Introduction

Small businesses are owned by and employ a wide variety of entrepreneurs—skilled trade technicians, medical professionals, financial consultants, technology innovators, and restaurateurs, among many others. Similarly, small family farms are important to the U.S. agriculture industry because they operate nearly half of America's farmland.¹ As do large corporations, many small businesses and small farms rely on loans (credit) to purchase inventory, to cover cash flow shortages (which may arise from unexpected expenses or periods of inadequate income), or to expand operations, among other things. Congress has demonstrated an ongoing interest in lending to small entities, viewing them as mechanisms for stimulating the economy and creating jobs. Congress's interest in small business credit access generally focuses on (1) whether small businesses can secure credit from private lenders and (2) whether small businesses can obtain such credit at fair and competitive lending rates. In addition, Congress is interested in fair lending issues, which are matters concerning illegal discrimination against certain demographic groups (e.g., race, religion, gender).

Congress has passed various legislation to address credit access for small business and small farms as well as to promote equal credit access. For example, the Small Business Act of 1953 (P.L. 83-163) established the Small Business Administration (SBA), which supports capital access for small businesses that are unable to obtain credit at reasonable terms and conditions from private sector lenders.² The Community Reinvestment Act (CRA; P.L. 95-128) encouraged banks to address persistent unmet small business credit demands in low- and moderate-income (LMI) communities.³ The Farm Credit Act of 1971 (P.L. 96-592) was amended in 1980 to consider the credit needs of young, beginning, and small farmers and ranchers.⁴ The Riegle Community Development Regulatory Improvement Act of 1994 (P.L. 103-325) established the Community Development Financial Institutions (CDFI) Fund, which certifies and provides financial awards for lenders with a primary mission of community development to support local credit needs.⁵

Past legislative actions have focused predominantly on making credit available in underserved communities that consist of LMI individuals rather than on certain demographic characteristics of individual applicants requesting credit. Furthermore, the Equal Credit Opportunity Act of 1974 (ECOA; P.L. 94-239) generally prohibits creditors from discriminating against applicants on the basis of race, color, religion, national origin, sex, marital status, or age or because the applicant receives public assistance. For these reasons, collecting these demographic characteristics about loan applicants is generally prohibited except under certain circumstances. For example, the Home Mortgage Disclosure Act of 1975 (P.L. 94-200) allows for the collection and reporting of demographic characteristics of mortgage applicants to increase credit access to creditworthy individuals and discourage discriminatory loan-markup pricing practices.⁶ Thus, statutory authority was necessary before the gathering of demographic information about small business loan applicants could occur.

Section 1071 of the Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act; P.L. 111-203) amended ECOA to give the Consumer Financial Protection Bureau (CFPB) the authority to compile data from small business lenders to identify the financing needs of small businesses, particularly those owned by women and minorities.⁷ The focus on lending gaps as they pertain specifically to women- and minority-owned firms is to facilitate the enforcement of fair lending laws, which are typically designed to protect consumers (as opposed to commercial businesses) from unfair and discriminatory practices.⁸

On March 30, 2023, the CFPB finalized the Section 1071 data collection requirements, and it will implement the law via Regulation B.⁹ Regulation B specifically requires financial institutions to collect and report certain data regarding the credit applications of women-owned, minority-owned, and other small businesses, which includes small farms. The data must be reported annually to the CFPB.

This report discusses the Section 1071 final rule. It first explains various challenges and considerations encountered by the CFPB while promulgating the rule. Next, it provides an overview of the data collection and reporting requirement for covered lenders under the final rule. Following a discussion of issues related to certain other data collection requirements and other collection matters, the last section highlights some observations regarding the SBA collecting of data on various SBA program borrowers that may arise with the collection of data under Section 1071. Specifically, SBA loan applicants may choose not to provide race and ethnicity, gender, and veteran status information. Loans that do not get reported—because either some small lenders will not meet the minimum loan origination thresholds to report or some borrowers choose not to provide the requested demographic information—could arguably contain useful insights that would be unobservable to policymakers.

Challenges and Considerations Promulgating the Final Rule

The CFPB took more than a decade before promulgating the final Section 1071 rule. Evaluating the extent of lending gaps—and specifically fair lending risks—in small business credit markets has complications. Dodd-Frank directed the definition of small business, discussed in the section of this report entitled "Summary of the Section 1071 Final Rule." Nevertheless, the CFPB's challenge was to design a dataset with the ability to conduct meaningful comparisons across loan products and over time given the various differences in small business types and models.

The Small Business Regulatory Enforcement Fairness Act of 1996 (P.L. 104-121) also requires the CFPB to address issues that could potentially have significant economic impacts on small entities subject to the Section 1071 rule.¹⁶ The CFPB had to consider, for example, key differences in the lending models of large and small lenders, which affect the type and cost of data that would be collected.¹⁷

First, large and small lenders often collect different types of data. Large lenders typically engage in lending to borrowers who possess more conventional financial metrics and documentation (e.g., sales fluctuations, costs of inputs, specific industry factors), which is considered hard information that can be used in automated and statistical underwriting methodologies to price loans.¹⁸ By contrast, small lenders typically engage in relationship lending, meaning that they must develop close familiarity with their customers to gather soft information, which contains circumstantial details about factors such as non-standardized business risks, insufficient collateral, or weak or thin (business) credit histories. Because of soft information, the loan underwriting process to determine more customized loan products and loan pricing is generally less algorithmic and more labor intensive.¹⁹

Second, the type of information collected, which varies among lenders, would also be expected to influence their reporting costs. For example, because hard information is already quite uniform, large lenders may already have adopted automated technological systems that can handle large volumes of standardized and digitized financial data. In these cases, reporting is likely to be less expensive per applicant. By contrast, soft information is more unique to applicant circumstances, infrequent, and localized such that standardization of the data for electronic collection and reporting purposes is challenging. The reporting cost per applicant is also likely to be more expensive for small lenders that lack the volume of applications to justify the costs to convert soft information to digital and secure formats. Therefore, data likely to be informative about lending gaps in the small business and farm credit markets may be more difficult to standardize and more costly to collect, especially if small lenders predominantly serve these markets.

The CFPB also had to consider how Section 1071 implementation requirements might affect the supply of small business loans. For example, some institutions might decide to offer more standardized, less tailored financial products to reduce their reporting costs. Some lenders might require minimum principal loan amounts (e.g., $100,000) to ensure that the loans generate enough revenue to cover the costs to fund and report data, thereby leaving gaps in credit markets for many businesses that are starting up or small. In short, Section 1071 implementation, which is designed to identify any lending gaps, could potentially exacerbate lending gaps in various credit market segments without careful consideration of the potential impact of its requirements.

Summary of the Section 1071 Final Rule

The CFPB finalized Regulation B after years of surveying stakeholders to determine what data collection efforts were already in place, for what purposes, and how to minimize redundancy in data collection while factoring in the collection costs on the smallest institutions. Highlights of the final rule appear below.²⁰

Definition of Small Business. The final Section 1071 rule generally incorporates the meaning of business concern and small business concern under the Small Business Act and SBA regulations.²¹ However, instead of relying upon the SBA's size standards, the final rule defines small business concern as a business that had $5 million or less in gross annual revenue for its preceding fiscal year.²²

Covered Lenders. A financial institution covered by Section 1071, by the time all compliance dates have passed, is one that originates at least 100 covered small business loans in each of the two preceding calendar years. (The final rule increased the minimum loan volume threshold of 25 in the proposed rule to 100 loan originations, which could potentially reduce reporting requirements for some small lenders.²³) These entities include, but are not limited to, depository institutions (i.e., banks, savings associations, and credit unions); online lenders; platform lenders; community development financial institutions (CDFIs); lenders involved in equipment and vehicle financing; Farm Credit System lenders; commercial finance companies; merchant cash advance providers; governmental lending entities; and nonprofit lenders that satisfy the minimum origination thresholds. Lenders must determine annually whether they are covered by Regulation B. Any financial institution, however, may voluntarily collect and report data. Small CDFIs and mission-oriented community banks, for example, which may fall below the loan volume thresholds, expressed a willingness to collect and report small business lending data.²⁴

Covered (Reportable) Applications. A covered application from a small business is an oral or written request for a covered loan (defined in the next paragraph). The CFPB, however, will not treat a reevaluation, extension, or renewal request on an existing business loan as a covered application. To reduce excessive reporting and data inaccuracies, business inquiries regarding the likelihood of qualification will not be considered formal applications. Similarly, solicitations, firm offers of credit, or other evaluations initiated by a lender will not be treated as a covered application until a business pursues the credit offer.

Covered Loan. A covered loan or covered credit transaction is a formal extension of business credit that can include loans, lines of credit, credit cards, merchant cash advances, and credit products used for agricultural purposes. The CFPB also lists various transactions that would not be considered covered transactions (e.g., mortgage loans that are reportable under a different set of requirements; a reevaluation, extension, or renewal request on an existing business loan).

Key Data Variables for Collection. The data variables below will be collected from covered applications:²⁵

• a unique loan identifier;²⁶
• the application date (month, day, and year);
• the application method (whether the borrower applied in person, by telephone, by mail, or online);
• the application recipient (whether the application was submitted directly to lender, a lender's affiliate, or an unaffiliated third party);
• the action taken by the covered financial institution on the application (whether the loan was originated, approved but not accepted, denied, withdrawn by the applicant, or incomplete);
• the action taken date (month, day, year);
• for denied applications, up to four denial reasons (from 10 choices);²⁷
• for applications (i) originated or (ii) approved but the borrower chose not to accept, pricing information that may include interest rate(s), total origination charges, charges by third parties, broker fees, initial annual charges, the difference between the amount advanced and the amount to be repaid for merchant cash advances and similar products, and prepayment penalties;
• the credit type (one credit type must be chosen from either nine credit products or 12 loan guarantee products, or the number of months for a loan product that does not conform to any of the credit or guarantee products must be provided);²⁸
• up to three options pertaining to the credit purpose from a list of 15 options;²⁹
• the amount applied for in the form of a range of numbers;³⁰
• the census tract based on an address or location provided by the applicant using a "waterfall" approach;³¹
• the gross annual revenue for the applicant's business for the preceding fiscal year;³²
• a three-digit North American Industry Classification System code for the applicant;³³
• the number of people working for the applicant in the form of ranges (rather than a specific number of workers);³⁴
• the applicant's time in business (i.e., less than two years, two or more years);
• the number of principal owners;³⁵
• self-identification as a women-owned business status, minority-owned business status, or LGBTQI+-owned business status if one or more individuals (in each category respectively) holds more than 50% of ownership or control and if more than 50% of the net profits or losses accrue to one or more such individuals—and respondents may elect to answer "I do not wish to provide this information," particularly for any business that cannot reach a legal conclusion;³⁶ and
• the ethnicity, race, and sex of the principal owner(s) applying for credit.³⁷

No Requirement to Collect Information via Visual Observation or Surname. Should applicants choose not to provide demographic information, the final rule does not require financial institutions to use visual observation, surname analysis, or any other basis to determine principal owners' ethnicity and race.³⁸ Instead, the questions are to be reported as unanswered.

Compliance Dates. Final rule compliance is based on a tiered compliance date schedule. Lenders that originate at least 2,500 small business loans annually must report data starting October 1, 2024. Lenders that originate at least 500 loans annually must report data starting April 1, 2025. Lenders that originate at least 100 loans annually must report data starting January 1, 2026.

Selected Data Collection Issues Under Section 1071

Prior to Section 1071 implementation, various types of lenders collected and reported small business data under different programs, as listed in the following examples:

• To comply with mission-based requirements, banks and CDFI-designated lenders must report data to demonstrate compliance with the CRA and CDFI Fund requirements, respectively. Both CRA and the CDFI Fund focus on lending in circumscribed geographical areas by small lenders.⁴⁷
• The purpose for reporting credit unions' business lending activities is to demonstrate compliance with statutory requirements that differ from banks' business lending activities.⁴⁸
• Reporting requirements in SBA lending programs are generally designed to collect data about whether the programs expand access to credit to entrepreneurs who cannot access credit elsewhere. For example, the Microloan program is statutorily targeted "to assist women, low-income, veteran … and minority entrepreneurs and business owners and other such individuals."⁴⁹
• The USDA reporting requirements are generally designed to collect data about whether its programs benefit less developed farms.

Section 1071's intended focus, however, is on the race and gender of credit applicants. For this reason, many lenders—even those that already report data for various other purposes—are likely to incur costs to comply with Section 1071 collection requirements. Notwithstanding harmonization efforts by various federal regulators to mitigate the compliance costs—particularly for small lenders—some data issues are likely to remain. This section discusses selected small business data collection activities that currently exist, particularly some issues that federal regulators have made efforts to address and those that may still remain.

Banks: Data Reporting Under the Community Reinvestment Act

Congress passed the CRA in response to concerns that federally insured banking institutions were not making sufficient credit available in the local areas in which they were chartered and acquiring deposits.⁵⁰ The CRA requires federal banking regulatory agencies—the Board of Governors of the Federal Reserve System, the Federal Deposit Insurance Corporation, and the Office of the Comptroller of the Currency—to implement and enforce the CRA by conducting examinations to assess whether banks are meeting local credit needs.⁵¹ The regulators assign CRA credits (i.e., points) where banks engage in qualifying activities in the areas where they have deposit-taking operations. Qualifying activities include mortgage, consumer, and business lending, among other activities that would benefit LMI areas and entities.⁵² CRA credits are subsequently used to issue each bank a performance rating.

For purposes of the CRA, banks must adhere to the following guidelines:⁵³ Under current CRA regulations, small business loan is defined as (1) having an original loan amount of $1 million or less, and (2) reported by the institution as either "secured by nonfarm or nonresidential real estate" or "commercial and industrial" on its call report.⁵⁴ Small farm loan is defined as having an original amount of $500,000 or less and reported as either "loans to finance agricultural production and other loans to farmers" or "loans secured by farmland" on the call report. In addition to these definitions, banks may still rely upon SBA's size standards for small business loans to negotiate for CRA consideration, particularly for originations that occur in LMI geographical locations and are guaranteed by the SBA. For each small business or small farm loan, a bank must collect a unique loan identifier, the loan amount at origination, the geographical location, and whether the loan was for a business or farm with gross annual revenues of $1 million or less.

On May 5, 2022, the three federal banking regulators issued a proposed rule that, among other things, would harmonize the CRA definitions of small business and small farm with those proposed in the Section 1071 rule.⁵⁵ Instead of using definitions based upon original loan amounts and the definitions of small business and small farm, the CRA reporting requirements are expected to adopt the finalized Section 1071 definition of "gross annual revenues of $5 million or less." Because the federal banking regulators allow the database created for the disclosure of mortgage originations to be used for CRA examinations, the use of Section 1071 will further standardize data collection and reporting efforts in the lending (i.e., banks and non-banks) industry, arguably resulting in greater data consistency, comparability, and possibly more uniform loan pricing.⁵⁶

Credit Unions: Definition of Member Business Loans

Although credit unions primarily engage in consumer and residential lending, some originate commercial business loans for members, which are also referred to as member business loans (MBLs). Congress established provisions in the Credit Union Membership Access Act of 1998 (P.L. 105-219) regarding MBLs, including a commercial lending cap that limits most credit unions to lending no more than 12.25% of their assets to small businesses.⁵⁷

The National Credit Union Administration (NCUA), the primary regulator of the credit union system, updated the regulations that define business loans to facilitate compliance with the act.⁵⁸ The revised definition, which differentiates MBLs from commercial loans, excludes business loans with balances less than $50,000 from being defined as either MBLs or commercial loans.⁵⁹ Based upon NCUA call reports (which capture credit unions' total loan originations rather than applications), the CFPB estimated in its Section 1071 rulemaking that approximately 400 credit unions would have been required to report under the initially proposed threshold of 25 loan originations.⁶⁰ The CFPB final rule does not exempt MBLs with balances less than $50,000. Thus, some credit union loans that would be exempt from the MBL cap would still be eligible for Section 1071 reporting. Whether some credit unions are likely to respond by making fewer of these loans to mitigate reporting costs is uncertain.⁶¹

Non-Depository CDFIs: Reporting of Business Lending Data

The CDFI Fund was created by the Riegle Community Development Regulatory Improvement Act of 1994 (P.L. 103-325) to promote economic development in distressed urban and rural communities.⁶² The CDFI Fund certifies financial institutions—depositories (i.e., banks, credit unions) and non-depositories such as nonprofit loan funds, microloan funds, and (for-profit and nonprofit) venture capital funds—that can demonstrate having a primary mission of promoting community development. After certification, CDFIs become eligible for financial awards and other CDFI Fund assistance that promotes community development in markets comprised of economically distressed people and places.

In 2016, the CDFI Fund required CDFIs to annually submit Annual Certification and Data Collection Reports (ACR).⁶³ The CDFI Fund uses ACRs to ensure that its awards are dispersed to intermediaries predominantly engaged with serving people and communities in a manner that is consistent with its mission. ACR data may also facilitate better understanding of the types of financial products obtained by customers and the development of a policy map comprised of CDFI target markets. In 2020, the CDFI Fund proposed data modifications to ACR and introduced the Certification Transaction Level Report, which is designed to standardize and automate the data collection process.⁶⁴

CDFI banks and CDFI credit unions already collect and report data for their call reports, but non-depository CDFIs may be subject only to CDFI Fund reporting requirements. The CFPB received comments from stakeholders who estimated that non-depository CDFIs may take up to three years to comply with Section 1071.⁶⁵ Using CDFI Fund ACR data from FY2019, the CFPB estimated that, of the 340 non-depository CDFIs engaged in small business lending, 240 of them would have met the original proposed reporting threshold of 25 covered loan originations in each of the two preceding calendar years.⁶⁶ In 2019, the CDFI Fund reported that a total of 934 CDFIs (or approximately 25%) would have been eligible for Section 1071 reporting under the proposed rule.⁶⁷ With the increase of the minimum threshold to 100 covered loan originations under the final rule, fewer CDFIs would be expected to report, particularly non-depository CDFIs. However, some CFDIs, including other mission-oriented community banks, indicated that they still intended to voluntarily report data, which may be more feasible, particularly for those institutions that already satisfy various data collection and reporting requirements.⁶⁸ Nevertheless, because overall CDFI lending represents a small percentage of the U.S. financial system, the potential loss of unreported CDFI commercial lending may not significantly affect what the CFPB learns about general credit access.⁶⁹

Agricultural Lending: Data Collection Issues

USDA and the Farm Credit System (FCS), a government-sponsored enterprise, collect data related to their lending activities to farmers and ranchers. Most agricultural lending (approximately 80%) is done by either commercial banks (36%) or the FCS (44%). As a direct lender, USDA has a share of the agricultural credit market of approximately 3%. It guarantees an additional 4% of loans, which are originated by commercial lenders or the FCS.⁷⁰

The CFPB includes agricultural lending in the scope of small business lending for implementing Section 1071 to improve the demographic data available about agricultural lending. The CFPB estimated, using the 2019 annual report of the Farm Credit Administration (FCA), that the FCS would have had 72 lenders eligible to report under its proposed rule (which had initially relied upon a minimum loan threshold of 25 loan originations). In addition, the CFPB noted that high fair lending risk may be present in agricultural lending.⁷¹ This section describes the level of information available about the demographics in agricultural lending at the time the rule was finalized.

Socially Disadvantaged Farmers and Ranchers (SDFR)

The Farm Service Agency (FSA), which is part of USDA, receives federal appropriations to make direct loans to farmers and issue guarantees on loans made by other lenders to farmers who are unable to obtain credit elsewhere. USDA FY2021 data indicate that FSA made 5,200 direct loans to SDFRs (24% of the number of loans) and issued guarantees for 984 SDFRs (14% of the number of guarantees). These represented $582 million (18%) of FSA's direct loans and $558 million (16%) of guaranteed loans.⁷² Given that the FSA is a federal agency, direct loans made by FSA would not be subject to the Section 1071 rule.

USDA administers various programs to SDFRs (7 U.S.C. §2003), which include individuals who belong to groups that have been subject to racial or ethnic prejudice (e.g., Black or African American, American Indian or Alaska Native, Hispanic or Latino, and Asian or Pacific Islander, and women).⁷³ In addition to SDFR targets, the farm loan program also reserves funds for beginning farmers and ranchers (7 U.S.C. §1994(b)(2)). Target levels of lending to SDFRs are determined at the county level based on local demographic information for the population of farmers and ranchers.

From 2015 to 2017, SDFRs represented 17% of farmers but accounted for 8% of agricultural debt.⁷⁴ The observed disparity may be difficult to explain because, according to the Government Accountability Office (GAO), the data on lending to SDFRs is limited.⁷⁵ It is also difficult to ascertain the effects of FCS outreach and engagement efforts. Furthermore, different SDFR data (collected pursuant to the 2018 farm bill, which required USDA to differentiate race, ethnicity, and gender) reveals that women accounted for about 60% of SDFR direct loan borrowers and 45% of SDFR guarantees in FY2019. By race and ethnicity, borrowers also identified as Black (8%), Asian-Hawaiian-Pacific Islander (4%), Indian-Alaska Native (21%), and Hispanic (17%). However, because borrowers may identify in multiple categories (e.g., non-White Hispanic or Black female), the sum of the demographic records exceeds the number of SDFR borrowers.⁷⁶ Given these SDFR data issues, Section 1071 data collection may facilitate identification of fair lending risks that the CFPB believes are heightened in agricultural lending.⁷⁷

Young, Beginning, and Small (YBS) Farmers and Ranchers

The FCS is a nationwide financial cooperative consisting of four district banks and member lending institutions that collectively operate as a government-sponsored enterprise.⁷⁸ After raising funds by selling bonds to private investors, the FCS may act as a direct lender to eligible individuals and businesses to make certain agricultural and rural loans to purchase land, livestock, equipment, and other supplies as well as to construct or improve buildings. In addition, the FCS provides loans to or participates with other financial institutions (OFIs). Some Native American CDFIs, for example, participate as OFIs with the FCS to obtain the low-cost funding available to its member institutions.⁷⁹

Credit access for start-up farms may be as challenging as it is for start-up commercial businesses, and small relationship lenders typically serve this niche market.⁸⁰ In 1980, the Farm Credit Act of 1971 (P.L. 96-592) was amended to require the FCS to focus on the credit needs of young, beginning, and small (YBS) farmers and ranchers.⁸¹ Specifically, regulation defines young farmers as those age 35 or younger, beginning farmers as those farming for 10 years or less, and small farmers as those having gross annual sales below $250,000.⁸² On February 14, 2019, the FCA announced its plan to improve its YBS data collection, evaluation, and reporting process.⁸³ The FCA is the primary regulator of the FCS.⁸⁴

Out of a total of 313,439 loans that FCS made in 2022, about 18% were to young farmers, 25% were to beginning farmers, and 41% were to small farms or ranches. Since YBS loans are typically smaller than the average FCS loan, the percentages of the $135 billion of loans made were smaller than the number of loans: About 10% of the amount of loans were to young farmers, 16% to beginning farmers, and 14% to small farmers.⁸⁵

Similar to the CRA, however, the YBS mandate does not compel the FCS to focus on any other demographic groups, meaning that its data collection and reporting is not delineated by race or ethnicity.⁸⁶ Similar to the CDFI Fund, the FCA's objective is to automate and standardize its data collection efforts while also reducing the regulatory burden for FCS member institutions.⁸⁷ Nevertheless, the CFPB does not exclude FCS lenders from the Section 1071 final rule, and it estimates that all loans made by FCS members to farmers will be covered.⁸⁸

A provision in the House Agriculture Committee–passed farm bill, H.R. 8467, Section 5503, incorporates the text of H.R. 2423 to (1) amend ECOA to exclude entities supervised by the FCA from Regulation B; (2) add to the Farm Credit Act that the FCA is the sole regulator of the FCS and that the definition of small farmer is one with annual farm sales less than $250,000; and (3) add a small farmer data collection requirement for FCS lenders to collect voluntary data about borrowers' race, sex, and ethnicity and report it to the FCA—replacing the ECOA data collection requirement. (A bill in the 117^th Congress, H.R. 7768, would have removed authority for CFPB and ECOA to implement policy affecting the FCS such as Section 1071.)

Reporting of SBA Lending Activities Under Section 1071

While SBA lenders would continue reporting data to the SBA for the various programs in which they participate, applications for SBA loans are generally covered for Section 1071 reporting requirements. The extent of SBA loan application data reporting, however, depends upon whether the lenders meet other final rule requirements such as the annual origination requirements. Some SBA lenders—particularly those Certified Development Companies (CDCs) that lend only through SBA's 504 loan guarantee program and those Microloan intermediaries that lend only through SBA's Microloan program—may not meet the annual origination requirements for required Section 1071 reporting. According to SBA program data, for example,190 CDCs originated at least one loan in the 504 program during FY2022, while 24 (13% of all CDCs) originated at least 100 loans in the 504 program.⁸⁹ Under the Microloan program, 137 intermediaries originated at least one microloan during FY2022, while nine (7% of all Microloan intermediaries) originated at least 100 microloans.⁹⁰

SBA lenders, however, typically do not limit themselves exclusively to SBA lending or participation in only one SBA lending program. Some CDCs or Microloan intermediaries may lend in multiple SBA programs or have non-SBA lending operations that would count toward determining their annual origination numbers. Lenders that engage in both SBA and non-SBA lending may also have enough origination volume to meet the minimum origination threshold. Furthermore, SBA lenders that may not meet the required reporting threshold may still choose to voluntarily report data if they support the purpose of Section 1071, as some CDFIs do.

Comparisons of data from various SBA programs and collected under Section 1071 will have challenges. For example, SBA program data are for approved loans, but Section 1071 data are for loan applications. SBA programs also have underwriting, servicing, risk grading, and loan liquidation requirements that often differ from conventional lending programs.⁹¹ These differences arguably can be identified under Section 1071 reporting requirements, thus enhancing the ability of fair lending examiners to make like comparisons.

SBA's Experience Collecting Demographic Data

The SBA began systematically collecting data on race and ethnicity in at least some of its small business loan programs in 1964.⁹² SBA's experience collecting demographic data, therefore, may preview some issues that CFPB might encounter while collecting data under Section 1071. Specifically, the next sections discuss questions that may arise when a business is asked to select its race, ethnicity, and gender and the potential prevalence of unanswered responses.

Demographic Data Categories and Definitions

The SBA currently collects demographic data from borrowers in the 7(a) and 504 programs during the SBA guarantee application process. After borrowers have completed the forms, the lenders report the information to the SBA. Specifically, SBA Form 1919 requests information about each owner's veteran status, gender, race, and ethnicity, as shown in Figure 1.⁹³ Information collection is similar for the 504 loan program.⁹⁴ As explained below, data collection in the Microloan and SBIC programs is different, and both programs have unique data challenges.

Figure 1. Business Owner Demographic Information Collection for 7(a) Borrowers

Source: SBA, "SBA 7(a) Borrower Information Form," Form 1919, revised September 2020, https://www.sba.gov/sites/sbagov/files/2022-06/Form%201919_10-21-2020-rev%20508%20r2_0.pdf.

Collecting and reporting data on the demographics of small businesses can be challenging. Shorthand expressions used by some observers can leave the impression that the business itself has a race, ethnicity, gender, and veteran status even if the business itself does not.⁹⁵ Instead, small business demographics generally refers to owners' demographics, which can be challenging to collect. While many businesses (such as sole proprietorships) have only one owner, other businesses (such as partnerships, limited liability companies, and corporations) may have multiple owners. In these cases, assigning demographics to the business requires first determining which aspect to measure—ownership, control, operation, or something else. Focusing on ownership often requires examining ownership shares.

The SBA has recognized these and other challenges and has attempted to address them. Table 1 shows the demographic group used in reporting SBA loan program data in its online Office of Capital Access reports. As mentioned above, demographic data collection is similar for the 7(a) and 504 programs, and the Microloan program uses different demographic groups for its data collection. One example of the SBA confronting the data collection challenge of businesses potentially having multiple owners is how gender data are reported. For the gender data, a loan is reported based on the share of the business that is owned by women. This allows a clearer understanding of loans made to businesses that are majority-owned by women,⁹⁶ but it does not provide information about businesses operated or controlled by women. (For many small businesses, there is likely to be substantial overlap involving ownership, operation, and control.) The SBA does not report race and ethnicity data in the same way. For the Microloan program, SBA officials stated that the SBA's online Microloan system allows entering the race and ethnicity of all owners. Loans are classified by the race and ethnicity of the primary borrower (the first borrower listed),⁹⁷ who may or may not own a majority stake in the business.

Percentages of Unanswered Responses

Some program participants choose not to provide race, ethnicity, gender, or veteran status data to the SBA because such reporting is optional, which poses another data challenge for the SBA. During the Obama Administration, for example, the SBA informed GAO that it could not require SBICs to report demographic data on their portfolio investments.

SBA officials told [GAO] that the Small Business Investment Act of 1958 may give [the SBA] the authority to collect demographic information on a voluntary basis but not to require SBICs to collect and report such data. The agency could require reporting of such information as a condition of retaining program benefits, [SBA] officials said, only if the agency needed it to assess program participants' compliance with statutory or regulatory criteria governing core program requirements, or to make licensing decisions. SBA officials told [GAO] their position was that demographic information was not necessary for program participation and that the agency could not require respondents to report such information because SBA did not have authority to assess compliance based on race, ethnicity, or gender factors.⁹⁸

Figure 2 shows the share of loans made to borrowers who did not provide race data. The average share of SBA loan borrowers who did not report race data over the period ranged from 16% to 18%, and was similar for all three programs. Whether certain racial and ethnic groups are more or less likely to decline providing the optional information is unclear.⁹⁹ On one hand, racial and ethnic minorities may be less likely to provide race and ethnicity data due to historical concerns about racial and ethnic discrimination in lending. On the other hand, because some SBA programs are targeted toward serving underserved and disadvantaged entrepreneurs, some White entrepreneurs may be less likely to provide race data due to fear of having their credit requests denied. Consequently, determining whether the unreported data are skewed—and, if so, the demographic group most underrepresented—is challenging.

Figure 2. Share of SBA Loan Borrowers Who Did Not Provide Race Data, FY2012-FY2023

Sources: Figure created by CRS using data from SBA. Data for FY2017-FY2023 from SBA data reports: "7(a) and 504 Summary Report," data as of July 20, 2023, https://careports.sba.gov/views/7a504Summary/Report?%3Aembed=yes&%3Atoolbar=no; "Microloans Summary Report," data accessed on July 21, 2023, https://careports.sba.gov/views/MicroloanSummaryReport/Report?%3Aembed=yes&%3Atoolbar=no. Data for FY2012-FY2016 from SBA Weekly Lending Report for September 30, 2017, https://www.sba.gov/document/report—2017-weekly-lending-reports. Notes: Data are not available for Microloan for FY2012-FY2016. Data for FY2023 are partial. For 7(a) and 504, it includes data through July 20, 2023. For Microloan, it includes all data reported online on July 21, 2023.

Missing data rates are following different patterns for the programs. For 7(a) and 504, the share of loans missing the borrower's race data was generally increasing from FY2012 to FY2020, when the share peaked for both programs. The share of 7(a) and 504 loans missing race data has fallen since. The partial FY2023 missing data rates for 7(a) loans are broadly similar to before the COVID-19 pandemic, while the missing data rates for 504 loans are below those of the years just before the pandemic. Conversely, the share of Microloans missing race data did not increase during the pandemic. Instead, missing data rates held broadly steady for FY2017-FY2021 and then increased in FY2022 and for part of FY2023.

Other SBA Demographic Data Issues

GAO has studied SBA data issues several times and provided recommendations regarding data collection methodology and guidance to program partners.¹⁰⁰ In November 2019, GAO found the existing Microloan program data on borrower characteristics (such as race, ethnicity, and gender) reliable but expressed concerns with other program data. For example, GAO noted that, at that time, SBA collected data from Microloan intermediaries about whether Microloan borrowers had "low income" but did not define that concept. GAO determined that the lack of SBA guidance on the definition of low income meant that "it is unclear what criteria [Microloan intermediaries] are using then they record a borrower as" low income. GAO recommended providing lenders with more information about the definitions of key data collection concepts such as low income.¹⁰¹

The SBA Office of Inspector General (OIG) also studied data issues in the Microloan program in September 2017. OIG selected a sample of 52 loans and analyzed whether the lender's documentation supported the data stored in the SBA's systems. OIG found several weaknesses. Of the 52 loans studied, 27 loans (52%) had non-matching information in the lenders' files and in the SBA's system. Additionally, OIG found that Microloan intermediaries did not have full documentation to prove that they originated and disbursed 44 of the 52 microloans (85%). OIG determined that the "deficiencies affect the reliability of the data reported to SBA by [Microloan intermediaries]" and that "SBA's ability to validate [Microloan] data … was impaired."¹⁰²

In January 2016, GAO determined that SBIC "data on minority, women, and veteran ownership of SBIC investments are unreliable, in part because SBA does not provide guidance to SBIC licensees on how to collect and report this information." GAO recommended that the SBA provide guidance on data collection concepts and methodologies to improve the quality of program data.¹⁰³

Conclusion

Without reliable data, the ability to evaluate the performance of various small business lending markets—specifically whether (1) a small business credit shortage exists, (2) pricing for loans to small businesses is significantly above the lending risks and funding costs, or (3) fair lending risks are present—is extremely challenging. The CFPB's Regulation B, promulgated under Section 1071 of the Dodd-Frank Act, gives the agency the authority to compile data from small business lenders. If it is possible to identify impediments to small business credit, policymakers may subsequently be able to better address these matters. Arriving at more definitive conclusions about the availability and costs of small business lending is more likely to occur with information such as the size and financial characteristics of the businesses that apply for loans, the types of loan products they request, the types of lenders to whom they applied, and which applications were approved and rejected. Nevertheless, the benefits of data collection also have costs and challenges that have been further illuminated given the challenges to implement Section 1071. Consequently, loans that do not get reported—either because some small lenders do not meet the minimum loan origination thresholds to report or because some borrowers choose not to provide the requested demographic information—could contain useful insights that would still be unobserved by policymakers.