Full text
Review of Quantitative Finance and Accounting (2025) 64:1783–1808
https://doi.org/10.1007/s11156-024-01348-3
ORIGINAL RESEARCH
Did Biden-Harris’s reforms on the Paycheck Protection
Program reduce racial disparities in lending?
Stefano Filomeni1 · Francesca Querci2
Accepted: 27 August 2024 / Published online: 15 October 2024
© The Author(s) 2024
Abstract
Introduced under the Trump-Pence Administration, the Paycheck Protection Program
(PPP) provided short-term relief loans to small American businesses during the peak of
the Coronavirus pandemic. The initial design of the PPP faced significant criticism from
researchers due to racial disparities, among other issues, in its lending process. Minor-
ity groups received smaller PPP loan amounts during the original two tranches released
in 2020. To increase equitable access for all, in February 2021 the Biden-Harris Ad-
ministration enforced swift changes to the initial PPP aimed at favouring access to PPP
loans for minority-owned small businesses that had been disadvantaged by the program’s
original design under the Trump-Pence Administration. By exploiting a granular dataset
of 1,759,270 PPP loans granted between Q2 2020 and Q2 2021 and by implementing a
difference-in-differences approach (DID), this paper provides novel evidence on the effec-
tiveness of the Biden reforms in reducing racial disparities within the Paycheck Protection
Program. Indeed, we observe a significant increase in the volume of PPP loans granted
to minority-owned businesses in the period following the Biden-Harris Administration’s
reforms. Furthermore, among different minority groups, the reforms appear most effective
for Native American minority groups (including American Indians, Alaska Natives, Na-
tive Hawaiians and/or Other Pacific Islanders), followed by Black Americans and Asian
business owners. Our findings offer novel contributions to the existing literature on in-
stitutional discrimination, particularly regarding the initial PPP design. Our findings are
especially valuable for policy makers as they underscore the importance of radical changes
in addressing racial disparities. Our paper also offers evidence of how a public credit guar-
antee program should be designed to empower and promote economic inclusion for all,
regardless of ethnicity, aligning with the UN’s Sustainable Development Goals.
Keywords Paycheck Protection Program · Lending · Minorities · Institutional
discrimination · Racial disparities
JEL Codes: G01 · G21 · G28 · J15
Extended author information available on the last page of the article
13
1784 S. Filomeni, F. Querci
1 Introduction
Due to the COVID-19 pandemic, economies globally suffered from one of the worst eco-
nomic downturns in recent history. Governments were forced to employ national lockdowns,
causing major shutdowns of key industries. Enforcement of self-isolation and state-level
lockdowns caused a 9.3% decrease in the U.S. GDP (The Bureau of Labor and Statistics,
2020). Such unprecedented protocols led to mass reduction of employees by businesses.
The U.S. unemployment rate peaked, rising in the early month of 2020 from 3.6 to 10.1%
(FRED, 2020). In response to the pandemic, the U.S. Congress passed the Coronavirus
Aid, Relief, and Economic Security (CARES) Act 2020. Signed by Former U.S. President
Donald Trump, the stimulus bill intended to cater for individuals, families, and small busi-
nesses who were the most severely affected to receive financial assistance. Solely dedicated
to helping small businesses, a federal fiscal policy called the Paycheck Protection Program
(PPP hereinafter) was introduced. The PPP intended to provide forgivable loans, guaranteed
by the Small Business Administration (SBA), to eligible firms. With current political and
cultural issues arising at the time, race, and ethnicity became the focal point of interest for
media and studies. In light of the program, questions were raised about the equitable fairness
of loan distributions. Countless studies found disparities within the program (e.g., Lederer
and Oros 2020; Wang and Zhang 2021; Kelly and van Holm 2021; Santellano 2021; Kickul
et al. 2021; Howell et al. 2022; Atkins et al. 2022; Chernenko et al. 2023; Lester and Wilson
2023; Howell et al. 2024; Chernenko and Scharfstein 2024; Kotomin et al. 2024; Lelo de
Larrea et al., 2024). The PPP only reached up to 20% of all eligible firms in rural-dominated
states with the highest concentration of Black-owned small businesses. An alarming statis-
tic when considering two thirds of such rural areas suffered the highest recorded number
of COVID-19 cases at the reported time, contradicting the primary target of the PPP and
the CARES Act 2020 which promised to help and protect small businesses most severely
affected by the Coronavirus pandemic (Mills and Battisto 2020).
After the 2020 presidential election, the Biden-Harris Administration set out new
reforms to the PPP. Amongst such changes, arguably the most important reform attempting
to address racial disparities was the inclusion of small business owners with non-fraudulent
felonies or delinquencies on federal student loans. This previously excluded group of own-
ers consisted of disproportionately large amounts of Black-American individuals. In theory,
the implementation of such reform alone should welcome more business owners to the pro-
gram who identify as part of minority groups. Accompanied by the rest of the reforms these
changes claim to improve the equitable distribution of forgivable loans to minority-owned
small businesses, according to official statements from the White House (2021).
The PPP intends to provide equitable, forgivable loans to small businesses most severely
affected by the Coronavirus pandemic. Following the first two tranches employed by the
Trump-Pence Administration, reports of lending discrimination towards minorities sparked.
Ever since, the program has undergone design reforms in an attempt to increase equal dis-
tribution of forgivable loans to U.S. small business owners of varying racial backgrounds as
well as industries. Henceforth, this paper investigates the true fairness of PPP loan distribu-
tion to small U.S. firms under the new design conditions. Our methodology is designed to
analyse the effect of reforms enacted by the Biden-Harris Administration in February 2021
to tackle racial disparities in lending under the PPP.
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1785
Conjointly, our paper intends to contribute towards existing literature proclaiming the
previous PPP design suffered from racial and/or implicit bias, causing racial disparities in
loan distribution to minority business owners as well as receiving lower loan amounts than
White American counterparts. From the perspective of this study, this is one of, if not, the
first study conducted on the PPP that takes into account reforms made to the federal fiscal
policy and its effects on equitable distribution to non-white business owners.
By way of preview, our findings confirm the existence of racial disparities in granting
PPP loans, consistent with the existing studies showing that racial minority groups received
disproportionately less than to their White counterparts as part of an ongoing issue with
institutional racism within the U.S. lending industry (Santellano 2021). Further to the point,
our results convey that the implemented reforms expanded the scope of racial minority own-
ers approved for forgivable loans as well as increased the loan amount permitted for certain
minority groups. Among present minority groups, we find that business owners belonging
to Native American minority groups (American Indian, Alaska Native, Eskimo & Aleut,
Native Hawaiian and/or Other Pacific Islanders), as well as Black Americans and Asian
business owners benefitted the most, in terms of access to PPP loans, from the introduction
of the Biden-Harris reforms. Our paper provides novel evidence on the effectiveness of the
reforms introduced by the Biden-Harris Administration in reducing racial bias in PPP lend-
ing to minority business owners, as compared to the previous US Administration.
This study is closely related to several papers showing disparities - whether based on
gender, race, or other borrower characteristics - existing within the PPP program, which was
designed to be accessible to businesses which were the most disadvantaged by the pandemic
(Howell et al. 2024). Previous papers considered discrimination based on lenders’ behaviour
(Lederer and Oros 2020; Kickul et al. 2021; Chernenko et al. 2023; Howell et al. 2022, 2024;
Atkins et al. 2022; Chernenko and Scharfstein 2024; Kotomin et al. 2024; Lelo de Larrea
et al., 2024), or discrimination based on applicants’ self-selection behaviour; Chernenko et
al. 2023; Chernenko and Scharfstein 2024), or inequality of opportunity (Wang and Zhang
2021; Kelly and van Holm 2021), as sources of disparities in credit access within the PPP.
Our research differentiates from the extant literature on PPP, as we examine institutional
discrimination, that refers to laws, policies or practices that are not necessarily discrimina-
tory in intent, but nevertheless contribute indirectly to favouring certain groups and penal-
ising others (Guimond 2023). In other terms, we consider a source of disparities that has
its origin in the way institutions, and not simply individuals, operate and design economic
policy measures.
Compared to previous studies that examine institutional discrimination within PPP,
our paper is unique in the external validity, in the scope of analysis and in the period of
investigation. While Santellano (2021) draws on interviews at Latino-owned coffee shops
in Los Angeles during Covid-19 to contend that the way PPP has played out is part of an
exclusionary history in entrepreneurship, we analyse a wider dataset on PPP loans gathered
from the SBA website to evaluate the effectiveness of the reforms enacted under the Biden-
Harris Administration in reducing race disparities. Moreover, while Lester and Wilson
(2023) examine the effectiveness of preliminary changes to the PPP before the beginning of
Biden’s mandate, we consider the subsequent, much broader, reforms of the program that
were introduced from February 24, 2021.
First, we show that the initial design of the PPP implemented under the Trump-Pence
Administration did not ensure an adequate reach of small businesses, penalising minority-
13
1786 S. Filomeni, F. Querci
owned firms. Second, we find that the radical reforms implemented by the Biden-Harris
Administration significantly reduced disparities among the different ethnic groups in obtain-
ing PPP loans. This is a distinct part of our research design that has, so far, not been inves-
tigated by the extant literature.
This paper adds insights into the causes and consequences of inequality, hence provid-
ing guidance to policy makers when conceiving business relief measures that also address
inequality and reduce disparities in resource allocation. In particular, our study takes up
the idea that in government efforts a greater concern must be put behind ensuring equity in
the distribution of funds (Kelly and van Holm 2021). More generally, this paper also offers
evidence of how a public credit guarantee program should be designed to empower and
promote the economic inclusion of all, irrespective of ethnicity, in line with the UN’s Sus-
tainable Development Goal 10 (‘Reduce inequality within and among countries’). Indeed,
inequality undermines long-term social and economic development, hampers poverty reduc-
tion and wrecks people’s sense of accomplishment and self-worth. This, in turn, can feed
crime, disease and environmental degradation. Greater efforts are needed to progressively
eradicate inequalities through inclusive political, economic, and social policies.
The remainder of the paper is structured as follows. Section 2 provides details about the
PPP and the reforms implemented by the Biden-Harris Administration to reduce racial bias.
Section 3 reports the related literature on racial disparities according to the PPP as well as
the hypotheses tested. Section 4 describes the data used for this study and documents our
empirical methodology. Section 5 describes the results. Finally, Sect. 6 concludes the paper.
2 The Paycheck protection program and the Biden-Harris
administration
Enacted by former U.S. President Donald Trump on 27th March 2020, the PPP paved the
way for a temporary solution for the growing issue of unemployment within American
small businesses. The original infrastructure for the program was to be funded $349 billion,
through the CARES Act 2020, and allocate firms with up to 500 employees with forgivable
loans no more than $10 million (SBA 2020). Funding utilized the E-TRAN system, devel-
oped by the SBA to reduce processing time of PPP loans for banks (SBA 2020; Bartik et
al. 2021). PPP loans were distributed through banking institutions as the assumption of pre-
existing relationships between business owners and lenders would increase disbursement
(Grandja, Makridis, Yannelis, and Zwick, 2021).
Applicable to the program, so-called “forgivable loans’’ add further incentives to bor-
rowers. To receive such incentives, eligible business owners within the first two tranches
must retain employee compensation identical to pre-pandemic levels, spend loan proceeds
on miscellaneous business expenses and utilize at least 60% of loan amount for payroll.
If satisfied 8 to 24 weeks following loan disbursement, the full amount is to be forgiven
(SBA 2020). Due to the large demand for these loans intensified by worsening COVID-
related events, the first tranche of PPP loans was exhausted within 2 weeks of its launch
(April 3–16, 2020). With virus infections continuing to soar and national restrictions crip-
pling small businesses, a second tranche of forgivable loans were introduced under the
Trump-Pence Administration through the enactment of the Paycheck Protection Program
and Health Care Enhancement Act. This act added $310 billion to be further allocated to
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1787
business owners, with or without a pre-existing PPP loan. The second tranche was available
from April 27th and was exhausted by the 8th of August 2020. The depletion of funds pro-
vided by the CARES Act and Health Care Enactment Act caused the PPP to discontinue the
distribution of forgivable loans to small business owners (Balyuk, Prabhal and Puri, 2021).
The introduction of the Consolidation Appropriations Act (CAA) 2021 allocated a third
tranche of forgivable loans worth $284 billion. Some changes were introduced compared to
the previous two tranches that responded to criticisms that the program did not adequately
reach small and minority-owned businesses. The first week of the third round was only
open to small lenders and participating SBA-approved Community Financial Institutions.
An amount of $35 billion was set aside for first-time borrowers and further $35 billion were
addressed to first- and second-draw borrowers with 10 or fewer employees or loans less
than $250,000 in a low-or moderate-income community. The third tranche, made available
from January 19, had an expiration date of May 31, 2021. At the end of its operation, the
PPP accounted for the distribution of $943 billion in forgivable loans to American small
businesses.
Following the successful 2020 presidential election, the Biden-Harris Administration
announced a plethora of reforms to the PPP advocating for equitable access to relief loans
and addressing higher rates of access to PPP funding for minority-owned small businesses
as compared to their White-owned counterparts (Fairlie 2020; Misera 2020). One of the
most important reforms introduced by the Biden-Harris Administration was the restructur-
ing of the previous loan calculation formula for self-employed individuals. Recognising that
businesses may have had negative net profits on their balance sheets, changes to the loan
calculations of the PPP allowed owners the choice between using their business’ net profit or
gross profit. The relaxed loan calculation meant, in theory, that small business owners would
receive more substantial PPP loan amounts compared to the previous design (O’Rourk and
Bobrosky 2021). In addition to the funding from the Consolidation Appropriations Act,
the Biden-Harris Administration allocated a further $1 billion for small businesses within
these business categories located in lower to middle income neighbourhoods. A welcomed
change to the program’s design as 70% of such businesses are owned by people of colour,
yet from previous tranches, numerous owners only received a couple of dollars as loan allo-
cations (Office of Advocacy 2018; The White House 2021). In addition, reforms removed
the exclusionary restrictions on individuals previously convicted of non-fraudulent felonies
as well as those with delinquencies on federal student loans. This is part of the PPP Second
Chance Act that the Biden-Harris Administration implemented with the objective to allow
more businesses to access PPP loans than the previous PPP tranches under the Trump-Pence
Administration. This is more significant for minorities such as Black Americans, who have
disproportionately higher amounts of delinquency on student loans as well as felony convic-
tions (The White House 2021). Our study analyses loans released before and after the new
reforms to the PPP introduced by the Biden-Harris Administration were in effect.
3 Related literature and hypotheses development
One of many major responses during an economic crisis is mass-scale public lending (Kelly
and van Holm 2021). Given its scope and the relevance of its purpose, the PPP has attracted
the interest of academic scholars, who examined the reach of this government-funded stim-
13
1788 S. Filomeni, F. Querci
ulus program (Schweitzer and Guo 2024), its effect on small business performance (Staples
and Krumel 2023), and on the risk-taking of US banks (Filomeni 2023). With relevant polit-
ical and cultural issues arising at the time of the PPP enactment, disparities became the focal
point of interest. Questions on the true equitable distribution of loan distributions to small
business owners were raised, race being one of the most critical topics of research within
the context of the PPP. Exploring existing studies on racial disparities is essential in under-
standing the level of inequality within the U.S. small business lending industry. Among the
countries whose population includes people with different origins and cultures, the U.S. is
a particularly interesting context for our analysis, since racial discrimination has very deep
historical roots.
Discrimination in access to the credit market is a widely studied phenomenon in the
United States, with reference to both the residential mortgage segment and small business
lending. Regarding credit to small businesses, the empirical evidence has shown that eth-
nic minority groups are less likely to receive loans (Bostic and Lampani 1999; Cavalluzzo
and Wolken 2002; Cavalluzzo et al. 2002; Blanchflower et al. 2003; Blanchard et al. 2008;
Asiedu et al. 2012; Mijid and Bernasek 2013; Rakshit and Peterson 2022), they are charged
with a higher interest rate (Blanchflower et al. 2003; Blanchard et al. 2008; Asiedu et al.
2012), or are even discouraged from applying for funding (Bates and Robb 2015a, b).
The literature on the PPP has confirmed these findings, attributing race disparities mainly
to lenders’ behaviour and their beliefs distorted by prejudice and hostility (taste-based dis-
crimination). Lederer and Oros (2020) conducted 63 fair lending matched‑pair audit tests
of financial institutions in Washington, DC, and found a difference of treatment in 43%
of cases with White testers being favoured over Black testers in the pre‑application stage.
Moreover, disparities also regarded the gender of testers. Differences emerged in the level
of encouragement, the products offered, information provided, and information requested.
More precisely, bank employees informed White testers that they would qualify for a loan at
a significantly higher rate than Black testers. Besides, women were offered less information
about the PPP products compared to men. Moreover, a statistically significant difference
was found in information discussed around the type of loan and purpose of the loan with the
bank employee discussing this information more with the White male tester compared to the
Black male tester and the female testers. In addition, the Black tester was asked more per-
sonal questions about identifying his business and was told less information such as fees and
interest rates. On the contrary, the White tester was told specific information without having
to verify any information about his business. Chernenko and Scharfstein (2024) highlighted
that Black and other minority restaurant owners in the State of Florida were 25% less likely
to receive PPP compared to White-owned restaurants. Moreover, minorities were 36.2%
less likely to even attain forgivable loans with respect to their non-minority counterparts.
The authors explained these disparities by pointing to the existence of racial bias in the way
banks processed PPP loan applications. Moreover, even if there was no discrimination in the
application process at banks, a legacy of past discrimination and poor treatment discouraged
minority-owned businesses from even approaching banks for a PPP loan. Thus, disparities
can be originated by both lenders’ conduct and borrowers’ self-selection behaviour.Similar
findings were reached by Lelo de Larrea et al. (2024) who analysed differences in PPP
approval amounts based on race and gender in the U.S. hospitality industry and found sig-
nificant differences between White- and Black or African American-owned businesses and
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1789
between male- and female-owned businesses, with the latter of these groups at a consider-
able disadvantage.
Examining the differences in PPP approval amounts based on the type of lenders, Howell
et al. (2022) observed that Black-owned businesses secured more PPP loans through non-
traditional lending institutions (such as FinTech lenders) than those who applied through
traditional banks relying more on soft information-intensive relationship banking. Indeed,
the authors attribute these results to the automation of lending processes that are likely to
lead to higher amounts of loan approvals for Black business owners (PPP loan approval
rates increased by 20% post-automation of existing loan approval systems). This, in turn,
reduces bias on race/ethnicity within banking institutions in their PPP lending activity. In
a subsequent study, Howell et al. (2024) found Black-owned businesses were about 12%
points more likely than other firms to receive their PPP loans from a fintech lender, mainly
due to preference-based discrimination. After conventional lenders automated their lending
processes, their rates of lending to Black-owned businesses increased substantially. Similar
findings are achieved by Atkins et al. 2022. However, in a study based on the 2020 Small
Business Credit Chernenko et al. (2023) found that approval disparities were similar in mag-
nitude at banks and fintechs. The authors argued that there was more scope for racial bias
to influence approval decisions at banks, since their PPP approval process was generally
more hands-on and interactive, particularly in more racially-biased locations. The approval
process at fintechs, by contrast, was far less personalised. However, the hands-on approach
better positioned banks to help applicants resolve documentation gaps and determine the
correct loan amounts. Conversely, fintechs did not have enough employees to help submit
a successful application. While fintech automation could be helpful in reducing taste-based
discrimination in small business lending, it could be ill-suited to helping firms through the
application process. More recently, Kotomin et al. (2024) highlighted that institutional and
racial disparities persist during the PPP loan forgiveness stage. Controlling for various loan-
and borrower-level characteristics, they demonstrated that relationship lenders — commu-
nity banks, credit unions, and farm credit institutions — are associated with higher rates of
PPP loan forgiveness. In contrast, automated lenders — fintechs and fintech banks — exhibit
the lowest forgiveness rates. Black borrowers experience the poorest forgiveness outcomes,
except for loans issued by non-depository fintech where they outperform White borrowers.
Some studies analysed the spatial distribution of PPP loans, finding that over-reliance
on established financial institutions led to disparities within financially under-connected
communities. In this case, disparities in access to PPP loans were not driven by lenders
or applicants’ behaviour but inequality of opportunity, deriving from circumstances over
which individuals have no control, such as unequal access to bank branches and ‘banking
deserts’. Wang and Zhang (2021) provided evidence that Zip codes with higher density of
Black-owned businesses received a lower chance of connecting to an enrolled bank branch.
Reasoning behind this was correlated to both lower numbers of bank branches in neighbour-
hoods with a predominant African American presence but also branches that were present in
the area had little to no prior relationships with the SBA. Kelly and van Holm (2021) found
that communities with more eligible firms, larger access to bank branches and areas experi-
encing higher rates of unemployment received more loans per capita. Supporting findings of
Wang and Zhang (2021), zip codes with predominantly white inhabitants received more and
larger amounts of PPP loans than neighbourhoods composed mostly of minority citizens.
13
1790 S. Filomeni, F. Querci
So far, only a few studies concerning the PPP have traced race disparities back to insti-
tutional discrimination. Santellano (2021) investigated how small businesses owned by
Latino individuals were able to access PPP loans in Los Angeles. Based on first-hand inter-
views with Latino coffee shop owners, the author provides evidence that a lower number
of PPP loans were distributed to small business owners of colour. This result is attributable
to the deepening institutional discrimination reportedly involved within the PPP applica-
tion process. More recently, Lester and Wilson (2023) compared PPP to existing residential
and small business lending patterns and found that the program mostly followed existing
mechanisms of capital flow that resulted in fewer resources reaching communities of colour
and low-income neighbourhoods. The policy changes after the second phase of the PPP,
including early access to SBA-approved Community Financial Development Institutions
and setting aside specialised loan pools for first-time borrowers and very small businesses,
corrected in part previous barriers. However, those changes did not seem to significantly
reduce disadvantage in majority Latinx areas.While most past studies on PPP lending high-
light the existence of racial disparities in the distribution of loans, Sant’Anna et al. (2023)
come to partially different results. They find that African American farmers received more
PPP loans both on the absolute amount and per employee basis than white farmers. Yet,
African American farmers living in rural counties or belonging to low- to moderate-income
groups received lesser amounts of PPP loans relative to African American farmers who did
not belong to those groups. The authors explain their findings presuming that approved
farmers had an established borrower-lender relationship with the lender they used to apply
for PPP. If true, a previous borrower-lender relationship would mean that these farmers were
potentially less risky, making their access to capital easier.
In general, the current literature agrees that minority-owned businesses were at a disad-
vantage in being granted PPP loans. The extant literature has provided a variety of explana-
tions for the presence of the racial disparities within the PPP. Following Santellano (2021)
and Lester and Wilson (2023), our study focuses on disparities deriving from institutional
discrimination that involves the procedures, routines, and organisational culture of any insti-
tution that, often without intent, contribute to less favourable outcomes for minority groups
than for the majority of the population, from the institution’s actions and services. The first
step of our analysis grounds on the theory of institutional discrimination that was originally
formulated by Carmichael and Hamilton (1967) to look at how institutional policies cre-
ate consequences that perpetuate and worsen racial disparities. Members of racial minor-
ity groups may be systematically discriminated not only by the wilful acts of individuals,
but also because of a range of policies and practices by state or non-state institutions, that
contribute – often unintentionally – to the structural disadvantage of members of certain
ethnic groups (Pager and Shepherd 2008; Fibbi et al. 2021). In the U.S., historical experi-
ences of economic, social, and political exclusion based on race (Watts et al. 1999), have
made inequality an enduring self-perpetuating phenomenon. The production and reproduc-
tion of inequalities are at the foundation of institutional discrimination. Because of their
cumulative disadvantage, minorities lag behind on almost every measure of prosperity (i.e.,
employment, criminal justice, economic resources, health, and education). Consequently,
the eligibility criteria for the PPP program under the Trump-Pence Administration indi-
rectly penalised minority-owned small businesses compared to White-owned counterparts.
Despite the preliminary changes introduced after the first two tranches under the Trump-
Pence Administration and the fact that fintech companies were already included among the
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1791
PPP lenders, racial disparities persisted, based on the overall design of the program that
reproduced and reinforced inequalities.
Our discussion on racial disparities in lending within the initial design of the PPP, i.e.,
from April 3, 2020, to February 24, 2021, the first phases implemented under the Trump-
Pence Administration, can be summarized in our first hypothesis (H1) as follows:
H1 Minority-owned small businesses received less PPP loan amounts than their White-
owned counterparts within the first two tranches.
The existing literature investigating the PPP has not examined the consequences of the new
reforms implemented by the U.S. Biden-Harris Administration in February 2021 designed
to allow equitable distribution of PPP loans to small businesses independently of the gender
or race/ethnicity of their owner. Sourcing from the organisational culture field and endog-
enous models of change (Morrill 2008), we look at Biden-Harris’s reforms of the PPP as an
example of institutional heterogeneity that lays the foundations for the search for affirmative
action policies aiming at equity, through critical questioning of past institutions’ actions.
The reforms of the PPP are an endogenous source of change of the previous institutional
procedures, routines, and organisational culture, which interrupted the nurturing of cumula-
tive disadvantage of minority-owned small businesses. As a second step of our analysis, we
aim at evaluating the effectiveness of the broad, radical reforms of the PPP enacted under
the Biden-Harris Administration, in reducing race disparities deriving from institutional
discrimination. In this regard, following the implementation of the latter reforms by the
Biden-Harris Administration, we expect minority-owned small businesses to receive more
PPP loan amounts with respect to the previous Trump-Pence Administration, in line with
further efforts to increase the equitable distribution of PPP loans to small businesses in need
of them to survive in period of acute financial instability. Our discussion on the effectiveness
of the Biden-Harris reforms in reducing racial bias in lending on the PPP can be summarized
in our second hypothesis (H2) as follows:
H2 The reforms introduced by the Biden-Harris Administration to the PPP design decreased
racial disparities in PPP loan distribution.
Surrounding the scope of racial disparity, the PPP reforms aimed to reduce such arbitrariness
for all racial minority groups. Nonetheless, the present literature does not highlight equal
effects of design changes across all racial groups. As mentioned, current affairs at the time
circulated racial issues and news items, more importantly towards the African American
community. According to Pew Research Center’s 2021 statistics, nearly half of Americans
(46%) thought there was a lot of discrimination against Black people, while about three-
in-ten saw a lot of discrimination against Hispanic people (30%) and Asian people (27%).
Higher exposure of racial issues for a specific minority group, compared to other racial
minority groups, could boast the higher possibility of resolution specific to their cases, leav-
ing other racial minority groups to potentially benefit less from design changes compared to
African American counterparts. This expectation is coherent with the notion that media play
a significant role in shaping the social context in which policies are developed, influencing
agendas, resource allocation and decision-making in ministries and agencies. When media
pressure and broad public support build up, action is particularly taken when the issues are
13
1792 S. Filomeni, F. Querci
deemed important by political actors (Tresch et al. 2013; Figenschou et al. 2019). Concern-
ing that point, we anticipate that small businesses owned by individuals of African Ameri-
can descent to obtain higher PPP loan amounts compared to other racial minority groups,
post-enactment PPP reforms between February 24 – May 31, 2021. The argument of racial
impartiality of the Biden-Harris reforms imposed on the PPP can be summarized in our third
and final hypothesis (H3), as follows:
H3 Black American small business owners received higher PPP loan amounts after the
enactment of the PPP reforms by the Biden-Harris Administration in comparison to other
racial minority groups.
4 Data and estimated models
To test our hypotheses, we use granular loan-level data on 1,759,270 PPP loans collected
over the whole PPP lending period, i.e., between the periods of April 3–16, 2020, April 27
– August 8, 2020, and January 19 – May 31, 2021, from the section of the Small Businesses
Administration (SBA)’s website that provides PPP loan data.
The SBA constructed multiple excel sheets containing all PPP approved loans pooled
from all periods mentioned above. Along with the ethnicity of the borrowing business owner,
data on other characteristics of the borrowers were available, as well as the loan amount
approved, the approval date, and the forgiven amount of the loan (i.e., the amount of the PPP
loan that did not have to be paid back). The loan amount approved is our main dependent
variable. We used the forgiven amount of the loan for robustness checks. The main inde-
pendent variable of interest for our analysis is represented by the ethnicity of the small busi-
ness owners. We built a large dataset comprising a total of 1,759,270 loans, spanning from
April 3, 2020, to May 31, 2021. The description of each variable is provided in Table 1. The
dependent and explanatory variables and their descriptive statistics, as well as the correla-
tion matrix, are reported in Tables 2 and 3, respectively. In Table 3, any pair-wise correlation
beyond ± 10% is significant at the 10% level at least. In this regard, Table 3 shows that the
majority of the pair-wise correlations in regression analysis are within acceptable bounds.
Moreover, in Table 4 we report the Variance Inflation Factor (VIF) generated by our baseline
regression model to show that multicollinearity does not affect our main regression results.
In this regard, Table 4 shows that the VIF values generated by the regression analysis are not
critical as they show a moderate correlation among the variables.
Our baseline model tests the existence of racial disparities in lending within the initial
design of the PPP. To this aim we run OLS regression by including only the observations
pertaining to the first phases of the implementation of the PPP under the initial Trump-Pence
Administration, i.e., until February 24, 2021. We measure racial disparity by testing the sig-
nificance of the effect of the borrower’s race (Minority), on the approved loan amount (PPP
loan amount), taken in natural logarithm. The latter represents our dependent variable in all
our regression models, while our main independent variable of interest Minority captures
the intensity of racial bias within the initial design of the PPP. In all our regression models,
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1793
Table 1 Variable description
Variables Definition
PPP loan Approved loan amount for the borrower.
amount
Forgiven Amount of the PPP loan that has been forgiven.
amount
Minority Dummy variable equal 1 if the owner racially identifies as part of a minority group
(Asian; Black; Latino; Native) ; 0 otherwise (White).
Post-PPP Dummy variable equal 1 if the loan was approved post-enactment of the Biden-Harris
reforms Administration’ reforms (February 24, 2021-May 31, 2021) on the Paycheck Protection
Program; 0 otherwise.
Area Dummy variable equal 1 if the business operates within a Rural area; 0 otherwise.
Hubzone Dummy variable equal 1 if the business operates within a Hubzone area; 0 otherwise.
LMI Dummy variable equal 1 if the business owner had Low to Middle Income (≤ 58,020 $)
Debentures at the time of loan approval; 0 otherwise.
Firm Size Number of workers employed by the business.
Asian Dummy variable equal 1 if the race of the borrower is Asian; 0 otherwise.
Black Dummy variable equal to 1 if the race of the borrower is Black; 0 otherwise.
Latino Dummy variable equal 1 if the race of the borrower is Latino; 0 otherwise.
Native Dummy variable equal 1 if the race of the borrower is Native; 0 otherwise.
Firm Age Categorical variable equal to 1 if the firm has experienced a “Change of Ownership”, to
2 if the firm is “Existing or more than 2 years old”, to 3 if the firm is a “New Business
or 2 years or less”, to 4 if the firm is a “Startup, Loan Funds will Open Business”, and
to 5 if “Unanswered”.
Payroll Numerical variable reflecting the loan proceeds that are used to cover the firm’s payroll
Expenses costs.
Utility Numerical variable reflecting the loan proceeds that are used to cover the firm’s utility
Expenses costs.
Loan Maturity Numerical variable reflecting the loan’s maturity expressed in months.
Loan Status Categorical variable equal to 1 if the loan is “paid in full,” which means the borrower
paid the money back or paid it partially back and received forgiveness on the rest, to 2 if
the loan is “exemption 4,” which means the firm received the money but the loan hasn’t
been forgiven (the loan could be in the process of forgiveness or could be “charged off,”
which means the lender has written it off as a loss because the borrower may have shut
down, filed for bankruptcy or died), and to 3 if the loan is “active un-disbursed,” which
means the loan was approved but the lender hasn’t reported it as disbursed.
we also account for several control variables that might affect PPP loan amount other than
our variable of interest1.
Following Atkins et al. (2022), our OLS regression model for racial disparities in lending
within the initial design of the PPP takes the following form:
z
PPP loan amounti,t = α + β0 Minorityi + δj Borrower’s characteristicsi,t
j=1 (1)
+ Business Type FE+Quarter FE+State FE+Industry FE+εi,t
where the subscript i indicates the given borrowing firm being granted PPP lending in quarter
t; Minorityi indicates the race of the owner of borrowing firm i; Borrower’s characteristicsi,t
1
Although no reverse causality-related issues are likely to affect our regression estimates, we cannot rule out
the possibility that the latter might be affected by endogeneity issues due to omitted variable bias to some
extent.
13
1794 S. Filomeni, F. Querci
Table 2 Descriptive statistics
Variables Obs. Mean Std. dev. Min. Max.
PPP loan amount 1,759,270 58,318.03 204,484.4 2,907 10,000,000
Forgiven amount 1,362,882 61,153.98 210,493.8 0 10,000,000
Minority 1,759,270 0.41 - 0 1
Race
Asian 161,329 0.0917 - 0 1
Black and African American 516,787 0.2938 - 0 1
Latino 123 0.00007 - 0 1
Native 50,530 0.0287 - 0 1
White 1,030,501 0.5858 - 0 1
Area
Rural 411,478 0.2339 - 0 1
Urban 1,347,792 0.7661 - 0 1
Hubzone
Yes 570,761 0.3244 - 0 1
No 1,188,509 0.6756 - 0 1
LMI
Yes 537,385 0.3055 - 0 1
No 1,221,884 0.6945 - 0 1
Firm Size 1,759,270 6.6259 20.41 1 500
Firm Age 1,759,270 2.13 - 1 5
Payroll Expenses 1,757,496 56,527.19 195,544.3 1 10,000,000
Utility Expenses 797,073 1,189.10 18,961.47 0 8,111,700
Loan Maturity 1,759,270 48.81 16.24 3 60
Loan Status 1,759,270 2.58 - 1 3
Notes Data are collected from the section of the Small Businesses Administration (SBA)’s website that
provides PPP loan data
is a vector of borrower-specific characteristics described at the end of this section; Business
Type FE controls for type of business fixed effects based on how the borrower has registered
his entity2; Quarter FE presents quarterly fixed effects; State FE controls for US State fixed
effects based on the State where the small business operates; and Industry FE indicates
industry fixed effects identified via the NAICS code associated with the loan approved3.
Secondly, to investigate the effectiveness of the Biden-Harris Administration’s reforms
in reducing racial disparity present in PPP lending, we employ a difference-in-differences
(DID) approach since minority-owned small businesses are most likely to be influenced
by the new reforms to the PPP implemented by the Biden-Harris Administration. Hence,
following the above argument, our empirical analysis is conducted by exploiting small busi-
nesses’ racial/ethnicity heterogeneity. To identify small businesses’ differential exposure
to the aforementioned reforms, we therefore implement a difference-in-differences (DID)
analysis by constructing the treated and control groups based on firms’ status of being
2
The degree of correlation between business type and industry fixed effects is low (i.e., -0.0246), thus ruling
out possible multicollinearity issues.
3
Given the cross-sectional nature of the dataset in which each borrower participating in the Paycheck Protec-
tion Program is observed once, adding borrower fixed effects for borrowers with a single observation would
be equivalent to dropping this borrower from our sample, thus making the inclusion of borrower fixed effects
not feasible.
13
Table 3 Correlation matrix
PPP loan amount Minority Area Hubzone LMI Firm size Business Type Industry State
PPP loan amount 1
Minority -0.1224*** 1
Area -0.0741*** -0.2635*** 1
Hubzone -0.0307*** 0.1665*** 0.0574*** 1
LMI -0.0207*** 0.2394*** -0.1275*** 0.4459*** 1
Firm size 0.5366*** -0.1167*** -0.0188*** -0.0267*** -0.0241*** 1
Business Type -0.2586*** 0.1344*** 0.0806*** 0.0712*** 0.0567*** -0.1716*** 1
Industry -0.0027*** 0.2354*** -0.2351*** 0.0363*** 0.0825*** 0.0190*** -0.0246*** 1
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce…
State -0.0341*** -0.0771*** 0.1277*** -0.0058*** -0.0429*** -0.0205*** -0.0061*** -0.0429*** 1
Notes Data are collected from the section of the Small Businesses Administration (SBA)’s website that provides PPP loan data. *** p < 0.01, ** p < 0.05, * p < 0.1, respectively
1795
13
1796 S. Filomeni, F. Querci
minority-owned. The main variable of interest is the coefficient β 1 of the interaction term
of Minority * Post-PPP reforms, that captures whether the minority-owned small businesses
in the post-PPP reforms period are able to improve their access to PPP lending, increasing
the approved loan amount, relative to their non-minority owned counterparts. Figure 1 veri-
fies the parallel trend assumption by providing graphical evidence that PPP loan amounts
display a similar trend over time for the treated and the control groups and that this trend
changed after the reforms introduced by the Biden-Harris administration. Moreover, the
sign and statistical significance of the correlation coefficients for the treatment and control
groups, before and after the introduction of the Biden reforms to the Paycheck Protection
Program, at -0.1227*** and 0.0239*** respectively, further underscore the effectiveness of
these reforms in reducing racial bias.
Following Bose et al. (2021)4, our DID analysis for racial disparities in lending after the
reforms introduced by the Biden-Harris Administration to the PPP takes the following form:
PPP loan amounti,t = α + β0 (Post − PPP reforms)
+ β1 Minorityi + β2 Minorityi × (Post − PPP reforms)
z
(2)
+ δj Borrower’s characteristicsi,t + Business Type FE
j=1
+ Quarter FE + State FE + Industry FE + εi,t
Finally, to examine the equality of the Biden-Harris Administration’s reforms in reducing
racial disparity for all involving minority communities, we employ further a difference-in-
differences (DID) approach, similar to H2. This model uses a dataset primarily consisting
of PPP loans obtained by racial minority business owners, since such communities were
the primary targets of the new program changes. Similar to the previous methodology, we
invoke a difference-in-difference (DID) method by producing the treated and control groups
with respect to the two minority groups previously mentioned. A unique feature of this
model is the inclusion of a dummy variable for each racial minority group, enabling the
model to recognise whether a certain racial minority group achieved greater advantage in
relation to other minority-owned firms.
Corresponding to Bose et al. (2021), our DID approach to investigate the impartiality
of the Biden-Harris Administration reforms to the PPP on all racial minority groups can be
summarized as follows:
4
We follow Bose et al. (2021) to set up our difference-in-differences (DID) methodology given the suitability
of this method for our empirical analysis, even if the latter manuscript employs this methodology in a differ-
ent context to analyse the impact of a new bankruptcy law on the performance of distressed firms in India.
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1797
Table 4 Variance inflation factor Variable VIF 1/VIF
(VIF)
Minority 2.58 0.388116
PPP_dummy 3.35 0.298554
PPP_minority 3.69 0.271149
Area 1.23 0.811049
Hubzone 1.29 0.773448
LMI 1.32 0.756862
Notes Variance Inflation Factor
Firm size 1.08 0.923202
(VIF) for the regression model
generated by Eq. (2) Mean VIF 2.07
Fig. 1 Parallel trend assumption (DID). Notes Data are collected from the section of the Small Businesses
Administration (SBA)’s website that provides PPP loan data
13
1798 S. Filomeni, F. Querci
PPP loan amounti,t = α + β0 (Post − PPP reforms)
+ β1 Asiani + β2Asiani × (Post − PPP reforms)
+ β3 Blacki + β4Blacki × (Post − PPP reforms)
+ β5 Latinoi + β6Latinoi × (Post − PPP reforms)
+ β7 Nativei + β8Nativei × (Post − PPP reforms) (3)
z
+ δj Borrower’s characteristicsi,t
j=1
+ Business Type FE + Quarter FE
+ State FE + Industry FE + εi,t
In all our regressions, to isolate the effect of race on PPP loan amount, we control for a
large number of possible confounding factors related to the characteristics of the given bor-
rowing firm. We control for the lower to middle income debentures that the borrower could
accrue prior to the inception of PPP (i.e., LMI), to account for the fact that businesses with
prior established relationships with the SBA, through the potential use of other programs
which invested in low-to-middle income communities, received higher PPP loan amounts.
In addition, we control for the rural or urban indicator which dictates the area in which the
borrower operates (i.e., Area), to account for the fact that business owners, regardless of
their race, received higher amounts of PPP loans in urban rather than in rural areas. We also
control for the Hubzone indicator (i.e., Hubzone), that indicates whether the borrowing firm
operates within an area which has been historically under industrialised. Lastly, we control
for the size of the business proxied by the number of employees reported to be on the payroll
(i.e., Firm size), where the maximum number of employees for a business to be eligible for
a PPP loan is 500 and the minimum value is 1.
5 Results
First, in Table 5, we report the regression results for racial disparities in lending within the
initial design of the PPP under the initial Trump-Pence Administration, therefore prior to
the introduction of the reforms to the PPP by the subsequent Biden-Harris Administration.
Racial bias hypothesis testing is conducted empirically by investigating the sign and signifi-
cance of the coefficient β0 in Eq. (1), which reflects the minority group belonging status of
the small business owner. If β0 is not statistically significant, then the race/ethnicity of the
business owner would not influence the PPP loan amount received and racial disparity in the
PPP lending would be absent. In contrast, a positive and statistically significant coefficient
for Minority, i.e., β0 > 0, would indicate that belonging to a minority group is associated
with higher PPP loan amounts, while β0 ˂ 0 would indicate that belonging to a minority
group is associated with lower PPP loan amounts being granted. We find a negative and
statistically significant coefficient β0 which indicates that belonging to a minority group
has a negative effect on the PPP loan amount received. Therefore, our results indicate that
White business owners received significantly larger PPP loan amounts compared to minor-
ity business owners prior to the PPP reforms implemented by the Biden-Harris Administra-
tion. These findings are supported by Chernenko and Scharfstein (2024) who illustrated
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1799
Table 5 PPP lending and minor- Variables (1) (2) (3) (4)
ity groups
OLS OLS OLS OLS
Minority -0.381*** -0.222*** -0.339*** -0.178***
(0.003) (0.003) (0.003) (0.003)
Area -0.298*** -0.173***
(0.003) (0.003)
Hubzone 0.014*** 0.035***
(0.003) (0.003)
LMI 0.026*** 0.036***
(0.003) (0.003)
Firm size 0.029*** 0.024***
Notes In all models the (0.000) (0.000)
dependent variable is PPP
loan amount, a continuous Observations 903,224 903,224 903,224 903,224
variable equal to the amount R2 0.02 0.28 0.32 0.47
of granted forgivable credit, Quarterly FE NO YES NO YES
taken in logarithm. Robust Business Type FE NO YES NO YES
standard errors are reported
Industry FE NO YES NO YES
in parenthesis. *** p < 0.01, **
p < 0.05, * p < 0.1, respectively State FE NO YES NO YES
that minority business owners experienced severe difficulties to be granted PPP loans as
compared to their White-owned counterparts under the initial Trump-Pence Administration.
Secondly, in Table 6 we report the regression results for the effectiveness of the Biden-
Harris Administration’s reforms in reducing racial disparities in PPP lending. Interest-
ingly and in line with our expectations, our findings suggest that the reforms to the PPP
introduced by the Biden-Harris Administration significantly and positively moderate the
negative effect of Minority on PPP loan amount. This result is empirically achieved by
investigating the sign and the significance of coefficient β2 associated with the interaction
term Minority * Post-PPP reforms in Eq. (2). This effect is significant at the 1% level. The
latter results highlight the successfulness of the Biden-Harris Administration’s reforms in
reducing racial bias in PPP lending. Therefore, the reforms to the PPP implemented by the
Biden-Harris Administration expanded the small business base eligible for the PPP by suc-
cessfully tackling institutional discrimination within the distribution of financial resources,
in a scenario where minority individuals could now be encouraged to apply for PPP lending
without being deterred by the fear of institutional racism (Santellano, 2020). As a result,
after the implementation of the aforementioned reforms, small businesses owned by minor-
ity individuals have benefitted from an increased amount of forgivable PPP loans received.
All in all, these findings contribute to proving the effectiveness of government intervention
in resolving institutional discrimination that was a significant source of disparities in credit
access within the PPP.
Finally, in Table 7, we provide the regression findings to evaluate the impartiality of
the Biden-Harris Administration’s reforms to reduce racial disparities within PPP lending,
within the context of all racial minority groups involved. Our results suggest that the PPP
reforms introduced by the Biden-Harris Administration moderated the influence of Minor-
ity on PPP loan amount for the firm owned by almost every minority group. This emerges
from the careful observation of the signs and the statistical significance of the coefficients
of the interaction terms for the different minority groups. Moreover, the results in Table 5
show that Black American small business owners received higher PPP loan amounts after
13
1800 S. Filomeni, F. Querci
Table 6 PPP lending reforms and Variables (1) (2) (3) (4)
minority groups
DID DID DID DID
Minority -0.381*** -0.260*** -0.320*** -0.212***
(0.003) (0.003) (0.003) (0.003)
Post-PPP reforms -0.744*** -0.248*** -0.522*** -0.214***
(0.003) (0.003) (0.003) (0.003)
Minority * Post- 0.429*** 0.460*** 0.374*** 0.403***
PPP reforms
(0.004) (0.003) (0.003) (0.003)
Area -0.251*** -0.155***
(0.002) (0.002)
Hubzone 0.016*** 0.042***
(0.002) (0.002)
LMI 0.010*** 0.026***
(0.002) (0.002)
Firm size 0.031*** 0.026***
Notes In all models the (0.000) (0.000)
dependent variable is PPP
loan amount, a continuous Observations 1,759,211 1,759,211 1,759,211 1,759,211
variable equal to the amount R2 0.07 0.28 0.32 0.44
of granted forgivable credit, Quarterly FE NO YES NO YES
taken in logarithm. Robust Business Type FE NO YES NO YES
standard errors are reported
Industry FE NO YES NO YES
in parenthesis. *** p < 0.01, **
p < 0.05, * p < 0.1, respectively State FE NO YES NO YES
the enactment of PPP reforms in comparison to other racial minority groups. Our hypoth-
esis H3 is thus confirmed. Small businesses owned by the Asian minority group also took
advantage from the PPP reforms implemented by the Biden-Harris Administration, but to a
relatively lower extent. An opposite result is achieved for Native-owned small businesses
that obtained a lower loan amount after the PPP reforms. Since a much smaller percentage
of such firms operated in the industries hardest hit by the COVID-19 pandemic in compari-
son with other racial groups (Atkins 2022), it is reasonable to presume that they requested
less credit through the PPP, which could explain why their PPP loan amount decreased even
following the program reforms. The coefficient of the interaction term Latino * Post-PPP
reforms is not statistically significant, but the very small number of Latino businesses in
our sample (0.007%) limits the detection of the effect of PPP reforms on that subgroup with
statistical significance.
All in all, these results further support the effectiveness of the reforms employed by the
Biden-Harris Administration in reducing racial bias within the PPP. The widened demo-
graphics of business owners eligible for forgivable loans undoubtedly encouraged more
racial minority owners to apply for the program. Nevertheless, the results obtained for
Native-owned small businesses suggest that the PPP reforms have contributed to mitigate
only partially racial disparities existing among minority groups.
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1801
Table 7 Biden-Harris reforms and PPP lending by minority group
Variables (1) (2) (3) (4) (5) (6) (7) (8)
DID DID DID DID DID DID DID DID
Asian -0.387*** -0.316***
(0.004) (0.004)
Asian * 0.111*** 0.068***
Post-PPP
reforms
(0.006) (0.005)
Black -0.075*** -0.048***
(0.004) (0.003)
Black * 0.386*** 0.350***
Post-PPP
reforms
(0.004) (0.004)
Latino 0.106 0.135
(0.153) (0.142)
Latino * 0.236 0.226
Post-PPP
reforms
(0.183) (0.168)
Native -0.074*** -0.072***
(0.007) (0.006)
Native * -0.119*** -0.115***
Post-PPP
reforms
(0.010) (0.009)
Post-PPP -0.083*** -0.065*** -0.171*** -0.149*** -0.066*** -0.054*** -0.064*** -0.051***
reforms
(0.003) (0.002) (0.003) (0.003) (0.003) (0.002) (0.003) (0.002)
Area -0.173*** -0.128*** -0.160*** -0.160***
(0.002) (0.002) (0.002) (0.002)
Hubzone 0.046*** 0.032*** 0.049*** 0.049***
(0.002) (0.002) (0.002) (0.002)
LMI 0.029*** 0.009*** 0.028*** 0.028***
(0.002) (0.002) (0.002) (0.002)
Firm size 0.026*** 0.026*** 0.026*** 0.026***
(0.000) (0.000) (0.000) (0.000)
Observations 1,759,211 1,759,211 1,759,211 1,759,211 1,759,211 1,759,211 1,759,211 1,759,211
R2 0.28 0.44 0.28 0.44 0.27 0.43 0.27 0.43
Quarterly YES YES YES YES YES YES YES YES
FE
Business YES YES YES YES YES YES YES YES
Type FE
Industry FE YES YES YES YES YES YES YES YES
State FE YES YES YES YES YES YES YES YES
Notes In all models the dependent variable is PPP loan amount, a continuous variable equal to the amount
of granted forgivable credit, taken in logarithm. Robust standard errors are reported in parenthesis. ***
p < 0.01, ** p < 0.05, * p < 0.1, respectively
13
1802 S. Filomeni, F. Querci
6 Robustness Tests
To draw stronger causal inference between the enactment of the PPP reforms by the Biden-
Harris Administration and the reduction of racial disparity in PPP lending, we carried out a
series of robustness checks on the DID estimation corresponding to Eq. 2.
First, we used an alternative dependent variable, that is Forgiven amount (i.e., the amount
of the PPP loan that did not have to be paid back) instead of PPP loan amount, taken in loga-
rithm. According to the Paycheck Protection Program (PPP), borrowers may be eligible for
loan forgiveness if the funds were used for eligible payroll costs, business mortgage interest
payments, rent, or utilities during either the 8- or 24-week period following disbursement.
Since for most borrowers, the loan amount forgiven proxies the initial principal amount
granted, the variable “forgiven amount” can be likely used as an appropriate substitute for
“PPP loan amount” for robustness purposes. The results are stable (Table 8), confirming the
findings of our previous analysis.
Second, considering that our models corresponding to Eq. 2 include many levels of fixed
effects (for business type, industry, State, and quarterly time), we ran the DID regressions
using the estimator of Correia (2016), which is an iterative process that can deal with mul-
tiple high dimensional fixed effects5. The results shown in Table 9 confirm the robustness of
our main empirical findings.
Table 8 Robustness: forgiven amount as alternative dependent variable
Variables (1) (2) (3) (4) (5) (6)
DID DID DID DID DID DID
Minority -0.324*** -0.054*** -0.378*** -0.287*** -0.337*** -0.240***
(0.002) (0.002) (0.003) (0.003) (0.003) (0.003)
Post-PPP reforms -0.788*** -0.258*** -0.542*** -0.226***
(0.003) (0.003) (0.003) (0.003)
Minority * Post-PPP reforms 0.460*** 0.502*** 0.394*** 0.441***
(0.004) (0.004) (0.004) (0.004)
Area -0.263*** -0.162***
(0.002) (0.002)
Hubzone 0.017*** 0.038***
(0.002) (0.002)
LMI 0.023*** 0.030***
(0.002) (0.002)
Firm size 0.030*** 0.026***
(0.000) (0.000)
Observations 1,362,835 1,362,835 1,362,835 1,362,835 1,362,835 1,362,835
R2 0.01 0.28 0.07 0.29 0.33 0.45
Quarterly FE NO YES NO YES NO YES
Business Type FE NO YES NO YES NO YES
Industry FE NO YES NO YES NO YES
State FE NO YES NO YES NO YES
Notes In all the models the dependent variable is Forgiven amount, a continuous variable equal to the
amount of the PPP loan that did not have to be paid back, taken in logarithm. Robust standard errors are
reported in parenthesis. *** p < 0.01, ** p < 0.05, * p < 0.1, respectively
5
Correia (2016)’s estimator addresses many shortcomings of existing estimators, which had slow con-
vergence properties, especially with large and complex datasets like ours. Therefore, it provides a suitable
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1803
Table 9 Robustness: estimator of Correia (2016)
Variables (1) (2) (3) (4) (5) (6)
DID DID DID DID DID DID
Minority -0.312*** -0.312*** -0.381*** -0.381***-0.320*** -0.212***
(0.002) (0.002) (0.003) (0.003) (0.002) (0.002)
Post-PPP reforms -0.744*** -0.744***-0.522*** -0.214***
(0.002) (0.002) (0.002) (0.003)
Minority * Post-PPP reforms 0.429*** 0.429*** 0.374*** 0.403***
(0.004) (0.004) (0.003) (0.003)
Area -0.251*** -0.155***
(0.002) (0.002)
Hubzone 0.016*** 0.042***
(0.002) (0.002)
LMI 0.010*** 0.026***
(0.002) (0.002)
Firm size 0.031*** 0.026***
(0.000) (0.000)
Observations 1,759,210 1,759,210 1,759,210 1,759,210 1,759,210 1,759,210
R2 0.01 0.01 0.07 0.07 0.32 0.44
Quarterly FE NO YES NO YES NO YES
Business Type FE NO YES NO YES NO YES
Industry FE NO YES NO YES NO YES
State FE NO YES NO YES NO YES
Notes In all models the dependent variable is PPP loan amount, a continuous variable equal to the amount
of granted forgivable credit, taken in logarithm. Robust standard errors are reported in parenthesis. ***
p < 0.01, ** p < 0.05, * p < 0.1, respectively
To further corroborate our empirical findings, we implement propensity score matching
(PSM) method according to the nearest neighbour matching approach with a caliper value
of 0.1. According to the latter method, matched samples are chosen where minority-owned
businesses (treated group) are matched to the nearest neighbour of non-minority-owned
businesses (control group) using all variables included in our baseline regression of Eq. (2).
Our PSM results, reported in Table 10, corroborate our main findings that minority business
owners received larger loan amounts after changes to the Paycheck Protection Program
were implemented by the Biden-Harris Administration. Specifically, Column (1) in Table 10
re-runs our baseline model generated by Eq. (2) using the restricted sample selected through
propensity score matching, while Column (2) re-estimates the causal effect by including
propensity scores as weights in our baseline model for further robustness. Overall, our PSM
robustness results suggest that the difference in PPP lending volume observed by minority-
owned businesses in the post-PPP period is not driven by existing differences in the two
groups, further corroborating our novel evidence that minority business owners received
larger loan amounts after changes to the Paycheck Protection Program were implemented
by the Biden-Harris Administration. Therefore, our main results prove to be robust also to
this matching technique.
Lastly, following Balyuk et al. (2021), we now re-run our baseline model specification
generated by Eq. (2) by incorporating several other control variables related to firm and loan
characteristics that may exert an influence on the amount of PPP loans. Specifically, our
method to assess the robustness of the estimation results in the context of our empirical analysis.
13
1804 S. Filomeni, F. Querci
Table 10 Robustness: propensity Variables (1) (2)
score matching (PSM)
DID DID
Minority -0.114*** -0.174***
(0.007) (0.003)
Post-PPP reforms -0.158*** -0.187***
(0.005) (0.004)
Minority * Post-PPP reforms 0.373*** 0.359***
Notes The Table reports (0.008) (0.004)
propensity score matching Area -0.174*** -0.143***
(PSM) method according to the (0.002) (0.002)
nearest neighbour matching Hubzone -0.004 0.016***
approach with a caliper value of
0.1 for our baseline regression (0.003) (0.002)
model generated by Eq. (2), LMI 0.032*** 0.036***
using all the variables for the (0.004) (0.003)
matching. Specifically, Column Firm size 0.028*** 0.030***
(1) re-runs our baseline model (0.000) (0.000)
of Eq. (2) using the restricted
sample selected through Observations 411,473 1,753,248
propensity score matching R-squared 0.44 0.45
while Column (2) re-estimates Quarterly FE YES YES
the causal effect by including Business Type FE YES YES
propensity scores as weights in
Industry FE YES YES
our baseline model specification
as a further robustness test State FE YES YES
baseline model now controls for the maturity of the given PPP loan (i.e., Loan Maturity), for
the PPP loan status (i.e., Loan Status), for the firm’s age (i.e., Firm Age), for the specific US
district in which the borrowing firm operates (i.e., Firm District), and for the firm’s payroll
and utility expenses (i.e., Payroll Expenses and Utility Expenses, respectively). Overall, the
new estimation results, reported in Table 11, further corroborate the evidence that minority
business owners received larger loan amounts after the Biden-Harris Administration imple-
mented changes to the Paycheck Protection Program.
7 Conclusions
The introduction of the PPP has been a key factor in providing short term aid for small busi-
nesses around the U.S., during the height of the Coronavirus pandemic. The implementation
of new reforms, enacted by the Biden-Harris administration towards the PPP, resulted in the
reduction of racial disparities within its loan allocation process.
Within this paper, we observed the implication of PPP reforms with regards to its effects
on the ability of minority business owners to achieve larger loan amounts compared to when
changes were absent. We utilized a granular dataset of 1,759,270 loans across Q2 2020 – Q2
2021 which confirmed our expectation that the introduction of the reforms would reduce
racial disparities in PPP lending. Overall, our paper provides a novel contribution to exist-
ing literature by demonstrating that minority business owners received larger loan amounts
after changes were implemented by the Biden-Harris Administration. This also indicates
the effectiveness of government intervention with regards to reforming racial bias within
the lending industry. Reforms enacted by the Biden-Harris Administration opened the PPP
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1805
Table 11 Robustness: additional Variables (1) (2) (3) (4)
firm- and loan-specific control
DID DID DID DID
variables
Minority -0.197*** -0.176*** -0.218*** -0.177***
(0.003) (0.003) (0.005) (0.005)
Post-PPP -0.308*** -0.190*** -0.380*** -0.190***
reforms
(0.003) (0.003) (0.005) (0.004)
Minority 0.253*** 0.314*** 0.255*** 0.282***
* Post-PPP
reforms
(0.003) (0.003) (0.005) (0.005)
Area -0.249*** -0.113*** -0.205*** -0.106***
(0.002) (0.002) (0.003) (0.003)
Hubzone 0.022*** 0.038*** 0.012*** 0.030***
(0.002) (0.002) (0.002) (0.002)
LMI 0.012*** 0.016*** -0.010*** 0.003
(0.002) (0.002) (0.002) (0.002)
Firm size 0.015*** 0.011*** 0.012*** 0.008***
(0.000) (0.000) (0.001) (0.001)
Loan maturity -0.010*** -0.012*** -0.010*** -0.014***
(0.000) (0.000) (0.000) (0.000)
Firm age -0.105*** -0.028*** -0.133*** -0.088***
(0.002) (0.001) (0.006) (0.006)
Loan status 0.031*** -0.029*** 0.048*** -0.034***
(0.002) (0.002) (0.002) (0.002)
Payroll 0.000*** 0.000*** 0.000*** 0.000***
Expenses
(0.000) (0.000) (0.000) (0.000)
Utility -0.000*** -0.000***
Expenses
(0.000) (0.000)
Observations 1,757,436 1,757,436 795,687 795,687
Notes In all models the
dependent variable is PPP R2 0.37 0.48 0.40 0.50
loan amount, a continuous Quarterly FE NO YES NO YES
variable equal to the amount Business Type NO YES NO YES
of granted forgivable credit, FE
taken in logarithm. Robust Industry FE NO YES NO YES
standard errors are reported
State FE NO YES NO YES
in parenthesis. *** p < 0.01, **
p < 0.05, * p < 0.1, respectively District FE NO YES NO YES
to more underserved small businesses than ever before, which in turn expanded the reach of
the PPP to more small business owners belonging to racial minority groups. Additionally,
we find that business owners belonging to the Black American minority group, as well as
Latin American and Asian business owners benefited the most, in terms of access to PPP
loans, from the introduction of the Biden-Harris reforms.
Our findings are beneficial to policy makers and lawmakers as we provide evidence that
drastic reforms implemented in credit programs could address the ongoing issue of insti-
tutional discrimination. We also show that when governments administer credit guarantee
13
1806 S. Filomeni, F. Querci
programs accounting for pre-existing inequities in those systems, specific reforms can be
enacted to avoid reproducing the same inequities when rebooting those programs.
Moreover, our analysis shows how an adequate design of a public credit guarantee pro-
gram can provide a substantial contribution to the reduction of racial disparities in the distri-
bution of financial resources, which in turn can favour the economic inclusion of all, moving
forward in the accomplishment of the UN’s Sustainable Development Goal 10 (‘Reduce
inequality within and among countries’).
Overall, this paper provides a glimpse into an ever-growing issue of institutional bias
within the finance industry and paves the way for further research into the topic.
Acknowledgements We are grateful to the Editor Cheng-Few Lee and to two anonymous referees for their
valuable insights and suggestions.
Declarations
Competing interests The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as
you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons
licence, and indicate if changes were made. The images or other third party material in this article are
included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material.
If material is not included in the article’s Creative Commons licence and your intended use is not permitted
by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
References
Asiedu E, James AF, Akwasi N (2012) Access to credit by small businesses: how relevant Are race, ethnicity,
and gender? Am Econ Rev, 102(3)
Atkins R (2022) A review of federal lending to small business owners of color during the COVID-19 pan-
demic, https://equitablegrowth.org/research-paper
Atkins R, Cook L, Seamans R (2022) Discrimination in lending? Evidence from the Paycheck Protection
Program. Small Bus Econ, 58
Balyuk T, Prabhala N, Puri M (2021) Small Bank Financing and Funding Hesitancy in a Crisis: Evidence
from the Paycheck Protection Program, Federal Deposit Insurance Cooperation, Center for Financial
Research. Available at SSRN: https://ssrn.com/abstract=3717259
Bartik A, Cullen Z, Glaeser E, Luca M, Stanton C, Sunderam A (2021) The Targeting And Impact Of Pay-
check Protection Program Loans To Small Businesses, National Bureau of Economic Research Working
paper No. 27623.
Bates T, Robb A (2015a) Has the community reinvestment act increased loan availability among small busi-
nesses operating in minority neighbourhoods? Urban Stud, 52(9)
Bates T, Robb A (2015b) Impacts of owner race and geographic context on access to small business financing.
Econ Dev Q, 30(2)
Blanchard L, Zhao B, Yinger J (2008) Do lenders discriminate against minority and woman entrepreneurs?
J Urban Econ, 63(2)
Blanchflower DG, Levine PB, Zimmerman DJ (2003) Discrimination in the small-business credit market.
Rev Econ Stat, 85
Bose U, Filomeni S, Mallick S (2021) Does bankruptcy law improve the fate of distressed firms? The role of
credit channels. J Corp Finance 68(2021):101826
Bostic RW, Lampani KP (1999) Racial differences in patterns of small-business finance: The importance of
local geography, Proceedings, Federal Reserve Bank of Chicago
13
Did Biden-Harris’s reforms on the Paycheck Protection Program reduce… 1807
Carmichael S, Hamilton CV (1967) Black power: the politics of liberation in America. Vintage, New York
Cavalluzzo KS, Wolken JD (2002) Small business loan turndowns, personal wealth and discrimination,
Finance and Economics Discussion Series No. 35, Board of Governors of the Federal Reserve System,
Washington, DC
Cavalluzzo KS, Cavalluzzo LC, Wolken JD (2002) Competition, small business financing and discrimina-
tion: evidence from a new survey. J Bus, 75
Chernenko S, Scharfstein D (2024) Racial disparities in the Paycheck Protection Program. J Financ Econ,
160
Chernenko S, Kaplan N, Sarkar A, Scharfstein D (2023) Applications or Approvals: What Drives Racial Dis-
parities in the Paycheck Protection Program? National Bureau of Economic Research Working paper
No. 31172
Correia S (2016) A feasible estimator for linear models with multi-way fixed effects. Preprint at http://scor-
reia.com/research/hdfe.pdf
Fairlie R (2020) The Impact of Covid-19 On Small Business Owners: Evidence Of Early-stage Losses From
The April 2020 Current Population Survey, National Bureau of Economic Research Working paper No.
27309.
Fibbi R, Midtbøen AH, Simon P (2021) Migration and discrimination. IMISCOE Research Series. Springer
Figenschou TU, Karlsen R, Kolltveit K, Schillemans T (2019) Mounting media pressure: push and pull
forces influencing agendas, resource allocation and decision-making in public bureaucracies. Eur J
Communication, 34(4)
Filomeni S (2023) The impact of the Paycheck Protection Program on the risk-taking behaviour of US banks.
Rev Quant Financ Acc, 62(4)
Guimond S (2023) Individual and institutional discrimination: theoretical and methodological contributions
of social psychology. Appartenances et Altérités, 3
Howell S, Kuchler T, Snitkof D, Wong J (2022) Automation in Small Business Lending Can Reduce Racial
Disparities: Evidence from the Paycheck Protection Program, National Bureau of Economic Research
Working paper No. 297364.
Howell S, Kuchler T, Snitkof D, Stroebel J, Wong J (2024) Lender automation and racial disparities in Credit
Access. J Finance, 89(2)
Kelly M, van Holm E (2021) Urbanization, community demographics, and the distribution of Paycheck Pro-
tection Program Loans Across counties. SSRN Electron J. https://doi.org/10.2139/ssrn.3801348
Kickul J, Griffiths M, Robb C, Gundry L (2021) All for one? The Paycheck Protection Program distribution
disparity. J Entrepreneurship Public Policy, 10(3)
Kotomin V, Frere W, Morr R (2024) Forgive me not? Racial and institutional disparities in the Paycheck
Protection Program loan forgiveness. Small Bus Econ. https://doi.org/10.1007/s11187-024-00952-w
Lederer A, Oros S (2020) Lending discrimination within the Paycheck Protection Program. National Com-
munity Reinvestment Coalition
Lester W, Wilson M (2023) The racial and spatial impacts of the Paycheck Protection Program. Econ Dev
Q, 37(3)
Lelo de Larrea G, Altin M, Koçak E, Okumus F (2024) Paycheck Protection Program and lending discrimina-
tion in the US hospitality industry. Tour Econ. https://doi.org/10.1177/13548166241232715
Mijid N, Bernasek A (2013) Decomposing racial and ethnic differences in Small Business Lending: evidence
of discrimination. Rev Soc Econ, 71(4)
Mills C, Battisto J (2020) Double Jeopardy: Covid-19’s Concentrated Health and Wealth Effects In Black
Communities, Federal Reserve Bank of New York Research Paper Series.
Misera L (2020) An Uphill Battle: COVID-19’s outsized toll on minority-owned firms. Fed Reserve Bank
Clevel Res Paper Ser
Morrill C (2008) Culture and Organization Theory. Annals Am Acad Political Social Sci, 619(1)
O’Rourk S, Bobrosky J (2021) SBA revises PPP loan amount calculation for filers of schedule C, www.
cohnreznick.com
Office of Advocacy (2018) A Look at Nonemployer Businesses, www.sba.gov/for-lenders
Pager D, Shepherd H (2008) The sociology of discrimination: racial discrimination in employment, housing,
credit, and consumer markets. Ann Rev Sociol, 34
Rakshit A, Peterson J (2022) Racial disparities in the small business loan market. J Small Bus Manage, 59(6)
Sant’Anna AC, Kim KN, Demko I (2023) Limits to capital: assessing the role of race on the Paycheck Protec-
tion Program for African American farmers in America. Appl Economic Perspect Policy, 46(1)
Santellano K (2021) Compounded inequality: how the U.S. Paycheck Protection Program is failing Los
Angeles Latino small businesses, Ethnic and Racial Studies, 44(5)
SBA (2020) PPP loan forgiveness, www.sba.gov/funding-program
SBA (2014) E-Tran: Electronic Loan Processing, www.sba.gov/for-lenders
Schweitzer M, Guo A (2024) Basic facts on the coverage of the paycheck protection program. Bus Econ, 59
13
1808 S. Filomeni, F. Querci
Staples AJ, Krumel TP (2023) The Paycheck Protection Program and small business performance: evidence
from craft breweries. Small Bus Econ, 6
The White House (2021) FACT SHEET: Biden-Harris Administration increases Lending to small businesses
in need, announces changes to PPP to further promote Equitable Access to Relief. The White House
Tresch A, Sciarini P, Varone F (2013) The relationship between media and political agendas: variations across
decision-making phases. West Eur Politics, 36(5)
U.S. Bureau of Labor Statistics (2020) U.S. Unemployment Rate, Federal Reserve Economic Data.
U.S. Bureau of Labor Statistics (2020) U.S. GDP, Federal Reserve Economic Data.
Wang J, Zhang D (2021) The Cost of Banking Deserts: Racial Disparities in Access to PPP Lenders and their
Implications, Department of Economics, Harvard University.
Watts RJ, Griffith DM, Abdul-Adil J (1999) Sociopolitical development as an antidote for oppression. Am J
Community Psychol, 27(2)
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Authors and Affiliations
Stefano Filomeni1 · Francesca Querci2
Stefano Filomeni
stefano.filomeni@essex.ac.uk
Francesca Querci
francesca.querci@unige.it
1
University of Essex, Essex Business School, Finance Group, Colchester, UK
2
Department of Economics and Business Studies, University of Genoa, Genoa, Italy
13