The week of September 16 saw the Federal Reserve Bank of New York inject funds into the repo market in response to an unusual spike in rates that was above the initial target range for the federal funds rate of 2.0–2.25%. (The target rate was lowered to 1.75–2.00% after the close of the FOMC’s meeting on Wednesday, September 18.)
The repo rate, which is usually closely linked to the federal funds rate, spiked to levels close to 10% on Monday that week, causing a squeeze on broker-dealers that finance their portfolios of Treasury and agency securities in the repo market.
On Tuesday, September 17, the Fed provided about $52 billion in funds to primary dealers (broker-dealers who regularly act as counterparties to the Fed in its open-market transactions), followed by $75 billion or more each subsequent day through September 25 and declining to $71 billion on September 27. In each case, the Fed offered to buy Treasuries and MBS overnight from primary dealers in return for funds.
It remains somewhat of a mystery as to why the sudden pressures have arisen in this market. Some observers point to the need for about $70 billion in corporate tax payments that had to be made, which drew down funds to some extent from money market mutual funds, and to the issuance by the Treasury that same day of about $50 billion in securities that had to be financed.
Both of these transactions ran through the primary dealers and arguably may explain some of the rate spike, or at least its timing. But to get a better understanding of what was going on and why the repo market was hit requires us to get into the weeds a bit in terms of the background and structure of that market and to understand who the main participants are before further exploring the possible causes of the problem.
The current repo transactions entered into by the Fed can be thought of as an extension of what was the Primary Dealer Credit Facility, which the Fed established in March 2008 in response to the financial crisis and problems in the tri-party repo market. The Facility was closed at the beginning of February 2010. As its name implies, the Primary Dealer Credit Facility was designed to support some of the important primary dealers, that is, specially designated counterparties who were authorized by the Federal Reserve Bank of New York to buy and sell securities with the Fed in connection with issuance and distribution of Treasury securities and the conduct of Federal Reserve daily open-market operations.
Later in October 2008, the Federal Reserve also created the Money Market Investor Funding Facility, designed to provide emergency liquidity to money market funds that were experiencing withdrawals and liquidity problems. That program was terminated on October 30, 2009. The repo transactions last week put the Fed once again in the position of being a significant lender to primary dealers. One way to think of the Facility is that it is similar to the discount window for banks, but is provided not to banks but to broker-dealers.
At its December 16, 2015 meeting, the FOMC authorized the New York Fed to supplement its normal open-market desk operations by initiating what it called Over Night Reserve Repos as a supplemental (but supposedly temporary) tool to control overnight money market rates. Authorized counterparties included certain banks with assets of more than $30 billion, GSEs, primary dealers, and money market mutual funds. This facility has been heavily used, until recently almost exclusively by money market funds, as the chart shows.
This history brings us to the last piece of institutional detail relevant to the current problems in the RRP market. First, in the normal operation of the Fed in the repo market, it sells securities overnight, and it calls that sale a reverse repo. The Fed contracts with the buyer (i.e. counterparty) to buy back the securities the next day, as is the case with an overnight repo. The Fed keeps the securities on its books, and on the liability side of its balance sheet it draws down the reserve account of the counterparty’s bank and increases an account that it calls “reverse repurchase” agreements. There is no impact on the size of either the Fed’s assets or its liabilities, just a change in liability composition.
The same is not true when, as was the case last week, the Fed purchased securities from the primary dealers to address their funding problem. In that case, the Fed bought securities overnight and recorded the transaction as an increase in “repurchase agreements” on the asset side of its balance sheet, and it also increased bank reserves of the primary dealers’ banks by the same amount on the liability side of its balance sheet. In this case, the transactions did increase the size of the Fed’s balance sheet.
Another element of the transaction needs to be clarified. Of the four categories of approved participants in the Fed’s repurchase program, only banks are permitted to hold deposits at the Fed and to receive interest on those funds. GSEs are permitted to hold deposits at the Fed, but they cannot receive interest on those deposits. Neither the primary dealers nor money market funds can hold deposits at the Fed, so if they need funds or are experiencing liquidity problems, they turn to the Fed through the repurchase market and transact through primary dealers.
Most of the primary dealers are broker-dealers and are affiliated with a bank or bank holding company. US law limits bank lending to subsidiaries or affiliates, essentially precluding a bank from borrowing funds at the discount window and lending those funds to its securities affiliate in times of stress. It also seems that banks are often reluctant to lend to the securities affiliates of other banks or bank holding companies. This may be one possible source of the recent stress in the market, especially if money market funds were liquidating securities holdings through the primary dealers.
Against this background, what do we now know about what might have happened to cause the spike in the repo rate? Simple supply and demand theory implies that when a rate such as the repo rate spikes, there can be three possible explanations: The demand for funds has suddenly increased; the supply of available funds has suddenly dried up; or a combination of both.
Some of the most insightful thinking about developments in the repo market over the past few months has been provided by Credit Suisse’s Zoltan Pozsar. He has been dissecting the evolution of the repo market for more than a year and was prescient in anticipating the kinds of problems that led to the Fed’s recent injection of funds. His analysis is detailed and nuanced, so that it is not possible to cover all that he details here.
What follows is a distillation of some of his key points and insights. His analysis suggests that supply and demand factors for both Treasuries and for short-term funding have been important in laying the groundwork, across numerous domestic and international short-term funding markets, for current pressures in the repo market.
In August, Pozsar noted that by the end of this year, the Treasury was on pace to issue more than $800 billion in new debt and to increase its deposits at the Fed by $200 billion, while at the same time, he suggested, primary dealers had a limited supply of funds to accommodate such an influx before being impacted by required leverage ratios.
Furthermore, Pozsar detailed a number of factors, all tending to reduce market demand for securities, that help explain what has been a steady rise in primary dealers’ inventories of securities from an estimated $75 billion in 2018 to almost $300 billion as of the second quarter of 2019. He argues that non-dealer demand for Treasuries has been dampened due to several factors, including shrinkage of off-shore corporate holdings of Treasuries and the inversion of the yield curve, making Treasuries less desirable investments.
Finally, it is also the case that foreign central bank usage of the Fed’s reverse repo facility has added to the volume of securities in the overnight market. Foreign central banks have been buying securities overnight from the Fed in return for a cash balance that shows up on the Fed’s balance sheet as a reverse repurchase agreement. Remember that such transactions reduce bank reserves (since foreign entities must transact through a bank with a reserve account) and increase recorded reverse repurchase agreements, thus acting to sterilize about $290 billion of reserves that might otherwise be available in the market. All told, it is estimated that total demand for US Treasuries has shrunk by about $800 billion since 2018, which has bloated dealer inventories and increased the demand for overnight funding by the primary dealers. At the same time, collateral supply is up by about $1 trillion that also needs to be financed.
Against this background, Pozsar also argues that there is in fact a hierarchy of both providers and users of both cash and securities in the repo market, depending upon whether or not they are members of the Fixed Income Clearing Corporation (FICC), who ultimately transact through the primary dealers. Cash lenders would include money market mutual funds and hedge funds that are not members of the FICC and that transact in the repo market at slightly different rates than do FICC member institutions such as banks, some GSEs, and branches of foreign banks.
Like the cash providers, there are also providers of collateral, some of which are members of the FICC and some not, and which consist of essentially the same types of institutions as the cash providers; and they too face different interest rates. The important point is that this hierarchy means that there is not one clearing rate in the repo market, but a sequence of rates depending upon whether FICC institutions and non-FICC members are trading with like member institutions or with non-member institutions. Shortages and surpluses of both cash and securities can exist in any of these markets, which can result in rate spikes and liquidity problems.
One last point. It is being reported that the Fed is considering increasing the size of its balance sheet so as to inject more reserves into the banking system as a way to deal with the problem in the ON RRP market. Presumably, this would endow banks with additional excess reserves, possibly making them more willing to lend to the primary dealers. Indeed, if the problems are in institutions that are not permitted to hold reserves at the Fed, then the private market must rely on banks’ willingness to lend reserves to address the needs for funds by either the primary dealers or money market funds.
The problem with proposals to increase the size of the balance sheet is that there is no guarantee that an increased amount of bank reserves would find their way into the various segments of the repo market where needed. Just increasing bank reserves does not guarantee that lending will take place. Moreover, there is little to guide the Fed on how much of an expansion of its balance sheet would be needed to ensure that the problem would not arise again.
One possible reform that would clearly address the problem would be to grant the primary dealers access to a permanent discount-window repo facility. A downside is that this would put the Fed in the position of lending to many subsidiaries of foreign institutions; plus, it doesn’t address the issue that the Fed’s foreign RP facility is uncapped and acts to reduce bank reserves as the facility expands. Another option would be to modify both its regular open market operations and the ON RRP facility so as to engage in continuous transactions during the day. This move would eliminate the need to try to guess at the beginning of the day what market funding needs might be by day’s end and ensure that rates were within the target range.
 The following discussion relies heavily upon institutional details and insights provided by James McAndrews, who was formerly head of the Open Market Desk at the Federal Reserve Bank of New York.
This commentary was written by Robert Eisenbeis, Vice Chairman & Chief Monetary Economist at Cumberland Advisors in Sarasota. Dr. Eisenbeis was formerly Executive Vice-President and Director of Research at the Federal Reserve Bank of Atlanta. This piece does not necessarily reflect the opinion of Hedgeye.