NCRC’s congressional testimony on fintech oversight
Statement for the record from John Taylor, president and CEO, of the National Community Reinvestment Coalition to the House Financial Services Subcommittee on House Financial Services Committee, Subcommittee on Financial Institutions and Consumer Credit
January 30, 2018
Chairman Luetkemeyer, Ranking Member Clay, and Distinguished members of the Subcommittee:
The National Community Reinvestment Coalition (NCRC) appreciates the opportunity to provide this written statement for the record of the January 30, 2018 hearing on, “Examining Opportunities and Challenges in the Financial Technology (“Fintech”) Marketplace.”
Because financial technology companies (fintechs) have significantly increased their market presence over the last several years, it is imperative to apply rigorous Community Reinvestment Act (CRA) and fair lending standards on them. The lessons of the financial crisis are clear: when one part of the market is required to comply with comprehensive regulations and the other part of the market does not, the un-regulated part of the market will compete by offering abusive and risky loans. Extensive research by Federal Reserve economists and academics demonstrates that CRA-regulated banks offered safe and sound loans to underserved populations while non-CRA covered independent mortgage companies offered high cost and predatory loans with high default rates. In order to avoid repeating the same mistakes of uneven regulation, federal agencies must apply comprehensive consumer protection regulations to fintechs now as they are increasing their market share.
A lack of data makes it difficult to know with precision the market share of fintechs. What is clear, however, is that fintech lending is increasing rapidly. Recent research documents that fintech consumer lending amounted to $21 billion in 2016, an increase of 17 percent from the previous year. According to a recent Federal Reserve survey, about 21 percent of small businesses with employees applied to fintechs for loans. Although a significant number of small businesses applied to on-line lenders, it is unclear how many eventually received loans from fintechs. One research report estimates that 143,344 small businesses received loans from fintechs during 2016, which is only about two percent of small business loans (6,106,355) reported by banks that year. This estimate of low market share, however, is not a reason for complacency among banks or regulators since it is apparent that fintech lending has been increasing significantly. Even a two percent market share can become 10 to 20 percent over the next decade.
While fintech lending has surged, evidence suggests that a significant amount of is likely to be abusive. Consumer and small business borrower satisfaction with fintechs is currently low because of opaque and unclear disclosure of loan terms and conditions and high costs. Accion Chicago, a Community Development Financial Institution, reports that 20 percent of its customers are seeking relief from problematic loans, many of which were made by fintechs. Likewise, Opportunity Fund (based in California) found that a large sample of loans from fintech lenders featured high Annual Percentage Rates (APR) and unaffordable monthly payments. Pricing differences between bank and on-line lenders are also common. 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. 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.
As well as guarding against high cost or abusive products, the regulation of fintechs presents an opportunity to address the digital divide. The Federal Communications Commission’s (FCC’s) 2016 Broadband Progress Report concluded that 10% of Americans across the country still lack access to adequate broadband Internet service. While only 4% of people living in urban areas lack adequate broadband services, this issue is particularly concentrated in rural areas and tribal lands, with 39% and 41% respectively, still lacking access. It is important for fintechs to acknowledge this “digital divide” and work toward closing the divide through their charter applications. The Federal Reserve Bank of Dallas extols using CRA as an effective means of closing the digital divide and providing low- and moderate-income individuals access to safe and sound banking products. Recently, the Federal Reserve Bank of Philadelphia released a working paper on fintech lending and found that fintechs can play a major role in filling a lending gap when bank branches close. In examining one online lender, the paper found that more than 75 percent of its newly originated loans in 2014 and 2015 were in areas where local bank branch locations were declining.
If executed thoughtfully, fintech entrance into banking could help narrow the digital divide by providing underserved populations and communities with increased virtual access to banking. However, if charter applications and CRA examinations of fintechs do not meaningfully assess whether fintechs are in fact narrowing the digital divide the result could be the exact opposite. The digital and banking divide would then be exacerbated and a great opportunity to narrow it would be missed.
Application of CRA
CRA requires that banks serve credit needs of low- and moderate-income communities consistent with safety and soundness. Since CRA is an affirmative obligation to lend responsibly, applying CRA or CRA-like requirements to fintechs will help curb abusive lending while also motivating fintechs to make more loans, investments, and services in low- and moderate-income communities. Three aspects of CRA need to be considered for fintechs: the role of CRA in the chartering process, ongoing CRA exams, and assessment areas. Each will be now considered in turn.
Chartering Fintechs – FDIC
During the summer and fall of 2017, two fintechs, Social Finance (SoFi) and Square, applied to the FDIC for an industrial loan charter (ILC). An industrial loan charter allows a depository institution to be owned by a non-bank. ILCs also typically have narrow product lines. NCRC opposes ILCs on safety and soundness grounds since federal regulatory agencies lack the authority to supervise ILC parents, which can include non-bank companies. During the financial crisis, two ILCs, Security Savings Bank, based in Nevada, and Advanta Bank Corp, based in Utah failed. In addition, a number of parents of ILCs, including Lehman Brothers, General Motors, Flying J Inc., Capmark Financial Group Inc., CIT Group Inc., and Residential Capital, LLC filed for bankruptcy.
If SoFi and Square had applied for regular depository charters, NCRC would not oppose the applications on safety and soundness grounds but would insist on rigorous CRA plans. The fintech applications to-date do not display full adherence to the FDIC’s chartering requirements of serving community needs. The FDIC’s Statement of Policy indicates that criteria for approval of charter applications includes serving convenience and needs. The Statement describes convenience and needs factor as the following:
The essential considerations in evaluating this factor are the deposit and credit needs of the community to be served, the nature and extent of the opportunity available to the applicant in that location, and the willingness and ability of the applicant to serve those financial needs.
SoFi did not indicate a willingness to employ its ability or business niche to serving the needs of communities. The CRA plan that SoFi described in its application was offering a secured credit card, purchasing affordable housing bonds from the Utah Housing Corporation (UHC), making vague promises of financial education and counseling, and providing a scholarship program offering few scholarships. SoFi’s business niche is to refinance student loan debt and SoFi markets itself to millennials. However, the CRA plan did not include marshalling this expertise to refinancing the student debt of lower income students, many of whom are burdened by high debt levels.
Instead the main retail product SoFi’s CRA plan envisioned offering lower income consumers was a secured credit card that is a higher interest rate product inferior to regular credit cards. The SoFi application states that “it is felt that revolving credit cards are not an appropriate credit instrument for a lower income community focus.” The notion expressed in the SoFi application that lower income customers can only handle secured credit cards rests on untested stereotypes of these customers as credit risks. This was emblematic of SoFi’s entire application which did not involve careful research documenting credit needs of low- and moderate-income consumers and how best to meet those credit needs. Moreover, SoFi’s geographical focus in its CRA plan was the Salt Lake City-Provo-Orem, Utah metropolitan area even though SoFi was a national lender (more on assessment areas below).
Square’s application was an improvement over SoFi’s. Importantly, Square indicated a desire to serve lower income, minority, and women businesses because small businesses are the main customers of Square. Square is a payment processor and has also started making small business loans based on data it obtains on processing payments. In its draft CRA strategic plan, Square outlines a numerical goal for community development finance expressed as a percent of assets. It also indicates that it will offer financial education on a national basis, provided it has met the needs of its assessment area.
While an improvement over SoFi’s application, Square did not establish goals for serving low- and moderate-income borrowers and communities for its two main products – small business lending and payment processing. Square took an initial step with small business lending, saying it would establish performance measures including the percent of small business loans in low- and moderate-income census tracts. Goal setting, however, would go further than this. After calculating the performance measures, a lender would compare itself against its peers and establish goals such as meeting or exceeding peer performance in the percentage of loans in low- and moderate-income communities. Lastly, Square also restricts its assessment area to the Salt Lake City, Utah metropolitan area.
OCC – Charter Applications and Proposed Charter for Fintechs
The OCC, like the FDIC, received a fintech application for a bank charter during the summer of 2017. Varo, an on-line company that focuses on deposit products and financial management services for consumers, applied for a bank charter to the OCC in the summer. Varo’s application exhibited shortcomings similar to those described above.
In addition to Varo’s application, the OCC undertook a significant endeavor spearheaded by former Comptroller of the Currency Thomas Curry to establish a national charter for non-bank fintechs. This proposal was multifaceted, complex, and controversial. For starters, it would allow non-banks with a national charter to preempt state law. NCRC has steadfastly opposed proposals to expand preemption powers to additional lending institutions. At a minimum, NCRC stated in comments about the OCC’s fintech charter proposal that fintechs should be required to demonstrate how they would either comply with state consumer protection law or adhere to the basic protections (such as clear disclosures to consumers) embodied in the state laws.
The OCC did not establish a CRA requirement for fintechs since fintech charters would not be charters for depository institutions. The OCC, however, did indicate that fintechs applying for an OCC charter would be required to develop financial inclusion plans stating how the fintech would serve underserved communities. The financial inclusion plans would represent an advance over CRA in one sense in that plans for serving minorities could be included. The OCC draft licensing manual for fintechs did not explicitly say “minority” but referenced underserved communities which presumably could include minority communities. CRA, in contrast, only describes obligations to low- and moderate borrowers and communities.
The financial inclusion plans would require fintechs to describe goals, establish performance measures, and milestones. The OCC states that community group input can help fintechs identify community credit needs. The fintech is to identify the geographical markets it will operate in, including underserved populations and geographical areas.
NCRC agreed with the OCC’s interpretation that the National Banking Act (NBA) provides the OCC with authority to implement CRA-like obligations for non-depository fintech companies. The NBA describes the procedure for chartering new banks and financial institutions, including the criteria to which the proposed charter must adhere.
The implementing regulations of the NBA, 12 CFR § 5.20, describe a number of community reinvestment and fair access considerations and requirements. In a subsection called “requirements,” the OCC states that it will assess if a need exists for the proposed institution in the community to be served and “whether there is a reasonable probability” of the institution’s “usefulness.” Whether a company is useful could be judged in part on the extent to which it will serve community credit needs. Another subsection (f) of § 5.20 called “policy” makes the reference to serving the community rather explicit. It states that a chartering consideration is whether the proposed institution will provide “fair access to financial services by helping to meet the credit needs of its entire community,” and whether the institution would promote “fair treatment of customers.”
Precedents for CRA and Financial Inclusion Plans
Additional models for financial inclusion plans are the OCC’s conditional merger approvals requiring CRA plans in the cases of Valley National Bank and Sterling Bank. These conditional merger approvals required marketing and outreach efforts which insured that low- and moderate-income consumers and communities were served in a fair and non-discriminatory manner. The CRA plans must also include annual goals and timetables and annual reporting to the OCC. The banks were required to also seek public input when developing their CRA plans.
More recently, NCRC has established community benefit agreements with several lending institutions including KeyBank and Santander establishing lending and investment goals to reinvest tens of billions of dollars in underserved communities. These community benefit agreements were often negotiated while a merger application was pending or by a bank desiring to improve its CRA performance. The performance measures committed to in these plans are similar to the ones discussed above and will result in improved performance and increases in lending and investing in future years. The community benefit agreements also involved meetings around the country with hundreds of community organizations to discuss how the banks could best respond to community needs. Many of the features of the agreements reflect commitments to address needs such as needs for low balance mortgage loans in areas of the country with depressed home values. The agreements also feature a monitoring mechanism consisting of advisory councils composed of community groups that review bank progress under the agreements and provide recommendations about how to improve performance.
Recommendations for Chartering Fintechs
As stated above, NCRC opposes the ILC charter and encourages fintechs and federal regulatory agencies to use charter authority that does not preempt state law and requires fintechs to adhere to comprehensive standards for serving communities in a safe and sound manner.
The concept of CRA strategic plans or financial inclusion plans as part of fintech applications is valuable and must be improved. Since fintech business models vary widely, a plan submitted as part of a charter application can allow for flexibility to accommodate the differences in the business models. For example, a plan for a payment processor will be different than a plan for a consumer lender. However, while allowing for flexibility, the plans must be rigorous with strong performance measures and geographical coverage (more on that below). The performance measures must not simply list numbers of loans or other products but include comparisons with industry peers so the public knows whether the fintech is proposing to be at or better than the level of its peers. The plans must also respond thoughtfully to community needs as identified through data analysis and discussions with community organizations.
A recent change in the Interagency Questions and Answers Regarding CRA (Interagency Q&A) can help inform the development of performance measures. The Interagency Q&A advises that CRA examiners will scrutinize whether a financial institution’s alternative delivery systems are effectively delivering services to low- and moderate-income populations by considering a variety of factors including: ease of access; cost to consumers; range of services delivered; ease of use; rate of adoption and use; and reliability of the system. Finechs should establish specific performance measures and goals for the low- and moderate-income community for each of these factors.
The FDIC’s language about ability and willingness of the applicant to serve community needs should guide the development of CRA and financial inclusion plans. NCRC has suggested a guideline in our comments on fintech charter applications of whether the fintech applications display a willingness to use their abilities to serve needs. In particular, the fintech should use its talent or expertise to develop products and programs that serve the needs of underserved populations. For example, a consumer lender refinancing student debt should not receive approval for a CRA plan that only offers secured credit cards to low- and moderate-income borrowers, and does not include this population in its plans to refinance student debt.
Describing a CRA or financial inclusion plan in a charter application is necessary but not sufficient for ensuring continued robust performance in meeting the needs of underserved communities in the future. To better ensure consistent and improving performance, federal examiners must periodically conduct evaluations and then provide written assessments of performance.
In the case of fintechs applying for depository charters, the federal agencies would conduct CRA evaluations. The fintechs would most likely opt to be evaluated under the strategic plan option. A strategic plan’s time period cannot exceed five years, and a bank generally develops measurable goals for lending, investing, and services. Public input is required. In the case of fintech charter applications, the public input occurs at the time of application. After the agency approves the plan, the fintech would operate under the strategic plan. Its next CRA exam would evaluate whether the fintech met the goals of the plan. CRA exam schedules are announced for the next two quarters (six months in advance) so community groups and other stakeholders have opportunities to comment on bank performance and possibly influence ratings. After the CRA exam, a fintech can then develop a new strategic plan and the CRA exam cycle repeats. Alternatively, the fintech could elect to be evaluated under one of the other CRA exam types that varies based on asset level.
In the case of the proposed OCC fintech charter, the frequency and type of evaluations of fintechs was still a work in progress. The OCC recognized that fintechs must serve underserved populations through the life of the charter and that the financial inclusion plans should be updated. However, how often and how the updates would be evaluated was left unclear. The OCC states:
The commitment to meet its financial inclusion goals, approach, activities, and milestones that support fair access to financial services and fair treatment of customers is ongoing through the life of the charter. For this reason, the OCC will require that the SPNB (fintech) update its financial inclusion plan (FIP) in appropriate circumstances. The FIP should address how the SPNB will continue serving the needs of the relevant market and community beyond the initial years after a charter is granted.
NCRC had urged the OCC to establish a periodic evaluation schedule similar to CRA exam schedules or other compliance exams that the fintech would undergo. As much as possible, the process should resemble CRA exams.
The fintech CRA strategic plans in charter applications to-date establish a fintech headquarter’s location as its single assessment area to be evaluated on its CRA exam. This is per current CRA regulatory procedures establishing assessment areas to include areas where an institution’s branches or deposit taking ATMs are located. Yet, the current CRA regulation procedure for fintechs results in a narrow assessment area that is not truly responding to credit and deposit needs where many fintechs conduct business. Narrow assessment areas will thus fall short of meeting the convenience and needs requirement for a charter application. The fintech applications to-date discuss how the fintechs serve the entire country. Accordingly, NCRC believes that the CRA plans ought to be national in reach.
It is a contradiction in terms for a branchless fintech to establish its assessment area where its headquarters is. In this case, the fintech is acting as if its headquarters location is a branch and as such, the headquarters location will make loans in its contiguous community. But the headquarters is not a branch and will not be used for making loans. This sleight of hand mocks the intention of CRA to serve credit needs wherever a lender is conducting business. To only establish one geographical area for a fintech’s only or primary CRA responsibilities is a ruse that will enable fintechs to avoid rigorous CRA responsibilities in all communities in which they conduct a substantial amount of lending. The regulatory agencies must not enable this behavior through blinkered application of CRA and banking regulation.
The CRA regulations do not prohibit a branchless bank from establishing assessment areas beyond its headquarters. Assessment areas can include areas where substantial amounts of lending activity occur.
Using loan data, NCRC believes that the agencies can require non-traditional banks and fintechs to create assessment areas that capture the vast majority of their loans. An example of lending by state for Lending Club during the time period of 2012 and 2013 shows that assessment areas can be meaningfully created for an on-line lender (a two year time period is a typical time period covered by a CRA exam). Lending Club makes data on its lending activity by state and for three digit zip codes publicly available, a practice NCRC recommends for other fintechs.
Several states have sizable numbers of Lending Club loans in this time period even before Lending Club’s substantial lending increases of more recent years. During 2012 and 2013, Lending Club made more than 188,000 loans; most of these were consumer-related loans and/or refinancing and consolidation of outstanding debt (see table below). Another table below on lending by state reveals that heavily populated states including California, New York, Texas and Florida had the highest percentage of loans. Ten states each had more than 3 percent of Lending Club’s loans. On the other end of the scale, 28 states each had less than 1.5 percent of Lending Club’s loans. In sum, it is quite feasible for at least the top ten or twenty states to constitute assessment areas; these states had high numbers of loans and reasonably high percentages of Lending Club’s loans. The top 15 states contain more than two thirds of Lending Club’s loans.
To further investigate how assessment areas would work for a non-traditional bank, NCRC tabulated loans by three digit zip code and metropolitan areas for Texas, one of Lending Club’s high volume states. We found five metropolitan areas with more than 1,000 loans each and one area, North Texas that could possibly be considered a rural area. The five metropolitan areas range in size and location across the state and include Houston, Austin, Ft. Worth, Dallas, and San Antonio. El Paso is the seventh largest area by loan volume with more than 500 loans. Using Lending Club as an example, designating metropolitan areas and counties as assessment areas for non-traditional lenders is feasible and can include a diversity of areas.
NCRC believes that assessment areas for fintechs must include rural areas. Populations in rural areas are less likely to be connected to the internet. If fintechs do not make efforts to serve rural areas, the digital divide disadvantaging rural communities will only widen.
Assessment areas must cover the great majority of fintech lending. NCRC’s research has documented that when the assessment areas of large banks cover less than 50 percent of their lending, the ratings of the large banks on their lending tests are higher. Higher ratings solely due to less coverage of lending on exams will ultimately cause lenders to relax their efforts to serve low- and moderate-income borrowers and communities. It is intuitive that less coverage of lending can lead to easier exams and inflated ratings. Lenders can focus their efforts to serve low- and moderate-income borrowers and communities in only those relatively few geographical areas and lending activities that are covered on CRA exams. It is easier to pass a CRA exam when less than 50 percent of an institution’s loans are examined than when the great majority of an institution’s loans are scrutinized.
Lending Club Loans 2012-2013
|Texas Zip Codes||# Loans||Percent|
|770, 72, 73, 74, 75 Houston||3,634||25.0%|
|750 North Texas||2,074||14.3%|
|760, 61, 62, 64 Ft. Worth, TX||1,836||12.6%|
|786, 87, 89 Austin, TX||1,360||9.3%|
|751, 52, 53 Dallas||1,215||8.3%|
|780, 81, 82, 88 San Antonio TX||1,084||7.4%|
|798, 99 El Paso, TX||527||3.6%|
|765, 66, 67 Waco, TX||455||3.1%|
|785 McAllen TX||361||2.5%|
|756, 57 East Texas||245||1.7%|
|793, 94 Lubbock, TX||231||1.6%|
|790, 91 Amarillo, TX||225||1.5%|
|769, 97 Midland, TX||208||1.4%|
|754 Greenville TX||198||1.4%|
|783, 84 Corpus Christi||195||1.3%|
|768, 78 Bryan, TX||155||1.1%|
|776, 77 Beaumont, TX||121||0.8%|
|795, 96 Abilene, TX||99||0.7%|
|759 Lufkin, TX||93||0.6%|
|763 Wichita Falls TX||77||0.5%|
|779 Victoria, TX||50||0.3%|
|758 Palestine, TX||42||0.3%|
|792 Childress, TX||14||0.1%|
|Three digit zip codes, some metro areas had more than one zip code, some zip codes are abbreviated|
Fair Lending Protections Must be Rigorous
Fintechs pose significant fair lending concerns. They use unorthodox data evaluation and underwriting methods. Some employ algorithms that appear to apply criteria in a neutral fashion but could result in disparate impacts disproportionately disadvantaging protected classes. Carol Evans, a fair lending expert of the Federal Reserve System, advises fintechs to carefully consider fair lending implications of their data and underwriting methods. She suggests factors such as which post-secondary school applicants attended are not really connected with creditworthiness and that fintechs may want to think twice before using these factors in their underwriting criteria. She states that “generally the more speculative the nexus with creditworthiness, the higher the fair lending risk.”
Another fair lending risk posed by fintechs is dual track lending in which a fintech purposefully or inadvertently uses zip codes or demographic characteristics to offer high cost products to underserved populations and more desirable products to white or male borrowers. Price disparities and discrimination of the type discussed above is also a pressing concern.
Given the complexity and wide variety of fair lending risks, fintech charter applications must include robust descriptions of fintech business models and how fintechs will comply with fair lending law and regulation. Subsequent exams including CRA and fair lending reviews must include rigorous evaluations of fair lending compliance.
There are fair lending frontiers that fintechs can help further develop. For example, federal law does not provide robust protections regarding disclosures of loans terms and conditions in small business lending. Fintechs can demonstrate voluntary but verified protections in this area. However, in their charter applications, some finetchs that make small business loans have not indicated whether they will adhere to the Small Business Borrower’s Bill of Rights, a check-list of compliance including transparent disclosures of loan terms and conditions that have been endorsed by many lenders.
Data disclosure laws must apply to fintechs so the general public and federal regulatory agencies can systematically verify that they are adhering to CRA and fair lending law and regulation. Data disclosure will enable federal agencies and members of the public to determine how the fair lending records of fintechs compare against traditional lending institutions such as banks and credit unions. If racial disparities in lending are more pronounced for fintechs than traditional lenders, further scrutiny of their underwriting, marketing approaches, and products would be warranted. Likewise, if fintechs are not as successful in lending to low- and moderate-income borrowers and communities, any CRA obligations for fintechs must be strengthened.
The Home Mortgage Disclosure Act (HMDA) regulations must cover fintechs, particularly if they make loans at levels similar to banks covered by HMDA. Likewise, when the Consumer Financial Protection Bureau (CFPB) develops regulations to enact Section 1071 of the Dodd-Frank Wall Street Reform and Consumer Protection Act, the CFPB must ensure that its regulation covers fintech small business lending. Finally, a number of fintech lenders are consumer lenders. When those fintechs apply to become CRA-covered banks, the federal regulatory agencies must require that their future CRA exams consider consumer lending and provide data on their consumer lending.
Fintechs are still relatively new entrants to the lending marketplace, but they are increasing their lending and market share at a rapid pace. On the one hand, fintechs have the potential to narrow the digital and banking divide by increasing access to credit for populations underserved by banks. On the other hand, fintechs can end up exacerbating the banking and digital divide if oversight of fintechs is lax and the agencies do not expect strenuous fintech efforts to serve underserved populations.
When fintechs apply for either bank charters or any new OCC fintech charter, the agencies must expect rigorous CRA and financial inclusion plans with measureable goals of performance. Periodic CRA exams and fair lending reviews must be rigorous in order to provide incentives for fintechs to meet and/or exceed their performance goals. Assessment areas must be created that cover the great majority of fintech lending in order to ensure that fintechs are making considerable efforts to serve low- and moderate-income borrowers and communities. Lastly, given the novelty of fintech underwriting and marketing, fair lending reviews must be comprehensive and ensure that fintechs are not blatantly or inadvertently discriminating in their lending.
Please contact Josh Silver, Senior Advisor, at firstname.lastname@example.org with any questions about this paper.
 Elizabeth Laderman and Carolina Reid, Federal Reserve Bank of San Francisco, “CRA Lending during the Subprime Meltdown” in Revisiting the CRA: Perspectives on the Future of the Community Reinvestment Act, a Joint Publication of the Federal Reserve Banks of Boston and San Francisco, February 2009, http://www.frbsf.org/publications/community/cra/cra_lending_during_subprime_meltdown.pdf. Also, see Governor Elizabeth A. Duke, at the Revisiting the CRA Policy Discussion, Washington, D.C., February 24, 2009
CRA: A Framework for the Future, http://www.federalreserve.gov/newsevents/speech/duke20090224a.htm
 Tania Ziegler, E.J. Reedy, Annie Le, Bryan, Zhang, Randall S. Kroszner, Kieran Garvy, 2017: The America’s Alternative Finance Industry Report – Hitting Stride, University of Cambridge and the University of Chicago, p. 15, https://www.jbs.cam.ac.uk/fileadmin/user_upload/research/centres/alternative-finance/downloads/2017-05-americas-alternative-finance-industry-report.pdf, p. 15.
 Consumer Financial Protection Bureau, Key Dimensions of the Small Business Lending Landscape, May 2017, p. 26, https://s3.amazonaws.com/files.consumerfinance.gov/f/documents/201705_cfpb_Key-Dimensions-Small-Business-Lending-Landscape.pdf
 Ziegler, et al., p. 15 for number of loans going to fintechs. For total bank loans see, Federal Financial Institutions Examination Council, CRA National Aggregate Table 1 available via https://www.ffiec.gov/CraAdWeb/pdf/2016/N1.PDF
 Eric Weaver, Gwendy Donaker Brown, Caitlin McShane, “Unaffordable and Unsustainable: the New Business
Lending on Main Street,” Opportunity Fund, May 2016
 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.
 Federal Reserve Banks, Small Business Credit Survey, April 2017, 17, https://www.newyorkfed.org/medialibrary/media/smallbusiness/2016/SBCS-Report-EmployerFirms-2016.pdf
 Federal Communications Commission, 2016 Broadband Progress Report, Jan. 29, 2016, retrieved at https://www.fcc.gov/reports-research/reports/broadband-progress-reports/2016-broadband-progress-report.
 Federal Reserve Bank of Dallas, Closing the Digital: A Framework for Meeting CRA Obligations, July 2016, retrieved at https://www.dallasfed.org/assets/documents/cd/pubs/digitaldivide.pdf
 Julapa Jagtiani and Catherine Lemieux, Fintech Lending: Financial Inclusion, Risk Pricing, and Alternative Information, Fed. Res. Bank of Philadelphia, Working Paper No. 17-17, July 6, 2017, 22, retrieved at https://www.philadelphiafed.org/-/media/research-and-data/publications/working-papers/2017/wp17-17pdf
 CRA statute, https://www.law.cornell.edu/uscode/text/12/2901
 Statement of Martin J. Gruenberg, Chairman, Federal Deposit Insurance Corporation on De Novo Banks and Industrial Loan Companies before the Committee on Oversight and Government Reform; U.S. House of Representatives; 2157 Rayburn House Office Building, https://www.fdic.gov/news/news/speeches/spjul1316.html
 SoFi Application, 13.
 Comptroller’s Licensing Manual, Draft Supplement: Evaluating Charter Applications from Financial Technology Companies, March 2017, p. 20, https://www.occ.gov/publications/publications-by-type/licensing-manuals/file-pub-lm-fintech-licensing-manual-supplement.pdf
 For summaries of the community benefit agreements, see https://ncrc.org/cra/
 Community Reinvestment Act; Interagency Questions and Answers Regarding Community Reinvestment Act Guidance, OCC, Board of Governors of the Federal Reserve System, FDIC, Fed. Reg. 81, 142 at 48506, https://www.gpo.gov/fdsys/pkg/FR-2016-07-25/pdf/2016-16693.pdf
 Interagency Q&A at 48542.
 See the OCC version of the CRA regulation, Section §25.27, Strategic plan via https://www.ecfr.gov/cgi-bin/text-idx?SID=d4ebf029aa7ade35dfb94a2ac306401b&mc=true&node=pt12.1.25&rgn=div5#se12.1.25_141
 Comptroller’s Licensing Manual, Draft Supplement, p. 22.
 See § 345.41 (c) (2), Assessment area delineation, of the FDIC CRA regulation via https://www.fdic.gov/regulations/laws/rules/2000-6500.html#fdic2000part345.41
 See https://www.lendingclub.com/info/statistics.action for summary data tables and to download data.
 These states are CA, NY, TX, FL, IL, NJ, PA, OH, GA,VA.
 NCRC, The Community Reinvestment Act and Geography, April 2017, available via https://ncrc.org/wp-content/uploads/2017/05/cra_geography_paper_050517.pdf
 Carol A. Evans, “Keeping Fintech Fair: Thinking about Fair Lending and UDAP Risks,” in Consumer Compliance Outlook – A Federal Reserve System Publication Focusing on Consumer Compliance Topics, Second Issue, 2017, pp. 4-5, https://www.frbsf.org/banking/files/Fintech-Lending-Fair-Lending-and-UDAp-Risks.pdf