RRP take-up reduces liquidity held by banks, but does not change the quantity of liquidity held by Non-Banks. This difference arises from our two tiered monetary system, where banks and non-banks hold different types of money. Banks lose reserves (money for banks) when they settle payments to the Fed on behalf of Non-Bank RRP participants. But from the perspective of Non-Banks, the RRP just replaces bank deposits (money for non-banks) with what are essentially secured deposits at the Fed. At the zero lower bound, the RRP is a cash equivalent and RRP take-up is largely a function of bank balance sheet constraints. In this post we walk through the balance sheet mechanics of RRP participation from the perspective of Non-Banks, Banks, and the Fed. The current context suggests the RRP is largely acting as a tool to manage the side effects of QE.
Non-Bank participation in the RRP is almost entirely through 2a-7 money market funds (“MMFs”). Cash investors place cash in MMFs, who in turn place it in the RRP when there is nothing else to invest in (Note MMFs charge a management fee, so cash investor receive around 0% even if RRP is 5bps). For example, suppose an investment in Treasury bills is repaid and all money market rates are below the RRP offering rate. Then the MMF will just re-invest the proceeds into the RRP. The MMF receives $100 in its bank account from the maturing Treasury, and then sends it to the Fed’s RRP. The Fed returns the money the next day.
Another way to think about this is that Non-Banks are simply depositing money at the Fed instead of a commercial bank. Non-Banks have the same amount of money, it’s just held at the Fed in the form of a overnight repo loan (functionally a deposit secured by Treasury collateral). They are free to spend that money as they wish when its returned the next day. QE created so many deposits that banks basically ran out of space to hold them, so Non-Banks are depositing the money at the Fed.
Banks don’t participate in the RRP (they receive interest on reserves at a rate higher than the RRP rate), but Banks must settle payments on-behalf of their Non-Bank clients. Remember, Banks and Non-Banks use different types of money. Non-Banks hold their money as deposits in a commercial bank (bank deposits) and Banks hold their money as deposits at the Fed (reserves).
When a Bank’s Non-Bank client sends an RRP payment to the Fed, the Bank must settle the payment on behalf of the Non-Bank. This payment must be settled in reserves, so the Bank deducts $100 in deposits from the Non-Bank’s account and sends over $100 in reserves to the Fed. RRP participation thus reduces the level of reserves held by the banking system. The banking sector as a whole shrinks – banks have fewer reserves in their Fed account and fewer deposit liabilities to Non-Banks. This can be helpful for some banks, as bigger balance sheets entail higher regulatory costs.
The RRP does not change the size of the Fed’s balance sheet, but its composition. The Fed expanded its balance sheet by creating reserves out of thin air to pay for QE purchases, and is now borrowing those reserves back via repo. In this example, the Fed goes from owing $100 in reserves to a Bank to owing $100 in repo loans to a money fund.
The wrinkle in this is that the Fed’s counterparty is no longer a Bank, but a Non-Bank (money fund). Generally speaking, only Banks are allowed to make deposits at the Fed. The RRP facility allows Non-Banks to also directly place money at the Fed in the form a repo loan. This sidesteps the banking system and gives Non-Banks direct access the Fed’s balance sheet. Note an overnight repo loan is not the same as a reserve, but in this instance they are functionally comparable as safe liquid assets.
The impact of these balance sheet changes depend in part on the motivation for the shift. At the end of the day Banks have less liquidity and Non-Banks have the same level but different composition of liquidity. If Non-Banks are attracted by higher RRP rates, then some banks would involuntarily lose liquidity and may be forced to raise deposit rates or borrow from the market. These actions would flow through the markets in the form of higher rates, with corresponding impacts on asset prices.
At the ZLB, the RRP offers basically the same return as leaving money in a checking account (1bps MMF share after management fees vs 0bps in a bank deposit). Some Non-Banks cannot take bank credit risk and must move money out of banks and into the RRP regardless of rates. But for others the move is also not rate driven, but because they were pushed out by their bank.
Banks operate under regulations that make it costly for them to accept too many deposits. At the same time, QE is pouring $120b of deposits into the banking system each month. As Banks run out of space to hold deposits, they are asking their clients to move money into a MMF. The MMFs in turn have received so much money that they have no where to place it but the RRP. Demand for money is low across all money market borrowers.
In this context the RRP is acting as a tool to manage the side effects of QE, with limited wider market impacts. In fact, higher RRP take-up may even be positive news. An economic recovery implies greater demand for private credit, and banks must make room on their balance sheets for those loans. In the absence of QT, higher RRP take-up could also be caused by increased bank lending.