The market to the Fed: Build the corridor!
admin | September 20, 2019
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors.
If nothing else, recent trading sessions highlighted the importance of the repo market to the Fed’s ability to run monetary policy. The Fed almost surely will have to build a mechanism to put an upper and lower bound around repo rates. It already has a corridor around rates on bank reserves, with IOER setting a lower bound and the discount rate at the Fed window setting an upper. With more than $4 trillion now running through the repo market, the Fed probably needs to complete a repo corridor as part of its full policy toolkit.
A Fed approach to the repo market will have to fit into a set of policy tools that have come a long way since the 2008 financial crisis. Prior to the crisis, bank reserves received 0% interest. Banks consequently spent considerable time and effort keeping their reserve levels as low as possible. Banks had to meet reserve minimums, but they held few reserves beyond that. In the months leading up to the crisis in 2008, required reserves amounted to a little over $40 billion and excess reserves were a mere $2 billion. Note that is billion with a B, not trillion!
The mechanics of the Fed hitting its funds rate target involved making an estimate of the needed amount of liquidity in the fed funds and reserves market each day and then conducting a repo operation to add liquidity or a matched sale operation to drain liquidity to get to the right number. Given that there were only a few billion extra reserves in the system on any given day, if the Fed miscalculated and injected too few or too many reserves, the funds rate would likely diverge from target.
Interest on reserves
The Financial Services Regulatory Relief Act of 2006 originally granted the Fed the authority to pay interest on reserves, but the effective date for that change was set at October 2011. When the crisis hit in 2008, Congress gave the Fed authority to do so as of October 2008. This ushered in a sea change for the way the Fed managed reserves in the banking system.
Even in the absence of the crisis, the mechanics of the Fed’s operating regime were going to change once the central bank was granted the authority to pay interest on reserves. Suddenly, excess bank reserves would be a competing investment for banks that held more liquidity than they needed.
Economists had long discussed what such a regime might look like. While the zero bound and QE complicated the transition, the general idea is that the Fed would hit its fed funds rate target using a corridor system. In effect, rather than having to hit an exact target for the quantity of reserves, the Fed could keep the funds rate close to its target by controlling the price of reserves. In effect, the Fed would stand as a lender or borrower of last resort for reserves. The discount window would be the mechanism for meeting any shortfall in reserves, while the Fed would pay interest on excess reserves, or IOER, for any extra reserves that banks were anxious to dump.
Note that for the corridor system to work, the discount rate would need to be at a higher level than market-driven money market rates, so that the Fed would only be the lender of last resort. Conversely, the IOER rate would need to be at the bottom of the money market rate complex to make the Fed the “absorber” of last resort for reserves.
Financial crisis changes the game
The financial crisis changed the structure of the money markets in four important ways. First, the Fed was forced to hit the effective lower bound for the federal funds rate target. As a result, rather than trying to calibrate to a precise rate target, the Fed was just attempting to keep the funds rate a little above zero in an attempt to prevent the entire money market fund industry from imploding. As part of that effort, rather than taking it all the way to zero, the FOMC decided to set the IOER rate at 25 bp— the top of the new target range for the funds rate. In essence, this made the Fed the best option for banks’ excess reserves rather than a backstop.
The second innovation was the Fed’s expansion of the balance sheet by implementing multiple rounds of QE. This grew the balance sheet by trillions of dollars cumulatively. As a result, excess reserves exploded to a peak of over $2.5 trillion. In that world, the Fed wanted the IOER rate to be at the high end of the funds rate target range because the Fed was looking to absorb all of that excess liquidity to avoid creating financial instability.
The third change fell out from the second. With trillions of dollars in excess liquidity at any given point in time, the federal funds market withered from one of the most important pieces of the money market universe to a relative backwater piece of the complex. In fact, the only reason that there was any trading in the fed funds market at all in those years was that not all financial entities—notably Fannie Mae and Freddie Mac—had access to the IOER rate.
Finally, the fourth major change is that Dodd-Frank and other regulatory changes since the crisis have forced banks to hold large amounts of high-quality liquid assets, or HQLA, to meet the liquidity and capital requirements established to prevent a repeat of the 2008 debacle. HQLA included reserves. As a result, the Fed has determined that the banking system will want to hold a large amount of excess reserves to meet its regulatory obligations and liquidity buffers. This is the main reason that the FOMC opted not to go back to the pre-crisis operating regime, as it knew that there would be a need for leaving ample reserves in the banking system. Instead, the Fed would settle on a new operating system that included a target range for the funds rate, a corridor rate system to enforce that range, and a significant amount of excess reserves, though the exact amount was less important than before the crisis, as long as the liquidity provided is sufficient to meet the needs of the system.
The long transition to normal
It has taken years to approach the new normal. The Fed finally raised rates away from the effective lower bound in late 2015 but only moved substantially away from zero in 2017. In late 2017, the FOMC also decided to begin shrinking its balance sheet, gradually draining off some excess reserves. The stock of excess reserves has been cut roughly in half from its peak to just over $1.3 trillion.
It has done so very gradually, but the Fed has been shifting the IOER rate down toward the bottom of the funds rate target range. The first 5 bp tweak was implemented in June 2018, followed by similar moves in December 2018, May 2019, and this week. The IOER rate is now just 5 bp above the bottom of the quarter-point funds rate target range, and I suspect that there may be one more tweak forthcoming at some point in coming months to finish the job.
At that point, there would be a corridor around the market for reserves, with the discount rate set at 50 bp above the top of the target rate for fed funds and the IOER rate at or close to the bottom of the range.
However, the evolution of the money market universe requires a second corridor. A key part of the construction of a robust post-crisis operating regime has been for the Fed to build up a formal interface with the repo market. As noted above, the fed funds rate market is now a relatively minor element of the money market complex today, even though the Fed has chosen to continue its use as the primary vehicle for the policy rate target. However, the funds rate market is largely driven by the far larger and more important repo market, where financial players borrow and lend securities for cash.
The Fed initially learned this lesson during the years after the crisis, when it was having trouble keeping the funds rate above zero because the massive amount of excess liquidity in the system was putting downward pressure on money market rates. Treasury bills traded at negative rates periodically in those years. The Fed addressed this problem by introducing a reverse RP facility to sop up some of that excess liquidity from financial firms, such as money market funds, that could not access the IOER rate. For years, the reverse RP operations absorbed huge amounts of liquidity, but as the Fed has shrunk the balance sheet, the take-up for this daily operation is now typically just a few billion dollars each day.
This week, for the first time since the crisis, the Fed was forced to conduct repo operations to add cash into the financial system. As soon as repo rates spiked early in the week, the funds rate traded at the top of the target range and ultimately above it, so the Fed had to provide cash to enforce its target. Chairman Powell indicated at his press conference on Wednesday that for the foreseeable future, the Fed thinks that ad hoc repo operations along the lines of what we have seen this week should be sufficient to manage the repo rate in order to keep the funds rate near the middle of the target range.
However, there is broad consensus on Wall Street that the Fed will ultimately need to add a standing repo facility to its arsenal, providing a bookend to the reverse RP mechanism. This would complete the double corridor system and provide the Fed with a full array of tools to enforce upper and lower bounds around both the cost of reserves and the larger and more important repo market.
Just to summarize, the Fed’s to-do list to get to where it ultimately needs to be with its operating regime entails some combination of the following three items:
- IOER rate. There should probably be one more 5 bp tweak to the IOER rate, bringing it in line with the bottom of the funds rate target range, though this step may not prove necessary, depending on how money markets behave.
- A standing repo facility. In theory, the Fed could conduct repo operations on an ad hoc basis, as it did this week, to address occasional upward spikes in repo rates, but a robust operating regime would include a permanent facility that was available every day, even when the take-up is minimal.
- Balance sheet size. The Fed stopped shrinking the balance sheet as of August. At this time, the size of the balance sheet is constant, while the composition of assets is gradually shifting, by up to $20 billion per month, from MBS to Treasuries. Over time, as the size of the economy grows, the amount of currency in circulation increases roughly in line. With a constant balance sheet, to the extent that the Fed is satisfying a growing demand for cash, excess reserves, the other main piece of the liability side of the Fed’s balance sheet, will decline gradually. In recent years, currency in circulation has risen by about $100 billion per year, so excess reserves would be falling at that pace.
At some point, the Fed will come to the conclusion that it has cut the level of excess reserves to the right number, the level that satisfies banks’ demand for reserves considering regulatory requirements. At that time, the Fed will begin to allow the balance sheet to expand again by doing Treasury coupon and bill passes, as it did regularly before the Crisis. Given that the Fed is already buying Treasury securities to replace its MBS holdings as they roll off, this would be a modest change, adding less than $10 billion per month to the current buying program that began last month.
Some analysts have speculated that the Fed may have already reached that point, as evidenced by the repo spike this week (which suggests tight liquidity). Fed officials were clearly not quite ready to make that determination this week, but declaring that reserves have reached the “right” number and the subsequent adjustment to the Treasury buying program will be one of the final steps to achieving the “new normal” for the post-crisis operating regime.