The FOMC’s Plan for Significantly Reducing the Federal Reserve’s Securities Holdings

Good afternoon. I thank Raphael Bostic for inviting me to participate in this panel focused on balance-sheet normalization. I have had the opportunity to attend and participate in this conference over the past few years, and it has never disappointed in bringing together interesting research and interesting policy views. This year’s conference is no exception.

In my brief prepared remarks, I will describe how the FOMC intends to significantly reduce the security holdings on its balance sheet. The views I present today will be my own and not necessarily those of the Federal Reserve System or of my colleagues on the Federal Open Market Committee.

Throughout the pandemic, in addition to the federal funds rate, the Fed has used its balance sheet as a policy tool, buying large volumes of Treasury securities and agency mortgage-backed securities. At nearly $9 trillion in assets, the Fed’s balance sheet is now about double the size it was before the pandemic. These asset purchases were an important policy response to the severe strains in financial markets seen early in the pandemic, and throughout the pandemic, they helped to support the economy in the midst of the unprecedented COVID shock. Markets are now functioning, a solid expansion is underway, and inflation is far above the Fed’s longer-run goal of 2 percent. So at its May meeting last week, the FOMC decided that it will begin reducing the size of its balance sheet in June. To lay the groundwork and to help prepare financial markets for this announcement, the FOMC, after its January meeting, released a set of guiding principles for the reduction, and the Committee gave more specifics about its plans in the minutes to its March meeting. The plans announced last week are consistent with the earlier communications.

Balance-sheet reduction will be done in a predictable manner, primarily by adjusting the reinvestment amounts of the principal payments the Fed receives on its assets. Starting in June, the Fed will allow up to $30 billion per month of Treasury securities and up to $17.5 billion of agency securities to run off the balance sheet. After three months, these caps will rise to $60 billion per month for Treasuries and $35 billion per month for agency securities. To the extent that maturing Treasury coupon securities are less than the monthly cap, Treasury bills will make up the rest of the runoff up to the cap.

The plan draws on our experience from October 2017 through August 2019 when the FOMC reduced the balance sheet of some of the assets that had been purchased in the wake of the Great Recession. But there are some important differences. Last time, runoff caps were initially set at $6 billion per month for Treasuries and $4 billion per month for agency securities, and the phase-in was very gradual, with caps increasing by $6 billion and $4 billion, respectively, every 3 months over the next 12 months, to $30 billion for Treasuries and $20 billion for agency securities. This time, the phase-in is only 3 months and the monthly runoff caps total $95 billion, almost twice the size as last time. Another difference is that, last time, balance-sheet reduction started almost two years after liftoff of the funds rate from zero; this time it is starting about 2-1/2 months after liftoff. I note, though, that even though the gap between liftoff and the start of balance-sheet reduction is much shorter this time, the level of the target range is not that different – only 25 basis points lower this time – because the funds rate rose very slowly last time.¹

These differences reflect the fact that the economy is in a very different place now than it was then. When the Fed began reductions in October 2017, balance-sheet assets had grown to about $4.5 trillion, or 22 percent of GDP; reserves were about $2.2 trillion, or 11 percent of GDP; the unemployment rate was about 4-1/4 percent; real output growth was near 2.7 percent; and inflation was still running slightly below 2 percent. Now, balance-sheet assets are about $9 trillion, or 37 percent of GDP; reserves have been averaging about $3.8 trillion, or about 15 percent of GDP; labor markets are very tight, with the unemployment rate at 3.6 percent; the economy grew 5.5 percent last year; and instead of being below our goal, PCE inflation is currently 6.6 percent, a 40-year high.

Let me finish by discussing two items that the plan announced last week did not address: asset sales and the size of the balance sheet when reductions will end.

The plan did not rule out asset sales. The FOMC has not discussed sales, but the minutes of the March meeting indicated that FOMC participants generally agreed that after balance-sheet reduction was well underway, it would be appropriate to consider sales of agency mortgage-backed securities. An important benefit of sales is that they would help to speed the return of our portfolio’s composition to being primarily Treasury securities, consistent with the FOMC’s stated desire to minimize the effect of the Fed’s balance-sheet holdings on the allocation of credit across economic sectors. One potential way to implement sales would be to sell agency securities up to the cap in any month in which principal payments were less than the cap. This is similar to our treatment of Treasuries. Another way to implement sales would be to set a monthly floor on reductions, which would be met first by principal payments received and then by sales. A potential drawback of sales is that, depending on the interest rate path, they could result in realized mark-to-market losses, which would lower the Fed’s remittances to the Treasury. Such losses would not entail any operational challenges for the Fed in setting monetary policy. However, they would pose communication challenges that would need to be appropriately addressed so that the public understands the benefits of returning the balance sheet to a more normal size and composition despite the losses.

Second, the plan did not indicate the size the balance sheet will be when the FOMC ends the reductions, but it did give some guidance. We are implementing monetary policy via an ample reserves operating regime in which reserve levels are ample enough that control over the federal funds rate and other short-term interest rates is executed primarily through setting the Fed’s administered rates and active management of the supply of reserves is not needed. The FOMC intends to slow and then stop the decline in balance-sheet assets when reserve balances are somewhat above the level it judges is consistent with ample reserves. Once runoff has stopped, reserve balances will likely continue to fall for a time, reflecting growth in other Fed liabilities, until the FOMC judges that they have reached the ample level. At that point, the FOMC will then manage its security holdings to maintain ample reserves over time. The ample level of reserves is uncertain. It will depend on the banking sector’s demand for reserves, as well as the distribution of that demand across institutions, which will evolve over time. During the last reduction episode, the FOMC ended runoff when reserve levels were about $1.5 trillion, or 7 percent of GDP. When stresses developed in short-term money markets in mid-September 2019, the FOMC judged that the level of reserves had fallen below the level consistent with ample reserves, and in October 2019, the FOMC began purchasing Treasury bills and conducting term and overnight repurchase agreement operations to maintain reserve balances at or above the level that prevailed in early September 2019.

This time, as the process to reduce the size of the balance sheet progresses, we will once again be monitoring developments in money markets to determine the appropriate level of reserves at which to end balance-sheet runoff, consistent with maintaining ample reserve balances over time.

This was a brief summary of the FOMC’s plan for significantly reducing the Federal Reserve’s balance-sheet assets. I look forward to hearing the other panelists’ remarks and participating in the discussion.