September 1, 2020
James McAndrews, TNB USA Inc., and the Wharton Financial Institutions Center
The Center for Responsible Lending reported in June 2020 that U.S. banks collected over $11 billion in overdraft fees annually, with most of those fees coming from only five percent of the banks’ depositors. Overdraft fees are an awful consequence of having a bank account, especially for those who have low-and-uncertain income. Overdrafts combine access to a transactions account with the provision of credit. For those with low-and-uncertain income, it is a toxic mix, one that traps people into a spending their resources on exorbitant fees rather than on helpful services. Overdraft fees concentrate our attention to the need for both transaction services and credit as necessary features in designing useful financial services for the poor.
An appropriate design for financial services for low-income people must address the entwined needs for both transaction services and credit access. The first step is to provide to people in need a “positive-balance account.” That is, an account whose balance is not allowed to drift into negative territory. Such accounts are now possible because “high-availability” network management has become widespread with the growth of the internet. As a consequence, an account provider can limit access to an account to real-time electronic means via debit cards used at the point-of-sale, online banking, and through automated teller machines.[1] At the same time, credit access needs to be improved for people with low-and-uncertain income. It is extremely costly to provide people with low-and-uncertain income credit privately because of the combined effects of the small amounts of credit involved, the largely fixed costs of underwriting, and the high risk of default. In the private sector, then, credit on reasonable terms is not available for those who are most in need of it. They rely on overdrafts, check-cashers, and pawn brokers; what John Caskey has called fringe banks, to supply their demands imperfectly at shockingly high interest rates.[2]
Surprisingly, positive-balance accounts are widely provided today in convenient forms through almost all government benefit programs, with the notable exception of federal tax refunds. Several easy-to-implement improvements in these services can transform benefit accounts into more all-purpose positive-balance banking accounts. Much more needs to be done by the public sector to subsidize credit access to people with low-and-uncertain income. The private sector, using conventional underwriting technology cannot provide small-loan credit at reasonable costs. I outline a new approach here. In it, the federal government would create a new federally sponsored enterprise to guarantee securities backed by credit card balances that are drawn on by owners of the eligible positive-balance accounts. Such credit can be limited in amount and partially secured by compensating deposits of savings by the borrower, greatly lowering underwriting costs and, for the borrower, preventing debt-spirals. With this approach, the cost of credit provision can be capped at reasonable levels. In these ways, individuals with low-and-uncertain income can have access to both deposit and borrowing services at low cost, and which use the dominant means available to higher-income people: debit and credit cards.
Positive-Balance Accounts
Three laws enacted between 1996 and 2000 are responsible for a vast expansion in the provision of positive-balance accounts to everyone who receives federal benefits and most state benefits as well. The laws mandated: first, that the federal food stamp program, administered by the states, provide all services electronically;[3] second, that all the states use methods that were portable and interoperable across states;[4] finally, that all federal benefits, excluding tax returns, be provided electronically.[5] These steps resulted in a nationwide system of “electronic benefit transfer” accounts for more than 40 million beneficiaries, accessible via debit cards (most affiliated with the Quest network), ATMs, and over the counter. Further, the federal program enables approximately 5 million beneficiaries to access their accounts by Direct Express cards, ATMs and over the counter services.[6] The accounts in the federal program, managed by a contracting bank, are insured by the FDIC and covered by the protections on fraud and loss outlined in Regulation E. The funds in the accounts cannot be garnished; the accounts are more available after disasters, such as hurricanes, less subject to loss or theft of funds, and more convenient than prior methods involving paper checks.[7]
The expansion of access to payment services provided by these accounts led the Federal Reserve, in 2016, to revise its definition of “transaction account” in the Survey of Consumer Finances. In that year’s survey, the Fed measured that approximately 93.2 percent of U.S. households had (traditional) bank accounts, while the 2017 FDIC Survey estimated that 93.5 had bank accounts, showing a high degree of coherence in measurement between the two separate surveys. But the Fed reckoned in 2016 that 98 percent of households had transactions accounts, where they included responses to questions about ownership of either reloadable prepaid debit cards or electronic benefit transfer and Direct Express cards. These positive-balance accounts provide many services that owners would otherwise obtain via a bank account, but without the risks of overdrafts, and significantly less risk of garnishment. Fear of garnishment of bank accounts is another disincentive to establishing bank accounts for those with outstanding debts or judgments. The costs to the account holder are low; in many systems, fees can be avoided altogether by using in-network ATMs and planning one’s withdrawals, while in others fees of a few dollars per month are imposed.
However, the existing positive-balance benefit cards have three limitations that should be overcome if we wish to better provide transactions services to those with low-and-uncertain income. First, the existing cards don’t allow owners to deposit value into their accounts—the accounts are restricted to providing government benefits to the recipient. Without the ability to deposit funds, for example, wages, into the account, the recipient’s transaction options are unduly restricted. Adding that functionality to the accounts would not add to the risks of their provision and would only add a modest amount of costs. Second, the services should be expanded by providing owners a savings account, one whose balances are not available to make purchases from a point-of-sale, but are transferable (with delay) into the transaction account at an ATM. As we will see below, funds placed in the savings account can provide the saver access to credit, so the account can serve as a commitment device to future repayment. These two expansions of services would make these benefit accounts true general-purpose accounts. Third, non-beneficiaries with limited income, for example up to 1.5 times the poverty level, should be able to establish accounts in these systems. We can call these expanded, buy-in accounts, public-option transaction accounts, or POTAs.
Part of the increased costs that would be incurred by allowing deposits to be made into the pre-existing benefit account and to allow non-beneficiary account holders is a result of the requirements to know your customer (KYC), engage in anti-money laundering (AML) actions, and conduct other due diligence under the Bank Secrecy Act. These requirements create costs that are difficult for banks to recover from accountholders that hold only modest balances. The administrators of government benefit programs, however, are well-positioned to economize on these costs as they already have good methods of identification of beneficiaries to match their benefit payments to the beneficiary.[8]
To sum up, I recommend that the current state and federal electronic benefit delivery systems be expanded to allow deposits, to create a separate account for savings, and to allow low-income non-beneficiaries to establish accounts in the system. Congress could accomplish this by requiring that the EBT program of the states and the Direct Express program of the federal government be jointly administered with those features. The legislation could appropriate money to subsidize, to a limited extent, these services. Accountholders could afford modest monthly fees, and non-beneficiary account holders could pay higher fees for access to the accounts.
Credit Access Corporation
Access to credit is severely restricted for people with low-and-uncertain income. That restriction is very costly as credit is so valuable to them. People with low-and-uncertain income have high demand for credit for obvious reasons. The banking sector cannot supply the credit at reasonable cost. The reason is that each loan must be underwritten, which takes time and labor. Because the amount of credit that is feasible to repay for people with low-and-uncertain income is low, and because people with uncertain income are at a high risk of default, the costs-per-dollar of credit provision are enormous. Consequently, the market fails to provide the credit that is needed by this population. Without access to credit from mainstream financial institutions and services, people respond by using overdrafts, payday loans, pawn brokers, and loan sharks, all of which are extremely costly and inflexible.
The key to surmounting this mismatch is to underwrite credit in a different way. Here I recommend a three-pronged approach. First, the federal government should create a government-sponsored enterprise, the Credit Access Corporation (CAC), that would guarantee the principal and interest paid on securities backed by credit-card balances of eligible participants. Second, owners of POTA accounts are eligible to be provided credit guaranteed by the CAC (and extended by a bank that services the debt). Finally, credit extended on these credit cards is limited to small amounts, say a maximum of $1000, and must be partially backed by a compensating deposit in a POTA savings account. For example, a borrower may be required to fully secure her borrowing to begin accessing credit. After some successful repayments, the borrower could ascend a ladder of credit, where she would only have to secure her credit by up to 75 percent, and so on. In other words, to get access to credit, an individual must save up amounts in a savings account. As she demonstrates success, she will have access to unsecured credit, which will be guaranteed by the CAC. This method of securing credit card debt is now common in the “secured credit card” market.[9] In this way, the interest rate at which the person can borrow can be capped at reasonable levels. If the borrower defaults, she loses the compensating balance, and is excluded from accessing credit for some period, such as two years, and thereafter must secure her credit fully for an extended period, but the funds in the POTA accounts (other than the compensating balance) cannot be seized by the lender.
Conclusion
The emphasis of this policy proposal is to build on the enormous success of electronic benefit transfer systems and the familiarity that the tens of millions of users have with those systems. These systems have extended access to electronic financial transaction services to millions of the most vulnerable Americans at low cost, low levels of fraud, high levels of user satisfaction, and low stigma of use. Expanding these services is likely to be the least expensive way of providing electronic financial transaction services to other Americans who need them. By combining deposit services with already established benefit accounts, owners experience a true convenience in managing their finances.[10] The federal and state systems could be managed jointly, by adopting the standards of the federal system, namely that the balances in the accounts are insured by the FDIC and subject to the protections against loss and theft by Regulation E. By having the federal and state governments contract for these facilities, competitive bidding can be used to provide these services efficiently.[11] Furthermore, the use of the debit card networks that are already in place and are the most widely used means of payment in the U.S. takes advantage of both familiarity and ubiquity to provide services to people with low-and-uncertain income.
By marrying the ownership of a POTA account to the option to have credit provided through a credit card, the design here leverages information contained in the deposit and savings behavior of credit recipients. In expanding credit, the design uses a currently successful service, secured credit cards, as a model. Rather than attempting complicated underwriting on a loan-by-loan basis, the underwriting is provided by the account owner saving some amount and being rewarded with credit. If the borrower keeps current on her payments, credit will continue to be available to her. Here again, by using credit cards and modern information systems, the cost of credit is kept low. The likely losses to the Credit Access Corporation would be quite low, given the incentives of borrowers to maintain their credit.
Consider the situation when the eligible owners of a POTA are adults with up to 1.5 times the poverty level of income. In 2018, that group consisted of about 35 million adults. If each had a maximum amount of credit outstanding on their cards, the maximum possible amount of securities in the program would be only slightly more than $35 billion, a small amount when compared with other government credit guarantee programs. Further, only a portion the outstanding credit would be unsecured, so expected losses in the program would be very small. The CAC could be created separately from the POTAs, but the use of the savings account enables the provision of secured credit cards, and it is likely that the two proposals have high complementarity as a result.
Significant progress has been made in recent decades in expanding the use of transaction accounts, with the Federal Reserve’s Survey of Consumer Finances showing that the number of American households with transaction accounts rising from 75 percent in 1970 to 85 percent in 1990, and to 98 percent in 2016. That these gains are not sufficient is clear from a simple review of the overdraft costs incurred by account owners. Checking accounts are not necessarily desirable for people with low-and-uncertain income, and banks don’t typically offer credit on reasonable terms to that population. The private sector alone is not well equipped to provide a minimum range of needed financial services. The positive balance accounts now in use, such as electronic benefit accounts, are limited in purpose, and should be expanded to offer deposit facilities, savings accounts, and the ability for low-income Americans to establish these accounts regardless of whether they are due federal or state benefits. These accounts could be built from the base of existing benefit accounts of states and the federal government, with Congressional legislation to provide only modest subsidies to the system. By linking ownership of broad POTAs with a secured credit card, one whose uncollateralized credit is guaranteed by a government-sponsored enterprise, low-income Americans would have access to the most widely used financial technologies in use today: debit and credit cards.
—–
[1] In general, because checks are paper-based instruments drawn without knowledge of the account provider, and because of the random delays between the check being written and being presented to the bank on which it is drawn, a check writer can easily and inadvertently overdraw her account
[2] John Caskey, Fringe Banks, Check Cashing Outlets, Pawnshops, and the Poor, 1994, The Russell Sage Foundation.
[3] The Personal Responsibility and Work Opportunities Reconciliation Act of 1996 (PRWORA) (P.L. 104-193) required states to implement Electronic Benefit Transfer systems before Oct. 1, 2002. The final implementation of the law was achieved in 2004.
[4] Electronic Benefit Transfer Interoperability and Portability Act of 2000, Public Law 106-171.
[5] A provision of the Debt Collection Improvement Act of 1996, EFT 99 requires federal agencies to use electronic funds transfer (EFT) for most payments, with the exception of tax refunds, starting Jan. 2, 1999.
[6] More specifically, each state manages its own EBT system. Those systems are all interoperable, in that a beneficiary may receive and spend benefits while out-of-state. Most of the systems are members of the Quest electronic funds transfer network. In most cases, the funds that are provided in the accounts of beneficiaries are direct liabilities of the state. As the states distribute food stamp or Supplementary Nutritional Assistance Program (SNAP) benefits, the number of recipients is so large that other state-specific benefits are also often distributed via these systems. The federal system is managed by a bank-contractor. The benefits provided in accounts are insured by the FDIC. The system distributes a variety of federal benefits including Veterans’ benefits, Social Security benefits, and others.
[7] Susan Herbst Murphy, “Conference Summary: Government Use of the Payment Card System, Issuance, Acceptance, and Regulation,” 2012, Federal Reserve Bank of Philadelphia. See pages 15-19.
[8] Another element of this recommendation is for Congress to amend the Bank Secrecy Act to create a tiered KYC regime that exempts electronic benefit accounts with balances under a certain limit.
[9] See Larry Santucci, “The Secured Credit Card Market,” November 2016, The Federal Reserve Bank of Philadelphia, and “Moving into the Mainstream: Who Graduates from Secured Credit Card Programs?,” May 2019, The Federal Reserve Bank of Philadelphia. The second paper finds that “Secured card graduation rates have accelerated in recent years across all credit score groups as well as within the initially unscoreable consumer population. Twenty percent of accounts in the 2012 cohort graduated by age 61 (months)…” Reporting of credit histories to credit reporting agencies can also provide incentives for borrowers to build a good credit history.
[10] This provides, in other words, an economy of scope in demand, that greatly reduces costs relative to establishing separate systems, one for benefits and another for other deposits, each with their own sign-up procedures, etc.
[11] There need not be a single nationwide provider. Instead, competition can be enhanced by providing services in different regions of the country, and maintaining the interoperability requirement that currently exists for state EBT programs.