Summer 2020
Public Money: Digital Dollars? Fed Acccounts? Postal Banking?

Public Money: Digital Dollars? Fed Accounts? Postal Banking?

Prompt for Discussion

Contributors: John Crawford, Morgan Ricks, Lev Menand, Aaron Klein, Robert Hockett, Abbye Atkinson, Leonidas Zelmanovitz, Bruno Meyerhof Salama, Sheila Bair, James McAndrews, Yesha Yadav, Sarah Bloom Raskin, Mehrsa Baradaran, Christopher Giancarlo, Saule T. Omarova, and Nakita Q. Cuttino.

The recently enacted CARES Act has exposed glaring problems in the U.S. system of money and payments. Delayed stimulus payments are costly for struggling families and for the economy as a whole. Unfortunately, the United States has one of the slowest payment systems in the developed world. On top of that, millions of Americans don’t have bank accounts. They must receive their stimulus dollars as physical checks, which are slow to arrive and often costly to convert into cash.

Growing awareness of these systemic defects has stimulated renewed interest in public sector solutions. When Democrats in the U.S. House of Representatives released their proposed stimulus legislation in March, they included a provision giving people the option to receive their stimulus as “Digital Dollars” through a new system of “FedAccounts” maintained at the Federal Reserve. While this provision didn’t make it into the ultimate legislation, Senator Sherrod Brown, ranking member on the Senate Banking Committee, later introduced separate legislation “to allow everyone to set up a digital dollar wallet, called a FedAccount.” Maxine Waters, chair of the House Financial Services Committee, did the same. And Representatives Rashida Tlaib and Pramila Jayapal included similar language in recently proposed legislation.

These proposals intersect with and complement proposals to implement postal banking as a way of serving un- and underbanked households. As these debates unfold in the United States, other central banks, including the Bank of China, are preparing to release their own central bank digital currencies (CBDCs) in the coming months.

In this roundtable, we invite participants to comment on these public-sector initiatives and what they mean for the future of money. Should the Federal Reserve issue a digital dollar, available to the general public? What problem would it solve or mitigate, and what new problems and risks would it create? Should central bank digital currencies take the form of “accounts” or should they try to emulate digital “tokens”? Can and should a FedAccount program be linked to or even merged with a postal banking initiative? Does maintaining the U.S. dollar’s status as the dominant global currency hinge on launching a digital dollar?


November 19, 2020
The Other Half of the FedAccounts Plan: What Happens on the Asset Side of the Fed’s Ledger?
Saule T. Omarova, Cornell Law School

October 16, 2020
Central Bank Digital Currency: the hidden agenda
Leonidas Zelmanovitz, Liberty Fund
Bruno Meyerhof Salama, UC Berkeley Law School

October 7, 2020
On Equity within Public-Sector Banking Initiatives
Abbye Atkinson, Berkeley Law

September 28, 2020
The Inclusive Value Ledger: A Public Platform for Digital Dollars, Digital Payments, and Digital Public Banking
Robert Hockett, Cornell Law School

September 1, 2020
Designing Financial Services for People with Low and Uncertain Income
James McAndrews, TNB USA Inc., and the Wharton Financial Institutions Center

August 27, 2020
What to Do While Waiting for Fed Accounts
Sarah Bloom Raskin, Duke University

August 18, 2020
How to Fix the Covid Stimulus Payment Problem: Accounts, Information, and Infrastructure
Aaron Klein, Brookings Institution

August 10, 2020
FedAccounts: Digital Dollars
Morgan Ricks, Vanderbilt University
John Crawford, University of California Hastings College of the Law
Lev Menand, Columbia Law School

A. Klein, How to Fix the Covid Stimulus Payment Problem: Accounts, Information, and Infrastructure

August 18, 2020

Aaron Klein, Brookings Institution

The financial response to the Covid crisis demonstrated the inability of the federal government to rapidly get money to people during times of crisis. It took between three weeks and three months for Americans to receive a single penny of their emergency impact payments, aka the Covid stimulus.

These delays impose significant hardship and costs on those most dependent on emergency funds, undermining the payments’ economic and social impact in a deeply regressive fashion. Half of American families live paycheck to paycheck, and cannot afford to wait for emergency assistance when a national emergency interrupts their income stream.

Nor is this a one-off problem: Covid stimulus payments were the third time in the past twenty years that the federal government has authorized direct stimulus payments to individuals (2001 tax cut, 2008 recession bill).

This problem is easily solved, just not in the manner frequently proposed. Creating a world where emergency financial assistance streams as quickly as Netflix is actually quite easy. The solution is three-pronged:

  1. Accounts: Americans need to have accounts to receive funds. 
  2. Information: The federal government needs to have the information to match people to their account.
  3. Infrastructure: The payment system’s infrastructure must be able to rapidly transmit funds from the government to people.

This paper examines each of these prongs. It proposes solutions that address the gaps that exist in each of them. A key takeaway is that the biggest impediments have to do with infrastructure and information. Thus, solutions that focus simply on universal accounts without fixing information and infrastructure will fail to improve delivery of stimulus payments.



The vast majority of American households have accounts: 93.5% of American families (more than 14 out of 15) have an account at a bank or credit union. This figure has grown a few percentage points over the past decade, demonstrating progress within the current framework.

Unbanked rates are much higher for Black and Latinx households, with account access at 84% and 86% respectively (roughly 6 out of 7). This highlights the importance of universal accounts for communities of color and is a good reason why disproportionate attention is paid to creating universal accounts. However, it is also a reminder that the vast majority of people of color have accounts.

The reasons why people are unbanked largely have to do with the cost of banking and anti-money laundering (AML) and fraud systems. These problems require different solutions from the commonly cited problems of limited physical branch locations and hours, lack of trust or faith in banks, and lack of financial “literacy” or education. Those issues are not major drivers of households being unbanked (see Figure 1).

The main problem for the majority of unbanked people is money, specifically how expensive accounts are and how little money they have. The cost of accounts and lack of consumer funds are cited by 4 in 9 unbanked households (44%) as the main reason they do not have an account. More than one-third of households cited lack of funds as the main reason for not having an account (and more than half cited it as one of the reasons). Another ten percent cited high or unpredictable fees. Combined, the cost of accounts and lack of funds explain the majority of unbanked individual’s decisions.

The solution is to make banking cheaper. Simple, no-cost bank account programs have been developed by the Cities for Financial Empowerment Fund’s Bank On initiative, which works in conjunction with financial institutions, and by the FDIC through its Safe Account program. Both have a series of features including low balance minimums, minimal fees, and (importantly) no overdrafts. Customers leverage technology to access the accounts, which function very much like traditional bank-transaction accounts, sometimes without a physical checkbook. These accounts have proven remarkably popular. For example, more than 1 in 5 of Citibank’s new accounts are structured as so-called Access Accounts modeled on this program.

All banks and credit unions are chartered by either the federal government or a state government and due to that charter and other benefits (including deposit insurance) have a fundamental duty to serve their community and their members. Congress or federal regulators should require all institutions to offer basic transaction accounts that meets these criteria. As many banks and credit unions have shown, this can be a profitable endeavor, particularly when AML compliance cost can be contained (discussed below). Regardless, Americans have a fundamental right to participate in economic life, and having a transaction account is necessary for full participation. This requirement tackles the main problem facing the unbanked: the high cost of banking for those with less money.

High AML costs and concerns about past consumer behavior are the next biggest obstacle to universal banking. There are 8.5 million households without bank accounts according to the FDIC. To apply for a bank account, an applicant goes through a screening system to see if they are on the “do not bank” list, a list comprised of people who have had accounts flagged for money laundering or fraud violations. Data on the ”do not bank” system are limited but the National Consumer Law Center report estimated that 2.3 million people applying were rejected as a result of this list over the course of a single year. A person is different than a household and every unbanked household may not apply for an account in a given year. However, if this report is even a rough indication, it shows that as many as 25 to 30 percent of those without bank accounts may be on the “do not bank” list. At the very least, the FDIC’s own survey indicated that over 15% of unbanked households had accounts involuntarily closed or refused due to credit or identification issues, setting a lower bound for the share of unbanked as a direct result of the “do not bank” list.

Solutions designed to provide universal bank accounts will fail unless they deal with the AML and fraud issue. For example, simply establishing a postal banking option to serve these Americans will not work unless the post office is exempt from AML requirements or required to provide services for people who have records of fraud.

Policy solutions need to stop the de-risking phenomenon whereby banks refuse to grant accounts to anyone on the do not bank” list. To start with, bank regulators, Treasury’s Financial Crimes Enforcement Network (FinCEN), and the Consumer Financial Protection Bureau (CFPB) should conduct robust regulation and oversight over the companies like ChexSystem that control these lists. The government likely has to do a better job of tailoring its AML regulation to catch bad actors without unfairly excluding innocent people from the banking system, while private industry has to provide consumers a fair method to contest involuntary account closure and inaccurate placement on the list.

When it comes to explaining why the unbanked don’t have bank accounts, the cost of accounts and AML obstacles dwarf other issues. Distrust of banks and privacy concerns combine to account for 15% of the unbanked not having accounts, while limited branch locations and hours only another 6%. Put a different way, only 1 in 5 unbanked people cite these concerns as the main reason why they do not have an account. That translates into about just over 1 percent of the entire population for whom distrust of banks and limited branch locations and hours is the main reason why they do not have an account. That constitutes fewer households (1.8 million) than people who are estimated to be on the “do not bank” list.



While most American families have bank accounts it turns out that Uncle Sam lacks the information to link people to their accounts. The Treasury Department simply does not know your bank account. It is a mistake to think that between all of the various arms of the government (IRS, Social Security, employment and food benefits, etc..) such information would be stored and easily accessible in times of emergency. The fastest way that Treasury could find people’s account was to look at those who had filed taxes and asked for their refund to be send via direct deposit. That constituted  only roughly half of taxpayers, in part because 20-25 million households filed using a tax advance product that did provided the Treasury department a bank account controlled by a tax preparation service, not by the tax filer.

Eventually, Treasury was able to coordinate with the Social Security Administration to get bank account information for social security recipients who had not filed taxes recently. For 70 million people Treasury had to simply do it the old fashioned way, write a check or issue a prepaid debit card.

The simplest solution would be to ask Americans to provide their bank account information to the federal government as part of filing taxes. At the same time, Treasury can use 1099 interest filings to pair individuals who do not file taxes with bank accounts registered in their name. While there are some privacy concerns, it should be noted that banks are already required to report 1099 interest income to the IRS.[1] Further, all bank accounts are already subject to AML requirements that allow bank regulators and FinCEN to identify individuals who hold personal accounts. Simply connecting that information to the Treasury Department’s payment arm should not worsen privacy concerns.

There are some individuals who are banked but only through non-interest-bearing accounts (including checking) and who do not file taxes. For those individuals, the Treasury Department could use other means to identify their accounts. First, some have already submitted that information as part of Treasury’s attempt to collect account information to send the COVID checks. This information should be kept on file and the ability to add to it should be made permanent. Second, numerous online systems such as Stripe, PayPal, and Plaid are able to identify individuals with their bank account. Treasury could easily engage these companies to identify individuals and link to their accounts, while allowing individuals to opt out. Finally, consumers who opt to use pre-paid reloadable debit cards as their primary ‘banking’ method could give Treasury their card information, either directly or through the pre-paid card company.

It should also be noted that bank accounts held by corporate entities with secret ownership are prime vehicles for money laundering, and bipartisan legislation has been proposed in Congress to require beneficial ownership information for all corporate bank accounts. That legislation is gaining traction, having been included in a recent National Defense Authorization Act, and may well become law soon.



The final issue, slow payments, is in some ways the simplest and in other ways the most difficult. The Federal Reserve has the legal authority to require immediate funds availability to customers for all payments, as well as for just Treasury payments. One of the Fed’s levers, the Expedited Funds Availability Act, empowers the Fed to set minimum times for funds availability for a variety of payment instruments, including Treasury payments (see EFAA Section 4002(b)). Other payments laws give the Fed broad authority to set availability schedules. Put simply, the Fed could solve this problem tomorrow with one regulation.

Why haven’t they? The problem is not technological; real-time payment systems already exist in the US. The problem is that the Fed is both regulator of all payment systems and operator of its own system, the Automated Clearing House (ACH). The ACH is incapable of transitioning to real-time, leaving the Fed with unappealing alternatives: either establish a regulatory standard its own system cannot meet, or build a new system. The Fed has chosen the later, announcing the creation of FedNow which they hope to have operational sometime the mid 2020s.

The Fed’s failure to upgrade the U.S. payment infrastructure provides tremendous profit opportunity for banks in the form of overdraft fees. Banks and credit unions earn as much as $35 billion a year from these fees, which are economically equivalent to the extension of credit at annualized percentage rates in the hundreds to thousands. Some banks have become so dependent on overdraft income that it constitutes 50 percent or more of their total profit (see Armed Forces Bank). Reducing overdraft fees would reduce banks’ profitability and could impact the basic safety and soundness of a handful that have basically become check cashers and payday lenders with charters.

These costs are born by consumers who by definition have nothing left in the bank. Eight percent of all bank customers are ’heavy overdrafters’ who pay $300 a year or more in overdraft fees, according to the CFPB. This is more than the number of people who are unbanked. Reducing the delay between incoming payments (credits) and debits would substantially cut the incidence of overdrafts, both directly through the faster flow of funds and indirectly by providing consumers more information about the actual money available in their accounts. Wells Fargo alone reported over 2 million customers had overdrafts attributable to only a 24 hour delay in direct-deposit payments, which like all ACH payments are not instant. (Wells has instituted a program to forgive those overdrafts—an indicator of the size of this problem).

Returning to the specific example of Covid stimulus payments, the first batch of Treasury’s Covid payments were sent on Good Friday but were not available to consumers until the following Wednesday. Those six days cost families dearly if they over-drafted or went to a payday lender to make up the difference.



To improve its ability to deliver economic impact payments to individuals, the federal government should do the following:

1) Require all financial institutions to offer no/low-cost basic banking services, similar to the Bank On initiative and the FDIC’s Safe Account programs. The main cause of the problem of unbanked households is the cost of basic banking. Banks are governmentally chartered entities tasked with serving public purposes as well as making private returns; there is no reason why they cannot be required to provide simple, no/low-cost accounts. With technology and with reductions in the cost of AML compliance, these accounts can be profitable or, at least, impose only modest costs on institutions. The financial system has spent too long using the poor to cross-subsidize the rich through payments and fees; providing this basic universal service is not too much to ask.

2) Empower consumers and regulators to vet the “do not bank” list to enhance access. Offering universal accounts will not help those who are blackballed. Insufficient incentives and structures exist to fix these systems. Unless they are addressed, millions will be financially marginalized regardless of the type of universal account system created.

3) Connect existing information between silos to effectuate payments, gathering and retaining information where it is lacking. Everyone has a financial relationship with Uncle Sam, and having account information on file will allow for rapid payments in emergencies.

4) Require immediate funds availability for all Treasury payments and, even better, for all payments. Legislation proposed by Senators Van Hollen (D-MD) and Warren (D-MA) would accomplish this. Their legislation would also require the Fed to build a real-time payment system, but this is not essential. What is crucial is that the legislation require banks to make funds available immediately. Consumers should not have to continue to suffer while a new system is built.


[1] Note this covers accounts that generate over $10 of interest income, thus exempting accounts some accounts.