What is the Fed's RRP and What Does It Do?

In the time before the GFC the US Dollar banking system functioned with far less banking reserves than it does today. For example, as of today there are approximately $3.2 Trillion in banking reserves. Back in 2003 on the other hand, the amount of banking reserves was $43.6 Billion. 

This incredible expansion in reserves can be explained by none other than Quantitative Easing, i.e. the process by which the Federal Reserve creates new bank reserves in order to pay for debt securities like US Treasury bonds and Mortgage-Backed Securities (MBS). QE lowered the cost of borrowing for the US government, corporations, and consumers by putting a floor under long duration debt. 

This caused a major change in the Fed’s methods. You see, back when bank reserves were in the Billions and not Trillions, the Fed controlled short term interest rates by buying or selling Treasury securities to the banks. If the Fed wanted Fed Funds to rise, it would sell securities, making bank reserves more scarce and driving up the rate. On the other hand if it wanted rates to fall, it would buy securities from the banks, increasing bank reserves and lowering Fed Funds. 

Now that there are Trillions in liquidity, the Fed can no longer control front-end rates in the same way; there are simply too many reserves and the Fed would have to buy or sell huge quantities of securities to influence short term rates directly, which would no doubt cause the very dislocations in the funding markets the Fed is trying to avoid. 

In order to put a floor under interest rates, the Fed decided to start paying interest on bank reserves. Before this bank reserves were interest free loans to the Fed. Now that the Fed was paying interest however, no bank would make an unsecured Fed Funds loan for less than they could get, risk free, from keeping reserves in their own account at the Fed. This put a floor under Fed Funds as the Fed began the “lift-off” from ZIRP. 

However, it turns out that this wasn't enough to make sure that non-banks didn’t lend out for less than the Fed’s target range. In particular money market funds, who could not access accounts directly at the Fed, were facing a crisis. Interest rates were so low that they risked going negative. In order to avoid this the Fed decided to let MMF’s access interest bearing accounts directly at the Fed through what is called the RRP facility.

The Reverse Repurchase facility is a daily window where non-banks e.g. money market funds can stash their cash overnight at the Fed. While the mechanics are slightly different under the hood, you can think of the RRP as a bank account at the Federal Reserve for non-banks. 

This makes the RRP pretty special, as it occupies the same place in the hierarchy of money right next to bank reserves. Funds locked in the RRP overnight provide the same systemic buffer as reserves; in the context of the Fed’s current ample reserve framework, it’s appropriate to view the total amount of effective reserves in the banking system as equal to reported RRP + Reserve Balances.   

This abstraction, that the RRP is simply a bank account for non-banks at the Fed, of course, hide’s some complexity. The actual mechanics might look like:

• MMF has money at private bank e.g. JPM or BNY (i.e. the money is in Banking Reserves on the Fed’s liability side of balance sheet)

• MMF wants to earn overnight interest rate on their cash

• MMF subscribes to Fed’s RRP facility:

  • Cash goes from private bank (reserves) to the RRP account

  • Fed sells collateral from SOMA to the MMF. This collateral is in the form of treasury bills, notes, etc picked from the Fed’s extensive portfolio of assets 

  • Fed buys back collateral from MMF’s next day at predetermined price, the difference being the interest rate

Ok so now that we get what the RRP is doing, what function does it play in the economy? What makes it move one way or the other? Does it matter whether it's high or low?

We said earlier that RRP functions like reserves, so let’s start first by reviewing the function of banking reserves in general. Banking reserves can be thought of as private banks’ bank accounts with the Fed. 

We all know that banks create money: whenever a bank makes a loan they literally create money out of thin air using the magic of their balance sheet. There is a limit to how much money they can safely create, however, and that limit is governed by the size of their reserves at the Fed. 

When someone sends money from a bank account at Bank A to Bank B, this causes a ripple throughout the hierarchy of money. When Alice sends Bob $100 from her account at Bank A to Bob’s at Bank B, that causes a chain reaction in multiple balance sheets. The “choke point” in the transaction occurs within the Fed’s balance sheet when Bank A has to actually transfer $100 in banking reserves to Bank B. If Bank A lacks sufficient reserves to make the transfer it must borrow them from a bank that has excess reserves. This is the whole point of the Fed Funds market. If Bank A lacks the reserves and cannot borrow them for whatever reason, then the system breaks and some form of bail-out has to occur. 

Since effective reserves are really reported bank reserves + RRP, that implies that the size of the RRP facility is also a limiting factor on the amount of money that the system can create on the edge of the hierarchy. Money can flow freely from reported reserves to RRP and back again depending on how many MMF’s are participating in the facility. 

Between March 2021 and December 2022, the amount of money parked in RRP went from basically nothing to over 2.5 Trillion USD. As the Fed began lift-off and the yield curve went inverted, the best place for MMFs to park short term cash became the RRP. Why take risk and earn less than you can make risk-free at the Fed?

However, since the Fed began destroying money via QT the amount of money in the RRP has dropped to half of its peak (see chart). We can also see that bank reserves have been largely stable in 2023. The effect of QT’s money destruction and the US government replenishing the TGA after the debt ceiling debacle has been born almost entirely by MMFs. This shows up as a significant decline in RRP take-up: a drop of almost one trillion USD since May 2023.   

Could the RRP ever go back to zero take-up like before 2021? If the Fed sticks to their guns and/or inflation doesn’t come back down, then yes it's very possible that RRP take-up could dwindle back to nothing. This would cause front-end rates to spike as zero take-up means MMFs are getting a better yield elsewhere. It's also possible that a lack of RRP + reserves could lead to a funding crisis e.g. like the repo crisis of 2019. In response to the repo crisis, the Fed has a “standing repo facility” which does the opposite of the RRP: it lends cash in return for securities. Any take up of the SRF would be a powerful signal that reserves are no longer “ample” and could force the Fed to restart QE.