Thursday’s WSJ cited a draft paper by Brian Sack (former manager of the FOMC’s open market desk at the NY Fed) and Joseph Gagnon (a former senior officer at the Board of Governors). The paper discusses the open market desk’s experimental reverse repo program as one of several tools for managing the exit from the FOMC’s monetary accommodation. The authors kindly provided a draft of their paper, which offers a very useful discussion of alternative monetary policy instruments, how they have been developed to deal with the unique circumstances of the financial crisis, and the Fed’s accommodative monetary policy. It also provides an interesting and thoughtful policy proposal for using the reverse repo program and the associated repo rate as the substitute for targeting the federal funds rate going forward.
First, a brief primer on open market operations. Historically, as Gagnon and Sack (G&S) point out, the main liabilities on the Fed’s balance sheet, until the financial crisis, were currency outstanding and bank reserves, the latter being composed of required and excess reserves. The reserve components were very small. Right before the onset of the financial crisis they were less than $ 15 billion.
Putting aside the complex structure of how banks must meet their required reserves, banks’ deposits with the Fed fluctuate day-to-day depending upon how checks are cleared, the flows of government payments, Treasury refinancing needs, and transfers to and from Treasury tax and loan accounts at commercial banks, just to name a few factors. The open market desk engages in two types of transactions with primary dealers (a small number of banks and dealers authorized to deal directly with the desk). There are so-called permanent transactions designed to permanently add to or decrease the volume of bank reserves outstanding and geared to keeping the currency to GDP ratio relatively constant. There are also temporary transactions designed to dampen daily volatility in the amount of outstanding reserves in response to anticipated daily fluctuations in those reserves.
The size of these daily programs is based upon coordinated estimates by the desk, the Treasury, and the staff of the Division of Monetary Affairs at the Board of Governors of much funding needs to be added or subtracted from the system to keep the federal funds rate at the desired target set by the FOMC. The size of the proposed transaction is set and put out for auction at a specific time each morning, and bids close in about 10 minutes. Temporary transactions are normally conducted using repurchase agreements, where a dealer sells Treasury securities to the desk with the agreement to buy them back the next day (or whatever the agreed upon tenor is), and reverse repurchase agreements where the dealer buys Treasury assets in return for a deposit at the Fed (fed funds) with the agreement to sell them back to the Fed. Thus, from the dealer’s perspective repurchase agreements add to the volume of reserves outstanding on a daily basis while reverse repurchase agreements subtract reserves.
When the desk guesses right, the amount of funds supplied will keep the effective federal funds rate close to the target. Note that through the use of repurchase agreements and the associated repo rate, the supply of federal funds is either increased or decreased, thereby affecting the federal funds rate. Historically, this mechanism has also provided a close link between the overnight repo rate and the federal funds rate. Note that the federal funds rate is not constant but rather fluctuates during the day, because the desk only makes one transaction a day. If the desk continuously bought or sold federal funds, then the going rate would be the target rate.
Institutions borrow and sell federal funds to meet liquidity and other demands for funds on an intraday basis and overnight. Alfonso, Entz, and LeSueur (https://libertystreeteconomics.newyorkfed.org/2013/12/whos-lending-in-the-fed-funds-market.html) have estimated that the size of the fed funds market fell from over $200 billion prior to the crisis to a little over $60 billion at the end of 2012. G&S point to the fact that the historical spread between the federal funds rate and repo rate widened significantly (by as much as 15 basis points or more) during the financial crisis. They attribute this widening to several factors, including the introduction of payment of interest on reserves, which lowered the cost of keeping deposits at the Fed idle rather than lending them out in the repo market. An additional factor was that large volumes of funds were owned by institutions like the GSEs, money market mutual funds, hedge funds, and large international cash pools that were not eligible to hold deposits at the Fed and earn the 25 basis points interest on reserves, so they would lend them out in the funds market at rates below the rate that banks could earn on deposits held at the Fed. The resulting arbitrage enabled banks to borrow from such entities at rates below the rate the Fed pays on reserves and earn the risk-free rate by simply holding those borrowed funds at the Fed. Finally, there was also the fact that large US banks were required to pay a fee, anywhere from 10 to 15 basis points, based upon their size to the FDIC to compensate for the risks they posed to the deposit insurance fund. The fees provided an incentive for those institutions to reduce their holdings of balances at the Fed and not borrow in the federal funds market. One consequence of the FDIC requirement is that foreign banks, which are not subject to the FDIC fee, now account for about 42% of the volume of federal funds borrowing.
The combination of all these factors acted to drive a wedge between the repo rate and funds rate, and also to drive more and more transactions into the repo market. G&S go on to argue that the gap between the repo rate and funds rate suggests that the links between the two may have weakened, thus reducing the ability of the FOMC to affect other open-market and short-term rates via its funds rate target. (The evidence for this can be found in Bech, Klee, and Stebunovs, https://www.federalreserve.gov/pubs/feds/2012/201221/201221pap.pdf). Bech, Klee, and Stebunovs, however, also suggest that it is possible that the weakening of links between the funds rate and repo rate might only be temporary and could be restored once more normal times return.
Regardless, G&S make the interesting argument that the FOMC and open market desk, in order to keep up with the changing structure of the short-term money markets, should now target the repo rate rather than the federal funds rate, since repos are a much larger market and hence the rate in that market is more closely linked to other rates along the term structure than the federal funds rate is. In large part this is due to the growth of the non-banking sector, hedge funds, money market mutual funds, cash pools, and other segments of the shadow banking system, as well as the dominant role now played by foreign banks in the federal funds market.
The first step in such a change would be to expand, as the Fed has been doing through its experimental reverse repo program, the range of participants with whom it could and should conduct repurchase and reverse repurchase agreements. The current program is very small, but the range of institutions now eligible to participate is not only interesting but also raises substantial questions about who has access to the Fed and any implicit and explicit subsidies that access may entail. For example, the list of participants includes six US banks and eleven US offices of foreign banking institutions, four Federal Home Loan Banks, Freddie Mac and Fannie Mae, and nearly 100 money market mutual funds. This last group is sponsored by some 25 different managers, including about 6 sponsors who are banks and are themselves eligible to participate in the program directly. The list also includes three different Blackrock LLCs. Blackrock has an unusually close relationship with the NY Fed as manager of transactions related to Maiden Lane and AIG assets the NY Fed acquired during the crisis.
While the G&S proposal may make some sense when it comes to implementation of monetary policy, the reverse repo arrangements contemplated also provide back-door access by non-banks (both foreign and domestic) to the explicit and implicit subsidies and interest bearing liabilities that are not permitted directly to these entities and are perhaps even contrary to permissible arrangements contemplated by the Dodd-Frank Act and other legislation. Furthermore, because most of the repo transactions are conducted through the tri-party repo market that is operated by two major US banking institutions, the systemic interrelationships are significant and complex and would only be promoted by the recommended change in the target rate. The systemic web of those relationships is further complicated by the practice of re-hypothecation, or reuse of collateral pledged in support of a repo transaction by the bank or dealer who accepted the collateral. In short, there are substantially deeper issues behind the Fed’s experimental reverse repo program than the simple creation of another tool for the implementation of monetary policy. The issues point first to the need to address the regulatory structure, fees, and other impediments that may have created incentives for transactions to move from the federal funds market to the repo market before accepting the market’s existence as a fact to be accommodated.