The main surprise in the minutes of the FOMC’s May meeting, released last week, was the brief discussion of a staff presentation outlining a proposed approach to normalizing the Fed’s balance sheet when the Fed stops reinvesting maturing securities. Briefly, the proposal is to establish an announced, sequentially declining set of caps, or upper limits, on the amount of maturing Treasuries and agency securities that are permitted to run off each month. Any amount of maturing securities in excess of the caps would be reinvested; and as the caps are raised, the amount that needs to be reinvested will decline. The caps would be raised every three months from the pre-announced initial levels to their terminal values, which would be maintained until the size of the balance sheet is normalized. What is noteworthy about this general approach is how cautious, gradual, and transparent it would be, presumably to mitigate any market uncertainty and to avoid another so-called “taper tantrum” like the one that occurred in 2013, when the Fed hinted at a change in its policy.
The questions that remain are, how big is the problem and what reinvestment flows is the proposal designed to mitigate? The answers lie both in the composition of the Fed’s portfolio and in its maturity structure. As of May 24, 2017, the System Open Market Account (SOMA) held $4.2 trillion in assets, of which $2.3 trillion was in the form of Treasury notes and bonds, with most of the remainder being $1.8 trillion of agency MBS. However, the maturity structure of the assets is unbalanced. Most of the runoff in the rest of 2017 through 2023 consists of Treasury obligations of various kinds, amounting to about 70% of the SOMA’s Treasury holdings. Maturing Treasuries average about 1.1% per month, or about $27 billion in 2017-2018. Between 2024 and 2032, relatively few assets are scheduled to mature, but thereafter the maturing assets are main MBS and some long-term Treasuries.¹
Given the maturity details for the short run until 2023, a normal runoff rate would mean that, in total, about $541 billion of assets, mainly Treasuries, would run off by yearend 2018, that is, less than $30 billion per month; and a similar amount would mature each month in 2019 and 2020. This scenario suggests that the cap plan outlined in the FOMC’s May minutes envisions a very modest, gradual runoff of maturing securities, and it would take much longer than the end of 2020 under the plan to reduce the Fed’s holdings by even $1 trillion.
How long would it take for the normal runoff of the Fed’s security holdings to return the balance sheet to some semblance of equilibrium without the proposed cap program? Before the financial crisis, the Fed aimed to keep the currency-to-GDP ratio about constant, and that meant that the portfolio was about $850 billion. Because the economy has grown, the comparable number today is about $1.4 trillion. Assuming that both the economy and currency demand continue to grow at about 2% per year, then, as shown in the chart below, the normal runoff would not reach equilibrium until about 2038.²
But at that time, the Fed’s remaining assets would consist of about $515 billion in Treasuries and $1.5 trillion of agency MBS. Such a portfolio, as it evolved, would likely create a different set of issues if the Fed had to rely mainly upon MBS with which to conduct daily open-market operations.
All in all, the proposed plan for normalizing the Fed’s balance sheet appears to be extremely cautious and not nearly as aggressive as some have suggested it should be. The market impacts would likely be modest, and this outcome would be supported by a very transparent process. But the plan also suggests that returning to normal open-market operations and targeting the federal funds rate will not be the main tool by which the FOMC conducts policy. The burden falls, then, on the new corridor structure, which will likely be in place for some while.