109
The Financial Inclusion Trilemma
Adam J. Levitin
†
The challenge of financial inclusion is among the most intractable
policy problems in banking. Despite living in the world’s wealthiest
economy, many Americans are shut out of the financial system. Five percent
of American households lack a bank account, and an additional thirteen
percent rely on expensive and sometimes predatory fringe financial services,
such as check cashers or payday lenders.
Financial inclusion presents a policy trilemma. It is possible to
simultaneously achieve only two of three goals: widespread availability of
services to low-income consumers, fair terms of service, and profitability of
service. Thus it is possible to provide fair and profitable services, but only to
a small, cherry-picked population of low-income consumers. Conversely, it
is possible to provide profitable service to a large population, but only on
exploitative terms. Or it is possible to provide fair services to a large
population, but not at a profit.
The financial inclusion trilemma is not a market failure. Instead, it is
the result of the market working. The market result, however, does not
accord with policy preferences. Rather than addressing that tension,
American financial inclusion policy still leads with market-based solutions,
soft government nudges, and the hope that technology will transform the
economics of small-balance deposit accounts and small-dollar loans.
It is time to recognize the policy failure in financial inclusion and
consider to a menu of stronger regulatory interventions: hard service
mandates, taxpayer subsidies, and public provision of financial services. In
particular, this Article argues for following the approach taken in Canada,
the European Union, and the United Kingdom. This approach—the
adoption of a mandate for the provision of free or low-cost basic banking
services to all qualified applicants—is the simplest solution to the problem
of the unbanked. Addressing small-dollar credit, however, remains an
intractable problem, largely beyond the scope of financial regulation
because the challenge many low-income consumers face is solvency, not
liquidity.
† Carmack Waterhouse Professor of Law and Finance, Georgetown University Law
Center. Thank you to Sara Burriesci for empirical research assistance and to David Hyman,
Patricia McCoy, David Super, and the participants at a Georgetown Law Faculty Workshop for
comments.
Yale Journal on Regulation Vol. 41:109 2024
110
Introduction .................................................................................................. 111
I. The Financial Inclusion Problem in the United States ......................... 117
A. The Unbanked ............................................................................. 117
B. The Underbanked ........................................................................ 120
II. The Financial Inclusion Trilemma ........................................................ 124
A. Illustration #1: The Economics of Deposit Accounts .............. 125
1. Costs of Deposit Accounts ................................................... 125
2. Spread Income ....................................................................... 126
3. Indirect Fee Income .............................................................. 127
4. Direct Fee Income ................................................................. 128
5. Relationship Banking ............................................................ 132
6. Summary ................................................................................. 134
B. Illustration #2: The Economics of Payday Loans ..................... 134
1. Store-Front Payday Lending ................................................ 135
2. Bank Payday Products .......................................................... 137
3. Online Payday Lending ........................................................ 138
III. The Financial Inclusion Policy Playbook ............................................ 140
A. Existing Approaches to Financial Inclusion ............................. 140
1. Private Provision: Fintech, Crypto, and Deregulation ...... 140
2. Negative Service Mandates .................................................. 146
3. Soft Service Mandates ........................................................... 146
4. Modeling Pilot Programs ...................................................... 148
B. Other Potential Regulatory Interventions ................................ 152
1. Hard Service Mandates ......................................................... 152
2. Public Options ........................................................................ 155
3. Public Subsidies ..................................................................... 156
IV. Choosing the Optimal Regulatory Intervention ................................ 158
A. Interventions for the Unbanked ................................................ 158
B. Interventions for the Underbanked ........................................... 161
Conclusion ..................................................................................................... 163
The Financial Inclusion Trilemma
111
Introduction
For half a century, the United States has failed at financial inclusion.
Despite antidiscrimination laws, soft mandates to provide services to
underserved populations, a proliferation of financial technology, and a
generally booming economy, a sizeable share of the U.S. population
remains without access to mainstream financial services from banks.
Nearly one in twenty U.S. households are “unbanked,” meaning that
no member of the household has a bank account,
1
a much higher rate than
in other developed countries.
2
Additionally, many households that have
bank accounts are “underbanked,” meaning that they use alternative
financial services for payments and small-dollar, short-term credit, such as
check cashers, money orders, pawn shops, auto title loans, payday loans,
earned wage access, or tax refund advances.
3
Overall, over one in eight
U.S. households with bank accounts is underbanked.
4
The unbanked and underbanked populations in the United States are
predominantly lower-income, with the unbanked population concentrated
among those households earning less than $30,000 annually, and the
underbanked population concentrated among those households earning
less than $50,000 annually.
5
Given correlations between income and race,
it should be no surprise that a much higher percentage of minority
populations is unbanked or underbanked. Nearly one in nine Black
households and one in eleven Hispanic households lack a bank account,
6
and nearly one in four Black and Hispanic households are underbanked.
7
Yet even among lower-income populations, minority households are
unbanked at a much higher rate than white households. While 3.31% of
1. FDIC National Survey of Unbanked and Underbanked Households, FED. DEPOSIT INS.
CORP. 1 (2021) [hereinafter FDIC 2021 Study], https://www.fdic.gov/analysis/household-survey/
2021report.pdf [https://perma.cc/2VY4-PEG8].
2. Luca Ventura, World’s Most Unbanked Countries 2021, GLOB. FIN. (Feb. 17, 2021),
https://www.gfmag.com/global-data/economic-data/worlds-most-unbanked-countries
[https://perma.cc/9X5E-C8G8].
3. Economic Well-Being of U.S. Households in 2021, BD. OF GOVERNORS OF THE FED.
RSRV. SYS. 43 (May 2022) [hereinafter Fed Study], https://www.federalreserve.gov/publications/
files/2021-report-economic-well-being-us-households-202205.pdf [https://perma.cc/6QY2-
8QUZ]. The Federal Reserve uses a narrower definition of “underbanked” than the Federal
Deposit Insurance Corporation (FDIC), which includes international remittances and rent-to-own
services in its definition of alternative financial services. Report on the Economic Well-Being of
U.S. Households in 2018, BD. OF GOVERNORS OF THE FED. RSRV. SYS. 25 n.14 (May 2019),
https://www.federalreserve.gov/publications/files/2018-report-economic-well-being-us-
households-201905.pdf [https://perma.cc/WLU5-3Q4M].
4. Fed Study, supra note 3, at 44.
5. Id. at 44 tbl.11; FDIC 2021 Study, supra note 1 at 14, 77. The median U.S. household
income in 2021 was $76,330. Gloria Guzman & Melissa Kollar, Income in the United States: 2022,
U.S. CENSUS BUREAU 1 (Sept. 2023), https://www.census.gov/library/publications/2023/demo/
p60-279.html [https://perma.cc/SQ2R-F8NX].
6. FDIC 2021 Study, supra note 1, at 14.
7. Id. at 76. Because of the Federal Reserve’s narrower definition of underbanked, which
excludes remittances, see note 3 above, the Federal Reserve finds a lower underbanked rate for
Hispanic households. Fed Study, supra note 3, at 44 tbl.11.
Yale Journal on Regulation Vol. 41:109 2024
112
below-median-income white households are unbanked, 12.29% of below-
median-income Hispanic households and 14.1% of below-median-income
Black households are unbanked.
8
Black women, in particular, are more
likely to be unbanked than any other group.
9
Lack of a bank account limits these households’ ability to fully
participate in the modern economy and has significant repercussions for
racial equity. Without a bank account, it is more onerous and costly to
make and receive payment, meaning that the poor pay more,
10
exacerbating wealth disparities. Moreover, without a bank account, certain
transactions are all but impossible—purchasing airline tickets, renting a
car, and making any sort of online purchase.
Being underbanked also imposes substantial costs on households. The
cost of alternative financial services is substantially higher than that of
mainstream, bank-provided financial services. For example, while a credit
card might have an annual percentage rate (APR) of 24%, a payday loan
might have an APR as high as 662%.
11
Other types of short-term, small-
dollar credit are similarly expensive.
Because the underbanked pay more to borrow, their net financial
position is often worse off than if they had used mainstream financial
services, but such mainstream services are not available to many
underbanked households because of poor or limited credit histories. Once
again, the racial disparities are stark, although they do not correlate as
strongly with income. While 9.3% of white households are underbanked,
24.7% of Black households and 24.1% of Hispanic households are
underbanked.
12
Again, Black women are the most likely to be
underbanked, with nearly half using alternative financial services.
13
This Article argues that the United States has failed at financial
inclusion because it refuses to recognize an economic reality: it is not
possible for financial institutions to profitably provide low-income
consumers with financial services on non-exploitative terms at scale. The
problem the United States faces is what this Article terms the “financial
8. Author’s analysis of underlying data from FDIC 2021 Study, supra note 1.
9. Vicki L. Bogan & Sarah E. Wolfolds, Intersectionality and Financial Inclusion in the
United States, 112 AEA PAPERS & PROC. 43, 43 (2022).
10. See generally DAVID CAPLOVITZ, THE POOR PAY MORE: CONSUMER PRACTICES OF
LOW-INCOME FAMILIES (1967) (documenting how low-income households pay more for a range
of goods and services than middle- or higher-income households).
11. Red Alert Rates: Annual Percentage Rates on $400, Single-Payment Payday Loans in
the United States, CTR. FOR RESPONSIBLE LENDING 3 (June 2023), https://
www.responsiblelending.org/sites/default/files/nodes/files/research-publication/crl-red-alert-
rates-payday-ratecap-map-jun2023.pdf [https://perma.cc/CZ8E-R32X] (listing 662% annual
percentage rate (APR) for Texas).
12. FDIC 2021 Study, supra note 1, at 76.
13. Bogan & Wolfolds, supra note 9, at 44. There are also correlations of age, gender,
marital status, and education with the use of alternative financial services. Julie Birkenmaier &
Qiang Fu, The Association of Alternative Financial Services Usage and Financial Access: Evidence
from the National Financial Capability Study, 37 J. FAM. ECON. ISSUES 450, 450 (2015).
The Financial Inclusion Trilemma
113
inclusion trilemma.” A policy trilemma is a situation in which only two of
three policy goals can be simultaneously achieved.
14
The financial inclusion trilemma means that it is possible to
simultaneously achieve only two of three desired policy outcomes in the
provision of financial services:
(1) widespread availability of services to low-income consumers;
(2) fair terms of service; and
(3) stand-alone profitability of those services for service providers.
Thus, it is possible to provide fair and profitable services to a small,
cherry-picked population of low-income consumers. Conversely, it is
possible to provide profitable service on a wide scale, but only on
exploitative terms. Or it is possible to provide fair services to a large
population, but not with the service being profitable on a stand-alone basis.
The golden trifecta of policy goals, however, is simultaneously
unattainable.
The financial inclusion trilemma is not a market failure. Instead, it is
an example of markets working precisely as expected. The problem is that
the market, left to its own devices, will not produce the desired policy
outcome of fair and widely available services absent some form of
subsidization. To the extent there is a failure here, then, it is a failure of
government to intervene when the market fails to produce the desired
policy outcome.
15
Addressing this policy failure will involve making tradeoffs among the
policy goals. If low-income consumers are able to access fair products, it
will come at the expense of financial institutions, while if financial
institutions are able to be profitable, it will come at the expense of
consumers through either lack of product access or unfair product terms.
14. Other examples of policy trilemmas include one in international economics among a
fixed foreign exchange rate, free movement of capital across borders, and independent national
monetary policy, see Maurice Obstfeld, Jay C. Shambaugh & Alan M. Taylor, The Trilemma in
History: Tradeoffs Among Exchange Rates, Monetary Policies, and Capital Mobility, 87 REV.
ECON. & STAT. 423 (2005); health care’s “iron triangle” among access, quality, and cost
containment, see WILLIAM KISSICK, MEDICINE’S DILEMMAS: INFINITE NEEDS VERSUS FINITE
RESOURCES 2 (1994); and one in financial regulation among provision of clear rules, maintenance
of market integrity, and encouragement of financial innovation, see Chris Brummer & Yesha
Yadav, Fintech and the Innovation Trilemma, 107 GEO. L.J. 235, 242 (2019).
15. See Mark V. Pauly, Trading Cost, Quality, and Coverage of the Uninsured: What Will
We Demand and What Will We Supply?, in THE FUTURE U.S. HEALTHCARE SYSTEM: WHO WILL
CARE FOR THE POOR AND UNINSURED? 353, 364-65 (Stuart H. Altman, Uwe E. Reinhardt &
Alexandra E. Shields eds., 1998) (“[T]he worsening of the lot of the uninsured under market
competition, if it occurs and is not offset by government, would not be an example of market
failure. Rather, it would be an example of serious ‘government failure’ (at least in the sense of
citizens collectively making a bad decision), an example of political failure, and perhaps of moral
failure. Markets would be doing what they do best. It would be government that would be failing
to do what it should do. Market competition will have abolished a type of charity that citizens,
when faced with the challenge to pay for it explicitly and consciously, determined to be not worth
its cost.”).
Yale Journal on Regulation Vol. 41:109 2024
114
For too long, financial inclusion policy in the United States has
proceeded on the assumption that a market solution to the trilemma would
emerge, facilitated by technological advances and soft governmental
nudges. Unfortunately, this is an unrealistic, but politically convenient
conceit. Instead, this Article argues, if the United States wants to succeed
at financial inclusion, it needs to consider more muscular interventions:
hard service mandates, public provision, or taxpayer subsidies.
Hard service mandates, such as a requirement that banks offer free or
low-fee basic bank accounts to all applicants, achieve financial inclusion by
imposing a cross-subsidy on bank customers. Because banks are required
to offer a service that is not otherwise profitable, they must fund it through
either higher charges to other customers or accept lower overall profits,
meaning lower returns for their shareholders. Such hard service mandates
for basic bank accounts exist in Canada, the European Union, and the
United Kingdom.
16
Public provision entails either direct governmental provision of
financial services or the use of governmental contractors, while
subsidization means that taxpayers are funding private provision outside
the government contracting process. There are trade-offs among these
approaches, but all of them slice through the Gordian Knot of the trilemma
by eliminating the stand-alone profitability condition.
All of these interventions, however, look quite different when applied
to the problem of the unbanked, as opposed to the problem of the
underbanked. Service mandates or public provision of deposit accounts
raise materially different issues than service mandates or public provision
of credit. There is substantially less financial risk involved in the provision
of deposit accounts than the provision of credit. Moreover, the provision
of credit, even if intermediated, puts the government into the
uncomfortable and unsustainable business of allocating credit in the
economy. If government allocates capital, eventually political pressure will
push for unsustainable allocations to politically favored constituencies,
rendering the banking system a giant political slush fund. Additionally,
credit necessitates collections, which is an awkward fit for public provision,
because it puts the government in an adverse role to those very consumers
it seeks to help.
Interventions in the form of service mandates or public provision are
appropriate for deposit accounts as well as for large-dollar credit, such as
mortgages and student loans, which have positive social externalities and
can be net positive financial and human capital investments. Such
interventions are not appropriate, however, for small-dollar credit, like the
majority of credit products used by the underbanked. Small-dollar credit—
loans of hundreds or low thousands of dollars—is generally used for
16. See infra notes 167-180 and accompanying text.
The Financial Inclusion Trilemma
115
liquidity management and for covering unexpected expenses. It is not used
to pursue value-creating investments, nor does it have positive social
externalities. In small-dollar credit markets, the trilemma must hold,
forcing a policy choice between access to credit and fair terms of service.
While a large microfinance and development economics literature
exists on various aspects of financial inclusion, most of it focuses on
developing countries. There is a limited body of scholarship on financial
inclusion in the United States.
17
Within that limited literature, there has
been remarkably little introspection as to why financial inclusion policy in
the United States has overall been a failure over the past half century. To
be sure, there have been some modest gains in American financial
inclusion, particularly during several recent years when there was a
booming economy and high levels of employment, but the United States
still lags behind most of the developed world on financial inclusion
metrics.
18
The literature has mainly documented the financial inclusion
problem, proposed particular solutions, or considered the effects of
consumer protection regulation on deposit account and credit availability
without holistic consideration of the nature of the policy problem. There is
no literature that engages with tradeoffs among the different possible
approaches to financial inclusion in terms of private and public provision.
This Article represents a first attempt to address this gap in the literature.
In so doing, this Article makes three signal contributions to the
literature. First, it identifies the financial inclusion trilemma as the
fundamental nature of the policy problem. Once the trilemma is
recognized, the policy response menu shifts from reliance on the market
and light governmental interventions to more substantive governmental
interventions.
17. See, e.g., LISA SERVON, THE UNBANKING OF AMERICA: HOW THE NEW MIDDLE
CLASS SURVIVES (2018); JONATHAN MURDOCH & RACHEL SCHNEIDER, THE FINANCIAL
DIARIES: HOW AMERICAN FAMILIES COPE IN A WORLD OF UNCERTAINTY (2017); MEHRSA
BARADARAN, HOW THE OTHER HALF BANKS: EXCLUSION, EXPLOITATION, AND THE THREAT
TO DEMOCRACY (2015); MICHAEL S. BARR, NO SLACK: THE FINANCIAL LIVES OF LOW-INCOME
AMERICANS (2012); INSUFFICIENT FUNDS: SAVINGS, ASSETS , CREDIT AND BANKING AMONG
LOW- AND MODERATE-INCOME HOUSEHOLDS (Rebecca M. Blank & Michael S. Barr eds., 2011);
BUILDING INCLUSIVE FINANCIAL SYSTEMS: A FRAMEWORK FOR FINANCIAL ACCESS (Michael
S. Barr, Anjali Kumar & Robert E. Litan eds., 2007); Jennifer L. Dlugosz, Brian T. Melzer &
Donald P. Morgan, Who Pays the Price? Overdraft Fee Ceilings and the Unbanked, FED. RSRV.
BANK OF N.Y. (2023), https://www.newyorkfed.org/medialibrary/media/research/staff_reports/
sr973.pdf [https://perma.cc/5Y9W-THYJ]; Bogan & Wolfolds, supra note 9; Birkenmaier & Fu,
supra note 13; Mehrsa Baradaran, It’s Time for Postal Banking, 127 HARV. L. REV. F. 165 (2014);
Sherrie L.W. Rhine & William H. Greene, Factors That Contribute to Becoming Unbanked, 47 J.
CONSUMER AFFS. 27 (2013); Leslie Parrish & Josh Frank, An Analysis of Bank Overdraft Fees:
Pricing, Market Structure, and Regulation, 45 J. ECON. ISSUES 353 (2011); Michael S. Barr, Credit
Where it Counts: The Community Reinvestment Act and Its Critics, 80 N.Y.U. L. REV. 513 (2005);
Michael S. Barr, Banking the Poor, 21 YALE J. ON REGUL. 121 (2004); Jeanne M. Hogarth, Chris
E. Anguelov & Jinkook Lee, Why Don’t Households Have a Checking Account?, 38 J. CONSUMER
AFFS. 1 (2004).
18. Ventura, supra note 2.
Yale Journal on Regulation Vol. 41:109 2024
116
Second, the Article explains why financial inclusion faces a trilemma.
The problem stems from the small size of the financial transactions
undertaken by lower-income consumers: these consumers maintain small
deposit account balances, and their liquidity borrowing is for small dollar
amounts. It is difficult for financial services businesses to amortize fixed
and semi-variable expenses like overhead over small transaction amounts,
so their fees and charges are necessarily large relative to the transaction
amount. These high costs are the foremost barrier to financial inclusion.
Third, this Article underscores the critical difference in policy
approaches required for the problem of the unbanked and the problem of
the underbanked. The unbanked and underbanked are fundamentally
different populations, both in terms of their economic profile and the
nature of the inclusion problem. While the unbanked are all necessarily
underbanked, many of the underbanked are not unbanked. Instead, the
primary characteristic of the underbanked is that they have poor credit.
Unfortunately, both policy discussions and the scholarly financial
inclusion literature often lump the unbanked and the underbanked
together. For example, discussions of postal banking and proposed
legislative implementations are often framed as a policy solution to both
the problem of the unbanked and the high costs of payday lending,
19
but
this framing fails to recognize that it is impossible to get a payday loan
without a bank account. There is no such thing as an unbanked payday
borrower. Providing postal bank accounts to unbanked consumers would
not reduce payday lending, but might in fact increase it by making more
consumers eligible to get payday loans. Recognition of the substantially
different nature of the unbanked and underbanked problems is necessary
for tailoring effective policy solutions to each.
The Article proceeds in four parts. Part I reviews the nature of the
financial inclusion problem in the United States. In particular, it focuses on
the very different nature of two financial inclusion subproblems: the
unbanked and the underbanked. Frequently, policy discussions treat these
groups as interchangeable, when they are in fact substantially different
both in their economic situation and in the nature of the financial inclusion
problem they face.
Part II introduces the nature of the financial inclusion trilemma.
Using the economics of deposit accounts and payday loans, it shows that
the cost structure of small-dollar transactions makes it impossible to
profitably serve low-income consumers on scale, except by engaging in
exploitative practices.
19. See, e.g., Baradaran, supra note 17, at 167 (noting that postal banks could offer lower-
cost payday loans); Providing Non-Bank Financial Services for the Underserved, OFF. OF
INSPECTOR GEN., U.S. POSTAL SERV. 12-14 (Jan. 27, 2014), https://www.uspsoig.gov/sites/default/
files/reports/2023-01/rarc-wp-14-007_0.pdf [https://perma.cc/D62P-X5HE] (proposing short-term,
small-dollar credit provision through the post office); Postal Banking Act, S. 4614, 116th Cong.
(2020) (proposing provision of small-dollar loans via the post office).
The Financial Inclusion Trilemma
117
Part III turns to a consideration of the alternative approaches to
financial inclusion: reliance on private provision, including technological
advances; soft mandates; hard mandates (cross-subsidies); public options;
and public subsidies. The United States has historically relied on private
provision coupled with soft mandates, such as the Community
Reinvestment Act, and antidiscrimination laws. This approach has failed
because private provision, even bolstered by technological advances
through fintech, has been unable to materially change the economics of
deposit accounts or small-dollar lending. Financial inclusion is a problem
precisely because the market works as expected.
Part IV argues that achieving financial inclusion requires greater
regulatory intervention, but that success is possible only for the problem of
the unbanked. While a hard mandate of free or low-cost bank account
provision is the best solution to the problem of the unbanked, the problems
of the underbanked cannot be addressed through public provision or
service mandates. Instead, small-dollar credit policy requires making
difficult choices about which two goals of the policy trilemma ought to be
prioritized.
I. The Financial Inclusion Problem in the United States
The financial inclusion problem in the United States has two separate
aspects: the unbanked and the underbanked. It is important to understand
that these are distinct populations with different financial inclusion
problems.
A. The Unbanked
The unbanked lack any sort of deposit account relationship with a
bank. Bank deposit accounts are important to consumers for two reasons.
First, they are a vehicle for the safekeeping of funds. And second, deposit
accounts are the launchpad for many types of payments. A deposit account
is required for making payments by automated clearing house (ACH)
(including via Zelle), check, and debit card, because those payments are
directly debited from a bank account. A bank account is also generally
necessary for making payments by credit card because the credit card bill
must be paid, and cannot be paid in person with cash.
Similarly, electronic payment systems like PayPal and Venmo allow
funds to be transferred among users without requiring a bank account, but
the initial loading of funds must either be from a bank account, a credit
card, or a payment from another user. This means that unbanked PayPal
and Venmo users are limited to transacting with balances that they have
received from other PayPal and Venmo users, and that they can transact
only with those who accept PayPal or Venmo (respectively) as a form of
payment.
Yale Journal on Regulation Vol. 41:109 2024
118
When the unbanked want to receive non-cash funds, they must either
use a check casher or a prepaid card, both of which have comparatively
high fees. In contrast, there is usually no marginal cost to the consumer for
receiving a payment in a bank account. Likewise, when the unbanked wish
to send funds, such as for paying bills or for remittances to family abroad,
they must purchase money orders or use money transmitters like Western
Union. All of these services come with a fee. In contrast, there is no
marginal fee to send funds from a bank account via ACH. Receiving and
sending payments thus costs more for the unbanked.
Nor is the higher cost of payments the only cost of being unbanked.
Because credit cards, and secondarily ACH and debit cards, are the main
payment methods accepted for online transactions, unbanked consumers
have limited ability to transact online, meaning that they have limited
ability to participate in the modern commercial world. In practical terms,
being shut out of online commerce limits consumers to shopping at brick-
and-mortar retailers.
For low-income consumers, the situation is particularly problematic.
Transportation costs and time restrictions effectively limit most low-
income consumers to shopping primarily at local neighborhood retailers.
When low-income consumers live in neighborhoods served by only one or
two retailers, they are likely to have to pay supracompetitive prices to local
monopolies, like the only grocery store or hardware store in the
neighborhood—if they can find the goods they want at all. Being unbanked
can result in higher costs of goods and services because it limits the
competition for the consumer’s business.
There are numerous reasons that consumers are unbanked. The
Federal Deposit Insurance Corporation (FDIC) has repeatedly surveyed
consumers about why they are unbanked. The leading reason given by
respondents is not having enough money to meet minimum balance
requirements, followed by not trusting banks.
20
Other reasons given
include the desire to maintain privacy; excessively high or unpredictable
bank account fees; problems with personal identification, credit, or former
bank accounts; banks not offering the desired products or services; and
inconvenient bank locations.
21
Although the FDIC survey responses are the go-to source for the
reasons that consumers are unbanked, the survey responses should be
taken with a grain of salt. They are self-reported survey responses and thus
may reflect how respondents wish to present themselves, rather than the
truth. For example, a respondent with bad credit, who has been repeatedly
turned down in attempts to open a bank account, might not wish to admit
20. FDIC 2021 Study, supra note 1, at 3.
21. Id.
The Financial Inclusion Trilemma
119
to having bad credit,
22
but might instead point to other factors, like failure
to meet minimum balance requirements, high or unpredictable fees, or
mistrust of banks.
Additionally, the categories of responses in the FDIC surveys are
hardly exclusive. Minimum balance requirements, for example, are
typically tied to fees, so not having enough money to meet minimum
balance requirements is really a variation on fees being too high or
unpredictable and is likely correlated with having credit problems.
Still, taken as a whole, the FDIC surveys suggest that the phenomenon
of the unbanked is multicausal, but that consumers’ economic situation is
the primary driver. Indeed, the FDIC found that approximately half of the
decline in the unbanked rate from 2011 to 2021 was associated with
improvements in households’ socioeconomic circumstances.
23
This is not
to say that other issues, like lack of identity documentation or cultural
factors, do not play a role, but consumers’ economic situation appears to
be the single most important factor in the unbanked rate.
Not surprisingly, then, FDIC surveys indicate that Black and Hispanic
Americans—on average lower-income populations
24
—are substantially
overrepresented among the unbanked. Figure 1, below, shows that much
higher percentages of Black and Hispanic Americans are unbanked
compared to Asian or white Americans.
It is unclear whether these other factors are on the financial institution
side or the household side. For example, on the financial institution side,
factors that correlate with race could include discrimination against
minority applicants, preference for credit scoring of applicants, or lack of
service in minority communities. On the household side, correlations with
race could include cultural attitudes toward banking, lack of credit scores,
25
or simply lack of familiarity with banks. For example, a person whose
parents were banked is likely to be used to going to a bank, while a person
whose parents were unbanked may be uncomfortable using banks because
they are unfamiliar.
22. Banks are exposed to credit risk on deposit accounts due to the possibility of
overdrafts. As a result, many banks require a credit screening, typically using Chex Systems, prior
to opening an account.
23. FDIC 2021 Study, supra note 1, at 1.
24. Gloria Guzman & Melissa Kollarr, Income in the United States: 2022, U.S. CENSUS
BUREAU 3 (Sept. 2023), https://www.census.gov/content/dam/Census/library/publications/2023/
demo/p60-279.pdf [https://perma.cc/RZP4-3DCL] (listing median incomes for Asian households
at $108,700, non-Hispanic white households at $81,060, Hispanic households of any race at
$62,800, and Black households at $52,860).
25. See, e.g., Data Point: Credit Invisibles, CONSUMER FIN. PROT. BUREAU OFF. OF
RSCH. 17-18 (May 2015), https://files.consumerfinance.gov/f/201505_cfpb_data-point-credit-
invisibles.pdf [https://perma.cc/D3FV-5P69] (providing evidence that Black and Hispanic
consumers are more likely to lack credit records).
Yale Journal on Regulation Vol. 41:109 2024
120
Figure 1. Percentage of Unbanked Households by Race
26
B. The Underbanked
The underbanked have deposit accounts at banks, but they use high-
cost alternative financial services providers for credit and payments. In
particular, they use small-dollar, short-term lenders, such as payday loans,
vehicle title loans, pawn loans, tax refund anticipation loans, retail
installment credit, buy-now-pay-later, rent-to-own, and small signature
loans. These lenders charge high fees, but they also offer very fast or even
immediate funding. The underbanked may also use check cashers despite
having bank accounts due to convenience factors, such as longer business
hours and more accessible locations, faster funds availability, or to avoid
setoff of overdrafts or garnishment orders.
27
The unbanked are, by definition, also underbanked, but many of the
underbanked have bank accounts. Indeed, many alternative financial
services require the consumer to have a bank account. Payday loans, for
example, require the borrower to have a bank account as part of the
mechanism for collecting on the loans: the borrower gives the lender either
a post-dated check or the authorization for an ACH draw on the account.
26. FDIC 2021 Study, supra note 1, at 14.
27. Some definitions of “underbanked” also include the use of money transmitters for
foreign remittances. See, e.g., Report on the Economic Well-Being of U.S. Households in 2019,
Featuring Supplemental Data from April 2020, BD. OF GOVERNORS OF THE FED. RSRV. SYS. 27
n.26 (May 2020), https://www.federalreserve.gov/publications/files/2019-report-economic-well-
being-us-households-202005.pdf [https://perma.cc/4YMS-38Y5].
2.9%
2.1%
9.3%
11.3%
0%
2%
4%
6%
8%
10%
12%
Asian White Hispanic Black
The Financial Inclusion Trilemma
121
Likewise, some types of alternative lending, particularly online lending,
require a bank account where funds can be deposited.
The ultimate policy problem of the underbanked is substantially
different than that of the unbanked. Whereas the unbanked are simply shut
out of much of the commercial world, the underbanked are able to
participate, but are spending more for credit than fully-banked consumers,
which comes with a host of negative consequences.
The underbanked are not uniform in their reasons for using short-
term small-dollar credit. Economic factors, however, are the driver of the
underbanked phenomenon, even more than the unbanked phenomenon.
28
Because the underbanked (excluding the unbanked subpopulation) are, by
definition, users of banks, they are not using high-cost alternative financial
services because of lack of trust of financial institutions or lack of legal
documentation. Instead, the underbanked are often strapped for liquidity,
and alternative financial service providers are able to provide immediate
liquidity—walk to the payday lender, pawnshop, or check casher and walk
away with cash—in a way that traditional banks do not. Thus, several
studies have found that payday borrowers either lack available credit card
lines
29
or that they decrease their borrowing following a tax rebate.
30
For some borrowers, short-term, small-dollar credit products provide
a liquidity bridge that help them deal with unexpected and immediate
expenses that must be paid before income comes in. But for many
borrowers, short-term, small-dollar credit only exacerbates their financial
problems. These borrowers face solvency as well as liquidity problems.
31
As one study notes, even if the loan is offered for “free,” meaning with no
fee, “a typical borrower will be unable to meet his or her most basic
28. The possible exception here might be the use of check cashers to avoid garnishment,
but this is a relatively small part of the underbanked phenomenon.
29. Neil Bhutta, Paige Marta Skiba & Jeremy Tobacman, Payday Loan Choices and
Consequences, 47 J. MONEY, CREDIT, & BANKING 223, 234 (2015) (finding that payday borrowers
have generally exhausted their credit lines at the time of their first payday loan application); Susan
P. Carter, Paige M. Skiba & Jeremy Tobacman, Pecuniary Mistakes? Payday Borrowing by Credit
Union Members, in FINANCIAL LITERACY: IMPLICATIONS FOR RETIREMENT SECURITY AND THE
FINANCIAL MARKETPLACE 145, 148-49 (Olivia S. Mitchell & Annamaria Lusardi eds., 2011)
(finding that payday loan borrowers who belonged to a credit union had about one-eighth the
available liquidity of credit union members who did not take out payday loans, and that 70% of
the borrowers had no available line of credit at the time of the loan). See also Neil Bhutta, Jacob
Goldin & Tatiana Homonoff, Consumer Borrowing after Payday Loan Bans, 59 J.L. & ECON. 225,
227, 230 (2016) (finding that following payday lending bans consumers shift to other forms of high-
cost credit, but not credit cards, suggesting that they do not have available credit card lines). But
see Sumit Agarwal, Paige Marta Skiba & Jeremy Tobacman, Payday Loans and Credit Cards: New
Liquidity and Credit Scoring Puzzles?, 99 AM. ECON. REV. PAPERS & PROC., 412, 412 (2009)
(finding that payday borrowers have substantial line availability on their credit cards on the day
of borrowing).
30. Paige Marta Skiba, Tax Rebates and the Cycle of Payday Borrowing, 16 AM. L. &
ECON. REV. 550, 550 (2014) (finding that payday loan borrowing was reduced in the short term
following the liquidity infusion of a tax rebate).
31. See Payday, Vehicle Title, and Certain High-Cost Installment Loans, 82 Fed. Reg.
54472, 54570 (Nov. 17, 2017) (summarizing research regarding high delinquency rates on credit
cards and frequent nonsufficient funds fees incurred by payday borrowers).
Yale Journal on Regulation Vol. 41:109 2024
122
obligations and repay the payday loan debt in a two-week period.”
32
The
high cost of short-term, small-dollar credit only deepens borrowers’
solvency problems. This leads to borrowers having to make painful
consumption decisions, such as whether to repay a loan or meet basic
expenses, such as food, childcare, and health care.
As a result, borrowers with short-term, small-dollar debt frequently
“roll over” or extend their loans, or borrow in repeated sequences. Instead
of being short term, the debt effectively becomes longer term, but it still
bears the high cost associated with short-term lending. Ultimately, in part
because of the high costs of short-term, small-dollar credit, many
borrowers default on their obligations.
33
Defaults have collateral consequences for borrowers. They may result
in additional fees from the lender and/or result in the borrower’s bank
account becoming overdrawn—resulting in incursion of the bank’s
overdraft or nonsufficient funds fees—and possible account closure. Thus,
access to payday loans is associated with increased rates of involuntary
bank account closures.
34
Additionally, the high cost of servicing short-term,
small-dollar debt may result in the borrower becoming delinquent on other
obligations.
35
In short, for many borrowers, rather than helping, short-
term, small-dollar debt exacerbates their financial problems.
As with the unbanked, Black and Hispanic Americans are
substantially overrepresented among the underbanked. Figure 2, below,
shows that a much higher percentage of Black and Hispanic households
are underbanked compared to Asian or white households.
32. Susanna Montezemolo, The State of Lending in America & its Impact on U.S.
Households, CTR. FOR RESPONSIBLE LENDING 5 (Sept. 2013), https://
www.responsiblelending.org/sites/default/files/uploads/10-payday-loans.pdf
[https://perma.cc/377Z-JF7W].
33. See, e.g., Uriah King & Leslie Parrish, Payday Loans, Inc.: Short on Credit, Long on
Debt, CTR. FOR RESPONSIBLE LENDING 19 (Mar. 31, 2011), https://getoutofdebt.org/wp-
content/uploads/2011/04/payday-loan-inc.pdf [https://perma.cc/7Y3W-SY3G] (finding a 37%
default rate for borrowers within the first year and 44% default rate within two years); Paige Marta
Skiba & Jeremy Tobacman, Payday Loans, Uncertainty and Discounting: Explaining Patterns of
Borrowing, Repayment, and Default, tbl.2 (Vand. L. & Econ. Rsch. Paper, Paper No. 08-33, 2008),
https://ssrn.com/abstract=1319751 [https://perma.cc/X48G-ZB4A] (estimating a 54% default rate
at a large Texas-based payday lender); Susanna Montezemolo & Sarah Wolff, Payday Mayday:
Visible and Invisible Payday Defaults, CTR. FOR RESPONSIBLE LENDING 4 (Mar. 2015),
https://www.responsiblelending.org/sites/default/files/nodes/files/research-publication/
finalpaydaymayday_defaults.pdf [https://perma.cc/KJ8M-FPX9] (estimating that 39% of new
payday borrowers in North Dakota default within one year and 46% default within two years).
34. Dennis Campbell, F. Asís Martínez-Jerez & Peter Tufano, Bouncing out of the
Banking System: An Empirical Analysis of Involuntary Bank Account Closures, 36 J. BANKING &
FIN. 1224, 1225 (2012).
35. See, e.g., Agarwal et al., supra note 29, at 416 (finding that credit card users who began
to borrow from payday lenders were 92% more likely to become delinquent on their credit card
payments). It is unclear whether payday borrowing is the cause of the delinquency or if the turn
to payday loans reflects general deteriorating financial condition. See Brian T. Melzer, The Real
Costs of Credit Access: Evidence from the Payday Lending Market, 126 Q.J. ECON. 517, 534 (2011)
(providing evidence that individuals with access to storefront payday loans were 25% more likely
to have difficulty paying bills and 25% more likely to delay medical care than those without
access).
The Financial Inclusion Trilemma
123
Figure 2. Percentage of Underbanked Households by Race
36
It is important to note that the type of alternative financial services
relied upon may vary across races. In particular, the FDIC data includes
international remittance transactions,
37
which are most likely to be
undertaken by immigrant populations sending funds back to family in their
countries of origin. This contributes to a higher level of alternative
financial service use for all populations, but is likely to contribute more to
the figures reported by Asian and Hispanic households, partially
explaining the high use of alternative financial services even by above-
median-income Asian and Hispanic households.
Still, it is clear that Black and Hispanic households are far more likely
to use alternative financial services for credit and payments than Asian and
white households, even accounting for income. Yet controlling for income
is not the same as controlling for credit. Credit scores have only a weak
correlation with income,
38
but stronger correlation with race.
39
Thus, to the
extent that minorities have lower credit scores, this may prevent them from
obtaining bank credit and result in greater use of alternative financial
services for credit provision. This can become a vicious cycle, because the
higher costs of alternative financial services make defaults more likely on
36. FDIC 2021 Study, supra note 1, at 76.
37. BD. OF GOVERNORS OF THE FED. RSRV. SYS., supra note 27, at 27 n.26.
38. Rachael Beer, Felicia Ionescu & Geng Li, Are Income and Credit Scores Highly
Correlated?, FEDS NOTES (Aug. 13, 2018), https://www.federalreserve.gov/econres/notes/feds-
notes/are-income-and-credit-scores-highly-correlated-20180813.html [https://perma.cc/9UK2-
UJ7K].
39. See, e.g., CONSUMER FIN. PROT. BUREAU OFF. OF RSCH., supra note 25, at 17-18
(finding that Black and Hispanic consumers are more likely to be credit invisible).
9.3%
16.5%
24.1%
24.7%
0%
5%
10%
15%
20%
25%
30%
White Asian Hispanic Black
Yale Journal on Regulation Vol. 41:109 2024
124
consumer reporting obligations and thus diminish credit scores, pushing
the consumers back towards using alternative financial services.
II. The Financial Inclusion Trilemma
Financial inclusion—meaning access to the formal banking system—
has been recognized as a policy problem in the United States for decades,
40
although it only began to receive concentrated attention around the
beginning of the twenty-first century.
41
Financial inclusion is a distinct
policy problem from access to mortgage credit, which has been the focus
of federal regulatory interventions since the New Deal.
42
Mortgage credit
and homeownership have been a middle-class project that presupposes
households being banked. A mortgage requires a borrower to first save up
for a down payment, which is presumably kept in a deposit account, and to
then make monthly payments from a bank account. Instead, financial
inclusion as a policy matter has focused on the unbanked and the use of
check cashers, money transmitters, and short-term, small-dollar lenders.
As noted in the introduction, financial inclusion involves a policy
trilemma in which only two of three policy goals can be simultaneously
achieved: widespread access, fair terms, and profitable provision.
43
Regulatory approaches have generally targeted one or another element of
the financial inclusion trilemma without recognizing the nature of the
trilemma, namely that interventions targeting one of the three elements
force a choice between the other two elements.
For example, state laws that regulate the terms of financial service
contracts, such as usury laws and laws restricting rollovers and renewal of
loans, address the fairness problem. Such laws force a choice between
widespread availability of small-dollar credit and the stand-alone
profitability of those products. Generally, the result of increased regulation
has been a contraction in the availability of small-dollar credit.
44
Likewise,
deregulatory initiatives aimed at expanding access to credit inevitably
40. See, e.g., The Fair Access to Check Cashing Act: Hearing on S. 2110 Before the
Subcomm. on Consumer Affs. of the S. Comm. on Banking, Hous., & Urb. Affs., 100th Cong.
(1988).
41. A Google n-gram for “financial inclusion” shows that the term was virtually unused
prior to 2000.
42. See ADAM J. LEVITIN & SUSAN M. WACHTER, THE GREAT AMERICAN HOUSING
BUBBLE: WHAT WENT WRONG AND HOW WE CAN PROTECT OURSELVES IN THE FUTURE 41-57
(2020).
43. See supra note 14 and accompanying text.
44. See, e.g., Colleen Honigsberg, Robert J. Jackson Jr. & Richard Squire, How Does
Legal Enforceability Affect Consumer Lending? Evidence from a Natural Experiment, 60 J.L. &
ECON. 673, 675 (2017) (noting that lenders responded to a change in usury law by extending less
credit); Efraim Benmelech & Tobias J. Moskowitz, The Political Economy of Financial
Regulation: Evidence from U.S. State Usury Laws in the 19th Century, 65 J. FIN. 1029, 1029 (2010)
(finding that usury laws, when binding, reduced credit availability); John D. Wolken & Frank J.
Navratil, The Economic Impact of the Federal Credit Union Usury Ceiling, 36 J. FIN. 1157, 1157
(1981) (finding that a usury ceiling reduces lending).
The Financial Inclusion Trilemma
125
result in consumer protection problems because small-dollar credit cannot
simultaneously be priced fairly and profitably on a stand-alone basis.
The financial inclusion trilemma is a theory that, like gravity, cannot
be definitively proven, but its effects are nonetheless readily observable.
This Part provides a pair of examples—the economics of bank accounts
and of payday loans—that illustrate the challenge underlying financial
inclusion, which typically involves small-dollar transactions in which
economies of scale are not possible.
A. Illustration #1: The Economics of Deposit Accounts
Traditionally, deposit accounts were stand-alone financial products; a
consumer could have a deposit account at a bank without having any other
financial relationship with the bank. This enables an analysis of the
economics of deposit accounts on a stand-alone basis. This illustration
reviews the costs a bank incurs and the income it receives from a deposit
account. It shows that small-balance accounts can be offered profitably on
a stand-alone basis only if there is substantial direct fee income, for
example, from overdraft fees.
1. Costs of Deposit Accounts
Banks incur expenses in opening and maintaining deposit accounts.
Opening an account involves verifying the consumer’s identity to comply
with anti-money laundering regulations, running a credit check on the
consumer, providing the consumer with various disclosures, and
onboarding the consumer’s information. Cost estimates here are
necessarily imprecise and somewhat dated, but they still give a ballpark
sense. A 2020 estimate indicates that it costs banks on average $280 to
onboard a new customer through brick-and-mortar infrastructure and $120
to onboard a customer in a digital-only process.
45
These figures are roughly
in line with a 2010 American Bankers Association estimate that it cost
between $150 and $200 for a bank to open an account.
46
Maintaining an account also has costs. Some expenses are account-
specific: the bank must pay federal deposit insurance premiums, provide
periodic statements, make interest payments, and incur fraud losses. If a
consumer overdraws and then abandons an account, the bank is left with
the loss on the overdraft in the first instance; the ability to overdraft means
45. Vincent Bezemer, Digital Onboarding and Origination: The Cure for Banks’
Customer Acquisition Pains, ABA BANK MKTG. (Aug. 11, 2020), https://bankingjournal.aba.com/
2020/08/digital-onboarding-and-origination-the-cure-for-financial-institutions-customer-
acquisition-pains [https://perma.cc/AH6V-WBFM].
46. The Cost of a Checking Account, AM. BANKERS ASS’N (June 2010) (on file with
author).
Yale Journal on Regulation Vol. 41:109 2024
126
any payment or withdrawal transaction from a deposit account creates
credit risk for the bank.
Other expenses are in the nature of overhead. The bank must
maintain technology systems that enable transfers in and out of the account
and calculation of balances and funds availability; operate call centers,
brick and mortar branches, and websites; and ensure regulatory
compliance. That means paying for physical space and utilities, as well as
compensation for employees. Additionally, the bank must engage in
marketing to prospect for new customers.
Allocation of overhead expenses to accounts is again imprecise and
varies with economies of scale, such that the marginal cost of an additional
account might be substantially lower than the average cost.
47
The
American Bankers Association estimated in 2010 that it cost between $200
and $300 annually for a bank to maintain an account.
48
Other sources
estimate annual account maintenance costs ranging from $175 to $450,
including all possible overhead.
49
Even admitting the imprecision and age
of cost estimates of opening and maintaining a deposit account, and that
these are average, not marginal costs, it should be clear that these costs are
far from minimal.
2. Spread Income
Banks earn money on deposit accounts in three ways: interest rate
spreads, direct fees, and indirect fees. The interest rate spread—also
known as the “net interest margin”—is the difference between the interest
the bank pays the depositor and what the bank is able to earn by
reinvesting the deposit.
Banks consistently average net interest margins of around 3%.
50
This
means that for small-dollar deposits, the interest spread is insufficient to
cover the costs of maintaining the accounts. Lower-income consumers tend
to have low and often negative balances, such that a bank cannot profitably
offer them a deposit account based solely on spread income.
Consider an account with an average balance of $100 over a year. If
the bank makes a 3% net interest margin on the account, it will earn all of
47. Some overhead expenses are lumpy: a call center might have the capacity to handle
each of 100,000 additional customers with little marginal cost, but once it reaches capacity, a new,
expensive call center will be needed.
48. AM. BANKERS ASS’N, supra note 46. Other, older sources, however, put the price of
account maintenance significantly lower, between $48 and $145 per year. FED. RSRV. BD.,
FUNCTIONAL COST & PROFIT ANALYSIS 129 (1997) (on file with author) (finding that the cost of
a fully loaded account is $145). See also Ralph Haberfeld, Cognitive Dissonance, Microeconomics,
and Checking Accounts, BANKSTOCKS (Mar. 4, 2002), (on file with author) (estimating variable
costs of $48).
49. Bank Accounts: More Fees Are Coming. How to Fight Back—or Flee., CONSUMER
REPS. (Feb. 2012), https://www.consumerreports.org/cro/magazine/2012/02/bank-accounts/
index.htm [https://perma.cc/5Y9L-7VYT].
50. Quarterly Banking Profile: First Quarter 2021, 15 FDIC Q. 1, 1 fig.2 (2021).
The Financial Inclusion Trilemma
127
$3 of spread on the account over the year. Three dollars of annual net
revenue per account will come nowhere close to offsetting the bank’s
marginal or average operating costs per account. The monthly balance
statements that have to be mailed at $0.60 per stamp result in $7.20 in
annual postage costs alone. And this ignores the other costs associated with
the mailing—paper, ink, labor—much less the other account-specific
expenses or the share of overhead that must be attributed to the account.
Indeed, assuming a 3% net interest margin and $200 in annual
operating costs per account, a bank cannot profitably offer an account with
a balance of less than $6,667 solely on the basis of spread income. Account
opening expenses must also be considered. To recoup an additional $200
in account opening expenses over five years would require the consumer
to maintain an average balance of $8,000. In 2022, the median balance in a
transaction account among families owning at least one asset (virtually all
families) in the United States was $8,000,
51
so banks are unable to
profitably serve half of banked households solely on spread income. Even
if one assumes that three-quarters of the costs are attributable to overhead,
such that they do not figure in the marginal costs of an additional account,
the point still stands—a bank would not be able to recoup its costs based
on spread income alone over five years from customers with average
deposit balances of less than $2,000.
It should be clear, then, that spread income alone is insufficient to
enable a bank to profitably serve a low-balance customer. Instead, the
bank must look to supplement the spread income with indirect or direct
fee income or revenue from cross-selling other products. As the following
sections discuss, indirect fee income, even when combined with spread
income, is not sufficient to cover the costs of serving a low-balance
consumer, while both direct fee income and cross-selling income pose
fairness problems.
3. Indirect Fee Income
The funds in many deposit accounts are accessible through a debit
card. Offering a debit card imposes some costs on banks, not least the
provision of the card itself, but banks earn revenue in the form of
“interchange fees” on every debit card transaction. The interchange fee is
not charged directly to the consumer, but is instead a fee paid by the
merchant’s bank to the consumer’s bank.
52
Interchange fees are set by
51. Aditya Aladangady, Jesse Bricker, Andrew C. Chang, Sarena Goodman, Jacob
Krimmel, Kevin B. Moore, Sarah Reber, Alice Henriques Volz & Richard A. Windle, Changes in
U.S. Family Finances from 2019 to 2023: Evidence from the Survey of Consumer Finances, BD. OF
GOVERNORS OF THE FED. RSRV. SYS. 16 (Oct. 2023), https://www.federalreserve.gov/
publications/files/scf23.pdf [https://perma.cc/L8YB-B7V3].
52. ADAM J. LEVITIN, CONSUMER FINANCE: MARKETS AND REGULATION 351-52 (2d
ed. 2022).
Yale Journal on Regulation Vol. 41:109 2024
128
debit card network associations,
53
but for larger banks—those with over
$10 billion in assets—the fees are capped by law.
54
The cap is currently set
at $0.21 plus 0.05% of the transaction value and a possible $0.01 fraud
prevention expense adjustment.
55
Interchange fees have for years been at the center of a ferocious
antitrust fight between merchants and banks,
56
but unlike direct fees, they
do not present immediate consumer protection problems. Debit
interchange fees, however, are quite small per transaction. Given that the
average-sized debit transaction in 2021 was only $46,
57
the interchange
revenue on the transaction would be a bit over $0.24. Even if a consumer
completed forty debit transactions every month for a year, the total
interchange revenue on the account would be $116.64, still not enough, in
addition to spread income, to offset account operating costs, much less
opening costs.
4. Direct Fee Income
A banking business model that relies on direct fee income is
problematic because competitive dynamics incentivize banks to try to
charge consumers hidden or at least less salient fees. In other words, a fee-
driven business model incentivizes billing tricks and traps and hidden fees,
which can readily veer from the merely crafty to the outright deceptive.
The price a bank charges a consumer to maintain a deposit account
can readily be partitioned into several fees. For example, instead of a single
annual account fee of $100, partitioned pricing might consist of a $6
monthly maintenance fee, a $2 monthly paper statement fee, a $2 monthly
check-writing fee (reduced to $1 if the consumer has direct deposit), and a
$12 annual online banking fee.
Partitioned pricing has three effects. First, it makes the total cost of a
product harder for a consumer to understand—the consumer must add up
all the different components of the price. This can be challenging for all
consumers, not just for those with limited numeracy and mathematical
skills, because some fees might be behaviorally contingent, charged only
when the consumer does or does not do something. Second, partitioned
pricing impedes comparison shopping by making prices non-
commoditized. And third, partitioned pricing makes all the individual
pricing components look smaller than a single price even if they are larger
53. Id. at 352.
54. 15 U.S.C. § 1693o-2 (2018).
55. 12 C.F.R. § 235.3-4 (2023). The Federal Reserve Board has proposed lowering the cap
to $0.144 plus 0.04% of the transaction value and a $0.013 fraud prevention adjustment. Debit
Card Interchange Fees and Routing, 88 Fed. Reg. 78100, 78122 (proposed Nov. 14, 2023).
56. See, e.g., Adam J. Levitin, The Antitrust Superbowl: America's Payment Systems, No-
Surcharge Rules, and the Hidden Costs of Credit, 3 BERKELEY BUS. L.J. 265, 269 (2005).
57. Issue 1232, NILSON REPORT 11 (Dec. 2022) (on file with author).
The Financial Inclusion Trilemma
129
in the aggregate. This is particularly true if the different price components
are not all disclosed together in the same place. None of this is illegal, but
it facilitates supracompetitive pricing for the entire industry.
A bank that charges a fixed, upfront fee, such as a monthly account
fee, is at a competitive disadvantage to a bank that charges a contingent,
back-end fee because the contingent, back-end fee is less salient to the
consumer than a definite upfront fee.
58
Contingent fees will be less salient
to a consumer because they are, at least in theory, avoidable, and because
the consumer is unlikely to accurately estimate the frequency of their
occurrence. Consumers are likely to overvalue the fixed fee relative to the
contingent fee and therefore prefer the contingent fee, all else being equal.
Consider, for example, overdraft fees. Overdraft fees are a contingent
fee—the fee is only charged if the consumer overdraws the account. It is
difficult, however, for a consumer to accurately predict the likelihood of
overdrafting, much less how many fees will actually be applied. This is
because whether a consumer overdraws is dependent upon the order in
which the bank posts credits and debits of various sorts to the account.
Posting orders are frequently quite complicated and are ultimately set at
banks’ discretion, so long as banks reserve such discretion in the account
terms and conditions.
59
The lack of clarity about when an overdraft fee will
be charged makes such fees opaque and hard for consumers to estimate,
with the result that consumers are likely to simply disregard them.
This means that a bank that does not permit or charge for overdrafts,
but instead charges a fixed monthly account fee, say $10 per month, is at a
competitive disadvantage relative to a bank that charges a $35 overdraft
fee under an opaque overdraft fee policy, because the bank with the certain
monthly fee will appear to be more expensive than its rival with
behaviorally contingent overdraft fees, even if the fixed fee is in fact likely
to be less expensive for the consumer.
The trade-off between consumer protection and financial inclusion
with regard to overdraft fees is well illustrated by a New York Federal
Reserve Bank staff report on the effect of a 2001 Office of the Comptroller
of the Currency regulation that preempted state laws restricting overdraft
fees for national banks.
60
The report found that the exempted banks
increased both their overdraft fees and provision of overdraft credit.
61
The
exempted banks also expanded the deposit account supply by lowering
58. See Xavier Gabaix & David Laibson, Shrouded Attributes, Consumer Myopia, and
Information Suppression in Consumer Markets, 121 Q.J. ECON. 505, 509 (2006). Credit cards are
the classic example of this, having largely shifted away from annual fees (fixed, upfront) to late
and overlimit fees (contingent, back-end).
59. See Gutierrez v. Wells Fargo Bank, N.A., 704 F.3d 712, 723, 726, 729-30 (9th Cir. 2012)
(upholding liability for misleading statements about transaction processing order, but overruling
injunction of high-to-low posting order based on preemption principles without considering
whether the bank acted in bad faith).
60. Dlugosz et al., supra note 17, at 3.
61. Id.
Yale Journal on Regulation Vol. 41:109 2024
130
minimum balance requirements for deposit accounts, which corresponded
with an increase in the share of low-income households with checking
accounts.
62
The report’s findings suggests that there is a messy trade-off
between consumer protection and financial inclusion: lower minimum
balance requirements likely made it possible for more low-income
households to obtain checking accounts—on which they then paid more
and incurred higher overdraft fees that they could ill-afford. This episode
illustrates the trilemma in action: widespread financial inclusion can be
profitable to banks, but only at the expense of fairness.
Nor are such contingent “fees” on deposit accounts limited to formal
fees with precise dollar amounts. They also extend to other contractual
terms that shift value between consumers and banks, particularly dispute
resolution terms, such as jury trial waivers, class action waivers, and
binding mandatory arbitration. The right to bring or participate in a class
action or to have a dispute heard publicly by a court and before a jury is
potentially quite valuable, but only when a dispute arises. The availability
of such legal recourse is unlikely to be assigned much value by a consumer
when the consumer enters the contract because the consumer does not
anticipate a dispute that is serious enough to merit litigation; consumers
who anticipate such problems are likely to avoid the financial service
provider. A consumer is likely to anticipate only an infinitesimally small
chance of litigation and therefore will rationally discount the value of the
waiver of the right to a jury, to participate in a class action, or to have a
court hear their case in public.
While the chance of a dispute with any individual consumer is very
low, the bank is not concerned with the likelihood of a dispute with any
individual consumer, but with the aggregate likelihood of a dispute, which
is, by definition, greater. Accordingly, the bank will value the jury or class
action waiver or arbitration clause more highly than any consumer because
the possibility of a dispute with some customer is not so remote. The jury
or class action waiver or arbitration clause thus has the same effect of
shifting value from the consumer to the bank as a fee, but its monetary
value is completely hidden. While jury and class action waivers and
arbitration clauses are part of the cost of a contract, their cost is opaque
and not salient to consumers who cannot take an actuarial approach to
valuation because they only conduct single transactions, unlike a business
that conducts multiple transactions.
Competition punishes transparent upfront pricing and pushes the
entire market to opaque, back-end, behaviorally contingent pricing. Thus,
virtually all deposit account agreements waive the depositor’s right to a
jury trial, waive the depositor’s right to participate in a class action, and/or
require disputes to be resolved behind closed doors through binding
62. Id.
The Financial Inclusion Trilemma
131
mandatory arbitration.
63
Similarly, as Figure 3 shows, this competitive
dynamic was accompanied with rapid growth in total overdraft fees
charged, starting in 1992 and continuing until 2010, when regulations went
into effect prohibiting the charging of fees for overdrafting on ATM and
one-time (non-repeating) debit card transactions without affirmative
consumer opt-in.
64
Figure 3. Total Annual Bank Overdraft Revenue (in Billions of Dollars)
65
63. Hidden Risks: The Case for Safe and Transparent Checking Accounts, PEW
CHARITABLE TRUSTS HEALTH GRP. 18 (Apr. 2011), https://www.pewtrusts.org/~/media/legacy/
uploadedfiles/pcs_assets/2011/SafeCheckingPewReportHiddenRiskspdf.pdf
[https://perma.cc/P43Y-J2Y8].
64. See Electronic Fund Transfers, 74 Fed. Reg. 59,033 (Nov. 17, 2009) (effective January
19, 2010, with a mandatory compliance date of July 1, 2010) (codified at 12 C.F.R. § 1005.17(b)
(2023)). A number of smaller banks still rely on overdraft fees as the primary driver of their
profitability. See Aaron Klein, A Few Small Banks Have Become Overdraft Giants, BROOKINGS
(Mar. 1, 2021), https://www.brookings.edu/articles/a-few-small-banks-have-become-overdraft-
giants [https://perma.cc/G3DG-5WZA]. A 2021 CFPB study found that banks maintain “[d]eep
[d]ependence on [o]verdraft fees,” CFPB Research Shows Banks’ Deep Dependence on Overdraft
Fees, CONSUMER FIN. PROT. BUREAU (Dec. 1, 2021), https://www.consumerfinance.gov/about-
us/newsroom/cfpb-research-shows-banks-deep-dependence-on-overdraft-fees [https://perma.cc/
26QY-UJ8C], but a subsequent study found that overdraft feeds had fallen significantly since the
pandemic, in part because of reduction and elimination of fees at some banks, Overdraft/NSF
Revenue Down Nearly 50% Versus Pre-Pandemic Levels, CONSUMER FIN. PROT. BUREAU (May
24, 2023), https://www.consumerfinance.gov/data-research/research-reports/data-spotlight-
overdraft-nsf-revenue-in-q4-2022-down-nearly-50-versus-pre-pandemic-levels/full-report
[https://perma.cc/BLN7-GJLK].
65. Data procured from Moebs Services. All figures have been adjusted for inflation using
the Consumer Price Index in January 2000 as the baseline. Dollars are inflation-adjusted to 2020
dollars.
$0
$5
$10
$15
$20
$25
$30
$35
$40
$45
$50
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Yale Journal on Regulation Vol. 41:109 2024
132
The aversion to fixed, upfront fees relative to contingent, back-end
fees is particularly important for small-balance deposits because a
consumer with a $100 balance will likely balk at fixed, unavoidable fees
that start to approach that total balance, even if the contingent fees in total
could readily be greater.
Competitive pressure creates an arms race not just for price opacity,
but to actively make products look cheaper than they are. As banks try to
gain competitive advantage, making their product offerings look cheaper
by shifting from salient fixed fees to less salient contingent fees and
contract terms, banks have an incentive to engage in behavior that further
obfuscates the cost of their products, such as deceptive advertising and
misleading disclosures that downplay or hide costs or simply making terms
hard to access.
66
Such behavior might be unfair, deceptive, or abusive.
67
At
its extreme, it means literally covering up fees.
68
A reliance on fee income
incentivizes potentially problematic practices.
5. Relationship Banking
Over the past decade, another revenue model has emerged related to
deposit accounts: relationship banking. In the relationship banking model,
deposit accounts are not stand-alone products. Instead, they are offered as
a loss-leader for other tied services, such as loans, insurance, or securities
products.
69
Thus, rather than offering a stand-alone, free checking account,
a checking account is now “free”—provided that the consumer obtains
other services from the bank.
There are two consumer protection problems with this product tying.
First, the bundled pricing is much more complicated and opaque, making
it difficult for a consumer to tell if she is getting the best price on either the
deposit account or the bundled product. In this regard, bundling operates
as the inverse twin of partitioned pricing, an augmentation of what is
included in the price instead of a diminution. For example, if the bundled
product is a loan, the consumer could compare the costs of the loan to that
of other loans, but that comparison would neglect the “discount” the
consumer gets because of the bundled “free” checking account. The
comparison is now apples to oranges.
66. See, e.g., First Amended Complaint at ¶¶ 106-118, Consumer Fin. Prot. Bureau v.
TCF Nat’l Bank, No. 0:17-cv-001660-RHK-KMM (D. Minn. Mar. 1, 2017) (alleging abusive and
deceptive sales tactics to get consumers to sign up for fee overdraft protection).
67. Recognizing this, Congress has prohibited such behavior.12 U.S.C. § 5531 (2018).
68. See, e.g., Complaint at ¶¶ 20, 21, 67, Consumer Fin. Prot. Bureau v. All Am. Check
Cashing, Inc., No. 3:16-cv-00356-WHB-JCG (S.D. Miss. May 11, 2016) (alleging that defendant
physically covered up the fee amount on receipts, minimized the time the consumer had to view
the receipt, and interfered with the consumer’s ability to see the sign listing fees).
69. Final Nail in the Free Checking Coffin: What Can Replace Free Checking?, MOEBS
SERVS. 1 (Sept. 30, 2019), https://moebs.com/Portals/0/pdf/Articles/Final%20Nail%20in%20
the%20Free%20Checking%20Coffin.pdf [https://perma.cc/D4L5-922Q].
The Financial Inclusion Trilemma
133
This obfuscation of pricing is precisely what banks desire because loan
products in particular should be among the most commoditized products
around. Given miniscule profit margins, all that the bank is selling is
money, and money is completely fungible with money from other sources.
Avoiding commodification helps the bank, but at the expense of the
consumer.
The other problem with product tying in relationship banking is that
it incentivizes the bank to constantly attempt to cross-sell the consumer on
other bank products. At its extreme, this cross-selling incentive is what
produced the Wells Fargo fake account scandal: Wells Fargo employees’
compensation and employment depended on their ability to sell the
maximum number of products to consumers. When those incentives ran
into consumers’ limited demand for financial products, the result was the
creation of fake accounts.
70
While Wells Fargo might be the extreme case, it is hardly the only
example of pressures to cross-sell resulting in deceptive practices. The
Consumer Financial Protection Bureau (CFPB) entered into consent
orders with U.S. Bank for similarly opening unauthorized deposit and
credit card accounts.
71
It has also brought suit against Fifth Third Bank for
fake account creation.
72
Likewise, the FDIC and CFPB entered into a $214
million consent order with Discover Bank, which allegedly engaged in
deceptive marketing and sales tactics to get consumers to purchase certain
credit card add-on products that provided insurance-type services and
credit score tracking.
73
The sales tactics included using misleading language
depicting the products as free “benefits,” suggesting that there would be
an opportunity to review printed materials before being charged, omitting
key exclusions from the insurance-type product coverage, and speaking
unusually fast when disclosing product prices and terms.
74
The CFPB has
entered into similar consent orders with American Express
75
and Chase
76
for pushing “add-on” products onto their credit card customers. And the
70. Consent Order at ¶¶ 7-16, Wells Fargo Bank, N.A., No. 2016-CFPB-0015 (Sept. 8,
2016).
71. Consent Order at ¶¶ 24-31, Bank of Am., N.A., No. 2023-CFPB-0007 (July 11, 2023);
Consent Order at ¶¶ 7-12, U.S. Bank, N.A., No. 2022-CFPB-0006 (July 28, 2022).
72. Amended Complaint at ¶¶ 4, 20-39, 53, Bureau of Consumer Fin. Prot. v. Fifth Third
Bank, N.A., No. 1:21-cv-00262-DRC (S.D. Ohio June 16, 2021) (alleging that Fifth Third
employees opened accounts in consumers’ names without consumers’ knowledge because of cross-
selling pressure on employees).
73. Joint Consent Order, Order for Restitution, and Order to Pay Civil Monetary Penalty
at ¶ 4-5, Discover Bank, FDIC-11-548b, FDIC-11-551k, 2012-CFPB-0005 (Sept. 24, 2012).
74. Id.
75. Consent Order, Am. Express Centurion Bank, No. 2013-CFPB-0011 (Dec. 24, 2013);
Consent Order, Am. Express Bank, FSB, No. 2013-CFPB-0013 (Dec. 24, 2013); Consent Order,
Am. Express Travel Related Servs. Inc., No. 2013-CFPB-0013 (Dec. 24, 2013).
76. Consent Order, JPMorgan Chase Bank, N.A., No. 2013-CFPB-0007 (Sept. 19, 2013).
Yale Journal on Regulation Vol. 41:109 2024
134
CFPB sued TCF National Bank for deceptively pressuring its deposit
customers to opt into for-fee overdraft protection.
77
Outright deceptive practices are not the only consumer protection
concern in a cross-selling situation. When a bank cross-sells, it turns itself
into a data platform, using data from the consumer’s account to support its
marketing of other products. While this might result in the consumer
learning about beneficial products, it might also result in the bank targeting
the consumer for products that are suboptimal for the consumer, but more
profitable for the bank. The bank is in essence relying on its past
relationship of trust with the consumer to sell additional products. As with
a platform like Amazon, the consumer cannot opt out of such cross-selling,
because there are few restrictions on the solicitation of consumers’
business based on the sharing of consumers’ personal data among affiliated
entities.
78
6. Summary
Deposit accounts are simply not profitable on a stand-alone basis for
low-balance accounts absent direct fee income or cross-selling income.
Business models that depend on direct fees and cross-selling are not illegal,
but prime the ground for consumer protection abuses because competitive
pressure pushes banks to obfuscate costs and aggressively push other
products onto the consumer. This suggests that it is not possible, on a per
consumer basis, to profitably engage in widescale provision of deposit
accounts to lower-income consumers without risking serious consumer
protection problems. The result is that some banks offer accounts to
consumers expected to have small balances, but with pricing that raises
fairness concerns, while many others simply do not serve low-balance
accounts. Instead, as Part III discusses, low-balance deposit accounts can
only be both widely and fairly offered if subsidized, whether by taxpayers,
other customers, or bank shareholders.
B. Illustration #2: The Economics of Payday Loans
The economics of payday loans provide a further illustration of the
financial inclusion trilemma as applied to small-dollar credit. Payday
loans—sometimes called deferred deposit advances—are a type of short-
term, small-dollar loan. The terms of payday loans vary by state, depending
on statutory limits on the rate, duration, and loan size, but a typical payday
loan is a $400 loan for two weeks with a $50 fee.
77. First Amended Complaint, supra note 66 at ¶¶ 35-48, (alleging pressure on employees
to cross-sell overdraft protection to deposit customers).
78. 15 U.S.C. § 1681s-3(a) (2018) (requiring consumers to have a right to opt out of
solicitations based on consumer report data shared between affiliates).
The Financial Inclusion Trilemma
135
In a traditional payday loan transaction using these numbers, the
borrower would give the lender a post-dated check for $450 in exchange
for $400 in cash. The borrower must repay the lender $450 within two
weeks or the lender will deposit the post-dated check. The loan is for $400,
and the $50 difference between the check amount and the amount
disbursed is the loan fee. There is no explicit interest rate charged on
payday loans. But such a loan has an APR of 326%.
79
Consumer advocates have called for a national 36% APR usury rate,
80
arguing that loans with prices higher than 36% APR are inherently unfair
to borrowers.
81
The Military Lending Act of 2006 already imposes a 36%
APR rate cap on certain loans made to active duty military members and
their dependents.
82
Whatever the wisdom of such a 36% APR cap, it must
be recognized that it is not possible to profitably make a $400 loan to a
consumer for two weeks at a 36% APR on a stand-alone basis.
Achieving a 36% APR for a two-week loan of $400 would require
limiting fees to about $5.50. That $5.50 would need to cover the lender’s
credit losses, cost of funds, other variable per loan costs, and fixed
overhead costs, as well as a profit margin. As the following sections show,
some simple assumptions show that it is not possible to come anywhere
close to such a cost structure.
1. Store-Front Payday Lending
Assume an independent, store-front payday lender. It might operate
one or several locations. Its revenue comes solely from fees on its loans. It
makes a fee every time a loan is made and again whenever the loan is rolled
over (that is, refinanced with a new loan). Let’s assume that the loan is
subject to a 36% APR cap, so the lender cannot charge more than $5.50
for a two-week loan of $400. Can the lender make such a loan profitably?
Let’s start with labor costs. Assume that the lender’s employees make
$15 an hour, including benefits. An employee is capable of processing
several loans in an hour, but actual loan production per employee is
incredibly low in the payday industry because of the large number of
competitors. In 2014, there were 15,766 payday lending stores in the United
79. The APR for a single payment closed-end product with a term of less than a year is
calculated by taking the product of (a) 100, (b) the quotient of the days in a year (365) over the
number of days of the product term (14), and (c) the difference of the quotient of the total amount
required to be repaid ($450) over the amount advanced ($400) and one. Annual Percentage Rate
Computations for Closed-End Credit Transactions, 12 C.F.R. § 1026, Appendix J(c)(5) (Form 1)
(2023). Thus, here: APR = 100 x 365/14 x (450/400 - 1) = 324.89%
80. See Veterans and Consumers Fair Credit Act, H.R. 5050, 116th Cong. (2019).
81. See, e.g., Why Cap Interest Rates at 36%?, NAT’L CONSUMER L. CTR. (Aug. 2021),
https://www.nclc.org/wp-content/uploads/2022/09/IB_Why_36.pdf [https://perma.cc/TQ3H-
85BS].
82. 10 U.S.C. § 987 (2018). The APR under the Military Lending Act is calculated using
a broader definition of the finance charge than under the Truth in Lending Act. 10 U.S.C.
§ 987(i)(4) (2018).
Yale Journal on Regulation Vol. 41:109 2024
136
States, concentrated in thirty-six states.
83
The average payday lending store
makes between 265 and 6,327 loans per year, depending on the state,
84
with
a national average of 3,541 loans per year.
85
The average payday lending
store had three full-time employees.
86
Using these averages, we can see that on average it takes over 105
minutes of employee time for a payday lender to make a single loan.
87
To
be clear, the actual processing takes far less, but the lender has to pay its
employees regardless of whether there are customers with loans to process.
At this pace, the labor costs for making a single loan are $26.43. Recouping
these costs via a $26.43 charge on a $400 loan for two weeks would translate
to a 172% APR.
88
The labor costs alone mean that it is not possible for a
payday lender to profitably lend at anything close to a 36% APR (a $5.52
finance charge here). And this is not counting other loan-specific costs—
cost of funds and credit losses—or fixed and semi-variable expenses like
rent, utilities, insurance, technology systems, advertising, customer service,
and legal expenses, much less sufficient profit to attract investment in the
business.
Thus, although the pricing of payday loans is shockingly high, there
do not generally appear to be supracompetitive profits in the payday loan
industry. Barriers to entry are low: payday lending requires little capital
because the loans are so small. Although there are some large, publicly-
traded payday lending chains, many payday lenders are closely-held small
businesses. Moreover, the state licenses required are generally a pro forma
matter, in contrast to a banking charter. Therefore, even if there were
supracompetitive profits, low barriers to entry mean that competition
would quickly dissipate such an inefficiency. Indeed, low barriers to entry
actually prevent lenders from realizing economies of scale with the effect
that payday lenders cannibalize each other’s businesses, driving up the
83. Payday, Vehicle Title, and Certain High-Cost Installment Loans, 81 Fed. Reg. 47864,
47871 (proposed July 22, 2016) [hereinafter Payday Rule]. In contrast, there were 14,350
McDonald’s stores, spread across all fifty states that year. Id.
84. Montezemolo, supra note 32, at 26.
85. Id. at 26 n.2. See also Payday Lending in America: Policy Solutions, PEW
CHARITABLE TRUSTS 18 (Oct. 2013), https://www.pewtrusts.org/~/media/legacy/uploadedfiles/
pcs_assets/2013/pewpaydaypolicysolutionsoct2013pdf.pdf [https://perma.cc/5ZCB-BEAX]
(estimating fewer than 500 unique customers per store per year).
86. Payday Rule, supra note 83, at 47871.
87. The national average of 3,541 loans divided by 6,240 (the total employee hours
product of three employees working forty hours a week each for fifty-two weeks of the year) yields
approximately 0.57 loans per employee hour, which means that it takes 105.73 minutes of
employee time to produce a single loan. At $15 an hour, it takes $26.43 to produce that single loan.
88. Although the marginal costs of making an additional loan are small, payday lenders
cannot price according to marginal cost, as they charge all borrowers the same rate. Accordingly,
they have to price based on average cost.
The Financial Inclusion Trilemma
137
fixed costs per loan, which drives up loan pricing.
89
Payday lending presents
the unusual case where greater competition increases prices.
Again, the point here is not the specific numbers being used, but that
they illustrate the fundamental problem: even with assumptions of zero
credit losses and zero profit margin, the labor costs alone of making a
payday loan render it an uneconomical stand-alone product at a 36% APR.
Thus, high fees are required to make payday loans a profitable product.
High fees alone are insufficient, however, for payday loan
profitability. Rollovers are key to payday lender profitability. Payday
lenders incur credit losses estimated to be in the range of 50% to 67% of
dollars loaned.
90
Given such high credit losses, payday lending is only
profitable because of rollovers and renewals. When a borrower rolls over
an existing loan, the borrower does not receive an additional advance of
cash, but merely retains the cash already advanced for a longer period, in
exchange for paying an additional fee. Thus, while the lender’s total credit
exposure has not increased with a rollover, the lender has collected an
additional fee.
Not surprisingly, rollovers are the profit center for payday lenders.
The CFPB found that over the course of a year, 90% of all payday loan
fees came from consumers who borrowed seven or more times, and 75%
of fees came from consumers who borrowed ten or more times.
91
This
means that the payday loan product—marketed as a short-term loan
product to bridge liquidity gaps—is only profitable if it is a long-term debt
trap for a significant number of borrowers who face not liquidity, but
solvency problems. Both the high cost and the likelihood of a long
borrowing sequence raise considerable consumer protection concerns with
payday loans.
2. Bank Payday Products
The economics of payday loans change when they are offered not as
a stand-alone product, but as an additional product at a pre-existing
business. If only the marginal costs of the payday loans are considered, not
their pro-rated share of labor or overhead (which would be incurred even
if they were not offered), then payday loans can be “profitably” offered at
lower prices. Indeed, this is the conceit behind proposals for the United
States Postal Service to offer payday loans.
92
89. See Trial, Error, and Success in Colorado’s Payday Lending Reforms, PEW
CHARITABLE TRUSTS 5-7 (Dec. 2014), https://www.pewtrusts.org/~/media/assets/2014/12/
pew_co_payday_law_comparison_dec2014.pdf [https://perma.cc/CYN9-PUL2] (finding that
Colorado’s 2010 payday loan law reform resulted in the number of lenders falling by
approximately 50%, but because loan volume remained steady, per store borrowers roughly
doubled and loan prices fell).
90. Payday Rule, supra note 83, at 47874.
91. Id.
92. See supra note 19.
Yale Journal on Regulation Vol. 41:109 2024
138
Likewise, some banks and credit unions have been offering payday-
loan-type products to existing consumers.
93
These bank and credit union
payday products have a lower cost than independent payday loan products.
It is unclear if these institutions are able to offer their payday-type products
profitably on a stand-alone basis, or if they are cross-subsidizing from other
products or consumers. It seems unlikely, however, that these products are
able to generate enough revenue to cover their pro-rated share of labor
and overhead. At best, the revenue generated covers the products’
marginal costs.
Additionally, the banks and credit unions are likely able to offer these
products at lower cost because they benefit both from pre-existing
coverage of overhead costs and from a cream-skimming effect: their
borrowers are not typical payday borrowers, but are pre-existing (and
long-standing) depositors who pose less credit risk than the typical payday
borrower. Not only are the borrowers themselves unrepresentative of the
general population of payday borrowers, but by virtue of the pre-existing
relationship, the bank or credit union also has substantially more
information about the borrower and the ability to offset funds the moment
they come into the borrower’s account, beating out all competing creditors
for the funds. In contrast, a regular payday lender has to guess when there
will be funds in a customer’s bank account and compete with other
creditors to grab those funds. Moreover, the borrower with a pre-existing
banking relationship with the lender feels a greater relational pressure to
repay the loan, lest a default endanger the borrower’s deposit account
relationship (and perhaps other product relationships with the bank or
credit union).
3. Online Payday Lending
Online lending holds out the promise of a reduced cost structure,
particularly by way of reduced overhead. Yet online payday loans are not
cheaper than store-front products. In fact, they tend to be more expensive.
Part of this is a function of the adverse selection problem faced by online
lenders, which are likely to attract borrowers nationwide who are unable
to obtain local storefront credit.
94
But the cost of online loans is also a
function of online lenders’ much higher costs of customer acquisition.
93. See Ann Carrns, An Alternative to Payday Loans, but It’s Still High Cost, N.Y. TIMES
(Sept. 21, 2018), https://www.nytimes.com/2018/09/21/your-money/alternative-payday-loans-high-
interest-us-bank.html [https://perma.cc/9KLH-9LP9] (describing U.S. Bank’s “Simple Loan”
product, which loans between $100 and $1,000 at a cost of twelve dollars for each $100 borrowed);
12 C.F.R. § 701.21(c)(7)(iii)-(iv) (2023) (regulating payday alternative loans).
94. Online lenders have higher default rates than storefront payday lenders. Payday Rule,
supra note 83, at 47990 (noting a 41% default rate for online loans, compared to a 17% default
rate for storefront loans, 55% sequence default rate for online loans (an eventual default in a
sequence of rolled-over loans by the same borrower), and a 34% sequence default rate for
storefront loans).
The Financial Inclusion Trilemma
139
Storefront lenders generally rely on foot traffic to generate business.
For that reason, they are frequently situated at busy intersections.
95
Online
lenders are not able to attract customers through their physical location.
While some lenders make loans to consumers who access their websites
directly, most obtain customers through lead generators. As many as 75%
of online payday loans are originated through lead generators.
96
Payday loan lead generators enable otherwise competing small
lenders to pool advertising resources. A lead generator advertises for
potential borrowers, but does not make loans itself.
97
Instead, the lead
generator collects payday loan applications from borrowers and auctions
the “lead” and associated loan application off in real time to prospective
lenders. The winning bidder gets the right to contact the borrower using
the information in the loan application. If the winning lender cannot close
the deal, the lead is reauctioned as a “second look” at a lower price.
The lead generator system helps online lenders avoid a common pool
collective action problem in advertising. To beat out other lenders for
customers, an online lender will generally have to out-spend its
competitors for advertising. The result is an arms race in advertising
because online lenders have to spend more and more to have their ads be
the ones consumers see—the top search hit, for example. The result of such
an advertising arms race would be to reduce all lenders’ profit margins by
increasing everyone’s costs without necessarily expanding the pool of
potential borrowers.
The lead generator system has the effect of pooling advertising. The
lead generator does the advertising, rather than the lenders, and the pooled
resources enable better advertising than any individual lender could
afford. The pooling benefit comes at a cost, however. Because of the
competitive auction system, online lenders must pay a substantial amount
to win each lead. A high quality, first look lead runs in the range of $150 to
$200.
98
There is no guarantee, however that the lender will even be able to
close the loan; the lead is only a right to deal with the borrower. Many leads
purchased do not result in a loan, so online lenders have to purchase
several leads to make one loan. The costs of the lead acquisition plus credit
losses, operation costs, cost of funds, and any profit must be recouped from
only those leads that result in loans.
95. Sheila R. Foster, Breaking Up Payday: Anti-agglomeration Zoning and Consumer
Welfare, 75 OHIO ST. L.J. 57, 62 (2014).
96. Andrew R. Johnson, Middlemen for Payday Lenders Under Fire, WALL ST. J. (Apr.
7, 2014, 9:50 PM), https://www.wsj.com/articles/SB10001424052702304819004579487983000120324
[https://perma.cc/NB8H-6F98].
97. Led Astray: Online Lead Generation and Payday Loans, UPTURN 1 (Oct. 2015),
https://www.upturn.org/static/reports/2015/led-astray/files/Upturn_-_Led_Astray_v.1.01.pdf
[https://perma.cc/D4QE-QYQZ].
98. Payday Rule, supra note 83, at 47878.
Yale Journal on Regulation Vol. 41:109 2024
140
To illustrate, if an online lender spends $50 to acquire a lead, and only
one in five leads results in a loan, it will cost the lender $250 to acquire one
actual customer. If that customer is charged $50 on a $400 loan, the lender
cannot even recoup its cost of customer acquisition without four rollovers,
much less its other expenses and profit margin. The high cost of leads
ensures that online lenders must charge high fees and roll over loans. It is
not possible to profitably engage in online payday lending otherwise.
The illustrations of both deposit account economics and payday loan
economics point to the same problem with small-balance deposits and
small-dollar lending: the small size of the transaction means that it is not
possible to profitably offer these products as stand-alone products on a
wide scale and at fair terms. Only two of the three goals—profitability,
scale, and fairness—can be simultaneously achieved.
III. The Financial Inclusion Policy Playbook
The previous Part examined why the financial inclusion trilemma
exists. This Part turns to the policy playbook for addressing financial
inclusion. It reviews the different policy options and how they have been
used or ignored in the United States: private provision, including
technological advances; soft mandates; hard mandates (cross-subsidies);
public options; and public subsidies.
A. Existing Approaches to Financial Inclusion
1. Private Provision: Fintech, Crypto, and Deregulation
In the United States, private provision of financial services has long
been the primary means of pursuing financial inclusion. It has not worked.
As the previous Part has shown, the economics of small deposits and small-
dollar loans require fee-based income or extremely high interest rates,
both of which are problematic from a consumer protection standpoint.
Private provision will occur only if provision is profitable, so if policy
makers rely on private provision, the choice becomes one of access versus
fairness.
The result has been different in the deposit and credit contexts. In the
deposit context, deposit accounts for low-income consumers are not widely
available. To the extent accounts are available, banks cover their costs with
overdraft and other account fees. In the credit context, loans are widely
available, but at extremely high costs. In recent years, however, a
tightening of consumer credit regulation in a number of states
99
has
99. ARK. CONST. amend. LXXXIX, § 3 (establishing a 17% usury cap in the Arkansas
state constitution); 2019 Cal. Stat. ch. 708 (establishing a 36% usury cap over the Federal Funds
rate in California); COLO. REV. STAT. § 5-3.1-105 (2020) (establishing a 36% APR cap for payday
The Financial Inclusion Trilemma
141
resulted in lower availability of credit or substitution to other forms of less-
regulated credit, such as pawn, rent-to-own, and borrowing from family
and friends.
100
i. Fintech
Fintech—a portmanteau of “financial technology”—is the newest
guise of private provision of financial services. Fintech refers to
“technology-enabled innovation in financial services that could result in
new business models, applications, processes or products with an
associated material effect on the provision of financial services.”
101
Fintech
companies, or “fintechs,” are nonbank financial service companies that
rely on technologies, such as “web-or mobile-based consumer interfaces,
automated underwriting, neural network and other machine-learning-
based underwriting, and the use of nontraditional underwriting data
sources to provide financial services to consumers.”
102
A more recent twist
on fintechs has been the use of cryptocurrency and decentralized finance.
Fintech has always held out the promise of being the silver bullet for
financial inclusion,
103
although much of that promise is directed at the
developing world rather than the United States.
104
The argument presented
by fintechs is that they (1) are better able to connect with underserved
populations because of their online presence and (2) can lower the cost of
loans in Colorado); 815 ILL. COMP. STAT. 123 / 15-5-5 (2021) (establishing a 36% military APR
usury cap in Illinois); HAW. REV. STAT. § 480J-4 (establishing 36% usury cap in Hawaii); MONT.
CODE ANN. § 31-1-701 (2010) (establishing a 36% usury cap in Montana); NEB. REV. STAT. § 45-
918 (2020) (establishing a 36% rate cap in Nebraska); OHIO REV. CODE ANN. § 1321.40
(establishing 28% usury cap in Ohio); S.D. CODIFIED LAWS § 54-4-36 (2016) (establishing 36%
interest rate cap on consumer loans in South Dakota); 2020 Va. Acts ch. 1258 (establishing 36%
APR cap in Virginia).
100. Angela Littwin, Testing the Substitution Hypothesis: Would Credit Card Regulations
Force Low-Income Borrowers Into Less Desirable Lending Alternatives?, 2009 U. ILL. L. REV. 403,
405 (2009) (finding evidence that restrictions on credit result in either lower consumption or a shift
to borrowing from family and friends).
101. Financial Stability Implications from Fintech: Supervisory and Regulatory Issues that
Merit Authorities’ Attention, FIN. STABILITY BD. 7 (June 27, 2017), https://www.fsb.org/wp-
content/uploads/R270617.pdf [https://perma.cc/CG6L-6CCU]. See also Examining Opportunities
and Challenges in the Financial Technology (“Fintech”) Marketplace: Hearing Before the
Subcomm. on Fin. Insts. & Consumer Credit of the H. Comm. on Fin. Servs., 115th Cong. 86 (2018)
(statement of Professor Adam J. Levitin) (discussing definitional problems regarding “fintechs”
and noting that fintechs are distinguished by being nonbank financial services companies that use
various forms of digital technology to provide financial services to consumers).
102. Examining Opportunities and Challenges in the Financial Technology (“Fintech”)
Marketplace, supra note 101, at 87.
103. See, e.g., id. at 5; Kalin Anev Janse & Gong Cheng, Can Fintech Make the World
More Inclusive?, KNOWLEDGE AT WHARTON (Oct. 10, 2019), https://
knowledge.wharton.upenn.edu/article/can-fintech-make-the-world-more-inclusive [https://
perma.cc/M5JN-HDYA]; Rupert Shaw, Why Fintech is the Biggest Driver of Financial Inclusion,
FINTECH MAG. (Apr. 30, 2021), https://fintechmagazine.com/financial-services-finserv/why-
fintech-biggest-driver-financial-inclusion [https://perma.cc/AUB9-AVEJ].
104. Microfinance programs abroad are simply not transferrable to the U.S. situation.
The dollar scale involved is different, and microfinance abroad is business finance, enabling
productivity. U.S. financial inclusion is not about small business, but about living expenses.
Yale Journal on Regulation Vol. 41:109 2024
142
providing financial services because they lack the legacy cost structure of
traditional brick-and-mortar financial institutions. The idea is that
technological developments—particularly the widespread dissemination
of smart phones, the development of blockchain-based applications, and
the increased use of big data, alternative data sources, and artificial
intelligence—will enable cheaper customer identification verification,
cheaper and faster payments, and underwriting that serves more
consumers more cheaply.
105
Furthermore, because fintechs lack brick-and-
mortar operations, their advent holds out the promise of eliminating
discrimination based on physical observation of consumer
characteristics.
106
Fintech companies have leveraged this vague promise of future
financial inclusion benefits to obtain immediate regulatory relief through
“sandboxes”
107
—effectively regulatory exemptions—that give them a
competitive advantage over their non-exempt traditional institution rivals.
Fintech, however, has yet to deliver on its promise in the United States.
Fintech is about a decade old now, but it has not produced material results
in terms of financial inclusion.
To be sure, certain payment fintechs have been quite successful—
PayPal and Venmo, for example—but they have not moved the needle on
financial inclusion. Instead, they have primarily poached business from
traditional banks. As noted above, a PayPal or Venmo account, for
example, can be used much like a bank account as a store of value, but it
can only be used to make payments at a limited number of merchants and
requires funding to come either from a bank account, a credit card (which
generally requires a bank account), or other PayPal or Venmo accounts. In
105. It should be noted that technological solutions can actually reduce financial
inclusion, particularly for the elderly. Technological solutions frequently require consumers to
keep track of passwords, which can be a challenge for those with memory issues associated with
senescence. Likewise, navigating apps and websites can be a challenge for digital non-natives.
106. Thomas Philippon, On Fintech and Financial Inclusion 10-11 (Bank for Int’l
Settlements, Working Paper No. 841, 2020). While this appears to have been borne out in the first
generation of fintech (predating the term), namely the automated underwriting programs used by
Fannie Mae and Freddie Mac, LEVITIN & WACHTER, supra note 42, at 85-86, there are also risks
of bias in algorithmic underwriting or in machine learning. See, e.g., Robert Bartlett, Adair Morse,
Richard Stanton & Nancy Wallace, Consumer-Lending Discrimination in the FinTech Era, 143 J.
FIN. ECON. 30 (2022); CATHY O’NEIL, WEAPONS OF MATH DESTRUCTION: HOW BIG DATA
INCREASES INEQUALITY AND THREATENS DEMOCRACY (2016); FRANK PASQUALE, THE
BLACK BOX SOCIETY: THE SECRET ALGORITHMS THAT CONTROL MONEY AND INFORMATION
35, 38-42 (2016).
107. CONSUMER FIN. PROT. BUREAU, CFPB-2018-0042, POLICY ON NO-ACTION
LETTERS (2018); CONSUMER FIN. PROT. BUREAU, CFPB-2018-0042, POLICY ON THE
COMPLIANCE ASSISTANCE SANDBOX (2018); CONSUMER FIN. PROT. BUREAU, CFPB-2018-0023,
POLICY TO ENCOURAGE TRIAL DISCLOSURE PROGRAMS (2018). Several states have also
authorized regulatory sandboxes. See, e.g., ARIZ. REV. STAT. ANN. §§ 41-5601 (2023) (authorizing
a fintech regulatory sandbox in Arizona); 2019 Ky. Acts. ch. 147 (authorizing a regulatory sandbox
in Kentucky); UTAH CODE ANN. §§ 13-55-101 (LexisNexis 2023) (authorizing a regulatory
sandbox in Utah); W. VA. CODE R. § 31A-8G-1 (2023) (authorizing a regulatory sandbox in West
Virginia); 2019 Wyo. Sess. Laws ch. 61 (authorizing a regulatory sandbox in Wyoming); VT. STAT.
ANN. tit. 8, § 15a (2023) (authorizing a regulatory sandbox in Vermont).
The Financial Inclusion Trilemma
143
other words, PayPal and Venmo are not banking the unbanked so much as
providing an alternative transaction platform for consumers who are
already banked.
The uncomfortable truth about fintech is that it is unlikely to ever be
transformative with regard to financial inclusion in developed economies.
Whatever marginal efficiencies fintechs might realize are simply
insufficient to overcome the fundamental economic problems of small-
balance deposit accounts or small-dollar lending.
When fintechs are able to reach new customers, the economics of
small transactions often force problematic practices. Some of the leading
fintech lenders—Think Finance (ranked second in 2013 on Forbes’ list of
“America’s Most Promising Companies”),
108
Elevate Financial, and
OppLoans—have business models that depend on evasion of state usury
laws, resulting in them being sued by regulators for consumer protection
violations.
109
Indeed, it would be strange to think that fintech could successfully
address the financial inclusion problem in credit. Credit fintechs are built
upon economies of scale. The problem, however, is that economies of scale
work only when consumer relationships work predictably and are simple.
They aren’t. The Anna Karenina Rule of consumer law is that every
unhappy consumer is unhappy in his or her own way.
110
Because of the
unique nature of unhappy consumers, attempting to deal with them in a
cookie cutter fashion will inevitably produce poor results. This has been
shown repeatedly with problems in mortgage loan and student loan
servicing.
111
Low-income consumers, in particular, are likely to present an
Anna Karenina problem because of the volatility of their financial lives.
108. J.J. Colao, America’s Most Promising Companies: The Top 25, FORBES (Feb. 6, 2013,
10:00 AM), https://www.forbes.com/sites/jjcolao/2013/02/06/americas-most-promising-companies-
the-top-25 [https://perma.cc/N4L7-G8ZM].
109. See, e.g., Complaint, District of Columbia v. Elevate Credit, Inc., No. 2020-CA-
002697 (D.C. Super. Ct. June 5, 2020) (alleging violations of DC’s usury, lender licensing, and
consumer protection laws); Cross-Complaint, Opportunity Financial, LLC v. Hewlett, No.
22STCV08163 (Cal. Super. Ct. Sept. 30, 2022) (alleging violations of California’s usury and
licensing laws); Complaint, Pennsylvania v. Think Fin., Inc., No. 2-14-cv-07139-JCJ (E.D. Pa. Jan.
26, 2018) (alleging violations of Pennsylvania’s usury law); Complaint, Consumer Fin. Prot.
Bureau v. Think Fin., LLC, No. 17-cv-00127-BMM (D. Mont. Nov. 15, 2017) (alleging violations
of federal unfair, deceptive, or abusive acts or practices statute based on underlying violations of
state usury law); Second Amended Complaint, Zavislan v. Avant of Colo. LLC, No. 2017CV30377
(Colo. Dist. Ct. Nov. 30, 2018) (alleging violation of Colorado’s usury and licensing laws); Second
Amended Complaint, Fulford v. Marlette Funding, LLC, No. 2017CV30376, 2019 WL 4451038
(Colo. Dist. Ct. June 5, 2019) (alleging violation of Colorado’s usury and licensing laws).
110. LEO TOLSTOY, ANNA KARENINA 1 (Richard Pevear & Larissa Volokhonsky trans.,
Penguin Classics 2004) (“Happy families are all alike; every unhappy family is unhappy in its own
way.”).
111. Complaint at ¶ 4, Consumer Fin. Prot. Bureau v. Navient Corp., No. 3:17-cv-00101-
RDM (M.D. Pa. Aug. 4, 2017) (alleging rampant misservicing of student loans); Adam J. Levitin
& Tara Twomey, Mortgage Servicing, 28 YALE J. ON REGUL. 1 (2011) (detailing misservicing of
mortgage loans).
Yale Journal on Regulation Vol. 41:109 2024
144
ii. Cryptocurrencies
The latest supposed technological panacea for financial inclusion is
cryptocurrencies and decentralized finance (DeFi). Some promoters of
cryptocurrency have argued that it can substantially increase financial
inclusion.
112
While the underbanked appear to be overrepresented among
cryptocurrency users,
113
there is little evidence to date that
cryptocurrencies have materially improved financial inclusion.
114
Cryptocurrencies are a poor store of value, undercutting
cryptocurrency platforms’ value as an ersatz banking system.
Cryptocurrencies have extremely volatile prices,
115
making it possible for
investments in cryptocurrency to rapidly lose a substantial part of their
value. Consumers, particularly those with more limited wealth, are
generally ill-suited for handling financial volatility, as they lack the ability
to adequately hedge and diversify their exposures.
116
Additionally, the
bankruptcies of a number of leading cryptocurrency platforms have
saddled consumers who custodied their cryptocurrency with those
platforms with illiquidity at best and significant losses at worst.
Cryptocurrency is also an inefficient payment mechanism. Payments
in cryptocurrency are likely more expensive than fiat payments,
particularly when costs of converting from and to fiat currency are
included. For example, by one estimate, the cost to send $200 via Tether, a
popular stablecoin, from a U.S.-dollar-denominated bank account to a
Euro-denominated account using some of the most popular
cryptocurrency exchanges would be between $5.98 and $86.44.
117
In
contrast, sending $200 via Western Union would cost $4.88.
118
112. Michael Hsu, Acting Comptroller of the Currency, Off. of the Comptroller of the
Currency, Remarks Before the Blockchain Ass’n: Cryptocurrencies, Decentralized Finance, and
Key Lessons from the 2008 Financial Crisis, 1, 3, 7 (Sept. 21, 2021); Tonantzin Carmona,
Debunking the Narratives About Cryptocurrency and Financial Inclusion, BROOKINGS (Oct. 26,
2022), https://www.brookings.edu/articles/debunking-the-narratives-about-cryptocurrency-and-
financial-inclusion/ [https://perma.cc/AF9H-A85B].
113. Michael J. Hsu, Acting Comptroller of the Currency, Off. of the Comptroller of the
Currency, Remarks Before the Inst. of Int’l Econ. L. at Geo. Univ. L. Ctr.: Thoughts on the
Architecture of Stablecoins 1 (Apr. 8, 2022).
114. Alex Fredman & Todd Phillips, Claims That Crypto Bolsters Financial Inclusion Are
Dubious, CTR. FOR AM. PROGRESS (Mar. 25, 2022), https://www.americanprogress.org/article/
claims-that-crypto-bolsters-financial-inclusion-are-dubious [https://perma.cc/2GYA-Z7FJ]
(explaining why it is unlikely that cryptocurrencies contribute to financial inclusion); Hsu, supra
note 112, at 7-8 (questioning how cryptocurrencies expand access to banking and credit).
115. Carmona, supra note 112.
116. Stablecoins, a subset of cryptocurrencies, have values that are supposed to be pegged
to the value of a fiat currency, but they can lose their peg, and even when they do not, they are
primarily used for transactions between cryptocurrencies rather than for purchases of other goods
and services. Id.
117. Stablecoins: How Do They Work, How Are They Used, and What Are Their Risks?:
Hearing Before the S. Comm. on Banking, Hous., & Urb. Affs., 117th Cong. 10 (2021) (written
testimony of Alexis Goldstein, Director of Fin. Pol’y, Open Mkts. Inst.).
118. Id.
The Financial Inclusion Trilemma
145
Additionally, clearing speeds for cryptocurrency payments are
unpredictable, depending on network demand and transaction fees.
Unpredictable clearing times, combined with the lack of a bank guaranty
of payment, makes cryptocurrency impractical as a payment system for
spot commercial transactions. For example, a bicycle shop will not let a
customer ride off with a new bike based on a cryptocurrency payment that
has not yet cleared, because if the payment does not clear, the bicycle shop
with hall difficulty ever collecting from the customer. Likewise, the
economics of contract enforcement for small-dollar contracts makes
cryptocurrency impractical for forward transactions, especially if over the
Internet: a merchant that delivers goods before a cryptocurrency
transaction clears risks never getting paid, while a consumer who pays
before a merchant delivers risks that the goods will never be delivered.
Credit cards and electronic fund transfers (including debit cards) solve
these problems with a bank guaranty of payment that ensures that the
merchant and the consumer can engage in the transaction because they are
assuming the payment risk not of a stranger, but of a highly regulated
financial institution that is subject to various statutory requirements
regarding non-judicial error resolution processes.
119
With a credit card,
(but not for an electronic fund transfer) if the merchant does not deliver
the goods promised, the consumer can have the transaction reversed as a
“billing error,”
120
and with a credit card or an electronic fund transfer, if
the consumer lacks sufficient funds to pay for the purchase or simply
absconds, that is the bank’s problem, not the merchant’s, as the bank has
undertaken to pay the merchant when it authorizes the transaction. The
uncertainty of clear speed plus the lack of a bank guaranty means that
cryptocurrencies are not workable as a commercial medium.
121
The false messiah of technology has enabled policy makers to avoid
grappling with the uncomfortable realities of financial inclusion, namely
that private provision alone is insufficient. Financial inclusion is not
achievable without strong governmental interventions. At the same time,
fintech firms have exploited the promise of financial technology as a
deregulatory strategy that gives them a competitive leg up on traditional
rivals. Not surprisingly, those traditional rivals have responded by calling
for a level playing field, but the level playing field they seek is one that is
deregulated for all. In other words, traditional financial institutions use
119. 15 U.S.C. § 1666 (2018) (providing statutory error resolution requirements for credit
cards); 12 C.F.R. § 1026.13 (2023) (providing regulatory error resolution requirements for credit
cards); 15 U.S.C. § 1693f (2018) providing statutory error resolution requirements for electronic
fund transfers); 12 C.F.R. § 1005.11 (2023) (providing regulatory procedures for resolving error
resolution requirements for electronic fund transfers).
120. 15 U.S.C. § 1666(b)(3) (2018); 12 C.F.R. § 1026.13(a)(3) (2023).
121. Layer Two applications, like the Lightning Network, may improve the commercial
usability of cryptocurrency, but often at the expense of other problems.
Yale Journal on Regulation Vol. 41:109 2024
146
fintechs as the camel’s nose under the regulatory tent. Financial inclusion
has become a beard for deregulation.
2. Negative Service Mandates
The United States has long buttressed private provision with negative
mandates, specifically prohibitions on discrimination against certain
protected classes. The Equal Credit Opportunity Act of 1974 (ECOA)
prohibits discrimination in credit transactions on the basis of race, color,
religion, national origin, sex or marital status, age, or income derived from
public assistance.
122
ECOA, however, only extends to credit transactions.
It does not cover the opening of deposit accounts. There is no federal anti-
discrimination law that addresses the situation of the unbanked.
123
ECOA covers the underbanked, at least for credit transactions, but it
does not help when a consumer cannot obtain credit from a bank because
of poor credit quality. At most, ECOA lets the consumer know why her
loan application was rejected.
124
While that is helpful in policing
discrimination against protected classes, it is not generally relevant for
ensuring the provision of fairly priced small-dollar credit. Indeed, the
problem with short-term, small-dollar credit is that the terms are onerous
for all borrowers, irrespective of membership in a protected class. If
anything, the concern with predatory lending is “reverse redlining,” in
which minority communities are targeted for offers of high-cost credit,
rather than being denied credit. Once again, the trilemma holds: negative
service mandates help protect equality of access, but as with fintech, then
force a choice between fairness of terms and profitability. Negative
mandates do little work for financial inclusion.
3. Soft Service Mandates
In addition to public provision and a negative anti-discrimination
mandate, the United States also has a soft mandate for financial inclusion.
The Community Reinvestment Act of 1977 (CRA) requires bank
regulators to evaluate whether each bank is “meeting the credit needs of
its entire community, including low- and moderate-income
neighborhoods.”
125
The regulatory implementation of the CRA imposes
122. 15 U.S.C. § 1691(a) (2018).
123. CFPB Targets Unfair Discrimination in Consumer Finance, CONSUMER FIN. PROT.
BUREAU (Mar. 16, 2022), https://www.consumerfinance.gov/about-us/newsroom/cfpb-targets-
unfair-discrimination-in-consumer-finance [https://perma.cc/YVR7-4C4P] (noting that the CFPB
will treat discrimination in financial services other than credit as violating the prohibition on
“unfair” acts or practices). A federal court has enjoined the CFPB’s amendment of its Supervision
and Examination Manual that implemented this policy. Chamber of Commerce v. Consumer Fin.
Prot. Bureau, No. 6:22-cv-00381-JCB (E.D. Tex. Sept. 8, 2023) (order granting preliminary
injunction).
124. 15 U.S.C. § 1691(d)(2) (2018).
125. 12 U.S.C. § 2906(a)(1) (2018).
The Financial Inclusion Trilemma
147
different types of tests depending on the size of the bank and whether it is
a retail, wholesale, or limited purpose bank. Since 1995, however, the
largest retail banks—those with over $1.322 billion (as of 2023) in assets—
have been evaluated using three tests: a lending test, an investment test,
and a service test.
126
The lending test looks at the number, amount, and geographic
distribution of loans.
127
The analysis focuses on geographic area, not on
actual borrowers.
128
Thus, if a bank is lending to yuppies in a gentrifying
(but still low-to-moderate income) neighborhood, it could readily get CRA
credit for it.
The investment test is focused on the amount of qualified community
development investments made by the bank, such as investments related
to affordable housing, community services targeted to low- to moderate-
income individuals, small business investment, and neighborhood
revitalization or stabilization projects.
129
The service test is an imprecise standard that looks primarily at the
geographic distribution of bank branches, but not at the actual number or
volume of deposits.
130
Thus, while the CRA provides a general nudge for
banks to provide services to low- to moderate -income communities, it does
not actually ensure provision of services to low- to moderate-income
individuals.
Each test is scored with one of five grades, and each grade is assigned
a specified number of points. The points, however, are not the same for
each test. Whereas there are up to twelve points available under the
lending test, there are only six points available under both the investment
and service tests.
131
This scoring system means that provision of deposit
accounts—which, along with a number of other services, would fall under
the service test—is treated as much less important than lending—which
falls under the lending test. And even for lending, the emphasis is on large-
dollar loans, such as mortgage loans.
Perhaps more importantly, the CRA lacks teeth. A bank’s CRA
compliance rating is publicly disclosed,
132
and the bank’s CRA compliance
record is one of many factors taken into account when bank regulators
evaluate whether to approve the bank’s acquisition of another bank,
126. 12 C.F.R. § 25.12(u), 25.21(a) (2023).
127. 12 C.F.R. § 25.22(b) (2023).
128. 12 C.F.R. §§ 25.22(a)(1), 25.41 (2023).
129. 12 C.F.R. § 25.23 (2023).
130. Michael A. Stegman, Kelly Thompson Cochran & Robert Faris, Toward a More
Performance-Driven Service Test: Strengthening Basic Banking Services Under the Community
Reinvestment Act, 9 GEO. J. ON POVERTY L. & POL'Y 405, 414-16 (2002) (discussing elements of
the service test).
131. Community Reinvestment Act Regulations, 60 Fed. Reg. 22156, 22170 (May 4,
1995). The point system is only in the preamble of the final rule; it is not in the codified regulation.
132. 12 U.S.C. § 2906(b)(1)(A)(iii) (2018).
Yale Journal on Regulation Vol. 41:109 2024
148
mergers, or branch applications.
133
Beyond such publicity and a possible
impact on mergers, however, there are no legal consequences for a bank
that does not meet community credit needs. The main impact of the CRA
as currently implemented is to discourage redlining in mortgage lending;
there is no evidence suggesting that it has accomplished much in terms of
reducing the unbanked population or expanding small-dollar credit. As
with negative service mandates, soft service mandates help ensure products
are widespread, but do not resolve the fairness-versus-profitability tension.
4. Modeling Pilot Programs
The federal government has also attempted to expand financial
inclusion using pilot programs to test and model concepts. This has
generally involved relatively small-scale grants and experiments, none of
which have produced notable results. In 2000, for example, the Clinton
Administration announced the “First Accounts” initiative to “bring the
‘unbanked’ into the mainstream.”
134
The initiative consisted of a $30
million dollar set of grants for community groups to work with financial
institutions in expanding bank account services to low-income
consumers.
135
The initiative also encouraged banks to experiment with
placing no-fee ATMs in post office branches.
136
The initiative does not
seem to have had much result.
In 2008, the FDIC engaged in a two-year pilot program to see if banks
could profitably offer small-dollar loans as an alternative to payday
loans.
137
The pilot loans were for no more than $2,500 at APRs of 36% or
less and for terms of at least ninety days.
138
While the FDIC touted the
program as a success, its report noted that most of the pilot program
participants saw the small-dollar loans as a tool for building or retaining
otherwise profitable relationships with consumers, for creating community
goodwill, or for garnering CRA benefits; few saw them as profitable on
their own.
139
The report noted that because of the loans’ small size, the
interest and fees generated are not always sufficient to achieve robust
133. 12 C.F.R. § 25.21, 25.29 (2023).
134. President Clinton Unveils “First Accounts”: Bringing the “Unbanked” into the
Financial Mainstream, WHITE HOUSE OFF. OF THE PRESS SEC’Y (Jan. 13, 2000), https://
clintonwhitehouse3.archives.gov/WH/New/html/20000113_2.html [https://perma.cc/5L85-89BA].
135. Id.
136. Id.; Larry Rulison, Baltimore Gets ‘Free’ ATMs for Area’s Needy, BALT. BUS. J. (Jan.
24, 2000, 12:00 AM), https://www.bizjournals.com/baltimore/stories/2000/01/24/story6.html
[https://perma.cc/CB47-NGFG].
137. FDIC's Small-Dollar Loan Pilot Shows Banks Can Offer Alternatives to High-Cost,
Short-Term Credit; Results in Safe, Affordable and Feasible Template for Small-Dollar Loans, FED.
DEPOSIT INS. CORP. (June 24, 2010), https://archive.fdic.gov/view/fdic/4021 [https://perma.cc/
T6EQ-PHXE].
138. A Template for Success: The FDIC’s Small-Dollar Loan Pilot Program, 4, FDIC Q.,
28, 28 (2010).
139. Id. at 32.
The Financial Inclusion Trilemma
149
short-term profitability. Rather, most pilot bankers sought to generate
long-term profitability through volume and by using small-dollar loans to
cross-sell additional products.
140
In other words, it does not appear that the FDIC pilot program’s
small-dollar loans were profitable on a stand-alone basis at a 36% APR.
Instead, they primarily worked as relationship-building (i.e., loss leader)
products. Not surprisingly, few banks offer such products.
141
In 2011, the FDIC ran a pilot program at nine banks using its Model
Safe Accounts Template for low-fee deposit accounts.
142
Under the pilot
program, 662 transaction accounts and 2,883 savings accounts were
opened.
143
Over a year, 19% of the transaction accounts and 5% of the
savings accounts were closed,
144
suggesting lower credit risk than
anticipated.
145
The FDIC observed several business models emerging in the
program, including using the Safe Accounts as “second chance” accounts
for consumers with credit problems, but also a “Cross-Selling Model” that
used the offer of one type of low-fee account to offer another type of
additional account.
146
The FDIC’s standards provided the starting point for arguably the
most successful attempt to date at providing low-fee bank accounts to the
unbanked. In 2015, the nonprofit organization Cities for Financial
Empowerment Fund launched a project called “Bank On” to coordinate a
national standard for low-cost checking accounts.
147
The Bank On
standards were based on those of the FDIC’s Model Safe Accounts
Template. The initial results are encouraging; as of 2022, some 17.4 million
accounts had been opened.
148
Over a quarter of accounts are closed every
year, however, so as of 2022, only 8.1 million of those accounts remained
140. Id.
141. See id. at 34 (“Banks other than those in the [FDIC] pilot provide small-dollar loans,
but it is likely that most banks do not offer these loans.”). Based in part on the FDIC’s pilot
program, the National Credit Union Administration (NCUA) authorized federal credit unions to
make “payday loan alternatives” with a 28% APR rate cap plus an application fee of up to $20,
but only to existing credit union members. Final Rule, Short-Term, Small Amount Loans, 75 Fed.
Reg. 58285 (Sept. 24, 2010). In response to the modest uptake by credit unions, NCUA expanded
the allowed terms, enabling new members to immediately obtain the loans. Payday Alternative
Loans, 84 Fed. Reg. 51942, 51943 (Oct. 1, 2019) (codified at 12 C.F.R. § 701.21 (2023)). It is unclear
how much take-up there has been since.
142. FDIC Model Safe Accounts Pilot: Final Report, FED. DEPOSIT INS. CORP. 1 (Apr.
2012), https://www.fdic.gov/consumers/template/SafeAccountsFinalReport.pdf [https://perma.cc/
AK3J-DK4A].
143. Id. at 6-7.
144. Id.
145. Id. at 8.
146. Id.
147. Bank On Coalition Playbook: Equipping Bank On Coalitions for Local Banking
Access Success, BANK ON 5, 7 (July 2022), https://issuu.com/cfefund/docs/
bank_on_playbook_jan_2017 (on file with author).
148. Violeta Gutkowski & Lisa J. Locke, The Bank On National Data Hub: Findings from
2022, FED. RSRV. BANK OF ST. LOUIS (Nov. 8, 2023), https://www.stlouisfed.org/community-
development/bank-on-national-data-hub/bank-on-report-2022 [https://perma.cc/WZM8-HSPB].
Yale Journal on Regulation Vol. 41:109 2024
150
open.
149
Three-fourths of the accounts were opened at large financial
institutions, rather than at community banks,
150
and 98% of the accounts
open as of the end of 2022 were at large financial institutions.
151
Notably,
not all of these accounts were being offered to the traditional core
unbanked population; at least some were being used for “student”
accounts with the expectation that today’s students will graduate into
regular account holders. Thus although 85% of accounts were from
customers new to the financial institution,
152
this figure includes students
with their first bank accounts as well as unbanked adults.
It is possible to calculate the average revenue from Bank On accounts
and evaluate their profitability. The accounts had an average monthly
balance of $1,117,
153
which at a 3% net interest margin suggests annual
float income of $33.51. Additionally, the accounts have monthly fees of $5,
for another $60 in annual income. And the accounts come with a debit
card. The average number of debits per account per month was 27.1,
154
and
the average debit transaction amount was $41.
155
Given that almost all of
the accounts involved in Bank On are at large banks
156
that are subject to
the Durbin Interchange Amendment’s price cap on debit card interchange
fees,
157
the average interchange income for the banks per debit transaction
would be $0.24, resulting in average annual revenue of $78.05. Thus, in
total, the bank would earn $171.56 in annual revenue from such an account,
putting it at the lowest end of revenue needed to break even on an account
level basis. (This model will become even more stressed under the Federal
Reserve Board’s proposed reduction in the Durbin Interchange
Amendment price cap,
158
as debit interchange revenue would fall to $57.56,
meaning that total revenue from the account would be $151.07.)
Additionally, Bank On accounts exhibit a very high closure rate.
Around a quarter of accounts have closed in each program year.
159
The
data does not distinguish between voluntary closures, such as a consumer
graduating into another type of account or switching financial institutions,
149. Id.
150. Id. (finding that 3.9 million Bank On accounts were ever opened at community banks
out of 17.4 million Bank On accounts ever opened; community banks are defined here as banks
having less than $100 billion in assets).
151. Id. (finding that 167,000 Bank On accounts are currently open at community banks
out of 8.1 million Bank On accounts currently open).
152. Id.
153. Id.
154. Gutkowski & Locke, supra note 148.
155. Id.
156. See supra text accompanying notes 150-151.
157. 12 C.F.R. §§ 230.235.3-235.4 (2023).
158. Debit Card Interchange Fees and Routing, 88 Fed. Reg. 78100, 78122 (proposed
Nov. 14, 2023) (proposing reducing the base component from 21.0 to 14.4 cents, reducing the ad
valorem component from 5.0 basis points to 4.0 basis points, and increasing the fraud-prevention
adjustment from 1.0 cents to 1.3 cents).
159. Gutkowski & Locke, supra note 148.
The Financial Inclusion Trilemma
151
and involuntary closures because of failure to pay fees or other account
terms violations.
160
Nevertheless, this high level of churn suggests that the
costs of onboarding customers into Bank On, including customer
identification verification, will take longer to recoup.
Why, then, would a bank ever offer such an account? Bank On’s own
materials suggest three reasons. First is to develop a “sustainable consumer
base” by bringing unbanked consumers into the “financial mainstream.”
161
This suggests the future possibility of graduating consumers into other
types of accounts and/or cross-selling them other products. Second is CRA
credit.
162
And third are “community opportunities,” essentially positive
publicity for banks to get “public recognition both locally and
nationally.”
163
CRA credit and positive publicity suggest that the real
motivation in offering Bank On accounts is to curry regulatory goodwill.
Although a Bank On account might be a losing proposition on purely
monetary terms, the intangible benefit of regulatory goodwill might be
substantial enough for banks to offer the product if the total number of
accounts, and thus total cost to the bank, is relatively limited. A few million
dollars is a cheap way for a large bank to obtain regulatory goodwill, but
whether banks will voluntarily offer such a product on a larger scale is
unclear.
The most recent pilot program is one by the United States Postal
Service that allows consumers to cash payroll and business checks in
exchange for stored value cards.
164
The pilot program has been offered at
four post office branches. In its first four months, it undertook all of six (!)
transactions.
165
While it is not surprising that a little advertised program
offered at only four post office branches would get scant usage, the Postal
Service’s failure to generate interest in its offering does not generate
confidence in its ability to successfully offer financial services.
Thus far, federal government pilot projects for banking the unbanked
and creating affordable small-dollar loans have had little impact on the
scope of financial inclusion. Even by their own terms as proofs of concept,
most federal pilot programs have been unsuccessful. The FDIC-inspired
Bank On remains the exception, but it is unclear if it can continue to scale,
and it only breaks the trilemma for deposit accounts (not credit) because
it offers a way to purchase regulatory goodwill, operating as a type of
subsidy for banks.
160. Id.
161. Bank on Coalition Playbook, supra note 147, at 22.
162. Id.
163. Id.
164. See Jory Heckman, USPS Pilot Expands Postal Banking Services After Years of
Skepticism, FED. NEWS NETWORK (Oct. 4, 2021, 2:17 PM), https://federalnewsnetwork.com/
agency-oversight/2021/10/usps-pilot-expands-postal-banking-services-after-years-of-skepticism
[https://perma.cc/2EPF-GXHQ].
165. U.S. POSTAL SERV., ACR2021, RESPONSES OF THE UNITED STATES POSTAL
SERVICE TO QUESTIONS 1-2 OF COMMISSION INFORMATION REQUEST NO. 1, at 7 (2022).
Yale Journal on Regulation Vol. 41:109 2024
152
B. Other Potential Regulatory Interventions
The United States has mainly relied on private provision of deposit,
payment, and small-dollar credit services, supplemented with negative
mandates and soft mandates. The failure of this light touch approach
suggests that more muscular regulatory interventions should be considered
if the United States is serious about addressing its financial inclusion
problem.
There are three stronger regulatory approaches that could be taken:
hard service mandates, public options, and public subsidies. Each of these
can be understood as an approach that prioritizes widespread access and
fairness of terms over stand-alone profitability, but the way they operate is
quite different. This section reviews each in turn, but the basic assumption
of this section is that any solution to the trilemma must focus on loosening
the stand-alone profitability requirement.
The other two requirements—widespread access and fairness of
terms—are fundamental to the entire idea of financial inclusion. If access
is not widespread, it is not meaningful inclusion, and if terms are not fair,
the inclusion is not worth pursuing. That leaves the stand-alone
profitability requirement to consider.
Although the need for stand-alone profitability has long been a policy
assumption, it is not fundamental to the idea of financial inclusion, and
insisting upon it is what creates the trilemma. Instead, the stand-alone
profitability requirement reflects a general political discomfort with
subsidization of consumer financial services. Yet there is something pearl-
clutching in this unease given that the U.S. banking system hardly operates
as a “free market.” For example, regulators control entry into the market
as well as mergers and acquisitions, impose capital and liquidity
requirements, and (for many banks) require participation in a mutual
insurance program. Banks are able to function solely because of a highly
(if not always perfectly) regulated environment that makes it possible for
customers to rely on the safety-and-soundness of fractional reserve banks.
Thus, it is the stand-alone profitability requirement that should be
examined as the leg of the trilemma that can be potentially addressed
through a regulatory fix.
1. Hard Service Mandates
One stronger regulatory approach is a hard service mandate. In
contrast with a soft mandate like the CRA, a hard mandate would require
banks to provide certain services at certain terms, irrespective of their
stand-alone profitability, effectively imposing a cross-subsidization
requirement.
An example of a hard mandate would be a basic banking requirement
obliging all banks to offer low-fee or free checking accounts to certain
The Financial Inclusion Trilemma
153
consumers. Given that a special license is required to engage in banking
activities,
166
it is well within the purview of the government to condition
such a license on provision of services seen to be in the public interest.
The United States lacks any sort of hard service mandate,
distinguishing it from several other developed countries. From 2003 to
2018, Canada’s Bank Act required banks to open a “low-fee retail deposit
account” with no minimum balance requirement for any individual that
meets the regulatory requirements.
167
The only regulatory exceptions
related to fraud, illegality, and customer abuse; bankruptcy (and
presumably the consumer’s credit score) was not grounds for refusing to
open an account.
168
Such accounts had to allow at least twelve debit
transactions per month, at least two of which could be done in-branch, as
well as check-writing privileges.
169
For these low-fee retail deposit
accounts, banks could not charge for deposits, debit cards, pre-authorized
payments, monthly printed statements, or online check image viewing.
170
For most low-fee retail deposit accounts, a four Canadian dollar per month
fee was authorized, and other services may have been offered “for a
reasonable fee.”
171
Youths, students, the poor, seniors, and certain disabled
persons were eligible for free accounts under the implementing
regulations.
172
In 2018, Canada repealed this statutory mandate
173
and replaced it
with a looser one that merely requires the opening of an account without
an initial minimum deposit or minimum balance requirement upon
presentment of adequate documentation, subject to the same fraud,
illegality, and customer abuse exceptions.
174
Although Canada no longer
requires that the accounts be “low-fee,” the Canadian government has
entered into voluntary commitments with Canada’s ten largest banks to
offer such low-fee accounts on the same terms as before 2018.
175
166. As a technical matter, no federal law prohibits unlicensed deposit taking and
lending, only the misleading use of “national” in the name of a banking business. 18 U.S.C. § 709
(2018). Instead, federal law only prohibits unlicensed money transmission. 18 U.S.C. § 1960 (2018).
State law often prohibits unlicensed banking. See, e.g., CAL. FIN. CODE § 1005 (West 2023); N.Y.
BANKING LAW § 131 (LexisNexis 2023); MD. CODE ANN., FIN. INST. § 3-208 (West 2023); MASS.
GEN. LAWS ch. 167, § 37 (2023).
167. S.C. 1991, c 46, §§ 448.1-2 (Can.), repealed 2018, c.27 § 31 (Can.).
168. Access to Basic Banking Services Regulations, SOR/2003-184, § 3(1)-(2) (Can.).
169. Id.
170. Low-Cost Account Guidelines, FIN. CONSUMER AGENCY OF CAN. (Jan. 9, 2017),
https://www.canada.ca/en/financial-consumer-agency/services/industry/laws-regulations/low-cost-
account-guidelines.html [https://perma.cc/HPS2-BSRS].
171. Id.
172. Id.
173. S.C. 1991, c 46, § 627.17(1) (Can.), as amended by 2018 .c 27 ¶ 329 (Can.).
174. S.C. 1991, c 46, § 627.18 (Can.), as amended by 2018 .c 27 ¶ 329 (Can.).
175. Low-Cost and No-Cost Accounts, FIN. CONSUMER AGENCY OF CAN. (June 26,
2023), https://www.canada.ca/en/financial-consumer-agency/services/banking/bank-accounts/low-
cost-no-cost.html [https://perma.cc/76NA-RLD4]. See also, e.g., Low-Cost and No-Cost Account
Yale Journal on Regulation Vol. 41:109 2024
154
The European Union has had a fee-free banking account mandate
since 2015,
176
and the United Kingdom has one originally adopted in
anticipation of the EU mandate, but retained after the UK left the EU.
177
Notably, the EU and UK mandates do not require all banks to offer free
or low-fee bank accounts. Instead, the EU mandate requires EU member
states to apply the mandate to either all banks or to “a sufficient number
of credit institutions to guarantee access thereto for all consumers in their
territory . . . .”
178
Similarly, the UK mandate applies only to the nine largest
banks in the UK,
179
all of which offer free basic accounts, but without
overdraft credit facilities.
180
In the United States, such low-fee retail deposit accounts would be
money losers for banks. In order to offer them, banks would have to
engage in cross-subsidization, either by charging their other customers
more or by accepting reduced profitability. A hard mandate is thus a cross-
subsidization requirement. The details of the cross-subsidy are left up to
the individual bank, enabling the government to avoid the question of how
to distribute the cost of the mandate.
One concern with a hard service mandate is that banks will drag their
feet and attempt to create frictions to discourage consumers from seeking
out money-losing accounts with them.
181
Here, a benchmarking law like the
Commitment, BMO BANK OF Montreal 1, https://www.bmo.com/pdf/
9628183LowCostAccounts_en2.pdf [https://perma.cc/3JJ5-2HJE]. Canada has a much more
concentrated banking market than the United States. The country’s six largest banks hold around
93% of all Canadian banking assets. Ian Bickis, How Big Banks Dominate Canada’s Financial
Landscape, GLOBAL NEWS (Apr. 19, 2023, 6:40 AM), https://globalnews.ca/news/9634933/canada-
big-banks-analysis [https://perma.cc/VVD5-M8L7].
176. Council Directive 2014/92, art. 17-18 2014 O.J. (L 257) (EC) [hereinafter EU
Payment Accounts Directive].
177. Payments Account Regulation 2015, SI 2038, art. 20, ¶ 1 (requiring provision of fee-
free accounts); Basic Bank Accounts: January to June 2016, HM TREASURY 4 (Dec. 2016),
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file
/576033/Basic_bank_account_2016_dec_final.pdf [https://perma.cc/7WCH-3KU2] (discussing the
voluntary nature of the UK arrangement, but stating that it was made in anticipation of the EU
Payment Accounts Directive becoming effective in the UK); Basic Bank Accounts: July 2020 to
June 2021, HM TREASURY 3 (Jan. 2022), https://assets.publishing.service.gov.uk/
government/uploads/system/uploads/attachment_data/file/1049371/Official_Sensitive_-
_Basic_bank_account_report.pdf [https://perma.cc/E28E-2PX3] (discussing application of basic
bank account mandate to the largest UK banks, which collectively hold ninety percent of deposit
accounts in the country); Payment Accounts (Amendment) (EU Exit) Regulations 2018:
Explanatory Information, HM TREASURY (Oct. 31, 2018) https://www.gov.uk/government/
publications/draft-payment-accounts-amendment-eu-exit-regulations-2018/payment-accounts-
amendment-eu-exit-regulations-2018-explanatory-information [https://perma.cc/AA6L-AJ2N]
[hereinafter EU Exit Regulations Explanatory Information].
178. EU Payment Accounts Directive, supra note 176 at art. 16(1).
179. UK Compliance with the EU Payment Accounts Directive, HM TREASURY (Sept. 18,
2016), https://www.gov.uk/government/publications/uk-compliance-with-the-eu-payment-
accounts-directive/uk-compliance-with-the-eu-payment-accounts-directive
[https://perma.cc/E82F-HLHH].
180. EU Exit Regulations Explanatory Information, supra note 177.
181. See Lauren E. Willis, When Nudges Fail: Slippery Defaults, 80 U. CHI. L. REV. 1155,
1185-1200 (2013) (providing examples of how banks have found ways to discourage consumers
from not opting in to for-fee overdraft coverage).
The Financial Inclusion Trilemma
155
CRA could come into play, measuring banks’ success in providing low-fee
retail deposit accounts. Additionally, while the unbanked themselves
might find signing up for an account daunting, social workers and aid
agencies could be instrumental in getting unbanked individuals signed up
for accounts.
2. Public Options
Another path to financial inclusion is through public provision of
financial services—public options. Historically, the United States offered
a public option for deposit services. From 1911 until 1967, the United
States Postal Savings System (USPSS) offered interest-bearing passbook
savings accounts.
182
The USPSS was not intended to be a financial inclusion vehicle.
Instead, it was created as the Republican-favored alternative to federal
deposit insurance.
183
Hence the USPSS’s sole offering was a passbook
savings account,
184
, a relatively rare product today. In a passbook savings
account, the consumer does not receive periodic balance statements.
185
Instead, the consumer receives a passbook—a small, passport-like booklet.
Whenever the consumer wishes to transact, the consumer presents the
passbook to the bank, which records the transaction and the account
balance before returning it to the consumer, a process that requires the
consumer to complete all transactions in person at the bank. Passbook
savings accounts provide safekeeping services, but nothing more.
186
Passbook savings accounts are essentially piggybanks. They cannot be used
to make payments, so they do not actually connect unbanked consumers
to the modern commercial world.
In recent years, there has been a call from progressive academics for
pursuing financial inclusion through retail-facing public options.
187
182. Postal Savings Depositary Act of 1910, Pub. L. No. 61-268, 36 Stat. 814 (1910)
(creating Postal Savings System); Pub. L. No. 89-377, § 5225, 80 Stat. 92, 92 (1966) (terminating
Postal Savings System).
183. See Gerhard Peters & John T. Woolley, Republican Party Platform of 1908, AM.
PRESIDENCY PROJECT, https://www.presidency.ucsb.edu/documents/republican-party-platform-
1908 [https://perma.cc/4SWQ-5B89]. In contrast, Democrats supported a postal bank only as a
second-best alternative to federal deposit insurance. See Gerhard Peters & John T. Woolley, 1908
Democratic Party Platform, AM. PRESIDENCY PROJECT, https://www.presidency.ucsb.edu/
documents/1908-democratic-party-platform [https://perma.cc/CL2S-HGEL].
184. Pub. L. No. 61-268, 36 Stat. 814, 817 (1910).
185. 12 U.S.C. § 1693d(d) (2018); 12 C.F.R. § 1005.9(c)(1) (2023).
186. See LEVITIN, supra note 52, at 258.
187. The USPSS was a rare retail-facing public option, but there is substantial public
provision of financial services in secondary or wholesale markets. Thus, there is already federal
provision of mortgage insurance for lenders, deposit insurance for banks, and securitization
guaranties for mortgage-backed securities investors. There is also federal provision of payment
systems: the Federal Reserve System operates a check clearing network, a wire transfer service
(FedWire), and an automated clearinghouse network (FedACH). In 2023, the Federal Reserve
System added a real-time payment system called FedNow. None of these payment systems are
Yale Journal on Regulation Vol. 41:109 2024
156
Professor Mehrsa Baradaran has advocated for a renewed postal banking
system,
188
while Professors Morgan Ricks, John Crawford, and Lev
Menand have called for the provision of free bank accounts (FedAccounts)
by the Federal Reserve.
189
Both proposals are focused on the provision of
deposit services, but hold open the possibility of credit services as well. One
proposed legislative implementation of a renewed postal banking system
would have a postal bank offering loans at the one-month constant
maturity Treasury rate,
190
currently 5.53%, but as low as 0.01% in recent
years. As we have seen, however, there is no way to make such loans
profitably, so there would necessarily be a taxpayer subsidy. Once again,
widespread access and fair terms cannot be paired with stand-alone
profitability, only with subsidized economics.
3. Public Subsidies
A final potential approach to financial inclusion is through direct
public subsidies of banks. A subsidy would function to reimburse banks for
provision of services that they would not otherwise offer. A public subsidy
could be done on a stand-alone basis or be combined with a hard mandate.
A key difference in approaches among a hard mandate, public
provision, and a direct subsidy is who pays for financial inclusion. An
unsubsidized hard mandate imposes a progressive cross-subsidy from
banked to unbanked or underbanked consumers or from bank
shareholders to unbanked or underbanked consumers, per the business
judgement of each bank. Public provision places costs in the first instance
on the users of the service, but any shortfalls are necessarily borne by
taxpayers. In contrast, a public subsidy is distributed among taxpayers
through the Internal Revenue Code.
If implemented on a stand-alone basis, the subsidy would have to be
large enough to motivate banks to provide the service—that is, it would
have to make service provision profitable, and the subsidy’s continued
availability would have to be sufficiently credible for banks to be willing to
invest given their upfront costs. If combined with a hard mandate, the
subsidy could be smaller. Having a subsidy would at least partially offset
any need for cross-subsidization, which might make the hard mandate
accessed by retail customers—consumers and businesses—directly. Instead, they are all accessed
by financial institution intermediaries.
188. Baradaran, supra note 17, at 211-13. See also Adam J. Levitin, Going Postal:
Financial Inclusion via Postal Banking (Working Paper 2011) (providing a history of postal
banking).
189. John Crawford, Lev Menand & Morgan Ricks, FedAccounts: Digital Dollars, 89
GEO. WASH. L. REV. 113, 113 (2021); Morgan Ricks, John Crawford & Lev Menand, Central
Banking for All: A Public Option for Bank Accounts, ROOSEVELT INST. GREAT DEMOCRACY
INITIATIVE (June 2018), https://rooseveltinstitute.org/wp-content/uploads/2021/08/GDI_Central-
Banking-For-All_201806.pdf [https://perma.cc/CC9P-MPMC].
190. Postal Banking Act, S. 2755, 115th Cong. (2018); Postal Banking Act, S. 4614, 116th
Cong. (2020).
The Financial Inclusion Trilemma
157
more politically acceptable, because other bank customers would not
perceive themselves as bearing the cost of financial inclusion.
Federal subsidization is a well-established tool in consumer finance
markets. The federal government subsidizes mortgage insurance through
the Federal Housing Administration, Veterans Agency, and U.S.
Department of Agriculture Rural Development. It also subsidizes student
loans through the Stafford Student Loan Program. And it subsidizes rent
through Section 8 vouchers.
There is already small-scale subsidization of transaction accounts. The
Debt Collection Improvement Act of 1996 required that all federal
payments made after January 1, 1999 be made electronically, subject to
certain exceptions.
191
It also provided that Treasury ensure that federal
payment beneficiaries have access to an account “at a reasonable cost” and
“the same consumer protections with respect to the account as other
account holders at the same financial institution.”
192
For federal benefit recipients with bank accounts, this is easy enough;
Treasury simply transfers funds to their accounts via ACH transactions.
But for the unbanked, electronic payments are not possible. Treasury first
attempted to address the electronic payment requirement through
subsidized Electronic Transfer Accounts at banks and credit unions. The
program was not successful. In its first two years, it produced only 8,100
accounts at some 600 institutions.
193
Starting in 2008, Treasury retooled its approach.
194
Instead of making
subsidy payments to numerous banks based on account openings, it instead
contracted with a single bank (Comerica Bank) for the issuance of
reloadable prepaid debit cards under what is known as the Direct Express
program.
195
Direct Express disburses Social Security and Veterans benefits
to some 4.5 million unbanked consumers.
196
The benefit payments are
automatically loaded onto beneficiaries’ Direct Express cards, which can
then be used like a debit card at any store that accepts Mastercard
191. Debt Collection Improvement Act, Pub. L. No. 104-134, § 31001(x) 110 Stat. 1321-
376 (1996) (codified at 31 U.S.C. § 3332(f) (2018)).
192. Id. at 1377 (codified at 31 U.S.C. § 3332(i)(2)(B) (2018)).
193. Stegman et al., supra note 130, at 406-07.
194. U.S. GOV’T ACCOUNTABILITY OFF., GAO-17-176, REVENUE COLLECTIONS AND
PAYMENTS: TREASURY HAS USED FINANCIAL AGENTS IN EVOLVING WAYS BUT COULD
IMPROVE TRANSPARENCY 13 (2017).
195. Treasury’s Bureau of the Fiscal Service Selects Comerica Bank to Continue Prepaid
Debit Card for Unbanked, U.S. DEP’T OF THE TREASURY, BUREAU OF THE FISCAL SERV. (Jan.
7, 2020), https://www.fiscal.treasury.gov/news/comerica-bank-continues-debit-card-for-
unbanked.html [https://perma.cc/FA7Q-TLG6].
196. Id. Most of the benefits distributed are Supplemental Security Income or Social
Security Retirement benefits. A Look at Cardholder Demographics, DIRECT EXPRESS (June 13,
2019), https://directexpress.info/2019/06/13/a-look-at-cardholder-demographics [https://perma.cc/
49QT-GXMC].
Yale Journal on Regulation Vol. 41:109 2024
158
products.
197
The Direct Express card can also be used to make ATM
withdrawals or get extra cash back on purchases.
198
The card has no fixed
fees, but allows only one free ATM withdrawal per month.
199
It also allows
balance transfer inquiries and text messages when funds are deposited or
balances fall below a specified threshold.
200
Only Treasury, however, can
load funds onto the card; the cardholder cannot load the card with funds
from other sources.
Direct Express did not originate as a financial inclusion program, but
as a response to a congressional directive to move to electronic payments
of government benefits for efficiency purposes. Nevertheless, Direct
Express functions as a form of subsidized financial inclusion; in 2019, 70%
of Direct Express cardholders reported that they do not have a bank
account.
201
Treasury shoulders the cost of the Direct Express program,
which provides some level of financial inclusion for Direct Express
cardholders. It is not the perfect financial inclusion product, particularly
because it allows access solely to funds disbursed by Treasury, but it is an
important example of subsidized financial inclusion that could potentially
be expanded.
202
IV. Choosing the Optimal Regulatory Intervention
The previous Parts have argued that private provision and soft
mandates have proven insufficient for addressing the problem of financial
inclusion and that more muscular interventions—whether hard mandates,
public provision, or public subsidies—are required. A key point of this
Article, however, is that the unbanked and underbanked represent
fundamentally different public policy problems, such that they are unlikely
to be solved through the same approach. Instead, each problem needs to
be evaluated on its own terms, as this Part proceeds to do.
A. Interventions for the Unbanked
A hard mandate, a public option, or a public subsidy can all, in the
abstract, significantly reduce the number of unbanked households by
providing free or low-cost accounts, irrespective of household credit
197. Direct Express, U.S. DEP’T OF THE TREASURY, BUREAU OF THE FISCAL SERV.
(May 2, 2023), https://www.fiscal.treasury.gov/directexpress [https://perma.cc/J2LL-PQEH].
198. Id.
199. Id.
200. Id.
201. A Look at Cardholder Demographics, supra note 196.
202. Adam J. Levitin, Lindsey Owens & Ganesh Sitaraman, No More Bailouts: A
Blueprint for a Standing Emergency Economic Resilience and Stabilization Program, ROOSEVELT
INST. GREAT DEMOCRACY INITIATIVE 28 (June 2020), https://ssrn.com/abstract=3639607
[https://perma.cc/27XK-UA5S]; Prasad Krishnamurthy, Stimulus for All, HILL (May 14, 2021, 6:00
PM), https://thehill.com/opinion/finance/553620-stimulus-for-all [https://perma.cc/4UH4-LTR3].
The Financial Inclusion Trilemma
159
quality. Yet there are reasons to prefer a hard mandate for provision of
basic banking services over either public options or public subsidies.
A hard mandate for provision of basic banking services has the virtues
of being both the most direct tool and the one without any cost to the public
fisc or administrative costs. With a mandate for the provision of basic
banking services, the government can specify precisely the product terms
it wants and not worry about funding it; the funding will be left up to each
individual bank, which will have to decide whether and how to cross-
subsidize from other product offerings or shareholder surplus. Presumably,
the cross-subsidy will be borne by the least competitive market.
In contrast, a public option lets the government craft the terms on
which it wishes to offer services, but has a direct cost to the public fisc and
huge administrability challenges. While it might be possible to structure a
public option that resides off the federal balance sheet, such as through the
Postal Service
203
or the Federal Reserve System, this is an accounting
measure; the public still pays for it. The Postal Service receives a standing
appropriation of all revenues it receives
204
and has received supplemental
congressional appropriations,
205
while the Federal Reserve remits its
profits to Treasury’s general fund.
206
An increase in operating expenses for
either could necessitate further appropriations or result in a smaller
remittance to Treasury.
Additionally, any sort of public option would require substantial lead
time to implement and would pose enormous operational challenges. The
federal government generally does not provide retail financial services
itself but instead provides secondary market services. The major exception
is student lending, but even there, the loans are made by the federal
government and disbursed to schools instead of students,
207
the
disbursement is handled by a private contractor,
208
and the loans are
serviced by private servicers, rather than the Department of Education
itself.
209
Nor are direct student loans underwritten in any traditional sense;
203. 39 U.S.C. § 2009a (2018).
204. 39 U.S.C. § 2401(a) (2018).
205. CARES Act, Pub. L. No. 116-136, § 6001, 134 Stat. 281, 504 (2020) (providing a $10
billion loan appropriation for the Postal Service); Consolidated Appropriations Act, 2021, Pub. L.
No. 116-260, § 801, 134 Stat. 1182, 2119 (2021) (providing that the Postal Service is not required to
repay its borrowing under the CARES Act).
206. 12 U.S.C.§ 289(a)(3)(B) (2018) (requiring any surplus in the Federal Reserve System
above $6.825 billion to be remitted to Treasury’s general fund).
207. Receiving Financial Aid, U.S. DEP’T OF EDUC, FED. STUDENT AID,
https://studentaid.gov/complete-aid-process/receive-aid [https://perma.cc/ERJ9-235A].
208. See Accenture Awarded $966 Million Contract to Support Federal Student Aid
Programs, ACCENTURE (Apr. 16, 2015), https://newsroom.accenture.com/news/accenture-
awarded-966-million-contract-to-support-federal-student-aid-programs.htm
[https://perma.cc/9KA9-D84G].
209. John R. Brooks & Adam J. Levitin, Redesigning Education Finance: How Student
Loans Outgrew the “Debt” Paradigm, 109 GEO. L.J. 5, 50 (2020).
Yale Journal on Regulation Vol. 41:109 2024
160
pricing is one-size-fits-all.
210
Simply put, there is no federal experience in
managing a large-scale retail financial operation. This is not to say that it
cannot be done, but that it presents a significant complication for any
public option.
Another problem with a public option is that it puts the government
in a potentially adverse position to consumers. Because bank accounts can
be overdrawn, a public option means that the government is potentially in
the position of being a creditor to low-income households. This raises
thorny questions about how collection activities would work. For example,
would the government set off overdrawn accounts against earned income
tax credits? If the government is a service provider, it is unavoidable that
there will be some situations in which it is adverse to the consumer.
If government is accepting deposits, there is also an unavoidable
question of how those deposits will be invested. Historically, the USPSS
addressed the issue by requiring that deposits either be invested in
Treasury securities or redeposited in local commercial banks.
211
By being
a depository, however, the federal government would face a question
about how it would allocate capital—whether to itself or elsewhere.
Finally, a public option in banking threatens all manner of disruption
and disintermediation with attendant unintended consequences. If
government accounts have attractive features, money will flow out of the
private system to them. Whether this is a feature or bug depends on the
goals of a public option.
If the goal of a public option is to compete with private banks and set
a market benchmark, then disintermediation is a feature—it is exactly what
should happen until and unless private banks adjust. But if the goal of a
public option is merely to supplement the private banking system and serve
only otherwise unserved customers, then the disintermediation is a bug. To
the extent that a public option is really just about financial inclusion, not
reshaping the terms on which banks offer services generally,
disintermediation is a bug.
The risk of disintermediation is hardly speculative. For example, the
USPSS caused huge disintermediation of funds from building-and-loan
institutions during the Great Depression, exacerbating the collapse of the
housing market with its requirement of deposit reinvestment in
commercial banks (which did little real estate lending) or Treasuries.
212
When depositors fled building-and-loans for the safety of the USPSS, their
money left the housing finance system, pushing up mortgage costs and
210. Id. at 35-36.
211. Pub. L. No. 61-268, § 9 36 Stat. 814, 816 (1910).
212. Id.; Maureen O’Hara & David Easley, The Postal Savings System in the Depression,
39 J. ECON. HIST. 741, 746-49 (1979).
The Financial Inclusion Trilemma
161
pulling down housing values.
213
A public option could be majorly
destabilizing to financial markets because of such unanticipated dynamics.
The third possibility—public subsidies for bank accounts—could be
coupled with a hard mandate, but could also be pursued on a stand-alone
basis. On a stand-alone basis, it is the least attractive option. Public
subsidies do not guarantee the offering of any services unless the subsidies
are larger than the cost of offering the services. In other words, the
government would, by definition, have to overpay in order to be sure that
subsidies would be effective.
If the subsidies come out of annual appropriations, they would also
add an element of uncertainty because future appropriations would not be
guaranteed. This uncertainty would discourage financial institutions from
investing in the capacity to offer accounts that would qualify for the
subsidies. To be sure, such subsidies could come directly from fees or taxes
levied on banks through the chartering or insurance process, or could be
given as a credit against such fees. This would have much the same effect
as the cross-subsidization imposed by a hard mandate.
A subsidy system would also require some type of administrative
apparatus to ensure that the payment of the subsidies accorded with the
number of accounts provided. While a subsidy system would necessitate a
far smaller administrative apparatus than a public option, none is needed
for a hard mandate.
All of this counsels for a hard mandate for provision of basic banking
services as the prime policy move for addressing the problem of the
unbanked. As a political matter, such a mandate might be more likely to
be enacted if it exempted community banks or if it were coupled with an
offsetting subsidy, but that is a political point, rather than a fundamental
system design point.
While there are policy tools for substantially addressing the problem
of the unbanked, the underbanked are a different story, as the next section
addresses.
B. Interventions for the Underbanked
The key problem of the underbanked—the demand for short-term,
small-dollar credit—is not one that can generally be resolved by
government intervention. The demand for short-term, small-dollar credit
is partially a demand for consumption-smoothing liquidity, but it is also
partially reflecting consumers attempting to address the fundamental
mismatch between their income and their expenses.
To the extent that the demand is merely for bridge liquidity, there are
government interventions that can help. For example, allowing advances
213. O’Hara & Easley, supra note 212, at 748-49.
Yale Journal on Regulation Vol. 41:109 2024
162
of the earned income tax credit or earned wage access products
214
or more
regular disbursement of government benefits would provide a source of
liquidity smoothing.
215
To the extent the demand is driven by solvency problems, then
facilitating greater credit availability is not a solution.
216
Credit cannot
bridge a solvency problem. As long as the solvency problem persists,
expanding access to credit is likely to be welfare reducing for many
consumers.
217
Moreover, interventions like hard mandates, public subsidies, or a
public option put the government in the awkward position of either
directing credit provision (hard mandates and public subsidies) or
engaging in direct credit provision (public options). Either way risks the
politicization of government lending programs. There would be a constant
temptation for those in power to attempt to curry favor with voters through
easier lending terms, ultimately undermining the soundness of
underwriting. Additionally, to the extent the government makes loans, it
has to collect them, putting the government in an adverse position to the
very consumers it aims to help.
To be sure, these risks also exist for existing government credit
programs in the home mortgage and student loan markets. But there are
offsetting benefits for the public in those markets—the social stability that
comes from homeownership and the benefits of a better educated
population. Furthermore, in the housing market, the problem of
politicization is mitigated by having the government operate mainly as a
secondary market and insurance provider, rather than a direct credit
provider. It is hard to see any benefit to the public from having cheaper
sources of short-term, small-dollar lending that can be spent without
restriction.
Ultimately, household solvency problems can only be addressed by
secular changes in the economy that will result in greater income and lower
expenses for households and greater savings rates that can provide cushion
214. See Nakita Q. Cuttino, The Rise of “FringeTech”: Regulatory Risks in Earned-Wage
Access, 115 NW. U. L. REV. 1505, 1505-06 (2021) (proposing a regulatory framework to mitigate
risks in earned-wage access products).
215. See Yonathan A. Arbel, Payday, 98 WASH. U. L. REV. 1, 14-16 (2020) (discussing
how more regular payment of wages would help consumer liquidity management). See also Yesha
Yadav, FedNow or FedLater?, BANKING RISK & REGUL. (Aug. 10, 2023), https://
www.bankingriskandregulation.com/fednow-or-fedlater (on file with author) (suggesting that a
lack of accessible real-time payments in the United States imposes costs on the unbanked and
underbanked). Again, if the economic problem faced by the unbanked or underbanked is one of
lack of funds, rather than lack of liquidity, it is not clear why accessible real-time payments would
be a solution.
216. See Abbye Atkinson, Rethinking Credit as Social Provision, 71 STAN. L. REV. 1093,
1098-99 (2019).
217. Id.
The Financial Inclusion Trilemma
163
against unexpected expenses.
218
Regulation can play a key role in changing
the shape of the economy, but it is not primarily consumer finance
regulation that is involved.
At best, the role consumer finance regulation can play in small-dollar
credit markets is to lower costs through structural changes in markets, such
as moving borrowers away from short-term, non-amortizing rollover
products to longer-term, prepayable, amortizing installment loans.
219
Regulation can mitigate costs, not eliminate them, because it cannot
eliminate demand. This still leaves us, then, with the trilemma for small-
dollar credit. The terms of small-dollar credit can be made more palatable,
but the fundamental economics of small-dollar lending means that the
product will never be cheap.
Balancing the imperatives of access and fairness in small-dollar credit
markets remains one of the thorniest policy decisions in consumer finance.
Recognizing that there is no easy solution to small-dollar credit, we can
still aim to achieve greater financial inclusion in deposit account markets,
but doing so will require disentangling these two very different types of
financial inclusion problems.
Conclusion
Financial inclusion has been a policy priority in the United States for
decades, but there is still little to show for it: many households remain
unbanked, and many more are underbanked. The financial inclusion
problem cannot be solved by private provision, even with the deployment
of new technologies, because the economics of small-dollar deposits and
small-dollar lending make it impossible to simultaneously offer these
products on a wide scale, fairly, and profitably on a stand-alone basis. This
is the financial inclusion trilemma.
For the unbanked, the trilemma can be addressed by loosening the
profitability constraint through subsidization in one form or another—
whether directly or indirectly through appropriations or by mandating
cross-subsidization by private parties. For the underbanked, however,
financial inclusion remains a thornier problem that can only be fully
addressed through broader changes in the U.S. economy.
218. See generally ELIZABETH WARREN & AMELIA WARREN TYAGI, THE TWO INCOME
TRAP: WHY MIDDLE-CLASS PARENTS ARE GOING BROKE (2004); JACOB S. HACKER, THE
GREAT RISK SHIFT: THE NEW ECONOMIC INSECURITY AND THE DECLINE OF THE AMERICAN
DREAM (2d ed. 2008).
219. See PEW CHARITABLE TRUSTS, supra note 85, at 12 (noting how Colorado’s 2011
payday loan reforms resulted in lower costs to borrowers, despite increasing authorized fees, by
changing the structure of the industry so that lenders could better compete on price).
