (5): 666–74. https://doi.org/10.1111/puar.13094.
Guy, Mary E., and Sean A. McCandless. 2012. “Social Equity: Its Legacy, Its Promise.” Public
Administration Review 72: S5–13.
41 of 61
Harper-Anderson, Elsie, Jay Albanese, and Susan Gooden, eds. 2023. Racial Equity, COVID-19, and
Public Policy: The Triple Pandemic. New York: Routledge.
https://doi.org/10.4324/9781003286967.
Hattke, Fabian, David Hensel, and Janne Kalucza. 2020. “Emotional Responses to Bureaucratic Red
Tape.” Public Administration Review 80 (1): 53–63. https://doi.org/10.1111/puar.13116.
Heinrich, Carolyn. 2018. “"A Thousand Petty Fortresses”: Administrative Burden in U.S. Immigration
Policies and Its Consequences.” Journal of Policy Analysis and Management 0 (0).
https://doi.org/10.1002/pam.22046.
Heinrich, Carolyn J. 2016. “The Bite of Administrative Burden: A Theoretical and Empirical
Investigation.” Journal of Public Administration Research and Theory 26 (3): 403–20.
https://doi.org/10.1093/jopart/muv034.
Heinrich, Carolyn J., and Robert Brill. 2015. “Stopped in the Name of the Law: Administrative Burden
and Its Implications for Cash Transfer Program Effectiveness.” World Development 72
(Supplement C): 277–95. https://doi.org/10.1016/j.worlddev.2015.03.015.
Heinrich, Carolyn J., Sayil Camacho, Sarah Clark Henderson, Mónica Hernández, and Ela Joshi. 2022.
“Consequences of Administrative Burden for Social Safety Nets That Support the Healthy
Development of Children.” Journal of Policy Analysis and Management 41 (1): 11–44.
https://doi.org/10.1002/pam.22324.
———. n.d. “Consequences of Administrative Burden for Social Safety Nets That Support the Healthy
Development of Children.” Journal of Policy Analysis and Management n/a (n/a). Accessed
January 12, 2022. https://doi.org/10.1002/pam.22324.
Herd, Pamela, Thomas DeLeire, Hope Harvey, and Donald P. Moynihan. 2013. “Shifting Administrative
Burden to the State: The Case of Medicaid Take-Up.” Public Administration Review 73 (s1):
S69–81. https://doi.org/10.1111/puar.12114.
42 of 61
Herd, Pamela, and Donald P. Moynihan. 2018. Administrative Burden: Policymaking by Other Means. 1
edition. New York: Russell Sage Foundation.
Homonoff, Tatiana, and Jason Somerville. 2020. “Program Recertification Costs: Evidence from SNAP.”
w27311. National Bureau of Economic Research. https://doi.org/10.3386/w27311.
Johnson, Norman J., and James H. Svara. 2011. Justice for All: Promoting Social Equity in Public
Administration. 1st edition. London; New York: Routledge.
Keiser, Lael R., and Susan M. Miller. 2020. “Does Administrative Burden Influence Public Support for
Government Programs? Evidence from a Survey Experiment.” Public Administration Review 80
(1): 137–50. https://doi.org/10.1111/puar.13133.
Larsson, Karl Kristian. 2021. “Digitization or Equality: When Government Automation Covers Some, but
Not All Citizens.” Government Information Quarterly 38 (1): 101547.
https://doi.org/10.1016/j.giq.2020.101547.
Linos, Elizabeth, Allen Prohofsky, Aparna Ramesh, Jesse Rothstein, and Matthew Unrath. 2022. “Can
Nudges Increase Take-up of the EITC?: Evidence from Multiple Field Experiments.” American
Economic Journal: Economic Policy. https://doi.org/10.1257/pol.20200603.
Linos, Elizabeth, Lisa T. Quan, and Elspeth Kirkman. 2020. “Nudging Early Reduces Administrative
Burden: Three Field Experiments to Improve Code Enforcement.” Journal of Policy Analysis and
Management n/a (n/a). https://doi.org/10.1002/pam.22178.
Linos, Elizabeth, Vikash Reddy, and Jesse Rothstein. 2022. “Demystifying College Costs: How Nudges
Can and Can’t Help.” Behavioural Public Policy, March, 1–22.
https://doi.org/10.1017/bpp.2022.1.
Linos, Elizabeth, and Nefara Riesch. 2020. “Thick Red Tape and the Thin Blue Line: A Field Study on
Reducing Administrative Burden in Police Recruitment.” Public Administration Review 80 (1):
92–103. https://doi.org/10.1111/puar.13115.
43 of 61
Liu, and Parilla. 2020. “Businesses Owned by Women and Minorities Have Grown. Will COVID-19
Undo That?” Brookings (blog). April 14, 2020. https://www.brookings.edu/research/businesses-
owned-by-women-and-minorities-have-grown-will-covid-19-undo-that/.
Lopoo, Leonard M., Colleen Heflin, and Joseph Boskovski. 2020. “Testing Behavioral Interventions
Designed to Improve On-Time SNAP Recertification.” Journal of Behavioral Public
Administration 3 (2). https://doi.org/10.30636/jbpa.32.183.
Marx, Benjamin M., and Lesley J. Turner. 2019. “Student Loan Nudges: Experimental Evidence on
Borrowing and Educational Attainment.” American Economic Journal: Economic Policy 11 (2):
108–41. https://doi.org/10.1257/pol.20180279.
Masood, Ayesha, and Muhammad Nisar. 2021. “Administrative Capital and Citizens’ Responses to
Administrative Burden.” Journal of Public Administration Research and Theory.
https://doi.org/10.1093/jopart/muaa031.
Meikle, Paulette A., and Lekeitha R. Morris. 2022. “University Social Responsibility: Challenging
Systemic Racism in the Aftermath of George Floyd’s Murder.” Administrative Sciences 12 (1):
36. https://doi.org/10.3390/admsci12010036.
Merikangas, Kathleen Ries, and Kevin Paul Conway. 2021. “Landmark Study Documents the Mental
Health Challenges of Military Service in Canada.” The Canadian Journal of Psychiatry 66 (11):
934–36. https://doi.org/10.1177/07067437211039936.
Mertens, Stephanie, Mario Herberz, Ulf J. J. Hahnel, and Tobias Brosch. 2022. “The Effectiveness of
Nudging: A Meta-Analysis of Choice Architecture Interventions across Behavioral Domains.”
Proceedings of the National Academy of Sciences 119 (1).
https://doi.org/10.1073/pnas.2107346118.
Mook, Douglas G. 1983. “In Defense of External Invalidity.” American Psychologist 38: 379–87.
https://doi.org/10.1037/0003-066X.38.4.379.
44 of 61
Morel, Laura, Mohamed Al Elew, and Emily Harris. n.d. “Hundreds of Billions of Dollars in PPP Loans
Marred by Racial Inequity.” Reveal. Accessed January 21, 2022.
http://revealnews.org/article/rampant-racial-disparities-plagued-how-billions-of-dollars-in-ppp-
loans-were-distributed-in-the-u-s/.
Moynihan, Donald, Pamela Herd, and Hope Harvey. 2015. “Administrative Burden: Learning,
Psychological, and Compliance Costs in Citizen-State Interactions.” Journal of Public
Administration Research and Theory 25 (1): 43–69. https://doi.org/10.1093/jopart/muu009.
National Council on Disability. 2021. “2021 Progress Report: The Impact of COVID-19 on People with
Disabilities.” https://ncd.gov/.
Nguyen, Thu T., Shaniece Criss, Eli K. Michaels, Rebekah I. Cross, Jackson S. Michaels, Pallavi
Dwivedi, Dina Huang, Erica Hsu, Krishay Mukhija, and Leah H. Nguyen. 2021. “Progress and
Push-Back: How the Killings of Ahmaud Arbery, Breonna Taylor, and George Floyd Impacted
Public Discourse on Race and Racism on Twitter.” SSM-Population Health 15: 100922.
Nielsen, Morten Meyerhoff, Nuno Ramos Carvalho, Linda Gonçalves Veiga, and Luís Soares Barbosa.
2017. “Administrative Burden Reduction Over Time: Literature Review, Trends and Gap
Analysis.” In Proceedings of the 10th International Conference on Theory and Practice of
Electronic Governance, 140–48. ICEGOV ’17. New York, NY, USA: Association for Computing
Machinery. https://doi.org/10.1145/3047273.3047334.
Nisar, Muhammad. 2017. “Children of a Lesser God: Administrative Burden and Social Equity in
Citizen–State Interactions.” Journal of Public Administration Research and Theory 28 (1): 104–
19. https://doi.org/10.1093/jopart/mux025.
Office of Evaluation Sciences. 2021. “Challenges and Opportunities for Pursuing Equitable Distribution
of Small Business Relief.”
Office of Management and Budget. 2021. “Study to Identify Methods to Assess Equity: Report to the
President.” Pursuant to Executive Order 13985. https://www.whitehouse.gov/wp-
45 of 61
content/uploads/2021/08/OMB-Report-on-E013985-Implementation_508-Compliant-Secure-
v1.1.pdf.
Olsen, Asmus Leth, Jonas Hogh Kyhse-Andersen, and Donald Moynihan. 2020. “The Unequal
Distribution of Opportunity: A National Audit Study of Bureaucratic Discrimination in Primary
School Access.” AMERICAN JOURNAL OF POLITICAL SCIENCE, December.
https://doi.org/10.1111/ajps.12584.
Peeters, Rik. 2019. “The Political Economy of Administrative Burdens: A Theoretical Framework for
Analyzing the Organizational Origins of Administrative Burdens.” Administration & Society,
June, 0095399719854367. https://doi.org/10.1177/0095399719854367.
Peeters, Rik, and Sergio A. Campos. 2021. “Taking the Bite out of Administrative Burdens: How
Beneficiaries of a Mexican Social Program Ease Administrative Burdens in Street-Level
Interactions.” Governance 34 (4): 1001–18. https://doi.org/10.1111/gove.12534.
Pepin, Gabrielle, Christopher O’Leary, and Dallas Oberlee. 2021. “Nudges to Increase Completion of
Welfare Applications: Experimental Evidence from Michigan.” Journal of Behavioral Public
Administration 4 (2). https://doi.org/10.30636/jbpa.42.237.
Perry, Andre M. 2020. Know Your Price: Valuing Black Lives and Property in America’s Black Cities.
Brookings Institution Press.
Portillo, Shannon K., Nicole Humphrey, and Domonic A. Bearfield. 2022. The Myth of Bureaucratic
Neutrality. 1st edition. Routledge.
Ray, Victor, Pamela Herd, and Donald Moynihan. 2022. “Racialized Burdens: Applying Racialized
Organization Theory to the Administrative State.” Journal of Public Administration Research and
Theory, January, muac001. https://doi.org/10.1093/jopart/muac001.
Reijnders, Mark, Jelmer Schalk, and Trui Steen. 2018. “Services Wanted? Understanding the Non-Take-
up of Social Support at the Local Level.” VOLUNTAS: International Journal of Voluntary and
Nonprofit Organizations 29 (6): 1360–74. https://doi.org/10.1007/s11266-018-00060-w.
46 of 61
Ridder, Denise de, Floor Kroese, and Laurens van Gestel. 2021. “Nudgeability: Mapping Conditions of
Susceptibility to Nudge Influence.” Perspectives on Psychological Science, August,
1745691621995183. https://doi.org/10.1177/1745691621995183.
Robb, Alicia. 2018. “Financing Patterns and Credit Market Experiences: A Comparison by Race and
Ethnicity for U.S. Employer Firms.” SBA’s Office of Advocacy (blog). 2018.
https://advocacy.sba.gov/2018/02/01/financing-patterns-and-credit-market-experiences-a-
comparison-by-race-and-ethnicity-for-u-s-employer-firms/.
SCM Network. 2004. “International Standard Cost Model Manual: Measuring and Reducing
Administrative Burdens for Businesses.” The Organization for Economic Cooperation and
Development.
Sevilla, Almudena, and Sarah Smith. 2020. “Baby Steps: The Gender Division of Childcare during the
COVID-19 Pandemic.” Oxford Review of Economic Policy 36 (September): S169–86.
https://doi.org/10.1093/oxrep/graa027.
Shah, Monica, Muskaan Sachdeva, and Roni P. Dodiuk-Gad. 2020. “COVID-19 and Racial Disparities.”
Journal of the American Academy of Dermatology 83 (1): e35.
https://doi.org/10.1016/j.jaad.2020.04.046.
Small Business Administration. 2022. “SDVOSB Program Administration.” SDVOSB Program
Administration. 2022. https://www.sba.gov/partners/contracting-officials/contracting-program-
administration/sdvosb-program-administration.
Sobota, Joseph. 2017. “Veteran-Owned Businesses and Their Owners.” US Small Business
Administration Office of Advocacy.
Soss, Joe, Richard C. Fording, and Sanford F. Schram. 2011. Disciplining the Poor: Neoliberal
Paternalism and the Persistent Power of Race. Chicago Studies in American Politics edition.
Chicago ; London: University Of Chicago Press.
47 of 61
Stivers, Camilla, Sanjay K. Pandey, Leisha DeHart-Davis, Jeremy L. Hall, Kathryn Newcomer, Shannon
Portillo, Meghna Sabharwal, Eiko Strader, and James Wright II. 2023. “Beyond Social Equity:
Talking Social Justice in Public Administration.” Public Administration Review 83 (2): 229–40.
https://doi.org/10.1111/puar.13620.
Svara, James H., and James R. Brunet. 2005. “Social Equity Is a Pillar of Public Administration.” Journal
of Public Affairs Education 11 (3): 253–58.
Teeters, Jenni B, Cynthia L Lancaster, Delisa G Brown, and Sudie E Back. 2017. “Substance Use
Disorders in Military Veterans: Prevalence and Treatment Challenges.” Substance Abuse and
Rehabilitation 8 (August): 69–77. https://doi.org/10.2147/SAR.S116720.
Thaler, Richard H., and Cass R. Sunstein. 2009. Nudge: Improving Decisions About Health, Wealth, and
Happiness. Revised&Expanded edition. New York: Penguin Books.
Trochmann, Maren B., Shilpa Viswanath, Stephanie Puello, and Samantha June Larson. 2022.
“Resistance or Reinforcement? A Critical Discourse Analysis of Racism and Anti-Blackness in
Public Administration Scholarship.” Administrative Theory & Praxis 44 (2): 158–77.
https://doi.org/10.1080/10841806.2021.1918990.
Watkins-Hayes, Celeste. 2009. The New Welfare Bureaucrats: Entanglements of Race, Class, and Policy
Reform. Chicago: University of Chicago Press.
———. 2011. “Race, Respect, and Red Tape: Inside the Black Box of Racially Representative
Bureaucracies.” Journal of Public Administration Research and Theory 21 (suppl_2): i233–51.
https://doi.org/10.1093/jopart/muq096.
Weimer, David L. 2020. “When Are Nudges Desirable? Benefit Validity When Preferences Are Not
Consistently Revealed.” Public Administration Review 80 (1): 118–26.
https://doi.org/10.1111/puar.13114.
48 of 61
Wiley, Kimberly, and Frances Berry. 2018. “Compassionate Bureaucracy: Assuming the Administrative
Burden of Policy Implementation.” Nonprofit and Voluntary Sector Quarterly 47 (4_suppl): 55S-
75S. https://doi.org/10.1177/0899764018760401.
Wright, James E., and Cullen C. Merritt. 2020. “Social Equity and COVID-19: The Case of African
Americans.” Public Administration Review 80 (5): 820–26. https://doi.org/10.1111/puar.13251.
49 of 61
Tables and Figures
Table 1. Documentation Requirements Before and After Policy Change
Document Required before Required after
May 21st May 21st
Business Tax Registration Certificate (valid prior Yes Yes
to March 2020)
Breakdown of Sources and Uses (budget) to check Yes Yes
whether they were authorized expenses
Business Tax Returns Yes, for 1-2 years Yes, for 1-2 years
if available if available
Business Financial Statements Yes, for 3 months Yes, for 1 month
Business Bank Statements Yes, for most Yes, for most
recent 3 months recent 3 months
Personal Bank Statement Yes, for most Yes, for most
recent 1 month recent 1 month
Credit check authorization Yes No
Profit and Loss statement Yes No
Business Plan Yes No
Projections Yes (3-5 years) No
Commercial Lease Agreement Yes No
Commercial Liability Insurance Yes No
Workers Comp Insurance (if applicable) Yes, if applicable No
Personal Financial Statement Yes No
Personal Income Tax Returns Yes, for 2 years No
Resumes Yes No
Note. After May 21st, documents that were not required were optional; provision of these
documents had no effect on the probability of obtaining assistance.
50 of 61
Table 2. Measurement of Key Variables
Independent Variables Description
Non-Underserved businessxiv Dichotomous indicator coded as 1 for businesses that were not classified
as Women-owned, Minority-owned, Disabled-owned or Veteran-owned.
Otherwise, the variable is coded as 0. Therefore, this variable captures
White, male, non-disabled, non-veteran business owner.
After A dichotomous variable for whether the day in question is before (coded as
0) on or after (coded as 1) the documentation requirements changed on May
21, 2020.
Time trend Measured as the days since the program change.
Derived Control Variables Description
Day of the Week An indicator for day of the week, which will help account for potential
idiosyncratic temporal variation.
COVID-19 rates Calculated as the number of new COVID-19 cases and deaths in the County
for each day in the application period.
Outcome Variables Description
Proportion making it to the Calculated as the proportion of applications initiated on each day in the
submitting phase application period that ever make it to “submitted” status (within the pre-
and post-period).xv
Proportion making it to the Calculated as the proportion of applications submitted on each day in the
“under review” phase application period that ever make it to “under review” status (within the
pre- and post-period).
Proportion making it to the Calculated as the proportion of applications “under review” on each day in
funded phase the application period that ever make it to funded status (within the pre-
and post-period).
51 of 61
Table 3. Descriptive Statistics
Before Program Change After Program Change
Variable N Mean N Mean
Business-Level
Outcomes
Initiated, but never submitted 6,715 0.52 2,004 0.65
Submitted 6,715 0.48 2,004 0.35
Under Review 6,715 0.36 2,004 0.12
Funded 6,715 0.10 2,004 0.06
Business Characteristics
Women-owned 6,715 0.24 2,004 0.29
Minority-owned 6,715 0.23 2004 0.34
Disabled-owned 6,715 0.00 2004 0.01
Veteran-owned 6,715 0.01 2,004 0.01
Jobs retained 6,715 4.52 2,004 3.85
Number of employees 6,585 424087.20 1,893 5.29
Loan amount 6,103 $19,864.99 1,777 $17,288.87
Community Characteristics
Covid Cases and Deaths 6,714 668.71 1,988 1770.49
Percent Black in Census Block 5,038 0.08 1,402 0.10
Percent Hispanic in Census Block
5,038 0.36 1,402 0.39
Percent Low or Moderate Income
5,062 0.55 1,408 0.56
(LMI) in Census Block
Day-Level
Outcomes
Initiated, but never submitted 129 0.57 271 0.67
Submitted 129 0.42 271 0.33
Under Review 129 0.25 271 0.09
Funded 129 0.07 271 0.05
Business Characteristics
Jobs retained 129 3.99 271 3.80
Number of employees 129 321787.50 268 5.99
Loan amount 128 $19,060.38 268 $17,129.98
Community Characteristics
Covid Cases and Deaths 128 747.30 270 1649.00
Percent Black in Census Block 129 0.09 258 0.10
Percent Hispanic in Census Block 129 0.35 258 0.38
Percent Low or Moderate Income
129 0.55 258 0.56
(LMI) in Census Block
Note: The panel is unbalanced because on some days in the application period, there were no initiations among underserved or non-
underserved businesses. In other cases, there is missing data due to an inability to match businesses based on addresses to census data. N is
the number of non-missing observations; for each variable, 0=no, 1=yes, so the Mean is the proportion of observations in the "yes" category
52 of 61
Table 4. Descriptive Statistics, by Underserved Status
Before Program Change After Program Change
Underserved Non-Underserved Underserved Non-Underserved
Variable N Mean N Mean N Mean N Mean
Business-Level Dataset
Initiated, never submitted 2,179 0.64 4,536 0.46 915 0.65 1,089 0.65
Submitted 2,179 0.36 4,536 0.54 915 0.35 1,089 0.35
Under Review 2,179 0.21 4,536 0.42 915 0.09 1,089 0.13
Funded 2,179 0.05 4,536 0.13 915 0.04 1,089 0.09
Business Characteristics
Jobs retained 2,179 3.55 4,536 4.98 915 3.70 1,089 3.99
Number of employees 2,134 394675.10 4451 438188.60 879 4.60 1014 5.88
Loan amount 2,108 $18,918 3995 $20,364 878 $16,720 899 $17,843
Women-owned 2,179 0.67 4,536 0.03 915 0.62 1,089 0.02
Minority-owned 2,179 0.64 4,536 0.03 915 0.69 1,089 0.04
Disabled-owned 2,179 0.01 4,536 0.00 915 0.01 1,089 0.00
Veteran-owned 2,179 0.03 4,536 0.00 915 0.03 1,089 0.00
Community Characteristics
Covid Cases and Deaths 2,308 750.62 4,406 625.80 948 1805.36 1,040 1738.70
Percent Black in Census
Block 1,682 0.09 3352 0.08 688 0.11 699 0.08
Percent Hispanic in Census
Block 1,682 0.37 3352 0.36 688 0.40 699 0.38
Percent LMI in Census
Block 1,688 0.55 3370 0.55 691 0.57 702 0.56
Day-Level Dataset
Initiated, never submitted 64 0.59 65 0.56 132 0.67 139 0.67
Submitted 64 0.4 65 0.43 132 0.33 139 0.33
Under Review 64 0.2 65 0.3 132 0.08 139 0.11
Funded 64 0.04 65 0.1 132 0.04 139 0.06
Business Characteristics
Jobs retained 64 3.54 65 4.44 132 3.47 139 4.12
Number of employees 64 293,885 65 345,030 129 4.88 139 7.09
Loan amount 64 $18,536 64 $19,584 132 $16,619 136 $17,625
Community Characteristics
Covid Cases and Deaths 64 747 64 747 132 1,661 138 1,636
Percent Black in Census
64 0.09 65 0.08 130 0.11 128 0.08
Block
Percent Hispanic in Census
64 0.35 65 0.35 130 0.39 128 0.36
Block
Percent LMI in Census
64 0.56 65 0.55 130 0.57 128 0.54
Block
Note: The day-level panel dataset is unbalanced because on some days in the application period, there were no initiations among underserved or
non-underserved businesses. In other cases, there is missing data due to an inability to match businesses based on addresses to census data. N is
the number of non-missing observations; for each variable, 0=no, 1=yes, so the Mean is the proportion of observations in the "yes" category
53 of 61
Figure 1. Descriptive graphs showing changes over time, by underserved status
Note: the Y-axis scale changes across the different outcomes based on the distribution of the data.
54 of 61
Table 5. Change in Probability of Application Success for Underserved and Non-Underserved
Businesses
Submit Application Under Review Funded Application
After (Underserved Businesses) 0.261*** 0.0838** 0.0382**
(0.06) (0.04) (0.02)
Non-underserved Business -0.177*** -0.154*** -0.03
(0.06) (0.04) (0.02)
After*Non-underserved Business 0.157** 0.249*** 0.103***
(0.08) (0.06) (0.02)
Time -0.00450*** -0.00397*** -0.00162***
(0.00) (0.00) (0.00)
After*Time 0.00 0.00309*** 0.00167***
(0.00) (0.00) (0.00)
Non-underserved Business*Time -0.00619*** -0.00741*** -0.00232***
(0.00) (0.00) (0.00)
After*Non-underserved Business*Time 0.00663*** 0.00656*** 0.00162**
(0.00) (0.00) (0.00)
New COVID Cases and Deaths -4.06e-05** -2.37e-05* 0.00
(0.00) (0.00) (0.00)
Constant 0.216*** 0.05 -0.0543***
(0.06) (0.04) (0.01)
Day of the Week Fixed Effects X X X
Observations 398 398 398
R-squared 0.238 0.469 0.25
Note. The coding scheme (Table 2) means that the coefficient for After is the effect of the program change in
non-underserved businesses (i.e., an increase of 26.1 percentage points in submissions), which tests H1. Robust
standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
55 of 61
Table 6. Change in Probability of Application Success for Minority-Owned vs. Non-Underserved
Businesses
Submit Application Under Review Funded Application
After (Minority-owned Business)* 0.329*** 0.07 0.0463**
(0.08) (0.04) (0.02)
Non-underserved Business -0.148** -0.165*** -0.02
(0.07) (0.04) (0.02)
After*Non-underserved Business 0.08 0.257*** 0.0941***
(0.09) (0.06) (0.03)
Time -0.00567*** -0.00400*** -0.00190***
(0.00) (0.00) (0.00)
After*Time 0.00326* 0.00318*** 0.00197***
(0.00) (0.00) (0.00)
Non-underserved Business*Time -0.00524*** -0.00745*** -0.00205***
(0.00) (0.00) (0.00)
After*Non-underserved Business*Time 0.00585*** 0.00658*** 0.00135**
(0.00) (0.00) (0.00)
New COVID Cases and Deaths 0.00 0.00 0.00
(0.00) (0.00) (0.00)
Constant 0.162** 0.05 -0.0611***
(0.07) (0.04) (0.01)
Day of the Week Fixed Effects X X X
Observations 393 393 393
R-squared 0.211 0.447 0.22
Note: The coefficient for After is the effect of the program change in minority-owned businesses relative to non-
underserved businesses. Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
56 of 61
Table 7. Change in Probability of Application Success for Women-Owned vs. Non-Underserved
Businesses
Submit Application Under Review Funded Application
After (Women-Owned Business) 0.249*** 0.0833** 0.0340**
(0.07) (0.04) (0.02)
Non-underserved Business -0.174*** -0.150*** -0.01
(0.05) (0.04) (0.02)
After*Non-underserved Business 0.176** 0.254*** 0.104***
(0.08) (0.06) (0.03)
Time -0.00518*** -0.00453*** -0.00227***
(0.00) (0.00) (0.00)
After*Time 0.00263* 0.00373*** 0.00237***
(0.00) (0.00) (0.00)
Non-underserved Business*Time -0.00559*** -0.00689*** -0.00164**
(0.00) (0.00) (0.00)
After*Non-underserved Business*Time 0.00619*** 0.00588*** 0.00
(0.00) (0.00) (0.00)
New COVID Cases and Deaths -3.98e-05** -2.43e-05* 1.25e-05*
(0.00) (0.00) (0.00)
Constant 0.180*** 0.05 -0.0642***
(0.05) (0.04) (0.02)
Day of the Week Fixed Effects X X X
Observations 385 385 385
R-squared 0.257 0.477 0.282
Note: The coefficient for After is the effect of the program change in women-owned businesses relative to non-
underserved businesses. Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
57 of 61
Table 8. Change in Probability of Application Success for Underserved and Non-Underserved
Businesses, with Additional Control Variables
Submit Application Under Review Funded Application
After (Underserved Business) 0.211*** 0.0733** 0.0331**
(0.06) (0.04) (0.01)
Non-underserved Business -0.214*** -0.162*** -0.0300*
(0.06) (0.04) (0.02)
After*Non-underserved Business 0.170** 0.252*** 0.106***
(0.08) (0.06) (0.02)
Time -0.00368*** -0.00382*** -0.00158***
(0.00) (0.00) (0.00)
After*Time 0.00 0.00305*** 0.00169***
(0.00) (0.00) (0.00)
Non-underserved Business*Time -0.00691*** -0.00757*** -0.00239***
(0.00) (0.00) (0.00)
After*Non-underserved Business*Time 0.00749*** 0.00675*** 0.00167**
(0.00) (0.00) (0.00)
New COVID Cases and Deaths 0.00 -2.06e-05* 0.00
(0.00) (0.00) (0.00)
Loan Amount -1.01e-08** 0.00 -3.73e-09**
(0.00) (0.00) (0.00)
Number of Employees 1.33e-05*** 0.00 0.00
(0.00) (0.00) (0.00)
Constant 0.00 0.00 -0.0713***
(0.08) (0.05) (0.02)
Day of the Week Fixed Effects X X X
Observations 392 392 392
R-squared 0.288 0.471 0.254
Note. The coding scheme (Table 2) means that the coefficient for After is the effect of the program
change in non-underserved businesses (i.e., an increase of 26.1 percentage points in submissions),
which tests H1. Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
58 of 61
Appendix A. Semi-Structured Interview Protocol
As part of our work with the Small Business Administration, we used these informal interviews to
identify barriers to accessing small business relief funds during the COVID-19 pandemic, as well as scope
out opportunities to build evidence on tools for reducing barriers either in experimental or quasi-
experimental interventions. We partnered with colleagues at the Office of Evaluation Sciences and the
Brown Policy Lab to identify city contacts and contacts working on small business development in the
nonprofit sector in cities. We sent out an email to the contacts in the What Works Cities network, which is
a consortium of city officials committed to evidence building efforts.
Questions for City Partners
1. When you gave out grants and loans to small businesses before, how oversubscribed was it?
2. When are they planning to give out more relief?
3. Have they already decided on amount, grant/loan, allocation method, etc.?
4. Do they already have outreach programs for businesses? Is the outreach targeted to underserved
communities?
5. Which underserved communities are a focus for your city?
6. How busy were they with application assistance during previous rounds of funding? Were they
able to support everyone who needed help?
7. Guesstimate: What number/percentage of businesses failed to submit adequate documentation?
What were the key barriers for underserved businesses in the process of applying?
8. Did they advertise third-party assistance? How does this work?
9. What kind of data do they collect on applicants?
10. Do they have a checklist/template for documentation requirements?
11. What are you hoping to learn related to small business recovery and equitable access to small
business loans and grants?
12. Do you have any current evaluations ongoing in this space? How would this complement or add
to those efforts?
Based on the answers to questions 4-8, we constructed the following behavioral map shown in Figure A1
below that specifies the different burdens and barriers in the process of trying to access small business
relief funds during COVID-19, and why these barriers disproportionately fell on underserved
communities.
59 of 61
Figure A1. Structural Barriers, Behavioral Bottlenecks, and Proposed Interventions for Small Business
Relief Funding
60 of 61
i
We utilize this term to align with the recent Executive Order on Advancing Racial Equity and Support for
Underserved Communities. The executive order defines underserved communities as “populations sharing a
particular characteristic, as well as geographic communities, that have been systematically denied a full opportunity
to participate in aspects of economic, social, and civic life” (Biden 2021). This includes “Black, Latino, and
Indigenous and Native American persons, Asian Americans and Pacific Islanders and other persons of color;
members of religious minorities; persons with disabilities; persons who live in rural areas; and persons otherwise
adversely affected by persistent poverty or inequality” (Ibid., 2021).
ii
We include veteran-owned small businesses in the “underserved” category because they are frequently prioritized
for government services and are disproportionately likely to face mental health challenges. For instance, the 2021
$28.6 billion Restaurant Revitalization Fund prioritized the processing of applications veterans. Administrators in
the city we studied similarly included veterans in their list of underserved small business owners. This categorization
aligns with the high prevalence of disabilities, substance abuse disorders, and mental health conditions in veterans,
including conditions that often go undiagnosed and therefore would not be captured in the disability category
(Sobota 2017; Teeters et al. 2017; Burnam et al. 2009; Merikangas and Conway 2021).
iii
Another approach to burden reduction is the provision of personalized assistance, but this is a less prominent
intervention partially because of the substantial cost of these programs relative to nudges (especially when
considered at scale).
iv
Some scholarship in specific fields of policy, such as the work of Janet Currie, does examine heterogeneous
effects but this does not transfer over to public administration scholarship.
v
It is also worth noting that the interventions outlined above were generally tested in the context of entitlement
programs with benefits available to all eligible applicants, rather than oversubscribed benefit programs, such as
subsidized housing with long waiting lists, or small business grants and loans. For oversubscribed programs, efforts
to improve equity could take the form of compensating for unequal administrative burdens in application by giving
preference to applicants from underserved groups. However, that approach is typically seen as illegally
discriminatory (at least in the United States) (Cowley 2021). Since policymakers are largely blocked from using
explicit preferences to enhance equity for oversubscribed programs, it is particularly important to understand how a
general approach to easing administrative burdens for all applicants may affect equity.
vi
Veteran-owned small businesses have similar financing patterns and access to capital as other businesses (Sobota
2017), but given the service contributions veterans have made and the additional physical and mental health
challenges they face, the US government has prioritized support to these businesses, even creating set-aside
programs for service-disabled veteran-owned small businesses (Small Business Administration 2022).
vii
We preregistered all hypotheses in a pre-analysis plan posted prior to obtaining outcome data. The pre-analysis
plan is available upon request, but due to anonymity requirements, we exclude this link.
viii
Therefore, we only have information about businesses which at least start an application and are unable to capture
equity gaps prior to starting an application.
ix
There were additional statuses recorded by city employees, but we collapse the submission stages into 4 main stages:
initiated, submitted, under review, and awarded funding. This is for ease of interpretation, and because these are the
main stages where there is likely drop-off in the application process. In line with our pre-analysis plan, we exclude
data that does not include the timing of any application status, as this is essential to the analysis. Less than one percent
of the data was missing this information.
x
For example, if the first submission was after the policy change, this would be classified as After (see Table 2),
regardless of whether the first initiation was prior to or after the policy change. However, if they submitted prior to
and after the change, we would take the first submitted date, which would mean they are in the Before (pre-
treatment) group because they submitted an app when there was the longer list of documentation requirements.
xi
Approximately 30 percent or 5,000 out of 15,000 applications have missing data on these variables.
xii
We will reject the null hypothesis for the analysis that produces the smaller of the two p-values when that value is
< .025 (.05/2). If the smaller p-value is < .025 and the larger is < .05, we will also reject the null hypothesis for the
analysis that produced the larger of the two p-values.
xiii
All of the results reported are statistically significant after controlling for the familywise error rate using the
Holm-Bonferroni procedure.
61 of 61
xiv
We chose this coding approach so that the other coefficients (e.g, After) would represent the effect for the
Underserved businesses, the group of particular interest.
xv
Therefore, for any business initiating in the pre-period, we would be capturing the likelihood that they submit, go
under review, or are funded JUST in the pre-period. This means that for businesses that initiate in the pre-period, but
submit/go under review/are funded in the post period, they are not counted in the outcome variable. Only 16
businesses were in this category.
File and source
- File
- bell-2023-how-reducing-documentation-burdens-impacts-equity-in-access-to-small-business-co.pdf
- Size
- 1,462,836 bytes
- SHA-256
- 34e7ae04df5ab05f054794be6dacf1e395ba34dc0d19e12de01540f37781b84b
- Our copy
- bell-2023-how-reducing-documentation-burdens-impacts-equity-in-access-to-small-business-co.pdf
- Original
- researchonline.lse.ac.uk