One thing I have noticed throughout my years of learning, both in college and now when I attend conferences is that if I do not understand something and have a question, odds are, that other people are confused on the same thing. For full transparency, how exactly a repurchase agreement (repo) works has been one of those points of confusion for me. It is for this reason that I want to make a post specifically about each leg of the transaction and thinking through the transaction from both the perspectives of collateral and cash. We will quite literally be busting out the crayons for this one because sometimes you really need a picture to understand something. That’s how I thought it through, literally drawing a picture of a step-by-step repo transaction on the back of my planner.
A point of confusion for me arises when these repo transactions are done with the Federal Reserve and a commercial bank as the entities. With these specific entities, the cash involved are bank reserves and the movement of these reserves plays with the money supply within the banking system. That being said, this post will be a case study on this specific repo transaction. There are multiple other types of repos that vary in the specific entities involved, maturity of the agreement, etc. However, this post will be written exclusively for an overnight repo transaction between the Federal Reserve and a commercial bank.
Above if a diagram of the bilateral repo transaction. Even though it consists of a minority of the repo transactions today, it is preferable for our current purpose as the concepts of where the cash and collateral are flowing, and its effects are same as triparty repo. Using bilateral repo reduces the moving parts and provides for an easier illustration.
Opening leg of repo
In the case of a regular repo transaction with the Fed, the opening leg involves the Fed buying a security from a commercial bank. This means that the Fed plays the role of the cash investor in the above diagram. To remain consistent with the numbers provided within the diagram, $9 million in bank reserves is transferred to the commercial bank and $10 million in US Treasuries are transferred to the Fed. We will pick US treasuries as the securities used as collateral for our thought experiment because they are considered pristine collateral as well as those are also the securities the Fed bought during their quantitative easing programs.
So, from a bird’s eye view of the commercial banking system, upon opening this repo transaction, bank reserves were injected into the system and collateral (the treasuries) were removed from the system. This means that the bank has more reserves while this transaction is open and can do with them as they wish. On the other hand, there are now fewer treasuries in the open market, meaning their supply has decreased, therefore their price may be bid up due to lack of supply which in turn suppresses their yield (interest rate). This also means that there is less collateral in the system for the bank to use for other transactions.
Closing leg of repo
Upon the closing of the regular repo transaction, the commercial bank repurchases the treasury with interest, as can be seen by the $9.09 million cash payment in the above diagram. That is an interest payment of $90,000, meaning that $90,000 in bank reserves on net balance flow OUT of the banking system upon the closing of the repo transaction. There, on closing the repo transaction, the money supply has decreased by the amount of interest paid by the commercial bank.
From the perspective of the collateral, the $10 million in US treasuries is transferred back to the commercial bank, thereby replacing the deficiency in the supply of treasuries. Since the amount of treasuries available for the open market returns to its original level once the repo is closed, the effect of this type of repo transaction on the treasury market may be minimal. Unless there is a consistent and ever-increasing amount of these transactions being open every day. In that case there would be a significant amount of treasuries removed from the open market on a daily basis, and the phenomenon described in the previous section could become significant.
Opening leg of reverse repo
As the name suggests, reverse repo is the same thing as a regular repo, just in reverse. Keeping our example consistent, the federal reserve sells a security with a promise to buy it back in the future at a higher price. When the reverse repo transaction is opened, the Fed sells the security to a commercial bank, this means that the supply of that security increases in the open market because the bank will use it as collateral for other transactions. This is called rehypothecation. By rehypothecating these securities, the bank can make a much higher return than the interest rate on the U.S. treasuries and the ON RRP rate.
Also, the commercial bank bought these securities from the fed with bank reserves denominated in dollars, which means that bank reserves left the system while the reverse repo is open. However, this is typically just overnight as the majority of repo transactions mature the following day. Therefore, the decrease in bank reserves is temporary for as long as the amount of bank reserves used for reverse repo is relatively constant.
Closing leg of reverse repo
Upon the maturity of a reverse repo transaction with our given case, the Fed buys back the security at the higher repurchase price. Therefore, more bank reserves are being injected into the system than were removed in the opening of the reverse repo, thus creating a net increase in the amount of bank reserves in the banking system. This also means that the security that the commercial bank owned in between the opening and closing legs of the reverse repo, that they probably rehypothecated for other transactions, now goes back to the Fed.
Final words
Bear in mind that this post was written from the perspective of just one example of a repo transaction, specifically with the parties being the federal reserve and an entity that holds a reverse account at the Fed (i.e., commercial bank). In this case, the cash that is transferred would be bank reserves denominated in dollars, however, if a non-fed entity who does not have a reserve account at the fed takes the place of the commercial bank, then the cash would have to be the typically dollar-denominated bank liabilities that are so familiar to us regular people.
Another note worth mentioning is since the reserve requirement for banks is 0%, meaning they no longer need to hold bank reserves, it is arguable on how significant of an effect changing their amount would have on the monetary system. Also, commercial banks have so many bank reserves as it is, any change in their amount would essentially be negligible. At least according to the Fed, as they are currently operating in an “ample-reserves regime”. For more clarification on what exactly it means to have an ample amount of bank reserves and the different monetary policy regimes, (CLICK HERE).
To your wealth and future,
James Forsythe
For more on repo and other lending markets
https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr740.pdf
https://jamesdforsythe.com/explaining-the-repo-market-powerful-yet-fragile/
