Back in the 1980s Professor Edward Kane coined the term regulatory dialectic to capture the dynamics of how regulated financial institutions found innovative ways to circumvent regulations designed to restrict their behavior.

For example, banks adopted the one-bank holding company form in the late 1960s to avoid the restrictions on permissible activities. They subsequently used that device to create non-bank banks that also enabled them to cross state lines, avoiding the restrictions on interstate banking. The process led to a new era of nationwide banking that we now take for granted. Similarly, when Regulation Q restricted the rates that depository institutions could pay retail customers, money market mutual funds came into existence to solve that problem. And when federal law capped the amount of Federal Deposit Insurance an individual could receive to $250K per account, Promontory Financial Group (now a subsidiary of IBM), formed by former Comptroller of the Currency Eugene Ludwig and former Federal Reserve Board Vice Chairman Alan Blinder, found a way to pool accounts to expand deposit insurance coverage for large depositors far beyond the $250K limit.

The most recent example of the regulatory dialectic was the subject of an American Enterprise Institute program on December 2 in Washington, devoted to what is known as The Narrow Bank (TNB). The recently retired executive vice president of the Federal Reserve Bank of New York, Dr. James McAndrews, is now chairman of a newly licensed, uninsured, special-purpose wholesale bank in the State of Connecticut that has applied to the Federal Reserve Bank of New York for a master account. TNB would only do one thing: It would accept deposits from large accredited investors, namely the GSEs, Federal Home Loan Banks, money market mutual funds and selective other large investors; deposit those funds in the master reserve account at the Federal Reserve Bank of New York; receive an interest payment (known as interest on excess reserves [IOER]) from the Bank on those funds; and transfer all that interest to the depositing institutions, less a small fee for the service. So what is the regulatory avoidance? GSEs and Federal Home Loan Banks are permitted to hold deposits at the Fed, which serves as their fiscal agent, but they are not permitted to receive interest on those accounts.

In lieu of this option, what have the GSEs and Federal Home Loan Banks been doing with their deposits held at the Fed? They have been lending those funds out in the overnight market at rates somewhat below IOER, in part frustrating the Fed’s attempt to achieve its target for the federal funds rate with IOER operating as a floor on the funds rate. Whom have the GSEs been lending to? Interestingly enough, they have been lending to foreign banks, which borrow fed funds at a rate slightly below IOER and deposit the funds with the Fed to receive the IOER, earning a risk-free arbitrage. In fact, the terms are so attractive that about 35% of the excess reserves on deposit at the Fed are owned by foreign entities with US-chartered bank subsidiaries. This is despite the fact that such entities hold less than 10% of total deposits. These foreign institutions have some advantages over US banks. First, they are not subject to the hefty Federal Deposit Insurance rate charge to US banks on their total assets. Second, the reserves count towards the Liquidity Coverage Ratio stipulated in the Basel III regulations. Finally, the positive IOER of 2.2% is a risk-free alternative to the ECB’s negative 0.4% levied on reserves.

So what problem does TNB solve for the Fed? The main benefit is to bring the effective funds rate in line with the IOER floor. Of course, the federal funds rate target is above the floor, so what we are talking about is putting a downside limit on the intraday fluctuations in the federal funds rate.

The other issue is that TNB’s existence hinges critically upon there being a large volume of excess reserves. The Fed, as the FOMC’s most recent minutes, suggest, has not decided on the desired size of its balance sheet; nor has it determined to revert to the pre-crisis policy implementation strategy that targeted the federal funds rate in an environment with a very small volume of excess reserves. Its alternative is to continue with a much larger balance sheet and rely primarily upon its current reverse repo strategy for policy implementation. In the pre-crisis world, there would be only a limited opportunity for TNB, and there is the risk that its presence might serve to disintermediate funds, especially from smaller banks, and disrupt the credit process.

The threshold question the Fed faces as it decides whether to grant TNB a master account is what the size of its balance sheet should be. In the pre-crisis world, its balance sheet was determined primarily by the volume of currency outstanding; required reserves were largely met by banks’ cash holdings, and reserve balances were about $8 billion in 2006, or slightly less than 1% of the Fed’s total assets of $850 billion. As of the most recent data, from November, excess reserves were about $1.6 trillion, and assets were $4.1 trillion. If the Fed were to shrink its balance sheet to restore the old relationship between currency and GDP, it would need about $1.9 trillion and an additional $17 billion to account for excess reserves. Additionally, the Fed has other liabilities to the Treasury and to official foreign accounts, which total about $430 billion. Together, these liabilities would imply a balance sheet for the old regime of about $2.3–$2.4 trillion.[1] How big would the balance sheet have to be if the current reverse repo regime were to be followed? The answer hinges on the expected size of the reverse repo market. The market has shrunk considerably since the policy was put in place. The principal users are money market mutual funds. In 2017 the average daily transaction volume in the market was $143 billion, with a standard deviation of $60 billion. By comparison, through the first half of 2018, the average daily transaction volume was only $19 billion, with a standard deviation of $23 billion. These numbers suggest that a balance sheet of about $2.4–$2.5 trillion would enable the Fed to pursue either its historical or new policy regime with a much smaller balance sheet than is presently in place. Such a balance sheet would imply a small role for TNB (and any other copycats that might arise), since reserves would constitute only a small portion of the Fed’s liabilities.

[1] Data as of November 29, 2018.