Statement of the National Community Reinvestment Coalition: Data Drives Access to Credit and Capital for Small Business

CFPB Symposium on Section 1071

Introduction – Importance of Small Business Loan Data

The National Community Reinvestment Coalition (NCRC) appreciates this opportunity to comment on the importance of data in understanding and monitoring the small business lending market. NCRC is an association of 600 community-based nonprofit organizations dedicated to increasing access to credit and capital for underserved communities. Our efforts range from advocacy to counseling homebuyers and small business owners. NCRC and our members have viewed data as integral to the mission of increasing access to responsible credit. NCRC helped members of Congress draft Community Reinvestment Act Modernization legislation that included an earlier version of Section 1071 mandating improvements to publicly available small business loan data.

Section 1071 of the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010 requires the Consumer Financial Protection Bureau (CFPB) to enhance publicly available small business data to include the race and gender of small businesses applying for credit. Its purpose is to “facilitate enforcement of the fair lending laws and enable communities, government entities, and creditors to identify business and community development needs and opportunities of women-owned, minority-owned, and small businesses.”

In its white paper, Key Dimensions of the Small Business Lending Landscape, the CFPB surveys the current small business loan data and concludes that none of the current data sources “provides comprehensive information on the extent to which small businesses of various sizes and types are unsuccessful in obtaining financing. Our knowledge on these subjects are limited to occasional surveys containing incomplete information.”[1] The CFPB concludes that the current data is particularly incomplete in assessing the state of lending to minority-owned, women-owned, and other small businesses.

In order to further understand lending markets and boost lending to underserved small businesses, it is imperative to improve the quality of the publicly available small business loan data. As users of small business loan data, NCRC believes that data effectuates the purpose of Section 1071 by improving the efficiency and equity of lending markets. Data analysis identifies credit needs and gaps for different types of small businesses and therefore directs attention to further understanding and remedying these gaps. Data is also a great motivator for changing behavior. Data illuminates which financial institutions are effectively lending to traditionally underserved small businesses and which institutions lag their peers in serving these small businesses. It therefore motivates some leaders to maintain their lead in the small business lending market and the laggards to improve their performance. Publicly available data stimulates self-improvement efforts but when those efforts are not sufficient, data also enables public agencies to enforce fair lending laws in the cases of lending institutions that engage in discrimination and/or unfair and deceptive lending practices.

Data changes behavior by holding public and private sector entities publicly accountable. Writing nearly a century ago, the former Supreme Court Justice Louis D. Brandeis maintained that “Publicity is…a remedy for social and industrial diseases. Sunlight is said to be the best of disinfectants; electric light the most efficient policeman.” Justice Brandeis discussed how publicly available data on broker compensation enhanced the ability of investors to make prudent investment decisions and how data improved the ethics of public officials.[2] These insights apply directly to lending markets. When data is publicly available, responsible lending has surged to traditionally underserved borrowers. When data is absent or limited, a veil of secrecy clouds markets and facilitates declines in responsible lending and increases in abusive lending.

Small businesses play a vital role in the economy but difficulties accessing credit constrains their job creation potential. Small businesses employ about half the workers in the private sector and have created about two thirds of jobs since 1993.[3] In its working paper, the CFPB estimates that women-owned firms number almost 10 million and employ 8.4 million people.[4] Minority-owned firms total about 8 million businesses and employ 7.1 million people.[5] Studies have shown that access to credit enables small businesses to expand and employ more workers.[6] But as discussed below, barriers in access to credit have held back the growth of women- and minority-owned small businesses.

The experience of updating the Home Mortgage Disclosure Act (HMDA) data illustrates how enhancing the publicly available small business data will boost lending to underserved borrowers. After Congress updated HMDA data in 1989 to require information on the demographics of applicants, lending to minorities climbed in the 1990s before the advent of high-cost and abusive lending. For instance, from 1993 through 1995, conventional (non-government insured) home mortgage lending to African-Americans and Hispanics surged 70 percent and 48 percent, respectively. In contrast, the increase was just 12 percent for whites.[7]

After the lending increases in the 1990s, the home lending market adopted risky practices that the HMDA data was unable to monitor. High-cost and subprime lending riddled with obtuse and abusive terms and conditions surged in the early to mid-2000s. HMDA data could not keep track of these developments because it lacked robust data on pricing and loan terms and conditions. The Dodd-Frank Act remedied this deficiency in the HMDA data by requiring the addition of pricing and loan terms and conditions to the data. In future years, the enhanced HMDA data promises to enable members of the public and government agencies to more effectively monitor developments in the lending marketplace and to take corrective action to stem any abuses visible in the data.

The history of HMDA data informs the development of the small business loan data required by Section 1071. In addition to demographic information on applicants for small business loans, the data must contain carefully developed variables capturing pricing and key loan terms and conditions. NCRC believes it is possible to develop an effective database that motivates responsible lending at a low cost to financial institutions while preserving the privacy of borrowers.

Gender and Racial Barriers in Access to Credit and Capital

Small business is an engine for economic growth. However, our country’s ability to grow is slowed down by impediments to access to credit for small businesses. Research and first-hand experiences of NCRC and our member organizations show that women-, minority-owned, and very small businesses experience difficulties growing because of barriers in accessing credit.

Consider the following:

  • Women-owned firms are a significant force in the economy, with a higher rate of growth than other businesses, but they remain small. Ninety percent of women-owned small businesses have no employees other than the owner, a lower percentage than other small businesses.[8]
  • Part of the difficulty women face is a lack of credit and capital. On average, women start their businesses with half as much capital as men ($75,000 compared to $135,000). Also, just 5.5 percent of female-owned businesses use bank loans to start their businesses, compared to 11.4 percent of male-owned businesses.[9]
  • Women of color own business with lower average revenue than white women. The gap is largest for African-American women whose business average just $24,770 in revenue per year compared with $143,100 for all women-owned businesses.[10]
  • Small business ownership is an important wealth building strategy in minority communities, but minorities face difficulties starting and growing their small businesses. Non-minorities are twice as likely as minorities to own employer businesses (those with employees in addition to the owner). If minorities owned businesses at the same rate as non-minorities, our country would have one million additional employer businesses and more than 9.5 million additional jobs.[11]
  • Almost 60 percent of African-American entrepreneurs and 47 percent of white entrepreneurs did not seek small business financing because they thought they would not be approved by a lender. Minority-owned businesses are also more likely than white-owned businesses to report that their profits are negatively impacted by lack of access to credit. Less profits then translates into lower growth and fewer employers.[12]
  • NCRC research has found that the smallest businesses (those with assets less than $1 million) have the most trouble accessing credit. In 2010, just eight percent of these businesses received loans compared with about 20 percent of all businesses.[13]

Access to credit is critical for small businesses and regions of the country, yet inequalities in access to credit contribute to overall inequality.

  • NCRC found in a report conducted for the Appalachian Regional Commission that in 2010, the business lending rate in economically distressed counties in Appalachia was just 44 percent of the national rates.[14]
  • The Woodstock Institute found that in the Los Angeles and San Diego region, businesses in predominantly minority census tracts were 31.8 percent of businesses, but they received only 21.5 percent of the loans under $100,000. If those businesses had received loans in proportion to their share of businesses overall, they would have received 111,500 more loans amounting to $1.63 billion between 2012 and 2014.[15] In Illinois, for the period 2015 to 2017, 27.0 percent of business addresses were located in low- and moderate-income (LMI) census tracts, but they received only 19.3 percent of all CRA-reported loans under $100,000 and 17.8 percent of the total dollar amount of those loans. If Illinois businesses in LMI tracts had received CRA-reported loans under $100,000 in proportion to their share of business addresses overall, they would have received 46,648 more loans over the three-year period, totaling $618 million. Woodstock found similar disparities in tracts with more than 40 percent minorities and in some regions of the state like Moline-Rock Island and Peoria, the disparity between the share of loans and small businesses in minority tracts reached 15 percentage points.[16]

A variety of reasons exist for these racial, gender, and regional disparities. Some of it is due to the characteristics of the small businesses, such as lack of collateral or lower credit scores. Some of it is due to lack of access to bank branches. Research has demonstrated that higher numbers of bank branches in a geographical area boosts small business lending. Unfortunately, however, research also documents the persistence of lending discrimination.[17]

Recently, NCRC released a study called Disinvestment, Discouragement, and Inequity in Small Business Lending.[18] The study consisted of data analysis of lending trends in seven large metropolitan areas including Atlanta, Houston, Los Angeles, Milwaukee, New York, Philadelphia and Washington, D.C.

In addition, the study featured mystery shopping in Los Angeles in which fictitious African-American, Hispanic, and white male small business owners inquired about loans. NCRC’s testing focused on micro businesses, those with fewer than three employees, and smaller dollar loans, under $100,000, which are more likely to support smaller businesses and provide wealth-building opportunities. The profiles of all testers were sufficiently strong that on paper, all the profiles would qualify for loans. Furthermore, the black and Hispanic testers’ profiles were slightly better than their white counterparts in terms of income, assets and credit scores. For each test, three matched testers, one black, one Hispanic and one white visited the same retail bank branch location. NCRC selected 32 bank branch locations located in census tracts with less than 25% minority residents. A total of 180 interactions (tests) were compared for differences between testers. The purpose of the research was to determine the baseline customer service level that minority and non-minority testers received when seeking information about small business loans.

The NCRC study found from 2008 to 2016:

  • There were steep reductions in SBA 7(a) lending to black small business owners. This resulted in a decline from about 8% to 3% of loans during the Great Recession, a decline that has yet to recover.
  • Business owners in wealthier areas received the largest share of loans – 85% in Milwaukee. In fact, in six of seven metro areas analyzed, more than 70% of loans went to middle- and upper-income neighborhoods.
  • The number of bank branch locations declined 10% since 2009, likely affecting small businesses that are highly dependent on local-level banking relationships.
  • Banks have not reinvested the increased capital that they accumulated through deposits after the end of the Great Recession back into small businesses. The most significant difference between deposits and loans occurred in New York City metro area, where deposits increased by 100%, but lending decreased by nearly 40%.
  • There are tremendous gaps in black and Hispanic business ownership relative to their population size. Although 12.6% of the U.S. population is black, only 2.1% of small businesses with employees are black-owned. Hispanics are 16.9% of the population yet own only 5.6% of businesses.

Mystery shopping tests in 2018 revealed:

  • Bank personnel introduced themselves to white testers 18% more frequently than they did to black testers. White testers received friendlier service overall.
  • Black and Hispanic testers were requested to provide more information than their white counterparts, particularly personal income tax statements; Hispanic testers were asked to provide them nearly 32% and black testers 28% more frequently than their white counterparts.
  • While there could be legitimate business reasons for certain loans to require personal credit and financial information, the information requested from all potential well-qualified borrowers should be consistent. Bank representatives should not disproportionately tell Hispanic and black borrowers about the need for a credit report, inquire about credit card debt or request personal financial documents. There should be no reason to ask about a borrower’s education level (which occurred only for African-American testers) when applying for a loan for an established business. These actions can discourage borrowers from continuing the loan process and result in a fair lending violation for the banking institution.
  • White testers were given significantly better information about business loan products, particularly information regarding loan fees where white testers were told about what to expect 44% more frequently than Hispanic testers and 35% more frequently than black testers.
  • One area of customer service was significantly better for black and Hispanic testers – they received an offer to schedule an appointment to take their application more often, which happened 18% more frequently for black testers and 12% more often for Hispanic testers.

Multiple shortcomings with the existing small business loan data frustrate further investigation into the disparities uncovered by the NCRC report. The basic shortcoming is that the data collected per the Community Reinvestment Act (CRA) regulations do not contain the demographics of the small business borrower. In contrast to the CRA data, the SBA data does have information on the demographics of the borrowers, however, the small market share of SBA lending makes it impossible to use SBA data to assess the overall availability of credit to women- and minority-owned small business. In short, the publicly available data cannot be used to assess either overall access to credit for minority or women-owned business or to assess the performance of individual lenders. The research to date uses periodic national surveys or methodologies like mystery shopping. While valuable, this research cannot hold individual lenders or sub-groups of lenders like banks or fintechs accountable for their record in serving a diversity of small businesses.

It is precisely the data on individual lenders that has enabled HMDA data to foster increases in safe and sound lending to underserved populations. Individual lenders feel the impetus of public accountability. If they are not performing as well as their peers or if they are violating fair lending laws, they understand that an agency or a member of the public can call them to account for their shortcomings. No such powerful mechanism is stimulated by the current small business data. In fact, the lack of comprehensive data most likely significantly contributes to the persistent disparities and low levels of small business lending documented by NCRC’s studies.

Financial Institutions Engaged in Business Lending (Question 1, Panel 2)

Using authority granted in Section 1071, it is vital that the CFPB capture the vast majority of small business lenders in the Section 1071 data. If the CFPB omits lenders with a significant presence in the market, the data will be unreliable in that it is likely to over- or under-estimate the extent to which the lenders reporting data are serving various small businesses including minority- and women-owned small businesses. Also, the lenders omitted from the data could exploit the veil of secrecy and lack of agency oversight by peddling abusive loans to traditionally underserved small businesses.

In addition to the large banks currently reporting CRA small business data, the CFPB must require small banks, credit unions, non-banks, and on-line lenders to report data. Intermediate small banks (assets between $250 million and $1 billion) were previously required to report small business CRA data. These banks are especially important in rural communities and smaller cities. Using CRA data from 2003, one of the last years in which intermediate small banks reported data, NCRC estimated that these banks were between 15 to 20 percent of the market in Appalachian portion of states like Maryland and Virginia.[19] Also, the Federal Reserve survey reports that 42 percent of small businesses applying for credit applied to small banks.[20]

Likewise, on-line lenders are a significant force in the marketplace today and will likely increase their market share with each passing year. By 2020, Morgan Stanley forecasts online lenders or fintechs reaching $47 billion, or 16 percent of total U.S. small and medium enterprise approvals.[21] A recent Federal Reserve survey reports that the share of survey respondents applying for business loans from fintechs increased to 32 percent last year, up from 19 percent in 2016.[22] It is vital to capture fintechs in the data since the highest percent of applicants (53 percent) according to the Federal Reserve survey are unsatisfied with the high interest rates of their loans.[23] Finally, while credit unions have legal limits capping their small business lending, a number of credit unions have a significant presence in the small business lending marketplace. The Federal Reserve Banks find that 20 percent of business survey respondents with medium/high credit risk and with less than five years of operation sought financing from credit unions.[24]

The CFPB can strike a reasonable balance between the purposes of Section 1071 (of assessing credit needs and enforcing fair lending) and the objectives of reducing burden on smaller institutions. When the CFPB set the HMDA reporting threshold at 25 closed-end loans, it struck such a balance and exempted several lenders that made incidental amounts of home loans.[25] The CFPB can use the HMDA reporting thresholds as a guide to develop analogous thresholds that capture the vast majority of small business lending while exempting lenders that do not make meaningful amounts of small business loans.

Exempting certain lenders or loan types or purposes with sizable market share will provide a distorted view of the market and frustrate the purposes of Section 1071 to ascertain if credit needs are being met. For example, if the CFPB omits fintechs from data collection requirements, stakeholders will not be able to accurately estimate whether small businesses are receiving loans since fintechs most likely constitute about a third of the market as discussed above. Moreover, it will not be possible to determine with the Section 1071 dataset whether small businesses, especially the smallest and most disadvantaged, have access to reasonably priced loans or are disproportionately receiving higher cost financing offered by fintechs as suggested by the Federal Reserve survey.[26]

Just as exempting certain major lenders from data collection, omitting loan types and purposes would result in misleading Section 1071 data. For instance, merchant cash advance loans are higher cost and leaving these out of the data would frustrate the ability of stakeholders and enforcement agencies to monitor which lenders are offering responsible merchant cash advances and which ones are offering problematic cash advances. As discussed below, some form of pricing information is also imperative to ensure that vulnerable businesses are not subject to abuses with merchant cash advances or other forms of credit.

Data Points to Be Collected Must be Robust so Purpose of Section 1071 is Realized (Question 2 of Panel 2)

The purpose of Section 1071 is “to facilitate enforcement of fair lending laws and enable communities, governmental entities, and creditors to identify business and community development needs and opportunities of women-owned, minority-owned, and small businesses.” As the CFPB considers its discretionary authority to collect additional data, it must achieve the statutory purposes of the Section.

A robust enforcement of fair lending laws cannot rely solely on data regarding approval and rejection rates since discrimination and abusive lending often involve racial and gender disparities in pricing and loan terms and conditions. Additional characteristics of small business applicants need to be collected so fair lending analyses can determine if businesses of different races and genders are being treated similarly by lenders after controlling for size and capacity of the small business. Also, an analysis of whether lenders are legitimately meeting needs cannot rest solely on approval and rejection rates since lending needs to be safe and sound, affordable, and sustainable. Data on loan terms and conditions are imperative for a needs analysis since this data provides insights into whether lending is sustainable and affordable. Finally, data on different loan types and purposes is necessary to ascertain whether needs for various types of credit are being met.

The benefits will exceed the modest costs of collecting the discretionary data points. A number of these data points are likely to be collected in the underwriting process already such as whether the small business is pledging collateral. In addition, lenders must collect data as part of their compliance with the Equal Credit Opportunity Act (ECOA). NCRC therefore surmises that costs of small business loan data collection are modest and in line with the costs associated with data collection of Home Mortgage Disclosure Act (HMDA) data. The CFPB’s Notice of Proposed Rulemaking (NPR) regarding reporting thresholds in HMDA data calculated that most smaller lenders would save between $2,000 to $5,000 if they were exempt from either closed-end or open-end reporting.[27] Since costs are typically the highest for smallest lenders compared to their assets, these costs are indeed modest and would likely mirror the costs for collecting small business loan data.

While there are costs, bank behavior towards data disclosure illustrates a pragmatic acknowledgement of the benefits of data disclosure. In 2017, 177 banks had assets below the threshold that would require them to submit CRA data on small business loans. These 177 smaller banks voluntarily reported either because they wished to be evaluated on the large bank CRA exam or they saw some other benefit to data disclosure. Perhaps, they wanted to compare their performance with peers, in order to figure out how to be more competitive and/or to find some overlooked and profitable lending opportunities. Perhaps, they believed their performance was commendable and they wanted to disclose their data as a means of bolstering their reputations as responsive to the needs in their communities.[28] Overall, the benefits of data disclosure exponentially outweigh the modest costs. Data disclosure promotes more efficient and equitable markets, enabling more small businesses to acquire loans, expand their size, and create more jobs.

The NCRC comment regarding the CFPB’s request for information covered the methods and benefits of reporting both the mandatory data such as race and gender and discretionary data points.[29] This statement focuses on the discretionary data points in response to question two for panel two.

Employee Size

Employee size is a critical data field, which is often used to classify small businesses. Larger firms as measured by number of employees generally tend to be more successful in the credit markets. In a report conducted for the Appalachian Regional Commission, NCRC found that the number of loans fell in counties with higher numbers of the smallest businesses with 1 to 4 employees and rose in counties with higher numbers of businesses with 10 to 19 employees.[30] For these reasons, number of employees is a vital data point and should not be too burdensome for lenders to collect and report (the Census Bureau could be consulted as to how it collects employee information for its surveys of small businesses).

Years in Businesses

Start-ups and younger firms encounter more difficulties in accessing credit than more established businesses.[31] The Federal Reserve Banks report that their 2016 survey revealed that start-up firms have greater difficulty accessing credit than mature firms and that they are less likely than mature firms to receive the full amount of credit requested.[32] Therefore, years in business is an important control variable for explorations of gender and racial disparities in access to credit. It should also be relatively easy for lenders to collect and report. Related to years in business in terms of the sophistication of the business is the structure of the small business such as whether it is a sole proprietorship or a corporation. The CFPB should explore whether it is a straightforward exercise to add a categorical variable describing the structure of the small business.

Collateral  

A lack of collateral has been a barrier for minority- and women-owned businesses in obtaining small business loans. Collateral in small business lending is akin to downpayment requirements and savings lenders use to assess borrower capacity to repay in home lending. Collateral therefore is an important factor for Section 1071 data to capture. Options for the CFPB to consider is a categorical variable consisting of a “Yes” or “No” if collateral was pledged as part of the loan application, a variable capturing types of collateral, and/or a variable capturing the dollar amount of collateral.

Pricing information

The annual percentage rate (APR) expressed as a difference between the APR on the loan and an average APR obtained via surveys will be a critical variable in HMDA data for all loans. The CFPB must determine whether pricing information can feasibly be collected on small business loans given concerns and evidence of racial and ethnic pricing disparities. In a paper published by the Harvard Business School, Mills and McCarthy endorse price reporting for business loans.[33] Pricing differences between bank and on-line lenders are also common.[34] Thirty-three percent of respondents to a Federal Reserve survey reported that they were not satisfied with interest rates of on-line lenders compared to just three percent and six percent reporting dissatisfaction with small bank and large bank interest rates, respectfully.[35] Presumably, the higher rates of dissatisfaction associated with interest rates on the loans of on-line lenders corresponds to higher interest rates than those on bank loans. In addition, credit card lending is higher interest rate than term lending and is used disproportionately by minority-owned businesses.[36] Credit card rates average around 12.85 percent in comparison to 5 or 6 percent that is traditional for small business loans. Small businesses credit card spending rose by $215 billion between 2006 and 2015.[37]

As described in NCRC’s white paper, the Community Development Financial Institutions (CDFI) Fund has had experience collecting price information for small business loans made by CDFIs.[38] Section 1071 data will be incomplete if it does not report on the pricing offered by various types of lenders. Price disclosure, in and of itself, will most likely discourage extreme differences in pricing, particularly by race or gender, unrelated to creditworthiness of borrowers.

Definition of Small Business Must Focus on the Smallest Businesses (Question 5 of Panel 2)

NCRC recommends that the CFPB’s definition of small business cover most firms in the market and that the definition does not entail burdensome and hairsplitting decisions for lenders about when to report data. Since the vast majority of businesses in the United States are small, an expansive definition of small businesses is necessary so that the data captures their experiences in the lending marketplace. As the CFPB documents, about 20 million firms or 76 percent of all firms have annual receipts of under $100,000. An additional 5.2 million or 19 percent of all businesses have receipts between $100,000 and $999,999.[39] Together these two categories of businesses contain 95 percent of all businesses in the United States. At a minimum, the definition of small businesses must include businesses with revenues or receipts under $1 million so that policymakers and members of the public can determine how these businesses are faring in the lending marketplace. Moreover, there must be enough detail in the data enabling stakeholders to determine whether there are differences in the experiences of businesses in various revenue size categories under $1 million.

Defining a small business as one with revenues below $1 million has ample precedent and would facilitate data analysis. The Community Reinvestment Act (CRA) requires banks with assets above $1 billion to report loans made to small businesses defined as those with revenues less than $1 million.[40] This reporting requirement has been in effect since 1996 for a wide swath of the banking industry. It is thus a well-accepted procedure and can be adopted readily by other lenders.

CRA also defines a small business loan as one below $1 million.[41] This makes sense as it is unlikely that a small business with relatively few employees and small revenue size would seek loans above $1 million. Again, this is a standard that has been in effect for several years and could be easily applied to other lenders besides large banks.

The CFPB’s request for information discusses size standards that varies by sector of business. The size standards include different number of employees corresponding to different sector categories of small businesses grouped by North American Industry Classification System (NAICS).[42] While the NAICs categories have important applications in federal contracting and other purposes, it would be cumbersome to include NAICs categories as determining whether businesses within a sector category is small or not for data reporting purposes. It could significantly increase the time lenders would need to spend parsing the data without any significant benefit arising from the parsing.

Using NAICs codes established pursuant to Section 3 of the Small Business Act would also dilute the focus on access to credit for the smallest businesses. The NAICs codes include businesses with thousands of employees and millions of dollars in revenue. For instance, businesses in tortilla or soft drink manufacturing can have 1,250 employees and qualify as small businesses. Masonry contractors can have revenues as high as $16.5 million and qualify as small businesses.[43] Using NAICs codes to establish a definition of small business would counter the purpose of Section 1071 which is to assess whether the smallest businesses, most disadvantaged businesses, and minority- and women-owned businesses are receiving loans.

A preferable approach to using NAICs codes and number of employees for defining small businesses is to include both items in the Section 1071 as data fields. A data field can capture the NAICs code for a small business applicant and another data field can record the number of employees of the business. If data fields included NAICs codes and number of employees, analysts could then determine any differences in loan approval rates, dollar amounts of loans, or other variables of interest for businesses in different NAICs categories or number of employees.

Privacy Interests can be Readily Protected Under a Regime of Robust Data Disclosure (Question 3 of Panel 2)

The CFPB is charged with balancing the benefits of data disclosure against the risks of privacy invasions and identification of specific borrowers. While not minimizing the seriousness of this issue, NCRC notes that in the more than 40 years of HMDA data reporting, the federal regulatory agencies have not reported any instances of privacy invasions in either Congressional testimony or other official reports or documents. Likewise, NCRC believes that Section 1071 can include robust data disclosure without risking privacy invasions.

Neither HMDA data nor the Section 1071 data pose significant privacy concerns because the data does not have personally identifiable information that bad actors can use to target specific borrowers. In contrast, the breach of Equifax data imperiled the privacy of up to 143 million consumers precisely because the Equifax data contains personally identifiable information such as Social Security numbers, street addresses, and birth dates. Large numbers of private sector companies including lenders use Equifax data to market products to consumers and borrowers.[44]

Concerns have been expressed that lending data enables predatory marketing. In this scenario, bad actors would use the publicly available data to identify vulnerable borrowers or applicants that have been denied loans and then peddle abusive products to them. In the HMDA context, these claims have been far-fetched. A veil of secrecy or a lack of data on the extent of abusive home lending was a major reason predatory lending was able to proliferate. If the Dodd-Frank HMDA enhancements had been implemented sooner, the public and regulatory agencies could have used the data to identify surges in problematic lending and taken steps to deter abusive lending before it reached crisis proportions. Thus, data disclosure is a tool to deter predatory practices, not enable them. Also, just as with HMDA data, NCRC believes that the CFPB can take steps to protect against any possibilities of predatory marketing resulting from privacy invasions.

As discussed in NCRC’s response to the CFPB’s Request for Information, increasing the number of small business loan reporters could significantly increase the number of loans on a census tract level, which would make it harder for bad actors to identify local businesses receiving loans. In addition, the CFPB should consider using data swapping to protect the privacy of loan applicants whose demographic data points are particularly unusual. Following the methodology used by the U.S. Census in its Public Use Microdata Sample, the CFPB could switch records for similarly-situated but demographically unusual loan applicants between nearby census tracts, making it nearly impossible to reconnect individual loan applicants with public records, but maintaining the utility of Section 1071 data including for users doing analysis at the neighborhood level.[45]Additional masking techniques include making some variables categorical variables instead of reporting specific values. For example, revenue size of the small business could be reported in categories rather than specific revenue amounts.

Considerations for Defining Application, Type and Purpose of Loan, and Protections in Underwriting Process (Question 4 of Panel 2)

The CFPB issued useful guidance to borrowers of home loans about the requirements for an application that can help inform how to define an application for small business borrowing purposes. In the guidance for home loan borrowers, the CFPB advises consumers to obtain a good faith estimate of loan costs[46]. Then a consumer needs to formerly request that the lender start the application process, pay application fees, and supply various financial information. It would seem that a similar application process would apply for a small business loan. Moreover, the CFPB should consult the Small Business Administration (SBA) regarding the SBA’s application process. The definition of an application must be crafted in such a manner that the borrower has protections associated with the Equal Credit Opportunity Act (ECOA).

The CFPB needs to define types and purposes of loans in such a manner so as to ensure that Section 1071 covers the entire lending marketplace. As mentioned above, if significant gaps in coverage of the market occur, the purposes of the statute are violated since the gaps will frustrate fair lending enforcement and prevent an accurate assessment of whether credit needs are being met. The Federal Reserve Bank survey reports that respondents used loans or lines of credit 55 percent of the time, credit cards 52 percent of the time, trade credit 13 percent of the time, leasing 9 percent of the time, merchant cash advance 6 percent of the time, and factoring 3 percent of the time.[47]Omitting any of these significant sources of credit will create misleading data regarding fair lending and needs assessments.

It is possible for smaller lenders that loan underwriters may collect data required by Section 1071. The statute stipulates that in these cases the borrower is supplied a notice that the underwriter has access to the information required by Section 1071 and that the lender cannot discriminate on the basis of such information. It would seem that such a notice to a borrower would be sufficient to inform him or her of his or her rights and recourse under ECOA. A model notice could be helpful. Similar notices about the use of HMDA data for fair lending enforcement are provided to borrowers of home loans.

Process and Technology for Facilitating Data Submission (Question 6)

The new HMDA submission process developed by the CFPB in 2018 replaced the antiquated systems of previous years.  Now, lenders can create a Loan Application Record (LAR) submission file using normal desktop software and upload it to a CFPB secured portal.  For larger lenders, specialized HMDA compliance software is used to create the file, check it for errors, and upload it to the CFPB.  The CFPB has developed training materials and offers support to help users with their submission.

A step-by-step guide with a template file would facilitate submission of the Section 1071 files. For financial institution staff not familiar with the pipe delimited format used for HMDA text files, the CFPB should develop excel file templates. Excel is one of the most heavily used software programs and is readily understandable since columns correspond to data points and rows correspond to data for individual applications. Also, the general public can download HMDA data into excel so using excel files for data submission for Section 1071 can likewise make submission easier. For each data point, a guidebook can describe whether it is numeric, categorical, or alphanumeric. It can also describe how to deal with any complications or close calls associated with a data point. Clear data submission materials should also be provided to the general public so the general public can more readily understand the Section 1071 data.

Conclusion

Section 1071 data is vital for promoting fair lending and increasing access to responsible lending for minority-owned, women-owned, and other small businesses. By holding institutions accountable, public data disclosure effectively motivates financial institutions to increase their lending to traditionally underserved small businesses. Robust data also more precisely identifies the causes of unmet credit needs and therefore helps stakeholders to more efficiently craft policies and programs to meet the unmet needs. The above recommendations provide detail on how a robust database can be developed to meet the purposes of Section 1071 while minimizing burden. NCRC urges the CFPB to expeditiously proceed with this important rulemaking.


 

[1] Consumer Financial Protection Bureau (CFPB), Key Dimensions of the Small Business Lending Landscape, 39-40,  May 2017, https://www.consumerfinance.gov/data-research/research-reports/key-dimensions-small-business-lending-landscape/

[2] Other People’s Money, from Writings by Louis Brandeis, collected by University of Louisville,  https://louisville.edu/law/library/special-collections/the-louis-d.-brandeis-collection/other-peoples-money-chapter-v

[3] JPMorgan Chase Institute, The Ups and Downs of Small Business Employment (January 2017), available at https://www.jpmorganchase.com/corporate/institute/report-small-business-payroll.htm, U.S. Small Business Administration, Frequently Asked Questions, March 2014, available via https://www.sba.gov/sites/default/files/FAQ_March_2014_0.pdf

[4] Consumer Financial Protection Bureau (CFPB), Key Dimensions of the Small Business Lending Landscape, 13,  May 2017, https://www.consumerfinance.gov/data-research/research-reports/key-dimensions-small-business-lending-landscape/

[5] CFPB, Key Dimensions, 15.

[6] Luz Gomez and Elaine L. Edgcomb, FIELD at the Aspen Institute, A Newly Crowded Marketplace: How For Profit Lenders are Serving Microentrepeneurs, 2011, available at http://fieldus.org/publications/ForProfitLenders.pdf

[7] NCRC, Home Loans to Minorities and Low- and Moderate-Income Borrowers Increase in the 1990s, but then Fall in 2001: A Review of National Data Trends from 1993 to 2001. Available upon request.

[8] Women-Owned Businesses: Carving a New American Business Landscape, U.S. Chamber of Commerce Foundation, https://www.uschamberfoundation.org/sites/default/files/Women-Owned%20Businesses%20Carving%20a%20New%20American%20Business%20Landscape.pdf

[9] National Women’s Business Council – Problem: Women Entrepreneurs Need Greater Access to Capital, https://www.nwbc.gov/sites/default/files/Access%20to%20Capital%20Fact%20Sheet.pdf

[10] The 2018 State of Women Owned Businesses Report, Commissioned by American Express, p. 5, https://about.americanexpress.com/files/doc_library/file/2018-state-of-women-owned-businesses-report.pdf

[11] Kauffman Foundation, State of Entrepreneurship 2017: Zero Barriers – Three Mega Trends Shaping the Future of Entrepreneurship, http://www.kauffman.org/~/media/kauffman_org/resources/2017/state_of_entrepreneurship_address_report_2017.pdf

[12] Kauffman Foundation, ibid.

[13] NCRC, Access to Capital and Credit in Appalachia and the Impact of the Financial Crisis and Recession on Commercial Lending and Finance in the Region, commissioned by the Appalachian Regional Commission, July 2013, http://www.ncrc.org/images/accesstocapitalandcreditInappalachia.pdf

[14] NCRC, ibid.

[15] Woodstock Institute, Patterns of Disparity: Small Business Lending in the Chicago and Los Angeles-San Diego Regions, January 2017, http://www.woodstockinst.org/research/patterns-disparity-small-business-lending-chicago-and-los-angeles-san-diego-regions

[16] Woodstock Institute, Patterns of Disparity: Small Business Lending in Illinois, August 2019, https://woodstockinst.org/wp-content/uploads/2019/08/Patterns-of-Disparity-Small-Business-Lending-in-Illinois.pdf

[17] NCRC’s report for the Appalachian Regional Commission has a chapter summarizing the research on disparities in small business lending.

[18] NCRC, Disinvestment, Discouragement, and Inequity in Small Business Lending, Summer 2019,  https://ncrc.org/disinvestment/

[19] NCRC, Access to Capital and Credit for Small Businesses in Appalachia, May 2007, 40.

[20] Federal Reserve Banks, Small Business Credit Survey, April 2017, 14

[21] Karen Gordon Mills and Brayden McCarthy, The State of Small Business Lending: Innovations and Technology and the Implications for Regulation, Harvard Business School, 48, 2016.

[22] Federal Reserve Banks, Small Business Survey, 2019, executive summary, https://www.fedsmallbusiness.org/medialibrary/fedsmallbusiness/files/2019/sbcs-employer-firms-report.pdf

[23] Federal Reserve Banks, Small Business Survey, p. 20.

[24] Federal Reserve Banks, Small Business Credit Survey: Report on Startup Firms, August 2017, 15.

[25] The Bureau estimates that the new threshold of 25 closed-end mortgage loans will reduce the number of reporting institutions by approximately 1,400. See Federal Register, Vol. 80, No. 208, Wednesday, October 28, 2015, Table 5 at 66279, https://www.gpo.gov/fdsys/pkg/FR-2015-10-28/pdf/2015-26607.pdf

[26] Small businesses applicants posing medium or high credit risk were more likely to apply to fintechs according to the Federal Reserve survey, p. 16.

[27] CFPB, Proposed Rule, Docket No. CFPB-2019-0021, pp. 119-123, original double-spaced version posted on agency website, https://files.consumerfinance.gov/f/documents/cfpb_nprm-hmda-regulation-c.pdf

[28] See FFIEC report on CRA lending, https://www.ffiec.gov/hmcrpr/cra_fs18.htm

[29] See NCRC comment in response to CFPB RFI, September 2017, https://ncrc.org/wp-content/uploads/2017/10/ncrc_comment_re._request_for_info_on_small_bus_market2.pdf

[30]  NCRC, Access to Capital and Credit for Small Businesses in Appalachia, May 2007, 74 and 75, prepared for the Appalachian Regional Commission, http://www.ncrc.org/images/stories/mediaCenter_reports/ncrc%20study%20for%20arc.pdf

[31] NCRC, Small Business Loan Data: Recommendations to the CFPB, 32.

[32] Federal Reserve Bank of New York, Small Business Credit Survey: Report on Startup Firms, August 2017, iv, https://www.newyorkfed.org/medialibrary/media/smallbusiness/2016/SBCS-Report-StartupFirms-2016.pdf

[33] Karen Gordon Mills and Brayden McCarthy, The State of Small Business Lending: Innovations and Technology and the Implications for Regulation, Harvard Business School, 2016, 19, http://www.hbs.edu/faculty/Publication%20Files/17-042_30393d52-3c61-41cb-a78a-ebbe3e040e55.pdf

[34] Eric Weaver, Gwendy Donaker Brown, Caitlin McShane, Unaffordable and Unsustainable: the New Business

Lending on Main Street,  Opportunity Fund, May 2016 documents the high cost of on-line fintech lending.

[35] Federal Reserve Banks, Small Business Credit Survey, April 2017, 17, https://www.newyorkfed.org/medialibrary/media/smallbusiness/2016/SBCS-Report-EmployerFirms-2016.pdf

[36] Small Business Administration, Office of Advocacy, September 2011, Frequently Asked Questions about Small Business Finance, https://www.georgiasbdc.org/pdfs/SBAFinanceFAQ.pdf, 4.

[37] Ruth Simon, Big Banks Cut Back on Loans to Small Business, The Wall Street Journal; citing PayNet studies on Small Business Lending, available at https://www.wsj.com/articles/big-banks-cut-back-on-small-business-1448586637

[38] NCRC, Small Business Loan Data: Recommendations to the Consumer Financial Protection Bureau for Implementing Section 1071, 31.

[39] CFPB, Key Dimensions, 10.

[40] CRA regulation, see § 345.42 (b)(iv) Data collection, reporting, and disclosure, available via https://www.fdic.gov/regulations/laws/rules/2000-6500.html

[41] CRA regulation defines a small business loan according to the instructions for preparation of the Consolidated Report of Condition and Income. See the CRA regulation at §345.12 Definitions (v), available via https://www.fdic.gov/regulations/laws/rules/2000-6500.html, and Instructions for Preparation of Consolidated Report of Condition and Income, RC-C-30, available at https://www.ffiec.gov/pdf/ffiec_forms/ffiec031_041_200503_i.pdf

[42] CFPB, Key Dimensions, 6.

[43] See table “Small Business Size Standards by NAICs Industry,” in §121.201   What size standards has SBA identified by North American Industry Classification System codes?  https://www.ecfr.gov/cgi-bin/text-idx?SID=b919ec8f32159d9edaaa36a7eaf6b695&mc=true&node=pt13.1.121&rgn=div5#se13.1.121_1201

[44] Washington Post, Equifax Breach Hits Credit Data of Millions, September 6, 2017, https://www.washingtonpost.com/news/the-switch/wp/2017/09/08/after-data-breach-equifax-asks-consumers-for-social-security-numbers-to-see-if-theyve-been-affected/?hpid=hp_hp-more-top-stories_equifax-1145am%3Ahomepage%2Fstory&utm_term=.bf6cc31f537d

[45] http://www.stat.cmu.edu/~fienberg/DLPapers/Fienberg-McIntyre-LNCS-2004.pdf

[46] CFPB September 2017, https://www.consumerfinance.gov/ask-cfpb/my-loan-officer-said-that-i-need-to-express-my-intent-to-proceed-in-order-for-my-mortgage-loan-application-to-move-forward-what-does-that-mean-en-1989/

[47] Federal Reserve Banks, Small Business Survey, 2019, p. 7. Percentages exceed 100 percent because borrowers can use more than one source of credit.

Print Friendly, PDF & Email