The April FOMC minutes suggest the Fed is planning a new standing repo facility (“SRF”), which would be the Fed’s third SRF. The contours of the new SRF are still up in the air, but we can infer them from a stated desire to backstop Treasury repo and the coverage gaps of the other SRFs. The Fed currently operates each day a de facto SRF for primary dealers and another for foreign central banks. This leaves non-primary dealers and investment funds as the obvious candidates for a new facility, as the two are the remaining active participants in Treasury repo. In this post we describe what a SRF is, how the current SRFs operate, and suggest that a new SRF would have a very limited market impact.
What is a Standing Repo Facility?
An SRF allows eligible counterparties to borrow from the Fed against eligible collateral at a fixed rate. The bolded parameters are design choices that are guided by the intention of the facility. For example, a Fed SRF was first floated by two senior Fed economists as a tool to facilitate balance sheet normalization. The proposed SRF would have smoothed the transition into a lower reserves regime by making Treasuries and reserves more fungible – banks finding themselves low on reserves could use the SRF to borrow more reserves against Treasury collateral. In light of that goal, the proposed SRF would have been open to banks, accept Treasury collateral, and lend at slightly above market rates (act as a backstop).
Another way to think about the SRF is as a type of Discount Window. Historically, the Fed acts as lender of last resort to banks by offering them loans through the Discount Window against a very wide range of collateral (with appropriate haircuts) at a fixed rate (top of Fed target range, higher for weaker banks). An SRF is also a facility that offers loans, but to a wider set of counterparties and against a more restrictive list of eligible collateral. SRFs expand the Fed’s safety net, improving liquidity and controlling rates. The current SRFs perform this function by offering loans to primary dealers and foreign central banks against certain credit risk free collateral.
The Current Standing Repo Facilities
In September 2019 repo rates went from trading around 2% to as almost 10% intraday as dealers were scrambling for cash. The Fed was worried it was losing control over overnight rates and quickly began making repo loans in size to primary dealers against Treasuries/Agency MBS collateral at IOR. Those trades became daily and evolved into a ‘de facto’ SRF for primary dealers designed to put a soft ceiling on repo rates. The facility helped control repo rates during the March Covid shock by lending as much as $450b in into the repo market at rates around IOR.
The second SRF, the FIMA Repo Facility, was set up in late March 2020 in response to volatility in the cash Treasury market. At that time, Treasury market liquidity was under stress as many investors were desperate to liquidate their Treasuries for cash. The Fed sought to support the cash market by offering foreign central banks the option of obtaining cash by borrowing from the Fed with their Treasuries as collateral, instead of selling them outright. Foreign central banks hold around $3 trillion Treasuries in Fed custody, so the facility had potential to moderate selling pressures. However, the FIMA Repo Facility was never meaningfully used (max participation $1.4b), in part due to its punitively high offering rate of IOR + 25 basis points (market rates were below IOR).
The two current SRFs help backstop the Treasury repo market, but do not cover all repo market participants. Historically, the Fed works through primary dealers, who intermediate by borrowing from the Fed and then lending to other repo market participants. These market participants are predominantly investment funds and non-primary dealers. Last March there was concern that the transmission mechanism was not working well as primary dealers had some trouble intermediating, potentially due to regulatory constraints (leverage ratio) or simply out of prudent internal risk management policies amidst volatile markets.
An SRF that would bypass primary dealers and lend directly to non-primary dealers and investment funds would further strengthen the FOMC’s control over Treasury repo rates. It would be logical to include all non primary dealers, but more thought would be needed on the types of investment funds to include. Investment funds are a broad group, and include the opaque hedge fund community as well as SEC regulated public funds like mutual funds. [Note: FOMC minutes also suggest banks as a SRF counterparty, but in practice banks are not active borrowers in repo. They are also sitting on an ever increasing pile of liquidity.]
The New SRF Is Unnecessary
Today’s financial landscape is one of superabundant liquidity ($300b+ earning 0% in the ON RRP) and broadening access to that liquidity. Sponsored Repo, an innovation by the DTCC, now allows cash to move from cash providers to investment funds/dealers in a manner that both reduces balance sheet costs and counterparty risk. The binding constraint in a world of super abundant liquidity has been dealer intermediation capacity, which are like the pipes through which money flows. Sponsored Repo offers a way around those limits by moving all legs of a matched book repo transaction into a central clearing house, the FICC. This allows dealers to net down their balance sheet exposure, significantly increasing the potential level of intermediation.
Sponsored Repo volumes have declined from high levels last March along with overall repo market volumes, but the list of participants in the service continues to grow and is now around 1800. This amounts to a substantial broadening of the pipes through which the $4 trillion in money fund assets and $4 trillion in bank reserves can flow to cash borrowers. As long as QE continues, Treasury repo will remain zombified.
In this context, market participants will have all the Treasury repo they want for the foreseeable future. That means a new SRF would end up unused like the other two SRFs. But the new SRF would be an evolutionary step in a gradual but very clear trend in the Fed’s toolkit. Pre-Crisis the Fed was the lender of last resort to banks, then over the past decade it also became the lender of last resort to primary dealers, money market funds, foreign banks, foreign central banks, state and local governments, and even domestic corporations. Fed has also lent indirectly to mom and pop shops via the Mainstreet Lending Program. The trend seems to be direct access to Fed lending by everyone – perhaps that perfects the transmission of monetary policy. At this pace, the Fed may only be one crisis away from reaching that goal.
Note: Brent Donnelly of HSBC included me in his list of top Fed experts on FinTwit. Thank you Brent and thanks everyone for your voting for me. It takes a lot of time to write these posts, but your support makes it all worthwhile. I am grateful for your readership – thanks very much!
6 comments On Another Standing Repo Facility
Very interesting article. Question from a layperson: Is there a difference between the reserves the Fed prints out of thin air and credits the dealers in exchange for assets in a typical QE program, and the funds the Fed is loaning out in Repo loans to the dealers?
If the dealer is essentially acting as an intermediary between the Federal Reserve and a hedge fund, for example, is the dealer transferring the reserves directly to the hedge fund, or crediting it with cash based on the reserves it receives for the bonds? Another financial commentator speculated that these loans are operating on a different circuit, suggesting the source of the Fed’s loans was its own cash. Is this even possible, or is the Fed’s SRF just QE by another name?
In QE the Fed makes an outright purchase of Treasuries with the intention of putting downward pressure on rates. In Fed Repo the the dealer exchanges Treasuries for cash on a temporary basis. Dealer already owns the Treasury and the Fed allows the dealer to put forth Treasury collateral of any tenor. In that case there wouldn’t be a direct impact on rates. Fed repo transaction would unwind the next day, while QE purchases are held to maturity so reduce the stock of Treasuries available to the private sector.
With respect to the reserves operating on a different circuit – that has to do with our two tiered monetary system. See this post for details.
Hi Joseph, congrats!
I am wondering about the JPM chart of traditional repo vs sponsored repo. Why is traditional dealer intermediation balance sheet intensive? Isn’t the dealer finding a HF that needs cash and has a Treasury, and then after the repo, takes that Treasury and repos its out to the MMF? In that case, is it because the repo loan and reverse repo are an asset and liability on the dealer balance sheet perhaps? And if so, I still fail to see how the sponsored repo case is different. Why does trading with the FICC automatically reduce balance sheet exposure? Is it because the FICC on either leg of the repo is the same entity and thus you net it out? Technically it is the same gross though no?
Love it. Great stuff, clearly explained. Thank you for your work.
My question – who are the ‘primary dealers’? Is this just a separate subsidiary of the commercial banks itself, or are they completely different?
They are securities dealers with the privilege of trading directly with the Fed. Usually dealers affiliated with big banks. See here.
Hey Joseph, super basic question — what do non-primary dealers typically use repo loans for? I specify non-primary dealers because I assume primary dealers primarily use it to finance QE stuff these days.