Usefulness: this article attempts to explain the tensions observed in the US money market in September by placing them in a long-term context and shedding light on the decisions taken by the Federal Reserve governors. It distinguishes between traditional monetary policy tools (open market operations) and less conventional operations (QE). Finally, this article provides an overview of the options available to the Federal Reserve as it continues to normalize its balance sheet.
Summary:
- US money markets experienced a period of significant volatility in September 2019, raising fears that the Federal Reserve had lost control of its monetary policy target.
- This renewed tension is part of the normalization of the Federal Reserve’s balance sheet following the massive injections of liquidity during the financial crisis through various quantitative easing programs.
- The stress on money markets stems from the difficulty of estimating both a comfortable level of liquidity available to banks and its composition, which is affected by regulatory constraints.
- In this context, the injection of reserves, notably through purchases of US T-bills worth USD 60 billion per month, should enable the Federal Reserve to limit the risks associated with the normalization of its balance sheet by resuming its organic growth.
US money market rates experienced a period of volatility in September, marked by a significant rise in the overnight interbank rate (the effective Fed fund rate stood at 2.7%, far from its target of 2.00%) and the repo market rate (the market for repurchase agreements, which are used to finance various transactions and are crucial to the functioning of financial markets and the economy in general[1]), which reached levels close to 10% on September 17 before closing at 6.00%. These episodes of volatility prompted the US Federal Reserve to intervene, with its New York branch urgently injecting liquidity into the interbank market via overnight and forward auctions.
These tensions are in fact part of a broader context of normalization of the Federal Reserve’s balance sheet following its post-crisis expansion, and reflect the difficulty of estimating the level of reserves necessary for the proper functioning of the money market. Multiple regulatory, monetary, and financial factors justify maintaining the Federal Reserve’s balance sheet at a size well above its pre-crisis level (a universe of « ample reserves, » as described by Jay Powell).
In this context, the FOMC’s October announcements (purchases of US Treasury bills worth USD 60 billion per month, in principle until the end of the first half of 2020, and the continuation of temporary liquidity injection operations) mark the start of the resumption of organic growth in its balance sheet, in a very different perspective from that of previous quantitative easing programs.
1/ The normalization of the Federal Reserve’s balance sheet is in line with the normalization principles developed and regularly updated by the FOMC
In response to the 2008 financial crisis, the US Federal Reserve implemented several programs to inject liquidity and expand its balance sheet through the acquisition of Treasuries, US Treasury bills, mortgage-backed securities (MBS)[2], and securities guaranteed by Fannie Mae and Freddie Mac. The Fed thus conducted threelarge-scale asset purchase programs (LSAPs) between November 2008 and December 2014, as well as an operation known as Twist[3] (maturity extension program) aimed at increasing the average duration (i.e., limiting interest rate risk, or the sensitivity of interest rates to changes in bond prices) of its outstanding securities holdings. The total assets acquired under the three LSAPs thus reached a peak of USD 4,230 billion, including USD 2,300 billion in US government securities (excluding TIPS[4]).
More specifically, these quantitative easing (QE) programs were aimed in particular at reducing long-term interest rates[5], notably by extracting duration[6] from the market, i.e., interest rate risk. From a financial perspective, this extraction of duration reduces the risk for investors of holding long-term securities and, as a result, reduces one of the two traditional components of a long-term interest rate, namely the « term premium »[7]. The second component of long-term interest rates, the « short-term rate anticipation » component, remains firmly anchored and stable, notably through the signal effect that QE sends on the direction of monetary policy.
The aim here is to reinforce market expectations of a sustained accommodative monetary policy, as the central bank has no interest in raising its rates while it holds a very large stock of securities. Overall, econometric studies on the subject consider that the US Federal Reserve’s quantitative easing programs have led to a cumulative decline in the term premium of nearly 100 basis points in the United States.
In 2014, noting progress toward its growth and inflation objectives, as defined in its mandate, the Federal Open Market Committee (FOMC, the Federal Reserve’s decision-making body) decided to begin normalizing its monetary policy, both conventional and unconventional tools. To this end, the Federal Reserve published its » normalization plans and principles » in September 2014, providing for a gradual reduction in the size of its balance sheet by very gradually reducing the amount of reinvestments of securities held in the QE portfolio (i.e., SOMA) and gradually increasingthe interest rate on excess reserves (IOER), in other words, the key interest rate that corresponds to the floor of its interest rate corridor. This rate remunerates excess reserves, i.e., the central bank money placed with the Federal Reserve by US banks in addition to their required reserves.
These excess reserves, which are recorded as liabilities on the central bank’s balance sheet, peaked at almost USD 2.5 trillion in 2014-2015 as the Federal Reserve injected liquidity, before gradually declining when the Federal Reserve began to normalize its monetary policy (see Chart1 below). This concept, which is relatively new in monetary policy since there was no such surplus before the financial crisis, helps determine the level of money market rates and, in this case, the Fed funds rate. It is believed that the higher the level of excess reserves, the closer money market rates are to the lower bound of interest rates, specifically the IOER. As a result, the Federal Reserve operates within a so-called floor system, in which the high level of excess reserves compresses monetary rates (Fed funds, repo, etc.) to the level of the IOER.
Chart 1. Level of excess reserves (sources: Federal Reserve Bank of St. Louis)
2/ While the normalization of its policy has been interrupted in the context of Sino-American trade tensions, the Federal Reserve is faced with the question of the minimum level of reserves that can be considered « comfortable. »
Following the structural normalization of its balance sheet through caps on the reinvestment of securities held under its QE program, and faced with new risks to its inflation target given the economic downturn, the Fed decided to halt its policy of raising interest rates in order to protect the US economy from the risk of recession. It has therefore cut interest rates three times, bringing the IOER down to 1.8% from a high of 2.4% at the beginning of 2019. This has resulted in a decline in the effective Fed funds rate, which has gradually followed the IOER (see Chart 2 below).
Chart 2. IOER and effective Fed funds rate (sources: Federal Reserve Bank of St. Louis)
However, while the level of excess reserves seemed comfortable enough (USD 1,500 billion in March 2019) to keep effective Fed funds close to their floor, money market rates became more volatile again from April 2019 onwards. This volatility can be explained by various factors affecting excess reserves. In central bank accounting, excess reserves are affected by various autonomous factors that reduce them substantially.
The first is the growth in coins and banknotes in circulation in the real economy and on the liabilities side of the central bank’s balance sheet, which changes in line with US demand for money. Another factor, the change in the US Treasury’s account with the central bank, affects the composition of the balance sheet assets, with the US Treasury increasing or reducing its position with the Fed in order to finance its activities.
While excess reserves reached their lowest level since 2012 (but which, at the time, seemed comfortable, particularly compared to the levels reached in 2007, when much lower reserve levels did not cause any tensions in the money market), it appears that these two more temporary factors indirectly contributed to exacerbating volatility in the interbank market by reducing the amount of these reserves, starting in April 2019. Despite the fact that the Fed had announced in March its intention to resume asset growth by stopping the capping of reinvestments of maturing Treasuries, tensions began to emerge in the money market in October. The renewed volatility observed in September 2019 can thus be explained by the settlement of a large US Treasury auction (reducing excess reserves) and the simultaneous payment of part of corporate taxes (payment in reserves, from the US Treasury account to the Federal Reserve). This led to a sharp rise in money market rates, caused by the reduction in reserves and a surplus of securities.
More broadly, this raises questions about the Federal Reserve’s ability to estimate the comfortable level of reserves needed to keep money market rates close to their floor. This difficulty has raised fears that the Federal Reserve may lose control of its main operational target (the effective Fed funds rate, which is all the more important in the United States as its economy relies heavily on market financing).
To understand this difficulty in forecasting, other factors are highlighted, such as the composition of these excess reserves. It appears that these reserves are highly concentrated among the major US banks (GSIBs) and circulate little[8], making it more difficult for smaller banks to access these reserves and therefore to manage their liquidity, for example in the context of the LCR (Liquidity Coverage Ratio, which requires banks to have the necessary liquidity via safe and liquid assets to cope with 30 days of stress).
For example, a Federal Reserve survey shows that banks anticipated a comfortable level of USD 1,244 billion in January 2019. Another study by the Federal Reserve Bank of New York also estimates that the needs of the major US banks in the event of significant stress would be more than $930 billion, setting a floor on reserves for only a small number of banks. It therefore appears that the comfortable level of reserves would be substantially higher than initially anticipated, requiring intervention by the central bank to inject new reserves.
3/ The resumption ofopen market operationsaims to keep pace with organic growth in the monetary base
Faced with this renewed volatility and fears of losing control of its operational monetary policy target, the Federal Reserve intervened via its New York branch (in charge of market operations) by setting up auctions aimed at injecting reserves in exchange for collateral.
These temporaryopen market operations, which are very common in a conventional monetary policy framework, aim to inject cash into the financial system in order to provide banks with the liquidity they need for their operations and to comply with prudential constraints, thereby easing tensions. These are overnight operations, to limit the risks of intraday volatility, but also forward operations (over a period of two weeks, for example), to inject reserves over a longer period, particularly during periods of regulatory window dressing (e.g., at the end of the quarter).
In total, by mid-October, the Federal Reserve Bank of New York had injected more than USD 400 billion in temporary reserves – however, this is not a permanent or aggregate amount, as it takes into account the renewal of operations coming to maturity.
The October FOMC meeting decided to extend these temporary operations until at least the end of the year. In an accompanying publication, the governors also decided to permanently resume the expansion of the Fed’s balance sheet through a program to purchase USD 60 billion of Treasury T-bills per month until the end of the first quarter of 2020. The principle behind this permanent expansion is to inject sufficient reserves on a permanent basis to match the organic growth of the Federal Reserve’s balance sheet. Although the Federal Reserve Bank of New York is effectively purchasing securities, in this case US Treasury bills, this operation should be distinguished from previous QE programs.
The aim is not to reduce long-term rates via the duration extraction mechanism mentioned above, but rather to inject reserves into the financial system. Furthermore, the expected impact of these T-bill purchases on the yield curve (a steepening in the short term) is precisely the opposite of that expected from QE (a flattening in the long term). The FOMC’s communication on this subject is particularly clear, and the governors’ statements subsequently reinforced this clarification.
Figure 1. Stylized diagram of the expansion of the US Federal Reserve’s balance sheet
That said, other decisions could enable the Federal Reserve to avoid resorting to securities purchases to increase the size of its balance sheet. In the very short term, a clearer approach to the use of Treasuries (rather than primarily reserves) as part of banks’ liquidity cushions (LCR) could help to ease reserve holdings and promote liquidity in the repo market. Both assets are considered safe and high quality (HQLA), but according to the CEO of JP Morgan, there remains some doubt about how the Federal Reserve will treat them in their respective prudential frameworks.
In the medium term, the Federal Reserve Bank of St. Louis has proposed[12] introducing a permanent facility for exchanging Treasuries for reserves at a slightly punitive market rate, allowing US banks to hold Treasuries without fear of being unable to liquidate them against reserves or of seeing them suffer unfavorable prudential treatment. This proposal aims to reduce the comfortable amount of reserves for US banks by allowing them to quickly convert their Treasury holdings into reserves at a relatively low cost. This facility has many advantages, including reducing the comfortable amount of reserves, but also complementing the Federal Reserve’s corridor with an equivalent of the marginal lending facility, which exists, for example, within the Eurosystem. This standing lending facility would make it possible to form a rate corridor and thus move away from a single floor approach. Finally, this lending facility would reduce the amount of interest paid to banks under the IOER.
Conclusion
In short, the Federal Reserve appears to have succeeded in limiting tensions in the money market and regaining control of its operational monetary policy target. However, this episode illustrates how the financial situation has changed since before the crisis, with stronger-than-expected demand for reserves and imperfect substitutability between safe assets. Beyond the structural normalization of its balance sheet, this transition requires the Fed to fine-tune its management of reserve levels.
[1]De Fiore, Hoerova, Uhlig, The macroeconomic consequences of impaired money markets, 2019
[2]Securitized products based on a pool of real estate loans and partially guaranteed by the government agencies Fannie Mae and Freddie Mac, which were bailed out by the US federal government in 2008
[3]The aim is to increase the average duration of the portfolio of securities held under QE in order to extract more pronounced duration from the market; in other words, the Fed buys longer-maturity securities in order to reduce the overall interest rate risk in the market, thereby favoring risk-averse players.
[4]US federal government bonds indexed to US inflation.
[5]Bernanke, Monetary policy in a new era, 2017
[6]NB. There are other channels through which QE can be transmitted to the real economy, in particular the portfolio rebalancing channel, whereby the rise in the price of sovereign securities leads to the appreciation of other financial assets (bank bonds, corporate bonds, equities) via increased risk-taking by investors.
[7]This is the component of long-term rates that corresponds to the risk of holding a long-term rate rather than a series of short-term rates.
[8] For a more comprehensive approach to the concentration of excess reserves in the US vs. the fragmentation of excess liquidity in the euro area, see: B. Coeuré, A tale of two money markets: fragmentation or concentration, 2019.
[9]Federal Reserve, Senior Financial Official Survey, February 2019
[10]Liberty Street Economics, Stressed outflows and the supply of Central banks reserves, 2019
[11] These are periods during which credit institutions must meet their regulatory requirements and therefore adapt their exposures to regulatory constraints, which can generate volatility.
[12]Federal Reserve Bank of St. Louis, Why the Fed should create a Standing Repo Facility, 2019
