MARK GRIFFITHS, Principal, MG2 Investments LLC
MIKE MAULDIN, Director of the Excellence In Banking Program, Texas Tech University
DREW WINTERS, Pickering Chair in Finance, Texas Tech University

Phase 1 of the Paycheck Protection Program (PPP) has been completed with roughly $349 billion distributed to businesses across the country. The program was established as part of the $3 trillion coronavirus relief bill passed in March 2020. A second round of an additional $310 billion commenced in April while a third round is currently under consideration. PPP is administered by the Small Business Administration (SBA), which is a division of the U.S. Department of the Treasury. The initial idea of the program was that any business (for-profit or nonprofit, veterans organization or tribal concern) with 500 or fewer employees is eligible for a government-backed loan equal to eight weeks of its prior average payroll, plus an additional 25 percent of that sum to cover various other non-payroll operating expenses such as interest rent and utilities. The payroll amount was capped at a maximum of $100,000 (annualized) for each employee unless the total amounts to more than $10 million, which is the maximum for any individual firm. The loan was structured as having a two-year term and a 1 percent interest rate ^{[ 1 ]}.

Further, the loans were actually more like grants in that the loans would be forgiven provided the business did not decrease its full-time employee headcount or decrease salaries and wages by more than 25 percent for any employee who made less than the annualized $100,000 and if the business restored the full-time employment and salary levels by June 30, 2020. The general argument of various government officials was to support the small-business sector and to immunize it from insolvency until the coronavirus pandemic passed.

Demand greatly exceeded supply, to the point where the program ran out of money in 13 days. ^{[ 2 ]} As we are preparing this report, phase 2 of the PPP has just completed with some modifications. One reason for the modifications is that some large successful companies—e.g., Shake Shack, Ruth’s Chris Steakhouse, Potbelly, Sweetgreen, Axios and the LA Lakers^{[ 3 ]}—received funds from phase 1 of PPP. To respond to these problems, the House Select Subcommittee on the Coronavirus Crisis launched an investigation into the implementation of the PPP in June to determine whether larger companies were receiving loans intended for small businesses. ^{[ 4 ]} The report goes on to state:

“Although PPP has enabled many small businesses to weather the pandemic, the program would be more effective if Treasury and SBA implemented it consistent with congressional intent. In the CARES Act, Congress specifically encouraged the administration to issue guidance ‘to ensure that the processing and disbursement of covered loans prioritize small-business concerns and entities in underserved and rural markets,’ including businesses owned by veterans, members of the military, socially and economically disadvantaged individuals, and women. ^{[ 5 ]} On May 8, SBA’s inspector general reported, ‘We did not find any evidence that SBA issued guidance to lenders to prioritize the markets indicated in the Act.’ ^{[ 6 ]} The inspector general also found that SBA failed to provide a demographic questionnaire with the PPP loan application, undermining the agency’s ability to determine whether lenders appropriately prioritized loans to underserved communities.” ^{[ 7 ]}

Accordingly, we ask the question: Did PPP funds get to small businesses as intended? Unfortunately, the data do not exist to answer this question directly, so we take an indirect approach and examine the participation of community banks in the distribution of the first round of the PPP funds. We chose to concentrate on community banks because of their commitment to reinvest local dollars back into the community and help create local jobs. For reasons of tractability (discussed below) we also decided to concentrate on the state of Texas, both due to its size and its large number of community banks. The relationship banking philosophy of community banks is designed to help small businesses grow and help families finance major purchases and build financial security. This aligns directly with the stated goal of the Paycheck Protection Program.

The Coronavirus Aid, Relief and Economic Security (CARES) Act provided financial relief for consumers, nonprofits and businesses alike through loan deferrals, PPP and economic injury disaster loans (EIDL). For the most part, the banks in our nation were called upon for quick delivery of this relief. Even though the EIDL was a direct program administered by the SBA, the majority of businesses reached out to their local banker for advice and assistance.

Prior to the funds being available for the first round of PPP, the majority of the banks completed their loan-extension process. The banks then answered the government call to begin lending under the PPP, even though the U.S. Treasury was still in the PPP rulemaking process. ^{[ 8 ]} Bankers were asked to move forward with uncertainty, upon a promise from the government that they would have a good-faith safe harbor. Since this pandemic began, bankers have been working, doing what they have done for years, serving their customers and community. We looked at a slice of the banking efforts during this crisis. Specifically, we examined the initial distribution of PPP funds for the state of Texas with a focus on community bank participation.

Data Description

We began with the list of Texas banks participating in the first round of PPP. We merged this list with an FDIC list of community banks to identify Texas community banks participating in the program. Then we hand-coded the other Texas first-round PPP lenders into the following categories: stress-tested banks ^{[ 9 ]}, credit unions, non-bank lenders and other banks. The category of “other” banks are the Texas PPP lenders that are not identified for our specific categories.

This dataset comprises 147,461 loans after deleting any loan in which the size was less than $5,000. ^{[ 10 ]} The data provided for the borrower includes city, state, zip code, NAICS code, business type (e.g., LLC, corporation, etc.), veteran status, race/ethnicity, gender, profit/not-for-profit status, jobs saved and congressional district. For loans under $150,000 (more than 75 percent of all the loans) the specific amount is provided, but the borrower name is omitted. For loans of $150,000 and larger, the specific amount is omitted (ranges are provided), but the borrower name is listed. Finally, the lender’s name is provided with each PPP loan.

The SBA does not make any representations about the accuracy or completeness of any information that borrowers provided to their lenders. Not all borrowers provided all information. For example, approximately 75 percent of all PPP loans did not include any demographic information because that information was not provided by the borrowers or recorded by the lenders.

Results Related to Lenders

In April, the National Association of Small Business (NSBA) conducted a survey of more than 980 small-business owners on how the coronavirus was impacting their business. Small-business owners are concerned about the pandemic’s effect on their businesses, with 77 percent stating that they are “very concerned,” and 47 percent anticipate the largest impact to their business will be economic related.^{[ 11 ]} The same survey reports 49 percent of small-business owners have been impacted by reduced customer demand and 33 percent by delays or closures in the supply chain. The tourism industry took the worst hit, with sightseeing transportation facing a 62.1 percent decline followed by amusement parks and casinos losing 59.9 percent of their jobs (Pietsch, 2020). However, the types of small businesses that are believed to be particularly vulnerable and at the highest risk of closing are hotels, food services, educational services, mining, and oil and gas.

Access to capital is a crucial need for small businesses, with bank lending determined to be the most significant source of external funding (Berger & Udell, 2002)^{[ 12 ]}. To investigate these early results, we commenced by examining which banks were lending in PPP¹, where the borrowers were and in which industries the borrowers operated. In the next section, we address the issue of access to capital for women and ethnic minorities.

Tables 1A and 1B provide a breakdown of the number of loans made by the different lender types. In both the under- and over-$150,000 loan amounts, the greatest number and amount of loans in Texas are made by community banks. In the under-$150,000 category, Texas community banks made 66,192 loans (69.4 percent of all loans), totaling $3,259,617,655. In the over-$150,000 category, these same banks made 28,195 loans (54 percent of all loans). Recall that we do not have specific loan amounts for this category, so we cannot make any statement on dollar amounts.

Table 1A: Loans Under $150,000 by Lender

Table 1B: Loans Over $150,000 by Lender

We used the following list of stress-tested major banks: JP Morgan Chase & Co., Citigroup, Bank of America Corp., Wells Fargo & Co., Goldman Sachs Group, Morgan Stanley, PNC Financial Services Group, US Bancorp, Bank of NY Mellon Corp., SunTrust Banks Inc., State Street Corp., Capital One Financial Corp., BB&T Corp., Regions Financial Corp., American Express Co., Fifth Third Bancorp and Keycorp. We define other bank types as non-stress-tested banks that are not defined as community banks.

Next, we provide an analysis of the major lenders, with a breakdown of those with 1,000 or more initial-round PPP loans. Table 2A provides the list of lenders for loans under $150,000 and Table 2B provides the list of lenders for loans above $150,000. The under-$150,000 category contains 15 lenders. The largest lender is Frost Bank (categorized as an “other” bank), which made 6,698 loans. Frost Bank is followed by eight community banks, one non-bank lender, one stress-tested bank^{[ 13 ]} and four additional “other” banks.

Table 2A: Lenders with 1,000 or More Loans Under $150,000

Table 2B: Lenders with 1,000 or More Loans Over $150,000

We used the following list of stress-tested major banks: JP Morgan Chase & Co., Citigroup, Bank of America Corp., Wells Fargo & Co., Goldman Sachs Group, Morgan Stanley, PNC Financial Services Group, US Bancorp, Bank of NY Mellon Corp., SunTrust Banks Inc., State Street Corp., Capital One Financial Corp., BB&T Corp., Regions Financial Corp., American Express Co., Fifth Third Bancorp and Keycorp. We define other bank types as non-stress-tested banks that are not defined as community banks.

In the over-$150,000 loan categories, Frost Bank is again the leading lender with 4,234 loans, followed by two stress-tested banks, two community banks and two “other” banks, all having made more than 1,000 loans.^{[ 14 ]}

Our analysis to this point shows that community banks are significant participants in the initial PPP loans in Texas. However, Texas is a large and geographically diverse state, and Congress wanted PPP to serve rural markets. So, next we examine whether the community bank lending covered the entire state or was localized in urban settings. Figure 1 depicts the distribution of participating community banks across the state of Texas. Zip codes with a community bank loan are shaded orange—almost the entire state is shaded orange. Figure 2 depicts the areas with first-round PPP loans that are not served by participating community banks. These zip codes are shaded blue—and very little of the state is shaded blue. These maps suggest that community banks enabled PPP loans to reach rural markets in Texas.

We can draw zip code maps for each lender, but with several hundred lenders it is not feasible to provide such maps here. We did draw lender maps for Frost Bank, Allegiance Bank and First Financial Bank to determine if these large PPP lenders covered the state. The answer is no, and with this, it became apparent to us that to reach rural Texas, a large number of community banks is required.^{[ 15 ]}

The purpose of the PPP program was to assist small businesses by targeting any business with 500 or fewer employees. The government-backed loan was equal to eight weeks of the business’ prior average payroll, plus an additional 25 percent of that sum to cover various other non-payroll operating expenses, such as interest rent and utilities. We pose the question: How many people were helped by this program? One of the fields (not mandatory) recorded in the SBA database was the number of jobs retained.

In the under $150,000 category (Table 3A) $4,725,437,983 in loans were made with a view to saving 731,538 jobs—an average of $6,460 per position. With the program being eight weeks long, the per-job average equates to $808 per week per job, which can equate to as little as $646 per job per week with the allowance for up to 25 percent of the loan for overhead.^{[ 16 ]} Community banks were responsible for distributing roughly 69 percent of the funds to save 66 percent of the jobs. In the over-$150,000 categories (Table 3B), community banks distributed funds with a view to saving 1,382,421 jobs—roughly 50 percent of all of the positions retained in the state through this program. Recall that the over-$150,000 categories do not report the exact amount of the loans, so we cannot calculate loan per job saved.

Table 3A: Jobs Retained with Loans Under $150,000 by Lender

 Table 3B: Jobs Retained with Loans Over $150,000 by Lender

Since the database records the zip code of the borrower, we were able to analyze the amount of the loans and the number of jobs saved by sub-regions of Texas. We report these results for the under-$150,000 category in Table 3C. In Table 3D, we present the numbers of loans and the jobs retained in the sub-regions of Texas for the loans in excess of $150,000. In both cases, the greatest number of jobs saved were in the urban centers of Houston and Dallas.

Table 3C: Jobs Retained with Loans Under $150,000 by Lender

Table 3D: Jobs Retained with Loans Over $150,000 by Lender

Our next concern centered on which industries were served by the community banks. That is, did community banks concentrated on only a relatively small number of industries and did the community banks participate at a lower rate than other financial institutions? We present this information in Tables 3E and 3F. We perform this analysis by examining the NAICS codes assigned to the borrowers.^{[ 17 ]} In the under-$150,000 loan category, community banks lend on a comparable basis to other banks and stress-tested banks in all categories except professional and technical services, and health care and social assistance, in which they provide roughly 50 percent fewer loan in these code groups relative to other banks and stress-tested banks. In the over-$150,000 category, community banks are again comparable in all but two code groups. They did not make as many loans to professional and technical services, but they led the way in lending to the construction trades.

Table 3E: PPP Loans Under $150,000 for Texas Industry frequency (5 percent for more or loans for community banks)

 Table 3E: PPP Loans Over $150,000 for Texas Industry frequency (5 percent for more or loans for community banks)

It could be argued that these analyses are too broad in that two-digit codes are very high level and thus could conceal considerable micro-level information. To address this issue, we examined the full six-digit codes for loans under $150,000 and present the data in Table 3G. As can be seen, while community banks do not participate in some industries, the results are generally consistent with the higher-level data that showed lower participation in the professional and technical code categories.

The National Association of Small Business survey indicates that the types of small businesses that are believed to be particularly vulnerable and at the highest risk of closing are hotels, food services, educational services, mining, and oil and gas. Table 3G shows that community banks were active lenders in all of these categories except educational services. NAICS code sector 61 covers educational services and all of the categories of bank lenders show very limited initial PPP activity in this NAICS sector. Interestingly, community banks in Texas have significant PPP lending to religious organizations (NAICS = 813110).

Table 3G: PPP Loans Under $150,000 for Texas

Six digits NAICS codes (more than 1 percent, by bank type)

Results Related to Gender and Race

Congress specifically encouraged the administration to issue guidance “to ensure that the processing and disbursement of covered loans prioritize small-business concerns and entities in underserved and rural markets,” including businesses owned by veterans, members of the military, socially and economically disadvantaged individuals, and women. Shortly after all the funds for the initial PPP loans were distributed, the media began reporting discrimination against minority businesses in the distribution of the funds. Accordingly, this last section of our analysis examines the data from PPP in Texas for women and minority-owned businesses.

The initial PPP loan data are for loans granted. This is an incomplete measure of the process as it only provides a measure of the outcome. There are no data on initial PPP requests and denials, which are the data needed to provide an informed analysis of discrimination. In addition, the borrower is not required to report demographic data and the lender is not required to collect it. The result is that the demographic data on the borrowers is mostly missing, so the available data cannot be considered a representative sample of all the initial PPP borrowers. Thus, the reader should not draw any conclusions from these data and we only present the analysis to provide an informed response to media reports of discrimination.

Research on the relationship between gender and capital access^{[ 18 ]} finds limited access for women-owned and minority-owned businesses.^{[ 19 ]} It is believed that this contributes to reduced opportunities for business growth.^{[ 20 ]} Women-owned firms in general are less likely to rely on bank financing for their businesses,^{[ 21 ]} perhaps due to perceptions of potential discrimination in the lending process.^{[ 22 ]} In the study by the Women’s Business Enterprise National Council (WBENC, 2018), 25 percent of women business owners sought business financing, compared to 33 percent of male business owners. However, Shepherd (2020)^{[ 23 ]} notes that 57.4 percent of the SBA micro-loan program’s loans went to women-owned or women-led businesses. Women entrepreneurs apply for approximately $35,000 less in business financing than male entrepreneurs, and men receive an average loan size of $43,916, while women receive an average loan size of $38,942.^{[ 24 ]}

Our initial findings reveal substantial differences between the amount and number of loans made to men and women. In the under-$150,000 category (Table 4A), women received 23.26 percent of the loans and men received 76.74 percent. The median initial PPP loan to women-owned businesses is about $36,000 to save a median of six jobs ($6,000 per job), while the median loan size for male-owned businesses is about $41,000 to save a median of six jobs ($6,833). We note that slightly more than 65 percent of the time the gender of the borrower was not indicated. In the over-$150,000 categories (Table 4B), women received 17.97 percent of the loans and men received 82.03 percent of the loans. These large initial PPP loans saved a median of 31 jobs for businesses owned by both genders. Again, caution is necessary in interpreting these statistics as in approximately 71 percent of the cases gender was not identified.

Table 4A: Distribution of Loans Under $150,000 by Gender

Note: As noted in the text, the high level of missing observations suggests that the reader should not draw any conclusions from these data and we only present the analysis to provide an informed response to media reports of discrimination.

Table 4B: Distribution of Loans over $150,000 by Gender

Note: As noted in the text, the high level of missing observations suggests that the reader should not draw any conclusions from these data and we only present the analysis to provide an informed response to media reports of discrimination.

The average minority-owned business in the U.S. operates with much less capital, even after controlling for factors influencing financing amounts. Further, these businesses tend to report less revenue and have less access to capital than non-minority-owned firms,^{[ 25 ]} and some minority-owned firms tend to be less profitable.^{[ 26 ]} Performance gaps can be attributed to factors that include obtaining sufficient financial capital to buffer losses, the achievement of efficient scale and exploitation of business opportunities.^{[ 27 ]} The Kauffmann Foundation reports that Black-owned and Hispanic-owned businesses experience less favorable loan application outcomes than do White-owned and Asian-owned businesses, even after controlling for firm- and owner-specific characteristics. The Kauffmann report concludes that Black and Hispanic entrepreneurs enter industries with low capital requirements and high failure rates, which weaken their firms’ abilities to buffer losses and financial growth if they survive in early stages.

In the under-$150,000 category (Table 5A), approximately 82 percent of the loans do not document the borrower’s race. Where race is stated, it appears that 0.36 percent were American Indian/Alaskan Native, 17.48 percent were Asian, 1.8 percent were Black or African American, 15.77 percent were of Hispanic descent and 64.59 percent were White. In the under-$150,000 loan category, Asian-owned businesses borrow a median of $37,000 and save a median of seven jobs. Black-owned businesses borrow a median of $42,800 and save seven jobs. Hispanic-owned businesses borrow a median of $42,500 and save seven jobs. White-owned businesses borrow a median of $44,100 and save six jobs. Note, these median loan amounts do not control for the industry of businesses. Recall that professional and technical services (NAICS = 54) along with health care and social assistance (NAICS = 62) had substantial numbers of the initial PPP loans and are industries in which small businesses likely pay higher salaries.

In the over-$150,000 category (Table 5B), approximately 85.69 percent of the loans do not document the borrower’s race. Where race is stated, it appears that 0.5 percent were American Indian/Alaskan Native, 9.82 percent were Asian, 2.31 percent were Black or African American, 17.54 percent were of Hispanic descent and 69.83 percent were White. In the over-$150,000 loan category, median jobs saved are as followed: Asian (37), Black (38), Hispanic (37) and White (30). Due to the large amount of missing data, we are unable to state whether any differences are significant.

Table 5A: Distribution of Loans Under $150,000 by Race/Ethnicity

Note: As noted in the text, the high level of missing observations suggests that the reader should not draw any conclusions from these data and we only present the analysis to provide an informed response to media reports of discrimination. 

Table 5B: Distribution of Loans Over $150,000 by Race/Ethnicity

Note: As noted in the text, the high level of missing observations suggests that the reader should not draw any conclusions from these data and we only present the analysis to provide an informed response to media reports of discrimination.

Our findings could be interpreted as support for the argument that women and minority-owned businesses did not receive the same levels of support in PPP. However, we again caution the reader against drawing such a conclusion due to the large amount of missing demographic data.

Conclusions

PPP is intended to provide short-term support to small businesses that are temporarily closed due to the COVID-19 virus and subject to the insufficient level of funding. It appears to have worked as intended and community bankers are significant contributors to the process.

It is not surprising to us that roughly 64 percent of the Texas-based first-phase PPP lenders are community banks. The need to provide national liquidity quickly to businesses not only illustrates the value of the community bank, but also the value of the banking system, with its diversity of size and location. Our results suggest that community banks in Texas were instrumental in distributing PPP loans to rural markets.

The National Association of Small Business survey indicates that the types of small businesses that are believed to be particularly vulnerable and at the highest risk of closing are hotels, food services, educational services, mining, and oil and gas. Our results suggest that these industry classes are well-served in Texas, with the exception of educational services.

The media has reported that initial PPP loans did not get to traditionally underserved business owners. The PPP data lean in the direction of the media reports, but the omission of complete data on owner’s demographics in the SBA database leads to a lack of understanding as to the actual demographics of the borrowers.

For years now, we have seen a rapid consolidation of the banking industry. The decline of the community banks continues to affect rural America and the country in many ways. In Texas, as we have shown, community banking covers the entire state in both rural and urban areas. Through this pandemic, we have seen the value of the banking system and the importance of the community bank on display as banks and bankers have delivered the monetary medicine to keep so many from financial ruin.

^{[ 1 ]} See Paycheck Protection Program (PPP) Information Sheet: Borrowers.

^{[ 2 ]} According to Intelligencer: “Bank of America was one of the few major lenders to participate in the PPP on opening day. Over the ensuing 72 hours, it received loan applications from 177,000 small businesses, which collectively requested $32.6 billion in financing. If those loans were all approved, a single lender will have wiped out nearly 10 percent of the bailout fund in just three days.”

^{[ 3 ]} See “Potbelly, Shake Shack, Axios: Here are All the Companies Returning PPP Money After Public Backlash” and “How, Exactly, Did the LA Lakers Get a ‘Small-Business’ Loan?”

^{[ 4 ]} Among other problems related to oversight, the analysis identified 10,856 loans totaling more than $1 billion granted to borrowers who received more than one loan and, more than 600 loans totaling over $96 million that went to companies forbidden from doing business with the federal government. See Memorandum on Preliminary Analysis of Paycheck Protection Program Data (September 1, 2020).

^{[ 5 ]} Coronavirus Aid, Relief and Economic Security Act, Pub. L. No. 116-136, §§ 1102, 1106 (2020).

^{[ 6 ]} Small Business Administration, Office of Inspector General, Flash Report: “Small Business Administration’s Implementation of the Paycheck Protection Program Requirements” (May 8, 2020)

^{[ 7 ]} See Memorandum on Preliminary Analysis of Paycheck Protection Program Data (September 1, 2020).

^{[ 8 ]} The PPP was prepared in such a rush that it was subsequently found to contain a  loophole that technically allowed firms to lay off their staff any time between February 15 and April 26—and still secure loan forgiveness so long as they rehired those workers by June 30. The solution was to adjust the terms of forgiveness. See Intellegencer, op. cit.

^{[ 9 ]} We used the set of large banks identified for the initial stress testing under Dodd-Frank Act. We collected this list from Allen, Cyree, Whitledge and Winters (Journal of Economics and Business, 2018, vol. 98, p. 19–31).

^{[ 10 ]} The purpose of this restriction was to eliminate potential errors in the data. For example, one loan was for $585, which supposedly retained 115 jobs. Similarly, a second loan of $4,400 was to support 158 positions. These small amounts could simply be “placeholders.” There were 1,069 such eliminations, all of which took place on either April 15 or 16, 2020. The placeholder argument may have some merit given 788 of the eliminations were for zero jobs or one job retained.

^{[ 11 ]} See NSBA’s COVID-19 Relief Survey (2020).

^{[ 12 ]} Berger, A.N.; and Udell, G.F. (2002). “Small-Business Credit Availability and Relationship Lending: The Importance of Bank Organizational Structure,” Economic Journal 112, F32–F53.

^{[ 13 ]} We used the following list of stress-tested major banks: JP Morgan Chase & Co., Citigroup, Bank of America Corp., Wells Fargo & Co., Goldman Sachs Group, Morgan Stanley, PNC Financial Services Group, US Bancorp, Bank of NY Mellon Corp., SunTrust Banks Inc., State Street Corp., Capital One Financial Corp., BB&T Corp., Regions Financial Corp., American Express Co., Fifth Third Bancorp and Keycorp.

^{[ 14 ]} It is worth noting that First Financial Bank—a community bank—was next on the list with 992 loans.

^{[ 15 ]} These maps are available upon request.

^{[ 16 ]} A 2018 Pew Research Center report says that 29 percent of Americans live in “lower class,” with a median income of $25,624 in 2016. Our weekly loan per job for PPP1 is $646, which annualizes over 50 weeks to $32,300. This suggests that PPP1 reached the intended workers.

^{[ 17 ]} A NAICS code is a classification within the North American Industry Classification System. The NAICS system was developed for use by Federal Statistical Agencies for the collection, analysis and publication of statistical data related to the U.S. Economy. NAICS is a self-assigned system; no one assigns you an NAICS code. What this means is that a company selects the code that best depicts its primary business activity and then uses it when asked for its code.

^{[ 18 ]} Carter, S.; Shaw, E.; Lam, W.; and Wilson, F. (2006). “Gender, Entrepreneurship and Bank Lending: The Criteria and Processes Used by Bank Loan Officers in Assessing Applications,” Entrepreneurship Theory and Practice 31(3), 427–444.

^{[ 19 ]} Orser, B.; Hogarth-Scott, S.; and Riding, A. (2000). “Performance, Firm Size and Management Problem-Solving,” Journal of Small Business Management 38(4), pp. 42–58.

^{[ 20 ]} Carpenter, R.; and B. Petersen (2002). “Is the Growth of Small Firms Constrained By Internal Finance?” Review of Economics and Statistics 84(2), 298–309. Haines, G.H.; Orser, B.J.; and Riding, A.L. (1999). “Myths and Realities: An Empirical Study of Banks and the Gender of Small-Business Clients,” Canadian Journal of Administrative Sciences, 16(4), 291–307.

^{[ 21 ]} Coleman, S. and Carsky, M. (1996) “Understanding the Market of Women-Owned Small Businesses,” Journal of Retail Banking Services 18(2), 47–49.

^{[ 22 ]} Coleman, S. (2000). “Access to Capital and Terms of Credit: A Comparison of Men- and Women-Owned Small Businesses,” Journal of Small Business Management, 38(3), 37–52.

^{[ 23 ]} Shepherd, M. (2020). “Women-Owned Businesses: Statistics and Overview,” Fundera.

^{[ 24 ]} Shepherd, M. (2020). “Women-Owned Businesses: Statistics and Overview,” Fundera.

^{[ 25 ]} Prakash, P. (2020). “Top Small-Business Loans for Minorities.”

^{[ 26 ]} Ortiz-Walters, R.; and Gius, M. (2012). “Performance of Newly Formed Micro Firms: The Role of Capital Financing in Minority-Owned Enterprises. Journal of Developmental Entrepreneurship 17(3).

^{[ 27 ]} Kauffmann Compilation: “Research on Race and Entrepreneurship” (2016) by the Ewing Marion Kauffmann Foundation.