This is the third installment examining the implications of the Fed’s experimental reverse repo facility. The issues surrounding this facility and its future role as a policy tool have only been heightened by the most recently released FOMC minutes. In fact some have argued that the associated repo rate may even replace the Federal Funds rate as the Fed’s main operating instrument. The objective here is to highlight some of the issues and risks that the Fed faces as it contemplates using this facility shrink its portfolio to a more normal size and extract the reserves injected into the financial system before they become the engine for accelerating inflation.
Briefly, the reverse repo transaction the Fed has been experimenting with is to temporally sell some of the assets in its portfolio overnight (or longer) under an agreement to repurchase those securities the next day. Technically, the Fed’s sale and agreement to repurchase securities from its portfolio is a repurchase transaction, but the convention at the open market desk is to refer to the transaction from the perspective of the Fed’s counterparties, so in this case the Fed’s repo is the counterparty’s reverse repo. For an excellent discussion of the Fed’s repo and related programs that also investigates the structural, competitive and other likely unintended consequences of the facilities see Michael Feroli (JPMorgan Chase, https://markets.jpmorgan.com/research/email/9ptonbem/GPS-1408971-0.pdf).
Putting the confusing terminology aside, the difference between the sales price and the slightly higher repurchase price is the overnight interest rate (repo rate). It is envisioned that the repo rate would lie within a corridor bounded by the interest rate the Fed pays on reserves (IOER) (right now the rate paid on excess reserves is the same as that paid on required reserves) and the discount rate. Were the repo rate to be set higher than the discount rate, this would set up an undesirable arbitrage possibility in which a bank could borrow from the window and simply hold the funds at the Fed earning a positive spread. But the corridor, including the repo rate would also have to high enough to overcome the forgone costs of deploying reserve funds in some other way.
There are several wrinkles to how the reverse repo transactions actually take place. First, the reverse repos are actually conducted in the tri-party repo market. The Federal Reserve and acceptable counter parties both maintain collateral and deposit accounts with the two clearing and settling instructions, JPMorgan Chase and Bank of New York Mellon, who operate the plumbing of the tri-party market. These two banks function as quasi-utilities that handle and value the collateral and make sure funds are returned on schedule as required. Providers of collateral and funds may establish collateral and deposit relationships with one or both of the institutions. Before recent revisions to the market’s operating procedures, these two institutions also provided trillions in intraday credit to market participants who needed temporary financing to hold securities that had been returned to them.
Prior to the financial crisis, the Fed transacted both repo and reverse repo transactions in this market, but only with primary dealers who were eligible to bid in the Fed’s daily open market transactions. If the Fed wanted to supply funds to the financial system, it would engage in a repo transaction with primary dealers who would sell eligible securities to the Fed in exchange for Federal Funds. Conversely, if the Fed wanted to extract reserves from the market, it would engage in a reverse repo transaction, selling securities overnight from its portfolio under an agreement to purchase them back the next day. Regardless of whether a repo/reverse repo transaction is just overnight or for a longer term, securities are returned each morning and settled. Even in the case of a term repo, the securities and funds are transferred back and forth each day.
Second, under the new facility, the Fed has expanded the eligible counterparties beyond just the primary dealers with whom open market repo transactions are conducted. These presently include some 18 banks, four Federal Home Loan Banks, Freddie Mac and Fannie Mae, 94 money market funds, and the 21 primary dealers. The complete list can be found on the Federal Reserve Bank of NY’s website along with current eligibility requirements. Except for the banks, the remaining counter parties are not eligible to hold interesting bearing reserve deposits directly with the Federal Reserve Bank of New York; however, the transactions being employed create a pseudo interesting bearing assets for the counter parties that are effective substitutes. Indeed, this is especially important to Freddie and Fannie. Before the establishment of the facility they collected funds from mortgage interest payments on loans that had been packaged into MBS that were then disbursed to securities holders monthly. Between the time the funds were collected and disbursed, Freddie and Fannie would lending them out in the Federal Funds market at rates below the interest rate being paid on reserves since they could not earn interest on deposits they were permitted to hold at the Fed. This practice explains in large part why the effective Federal Funds rate was below the rate the Fed was paying on reserves. The experimental facility effectively gets around that problem.
Third, a close look at the expanded counterparty list shows, eleven of the eighteen banks are U.S. affiliates or subsidiaries of foreign institutions. Six are government entities or are in conservatorship. Blackrock has three affiliates with 9 sponsored money market funds on the list. Similarly, Fidelity has 11 sponsored funds on the list, Goldman Sachs has four, JPMorgan has four, Morgan Stanley has five, and Wells Fargo has five. The point is that, of the 94 or so money market mutual funds now essentially having direct access to the Fed, they represent a much smaller group of sponsors, some of whom are primary dealers and a couple are foreign institutions. This access clearly puts these institutions and their sponsors in a privileged position and may bestow a competitive advantage relative to other institutions that could have unintended allocative effects among participants in the same markets.
What does this structure mean for the Fed in terms of its exit strategy? First, we must realize that regardless of whether the Fed engages in reverse repos overnight or for an extended term, it is simply a way of avoiding selling assets and recognizing a capital loss. Instead, the Fed pays the repo rate, which right now may be small, but should the system need to repo out a trillion dollars or so of securities to sterilize reserves, the interest payments will drastically reduce the net interest the Fed is receiving on its portfolio. For example, back of the envelope calculations as of 2013 suggest that the Fed earned about 2.3% on its portfolio and incurred .13% in interest costs. If a long term repo facility together with the rate paid on reserves were to increase significantly when rates rise, as they surely must, the Fed might even be thrown into an earnings loss position (not to mention a huge capital loss on its portfolio).
We don’t know what rate the Fed will have to pay or for how long. Because the present way the experimental repo rate is set as a markup over the interest rate paid on reserves, the possible cost issues the Fed faces involve both the interest rate paid on reserves as well as the repo rate markup. The level of both rates depend upon the elasticity of demand for short term liquid assets in the market place and the opportunity cost of deploying the funds elsewhere, such as making loans. Again, we have no idea what these elasticities are or how they might change if the economy takes off and market rates rise. Clearly, the equilibrium height of the corridor and the associated short term rates are much higher than it is presently.
Second, one of the rationales proponents put forward for this new facility is the fact that the Federal Funds rate has become decoupled from other market rates, probably because the Fed has pegged it so low for so long. But looking at the list of counterparties, it is hard to see how the repo rate will overcome this problem. A large portion of the excess reserves are being held by foreign institutions in their U.S. affiliates and subsidiaries. Roughly, foreign institutions have about 16-17% of bank assets but hold nearly 50 percent of the excess reserves. These institutions are not likely to be significant suppliers of funds to US businesses, especially small businesses.
One possible linkage may work through the commercial paper market. Because much of the Fed’s portfolio is in longer term MBS and Treasuries, the facility is essentially converting a long term asset into a short term asset. If the effective supply of short term assets accelerates because of the reverse repo transactions, this may drive up short term rates generally. More specifically because money market mutual funds typically play an important role in the short term asset-backed commercial paper market, they might find dealing with the Fed a superior, less risky way to deploy funds. This would thus divert funding from the commercial paper market and increase rates. At the same time, because of the declining budget deficit, the net new supply of longer term treasuries is shrinking, and this may bid up long term asset prices and put downward pressure on long term rates and/or flatten what might otherwise have been a steeper yield curve. It is even conceivable that should the volume of reverse repos become too large, the yield curve might even become inverted.
Third, with so many large foreign institutions as participants in the facility, this sets up an intra-firm arbitrage opportunity. They can repo out the securities obtained from the Fed, assuming hypothecation is permitted, either here in the US, to sovereign wealth funds or to foreign cash pools that have large demands for safe, liquid assets. The foreign institutions can use the funds obtain obtained in this way to purchase assets abroad on a highly leveraged basis. The possibilities expand as the term of the reverse repo transactions increase.
Finally, one concern about systemic risk is the scale of the interconnectedness among market participants. Expansion of the list of counterparties and the heightened importance of the tri-party repo market in the Fed’s operations strategy suggests that systemic risks may be enhanced. This facility, and its potential size, will certainly expand interconnectedness, since the facility is essentially a backdoor way of granting a wider range of non-banks access to the Fed. Should any of the Fed’s counterparties get into financial difficulty, especially the two banks that operate the plumbing of the tri-party repo market, the incentives to engage in bailouts will only be heightened, and this is especially the case with the money market mutual funds. A collapse in this market would eliminate the Fed’s key policy tool. We saw last time what happened when one money market fund broke the buck. Note that while Dodd-Frank Act may prohibit the Fed from bailing out a single firm, bailing out the whole tri-party repo market would not.
All of these concerns imply that even if the relatively small scale reverse repo facility is deemed a successful experiment that may tell us little about how the facility will work when the size of that market explodes, as it must if the Fed is to avoid shrinking its portfolio through asset sales. The alternative, as we have argued previously, is to simply raise reserve requirements now to sterilize most of the existing excess reserves. This, combined with a credible commitment to relax the requirements as the Fed’s portfolio shrinks consistent with a steady growth in the economy, has little current opportunity costs for either the Fed or the holders of excess reserves. The Fed can set up a system of differential reserve requirements under its existing authorities to avoid constraining small banks, or institutions with low levels of reserves. The Fed can continue payment of interest on reserves, and the cost would have to be lower than the costs associated with a large scale reverse repo facility. With such a system in place, the problem of managing the remaining excess reserves with either the Federal Funds rate and/or a much smaller reverse repo facility would make the whole exit problem much simpler and buy time to let the Fed’s portfolio run off as it matures.