Well, I’m by no means sure that anyone cares on a lazy, post-holiday-weekend Monday, but Stanley Fischer was out this afternoon droppin’ knowledge in a speech at Columbia.
If anything, his prepared remarks were notable for the taper tantrum references. Specifically, Stanley is pretty pleased with how stocks have held up in the face of rampant speculation about the pace of Fed balance sheet normalization.
“My tentative conclusion from market responses to the limited amount of discussion of the process of reducing the size of our balance sheet that has taken place so far is that we appear less likely to face major market disturbances now than we did in the case of the taper tantrum,” Fischer said.
That’s pretty amusing. I’m not sure what he’s been listening to, because from where I’m sitting the “amount of discussion” with regard to “the process of reducing” the balance sheet has been anything but “limited.”
Anyway, here are the notable bullets from Bloomberg:
- Fischer doesn’t comment about health of U.S. economy in his prepared remarks
- Fischer contrasts the market reaction to the so-called taper tantrum of mid-2013 with the more recent response to the Fed’s indication that it may start shrinking its balance sheet later this year
- “My tentative conclusion from market responses to the limited amount of discussion of the process of reducing the size of our balance sheet that has taken place so far is that we appear less likely to face major market disturbances now than we did in the case of the taper tantrum”
- “‘As we continue to discuss and eventually implement policies to reduce our balance sheet, we will have to continue to monitor market developments and expectations carefully”
- There are times when the Fed could be too predictable and when shocks hit the economy, it would be natural for a policy maker’s uncertainty about the path of rates to rise: Fischer
Happily, this does afford me an opportunity to highlight a few passages from a recent Goldman note I’ve been saving. Below find some useful color and visuals that may (or may not) interest those who have been following the “limited” discussion of SOMA roll off.
Q: Why is the FOMC planning to shrink its balance sheet?
The Fed’s balance sheet has grown from less than $900bn before the financial crisis to $4.5tn today, from 6% to over 23% of GDP. Currently, the Fed reinvests all principal payments from its Treasury, agency debt, and agency MBS portfolios, holding the nominal size of its securities portfolio unchanged. Reducing the size of the balance sheet is a natural step in removing accommodation from an economy that has now largely recovered. In addition, New York President William Dudley recently noted two additional reasons for shrinking the balance sheet: it would reduce the Fed’s footprint in the financial system and would give the Fed more room to expand the balance sheet in future downturns.
A slightly different question is why the Fed has turned its attention to shrinking its balance sheet recently, a bit earlier than markets had anticipated. We see two likely reasons. First, some Fed officials have called for a shift in the mix of tightening, relying on both rate hikes and balance sheet reduction to tighten financial conditions in a more balanced way. One key reason, highlighted by Boston Fed President Eric Rosengren, is that relying solely on rate hikes could put excessive pressure on the dollar, weighing disproportionately on the export sector and imposing downward pressure on inflation. Second, putting in place a framework for balance sheet normalization could minimize financial market uncertainty and ease the transition process if President Trump appoints a new Fed Chair next year.
Q: When will balance sheet normalization begin?
The March FOMC minutes noted that most participants judged that a change to reinvestment policy would likely be appropriate “later this year,” and that nearly all agreed that “the Committee’s intentions regarding reinvestment policy should be communicated to the public well in advance of an actual change.” President Dudley has argued that it would make sense to take a “little pause” at the time the Fed announces a decision on the balance sheet, meaning that it would forgo a rate hike at that meeting. These comments suggest to us that the December 2017 meeting would be a natural time to announce balance sheet normalization, and additional guidance about the process could be provided before then.
Q:Will the Fed reduce its holdings of both MBS andTreasuries?
The March minutes seem to have settled this question, noting that “participants generally preferred to phase out or cease reinvestments of both Treasury securities and agency MBS.” This approach implies that once the total balance sheet reaches its terminal size, the Fed will have to reinvest maturing MBS in Treasuries for some time in order to shift the composition of the balance sheet back toward Treasuries.
Q: How quickly will the Fed reduce the size of its balance sheet?
The FOMC intends for balance sheet normalization, once announced, to be essentially on autopilot and operating largely in the background, with short-term interest rates instead serving as the primary monetary policy tool. The speed of normalization involves two questions: whether the Fed will rely on passive runoff or active sales, and whether it will end reinvestment in a tapered or immediate fashion.
On the first question, the FOMC’s long-standing guidance suggests that active sales are very unlikely, unless future Fed appointees have substantially different views from the current leadership. In the more distant future, limited sales could be used to accelerate the reduction of residual MBS holdings, but this should not affect the total size of the balance sheet at that point.
On the second question, the March minutes noted that “participants agreed that reductions in the Federal Reserve’s securities holdings should be gradual and predictable, and accomplished primarily by phasing out reinvestments of principal received from those holdings.” While the minutes provided a surprisingly balanced take on the relative merits of a tapered phase-out versus an immediate cessation of reinvestment, Chair Yellen’s past comments that the impact of balance sheet policy is less predictable suggest to us that a tapered approach is much more likely. Our baseline projection is that the Fed will taper reinvestment over 10 months by first reinvesting 90% of maturing assets, then 80%, and so on.
Exhibit 2 shows the changes in the Fed’s balance sheet implied by our projection of a tapered end to reinvestment beginning in January of 2018. The pace of reduction accelerates over the course of 2018, resulting in a $260bn decline in Treasury holdings and a $125bn decline in MBS holdings over the year. In 2019, Treasuries decline by $350bn and MBS decline by $180bn. The Fed’s balance sheet reaches its terminal size in mid-2020, as discussed below. Thereafter, MBS continue to run off, but the Fed resumes Treasury purchases to keep the total size of its balance sheet stable as a share of nominal GDP (if the Fed accelerates this quite prolonged process through MBS sales, the quarterly reduction in MBS and purchases of Treasuries would each increase by the same amount).
Q: Mechanically, how does ending reinvestment result in a smaller balance sheet?
Conceptually, the mechanics of ending reinvestment could simply result in holdings of maturing securities being swapped for cash on the asset side of the Fed’s balance sheet. But because “cash” (whether physical currency or reserve balances) is a liability of the Federal Reserve, it would not record cash on both sides of its balance sheet, but would instead record a smaller overall balance sheet.
Digging deeper, the mechanics of a Treasury redemption would work as follows: (1) The Treasury security is extinguished on the asset side of the Fed’s balance sheet; (2) the Treasury general account (TGA) at the Fed is reduced by an offsetting amount on the liability side reflecting the maturity of the security; (3) the Treasury Department issues new debt to a non-Fed investor, replenishing the TGA; and (4) in the simplest case of a bank purchasing the new debt at auction, the bank’s reserve account is reduced by the amount the TGA is replenished. In this way, the reduction in the Fed’s Treasury holdings on the asset side is matched by a reduction in excess reserves on the liability side, resulting in a smaller overall balance sheet.
Q: What is the terminal size of the Fed’s balance sheet likely to be?
We recently projected the terminal size of the balance sheet under two scenarios, a “small” balance sheet scenario in which monetary policy returns to its pre-crisis operating framework, and a “large” balance sheet scenario in which reserve balances remain elevated at roughly $1tn so that the Fed can continue to operate under the current floor system. We estimate that the small balance sheet option would imply a terminal size of about 11% of GDP ($2tn today), while the large balance sheet option would imply a terminal size of 15.5% of GDP ($2.9tn today), as shown in Exhibit 3.
Since we wrote, President Dudley has commented that he “favor[s] a system more like the current system, which is called a floor system, where you have a sufficient amount of excess reserves in the system so that your interest rate control is really driven by the overnight RRP and the interest rate you pay on excess reserves.” He estimated that this system requires excess reserves of $0.5-1tn. Adding required reserves to the midpoint of his range for excess reserves implies a total close to the $1tn figure used in our large balance sheet scenario. The New York Fed traditionally provides leadership on policy issues related to the balance sheet. We therefore place considerable weight on Dudley’s comment and use the large terminal balance sheet scenario for our baseline projections. However, the FOMC has not yet provided official guidance, and a comment last week from San Francisco Fed President John Williams that the terminal size is likely to be $2tn and should be reached in 5 years—corresponding to our small balance sheet scenario—suggests that these questions are not yet settled.
Q: How long will balance sheet normalization take?
Under our forecast for runoff, the Fed would reach our “large” estimate of its terminal balance sheet size (15.5% of GDP, equal to $3.3tn at the time) in May of 2020, about 21⁄2 years after it began to taper, as shown in Exhibit 4. It would not reach the “small” terminal balance sheet size (10.8% of GDP, equal to $2.5tn at the time) until late 2022, about five years after it began to taper. In either case, shifting the composition of the balance sheet away from MBS and toward Treasuries could take considerably longer.