Policy Spotlight
The Narrow Bank

Author: Gavriel Schreiber

In a Nutshell

The Narrow Bank (TNB) is a state-chartered passthrough bank that proposes to hold only one asset: account balances at the Federal. Providing an ultra-safe option for large investors to hold their high-powered money is meant to increase financial stability and extend the benefits of Federal Reserve accounts, which are currently available only to depository institutions and select governmental entities, to a wider set of economic actors. Opponents worry that the spread of narrow banking, pioneered by TNB, would undermine the business model of existing banks, hamper economic growth, and amplify economic shocks.

The Problem: How Safe is High-Powered Money?

Small Investors/Depositors: Totally Safe

Small investors and depositors can hold cash – money that is recognized unconditionally as payment and is therefore “high-powered money (HPM)” – in bank accounts that are completely safe.

Consider a small investor/depositor with $100 of HPM. By placing her HPM in a commercial bank, she exchanges HPM for a bank deposit, which is a promise to pay $100 of HPM on demand.[1]

While the bank remains solvent, the investor/depositor can be certain that the bank will convert her deposit to HPM on demand.

If the bank goes bankrupt, the investor/depositor’s HPM remains safe. The Federal Deposit Insurance Corporation (FDIC) is a government agency that insures bank deposits up to $250,000. If a bank can’t return HPM to its depositors, the FDIC will do it instead. Relatively small amounts of HPM (under $250,000) can therefore be stored without any risk of loss.

Commercial Banks: Totally Safe

Commercial banks can also hold HPM in bank accounts that are completely safe.

Consider a commercial bank with $5 million in HPM (from depositors). This bank can place its HPM in a special Federal Reserve account available exclusively to commercial banks. Just like regular depository accounts, the Fed promises to pay its depositor HPM on demand. But because the Fed’s promise to pay is money (unlike deposit banks, who promise to pay with money), commercial bank reserves kept at the Fed are completely safe. HPM held by commercial banks, if deposited in the Fed, can therefore be kept completely safe.

Most of the time, commercial banks would prefer to invest their HPM rather than park it safely at the Fed. Nonetheless, commercial banks maintain positive balances in their Fed accounts for a few reasons:

First, the Fed serves as a type of “clearinghouse” with other commercial banks. When a depositor puts money in Commercial Bank A, they receive a deposit, essentially an IOU stating “Commercial Bank A promises to pay HPM to the depositor or anyone to whom the depositer transers the right to his HPM.” A depositor can use this IOU to buy goods, for example, by swiping his debit card at a coffee shop and transferring his right to the coffee shop. The right to the deposit (called a “check drawn against Bank A”) might then make its way into the hands of the coffee shop’s bank, Commercial Bank B, which can demand HPM from Commercial Bank A (the bank that issued the deposit).[2] Because Banks A and B (and all other commercial banks) frequently receive these checks written on deposits issued by one another, settling their accounts with money parked in their Fed accounts saves them the hassle of transporting cash back-and-forth.

The Fed also serves as a clearinghouse for deposits that banks issue to borrowers. When an individual borrows money from Commercial Bank A, they receive a deposit credit. Unlike depositors, however, borrowers provide the bank with a long-term IOU rather than HPM (e.g., borrowing $100 for a 5-year term means receiving a $100 deposit in exchange for a promise to pay the bank $100 plus interest at the end of five years).[3] As before, the depositor can write a check to a person with an account at Commercial Bank B, which can then demand HPM from Commercial Bank A (the bank that issued the deposit). This time, however, if Commercial Bank B demands dollars issued to borrowers who did not provide Commercial Bank A with HPM in exchange for their deposits, Bank A may not have enough HPM to meet its obligations to Bank B.[4]

This is where the Fed’s clearinghouse function becomes essential. Though Bank A’s lending leaves it open to demands for liquid resources that exceed its reserves, all commercial banks lend using this same model. Bank B is the recipient of checks drawn against Bank A, but Bank A is the recipient of checks drawn against Bank B as well. Their reciprocal deposit demands generally “cancel out,” leaving Bank A on the hook only for the difference.[5] This Fed-facilitated reciprocal cancelling-out undergirds our financial system by allowing banks to issue deposits beyond their reserves.[6]

The second set of reasons commercial banks keep some money in their Fed account relates to legal obligations and risk aversion. The financial crisis of 2008 highlighted the importance of holding sufficient liquidity at the Fed.

For the most part, however, ensuring sufficient commercial bank reserves (and thus sufficient liquidity) has been a higher priority for the government than for commercial banks. Historically, the government has mandated commercial banks keep a certain amount of money in their Fed accounts (Required Reserves, or RR). Banks had to simply absorb the costs of keeping their money at the Fed instead of investing.[7]

Around the time of the financial crisis, Congress authorized the Fed to pay banks interest on their reserves held at the Fed.[8] More recently, the Fed expanded this tool, giving interest on both RR[9] and additional interest on any reserves kept at the Fed above the RR baseline (this is called Interest on Excess Reserves, or IOER). This is a massive expansion—as of January 2020, excess reserves totaled $1,350 billion, dwarfing the merely $200 billion of RR. Furthermore, the IOER rate is not insubstantial—standing at the time of writing at over 1.5%.

It’s essential to bear in mind that the enviable position enjoyed by commercial banks—being able to deposit large amounts of cash safely and receive relatively high interest rates on it—is a result of policy choices made by the U.S. government. Guided by congressional directive, the Fed decides which banks receive Fed accounts and decides to give IOER (TNB claims that Congress mandates opening Fed accounts for state-chartered depository institutions. The Fed maintains it has some discretion). With a different set of policy decisions, our financial system could look much different.

Large Investors: Less than Totally Safe

In contrast to small depositors and banks, large investors like money market funds, pension funds, and businesses with large cash balances like Apple and Microsoft cannot store HPM in bank accounts that are completely safe.

These large investors cannot put their HPM safely in a bank, because the FDIC only insures deposits up to $250,000. Though they hold large amounts of unsecured deposits, if the bank collapses, these investors could be left with large losses. Nor can they put their HPM in ultra-safe Federal Reserve accounts, because the Federal Reserve Act only permits depository institutions and certain governmental entities to deposit money with the Federal Reserve Banks. Even though these funds are major economic players, they are still shut out.

By and large these investors turn to putting their funds into the money market, often using overnight general collateral repurchase agreements (repo). Repo is simply a secured loan, backed by collateral. In a repo, a large investor like a pension fund uses its HPM to buy an asset like a Treasury bond from a large financial institution like a broker/dealer or investment bank. In exchange, the financial institutions promises to repurchase that asset from the investor the following day. This sale-and-repurchase arrangement is functionally equivalent to a one-day loan from the buyer-lender (the pension fund) to the seller-borrower (the broker/dealer). The seller-borrower is happy–it gets to use the pension fund’s money to invest. The buyer-lender (the investor) is happy too–though its money isn’t totally safe, if the seller-borrower collapses and is unable to return the HPM, the investor will still have a Treasury bond (or other debt instrument) to sell. The investor then rolls these agreements over, day after day.

These repo agreements are a good option for large investors, but many would probably prefer to simply deposit their cash at the Fed (like commercial banks). TNB wants to help them do just that.

The Proposal

The Narrow Bank extends the unique benefit of Fed accounts, currently conferred only upon depository institutions and certain governmental entities, to large cash investors. TNB plans to take deposits from large investors and park them in its Federal Reserve account. That’s it. The money will earn interest, TNB will take a cut and pass the rest along to its depositors.

All money deposited at TNB will be in the Fed. Thus, there is no risk of TNB being unable to meet demands for deposits (unlike commercial banks, who could be unable to meet their obligations if they make poor investments). TNB deposits will be incredibly safe.

Arguments For

Advocates of TNB have said that…

  1. TNB will reduce the nation’s vulnerability to financial crises. Because TNB would store all deposits in the Fed, it will be immune to runs. Additionally, it will provide a floor to stabilize repo markets. Large investors will only use repo markets to store their money if the return offered by those markets is larger than the return offered by TNB. During times of economic uncertainty, large investors will be able to safely keep their money in the Fed (through TNB) rather than being forced to accept high risk and low returns in volatile markets.
  • TNB will provide an additional tool for the Fed to control inflation. Related to point (1), the “floor” set by TNB depends on the IOER rate set by the Fed. The Fed could raise or lower the IOER in order to encourage or discourage participation in repo markets. Alternatively, the Fed could further differentiate IOER rates and fine-tune liquidity flows between safe deposit facilities and competitive investment markets.
  • TNB will restore agency to investors. Large investors have little option but to maintain large balances with commercial banks, who often engage in risky investments. Some investors don’t mind the risk, others do. TNB provides a safe choice for investors who care about safety above all else.
  • TNB will reduce unfair government favoritism. The federal government advantages commercial banks to the exclusion of other major economic players in two important ways: First, it perversely subsidizes banks’ costs by giving them interest in excess of what they provide their depositors. Second, it allows them the security of Fed accounts. TNB, as a passthrough, will allow other investors to reap the same benefits afforded commercial banks.
  • TNB is not so different from a normal bank. Regular banks take deposits and invest them. Today, Federal Reserve accounts are an investment that many economic actors indirectly hold–from foreign banks to money market funds. TNB would do the same.
  • TNB would promote competition by providing their depositors higher interest rates than those currently offered by commercial banks.

Arguments Against

Opponents of TNB have said that…

  1. TNB will reduce economic activity. Commercial Banks, investment banks, and other financial entities rely on repo markets to fund trading and investment activities. With many institutional investors putting their money in TNB, a large swath of large investors will have less money with which to invest. This could raise the cost of credit for households, businesses, and other banks. Individuals and investors may be more risk-averse than necessary,[10] so employing a fractional reserve banking system forces people to use their money in productive ways rather than keeping it under the proverbial mattress.
  • TNB will reduce the Fed’s ability to implement monetary policy. Institutional investors depositing their money at TNB will leave short-term debt markets, making those rates more volatile and more difficult for the Fed to control. Specifically, TNB will render Fed’s Overnight Reverse Repurchase facility (ONRRP) obsolete. ONRRP, a tool of monetary policy implementation, offers to execute repo agreements at a certain rate. Because no money fund would lend repo to a broker-dealer for a rate that is lower than the one offered by the Fed, ONRRP sets a “floor” on money market lending rates. If TNB obviated the need for money markets and ONRRP, it would undermine the Fed’s ability to set the lower bound of its interest rate target range. TNB will also drastically increase the Fed’s balance sheet, which will limit its ability to implement monetary policy.
  • TNB will compound financial instability. In times of mild economic stress, investors will run from financial markets into TNB. Small shocks will turn into large ones.[11]
  • TNB will force taxpayers to subsidize money market funds. By increasing the Fed’s IOER payment obligations, TNB will reduce the Fed’s remittances to the Treasury. Taxpayers will have to make up the difference.
  • Commercial banks uniquely deserve the advantage of Fed accounts. Commercial banks do things that are socially necessary but less profitable (like providing credit to households and small businesses). By subsidizing commercial banks and using regulation to insulate competition, the Fed allows them to continue their socially beneficial activities.[12] TNB will benefit exclusively “the most financially secure institutions,” so it doesn’t deserve the same advantage.[13]
  • TNB represents a radical redesign of the monetary system that should be undertaken in a more thoughtful and centralized way. The Fed was designed to be a banker’s bank—commercial banks are arms of the central bank. If we want to adopt a new system of money augmentation, we do so by amending the Federal Reserve Act after broad and democratic debate, rather than letting one upstart profit-seeking company upend the system.

What’s the Status of TNB?

Connecticut granted TNB a bank charter. Normally, the Federal Reserve grants state-chartered banks accounts at the Fed as a matter of course. Here, the Fed has refused to act on TNB’s application.

TNB sued the Fed for refusing to act on its application. The Fed filed a motion to dismiss and then proposed a rule which would bar pass-through investment entities like TNB from receiving IOER. Below, a summary:

August 31, 2018 – TNB Files Complaint:

LINK: https://faculty.chicagobooth.edu/john.cochrane/research/papers/Complaint%20(filed).pdf


  1. Relief Sought:[14]
    1. TNB seeks declaratory relief and an injunction to compel the Federal Reserve Bank of New York (FRBNY) to open a Federal Reserve “Master Account” for TNB.
  • The applicable statute compels FRBNY to provide TNB a Master Account[15]
    • The statutory framework requires the Fed provide services to all depository institutions.
    • Fed services cannot be accessed without a master account.
    • TNB is a depository institution duly chartered in Connecticut.
    • Therefore, FRBNY must grant TNB a master account.
  • FRBNY has delayed and shows no intention of opening TNB a Master Account.[16]
  • TNB will have a positive effect on the Fed and the broader economy.[17]
    • TNB would pass along IOER rates to its depositors, setting a more solid interest rate floor than the current target federal funds rate.
    • TNB will place competitive pressure on all banks to raise depository interest rates for all depositors.
    • TNB will provide a similar function to the Fed’s current Overnight Reverse Repurchase Agreement Facility (“ON RRP”) and the Foreign Repo Pool (“FRP”).

March 8, 2019, Federal Reserve files Motion to Dismiss

LINK: https://www.tnbusa.com/wp-content/uploads/2019/03/2019.03.08-HNC-Motion-to-Dismiss-Brief.pdf


  1. TNB has no standing to sue.[18]
    1. They have not been denied an account, and therefore have suffered no injury.
  • TNB’s Claim is not Ripe.[19]
    • The Fed’s review of TNB’s application is ongoing, and so TNB cannot claim injury based on potential denial of their application.
  • The court should decline jurisdiction over the declaratory judgement request because the request is contrary to the public interest.[20]
    • The Fed is carefully evaluating the impact granting a Master Account might have on the economy, and the court should allow the Fed to finish its inquiry.
  • Providing Master Accounts is discretionary – TNB is not entitled to a Master Account.[21]
    • The Fed has discretion to reject deposits, and therefore reject granting Master Accounts.
    • An obligation to create Master Accounts would impermissibly undermine the Fed’s ability to execute its statutory mandate.

NOTE: the Federal Reserve Board of Governors requested permission to file an Amicus brief in support of the motion to dismiss. That request was granted, and their Amicus brief is forthcoming.

March 12, 2019, Federal Reserve Files Notice of Proposed Rulemaking[22]

LINK: https://www.federalregister.gov/documents/2019/03/12/2019-04348/regulation-d-reserve-requirements-of-depository-institutions



The comment period closed on May 13, 2019. As of January 1, 2020, the Fed’s decision was still pending.


Lower the IOER rate paid to institutions that hold a very large proportion of their assets in the form of balances at Reserve Banks. (Such institutions, like TNB, are termed Pass-Through Investment Entities or PTIEs).


  1. Large scale migration of short-term investments from federal funds markets to PTIEs would make other interest rates more volatile, making it more difficult for the Fed to implement policy.
  2. PTIEs would diminish the availability of capital for those institutions that benefit from short-term lending markets.
  3. The ultra-safe option provided by PTIEs would exacerbate economic downturns by fueling runs from money markets to PTIEs.

Further Reading


  • JP Koning provides insight about the genesis of TNB and highlights a concern about its long-term viability.

Advocates of TNB:

  • John Cochrane provides a robust defense of narrow banks generally and picks apart the Fed’s arguments against TNB.
  • The WSJ explains how narrow banks benefit depositors

Critics of TNB:

  • George Selgin defends the Fed’s decision to deny TNB a Master Account.

[1] See Andrew Jackson & Steve Dyson, Modernising Money: Why our Monetary System is Broken and How it Can be Fixed, 53 – 63 (Positive Money, 2012).

[2] See Jackson & Dyson, supra note 1 at 57 – 68.

[3] This is called “maturity transformation,” turning productivity in the future (i.e. an individual’s ability to pay $100 plus interest in five years) into money in the present (i.e. $100 in deposits). See C. Desan, Commercial Banking: Financing through Money Creation (aka Endogenous Credit Creation), A Short Overview 2–3 (Sep. 3 2019) (unpublished manuscript) (on file with Christine Desan, Harvard Law School).

[4] See id.

[5] A bank’s books must balance daily. If there is a shortfall, they have to pay the other bank or default. Generally they make up the sortfall by borrowing from other banks in the federal funds market or from the Federal Reserve Banks through the discount window. The repo market is also an option.

[6] See id. If a bank becomes a chronic “debtor” to other banks, lending more money than it receives in deposits, the Fed also facilitates inter-bank lending, whereby banks with too few depositors can borrow HPM from banks with too many. Id. at 3–4.

[7] Morgan Ricks, Money as Infrastructure, 2018 Colum. Bus. L. Rev. 757, 787 – 88 (2018).

[8] Id.

[9] See Gary Gorton, Clearinghouses and the Origin of Central Banking in the United States, 45 J. Econ. Hist. 277 (1985).

[10] Some scholars describe widespread investment as a type of collective action problem — society is better off when everyone invests, but the individual cost-benefit calculus promotes penny-pinching. Financial systems need to be complex in order to mask the true risks of investment to overcome individuals’ stingy inclinations.  See Steve Randy Waldman, Why is Finance so Complex?, Interfluidity Blog (Dec. 26, 2011) https://www.interfluidity.com/v2/2669.html.

[11] See Regulation D: Reserve Requirements of Depository Institutions, 84 Fed. Reg. 8829, 8829 – 31 (proposed Mar. 12, 2018).

[12] This is known as “cross-subsidization.” For a succinct explanation, read this piece by Scott Sumner. For a critique, check out this piece by Cochrane.

[13] Complaint for Petitioner TNB USA at 1, (S.D.N.Y. 1:18-cv-07978-ALC) (filed Aug. 31, 2018).

[14] Id. at 1 – 4.

[15] Id. at 5 – 10.

[16] Id. at 11 – 15.

[17] Id. at 15 – 22.

[18]  Law in Support of the Federal Reserve Bank’s Motion to Dismiss at 10 – 11, (S.D.N.Y. 1:18-cv-07978-ALC) (filed Mar. 8, 2019)

[19] Id. at 11 – 13.

[20] Id. at 13 – 14.

[21] Id. at 14.

[22] Regulation D: Reserve Requirements of Depository Institutions, 84 Fed. Reg. 8829, 8829 – 31 (proposed Mar. 12, 2018).