By Jeff Huther
ABA DataBank
Federal Reserve Governors Michelle Bowman and Christopher Waller have recently expressed support for the Fed’s balance sheet goal of returning its portfolio of securities to primarily Treasurys. Standing in the way of reaching that goal is the Fed’s $2.3 trillion in mortgage-backed securities, which are paying down more slowly than the Fed’s holdings of Treasury securities. The Fed has not announced a specific timeline to get to a primarily Treasury portfolio and, aside from statements by individual Fed governors, there is no indication that MBS sales are being considered. That said, the slow reduction in the size of the MBS portfolio raises the possibility of sales. This analysis considers the potential scale and implications of sales. If a sales program is implemented, banks would likely see a modest reduction in the values of their MBS holdings and, through pricing, be incentivized to increase their holdings.
The Fed as mortgage holder
“What if ….?” is an occasional series from the ABA’s Office of the Chief Economist that provides insights from the investigation of hypothetical but plausible policy and regulatory changes.
Figure 1: The Fed’s MBS Portfolio
Source: FRED
Those holdings were eventually allowed to decline, only to shoot up during the early days of the pandemic as the Fed sought to stabilize financial markets. The Fed’s continued substantial MBS holdings suggest that at least some of the Fed governors have not seen an urgent need to return to a Treasury-only portfolio, despite the views of the governors noted above.
The slow walk to a Treasury-only portfolio
The Fed currently lets a fixed amount of short-dated Treasury securities and MBS mature while reinvesting any residuals into new Treasury issues with maturities as long as thirty years. This approach is setting the Fed up to hold mostly long-dated Treasury securities and MBS. If this portfolio composition is ultimately viewed as too inflexible for monetary policy implementation, the Fed will need to adjust its purchases to better align with its stated target portfolio.
At the current paydown rate for MBS, it will take around two decades to get MBS holdings below $500 billion (see Figure 2). The Fed has capped the pace of MBS paydowns at $35 billion a month to avoid market disruptions, while actual monthly paydowns, slowed by reductions in mortgage prepayments, have been running closer to $15 billion. Since the Fed views the $35 billion a month paydown pace as non-disruptive, it could use that number as a basis for sales, either as a fixed amount (shown as the green dashed line in Figure 2) or a 19 percent fixed share of the portfolio (shown as a dashed orange line in Figure 2).
Figure 2: The Size of the Fed’s MBS Portfolio under Alternative Sales Paths
Source: Author’s calculations
As shown in Figure 2, sales could have a large effect on the Fed’s portfolio. With a $35 billion per month combination of prepayments and sales, the Fed would have no more MBS on its books by the end of 2030. A less aggressive paydown would be to maintain the current maximum rate of 19 percent of the portfolio (that is, $35 billion a month translates into an annual rate of roughly 19 percent of current holdings). Either approach would hasten the Fed’s exit from the MBS market at a pace already deemed acceptable.
When the Fed was buying MBS, Fed staff did not anticipate the sharp and persistent decline in mortgage prepayments due to much higher interest rates. In the New York Fed’s baseline in the spring of 2022, for example, the fed funds rate was projected to rise to 2.62 percent by the end of 2024. This compares to the current fed funds rate of around 5.33 percent. The slow projected increase in interest rates meant that the Fed’s holdings of MBS were projected to fall to $1 trillion by 2030. Instead, as of June 2024, the Fed held $2.3 trillion (par value) of MBS, declining at a rate of around $15 billion a month.
If the Fed sells, who buys?
A simple approach to thinking about who would buy securities sold by the Fed is to assume that current holders proportionally increase their positions. Complete liquidation of the Fed’s portfolio would increase all other investors’ positions by over 20 percent. For U.S. banks, the largest aggregate holders of MBS, a proportional increase in MBS would be $600 billion. (See Figure 3 for a view of the portfolio changes needed for proportional increases by investor type.) Other investor types might opportunistically increase their holdings so banks might not have to fully absorb a proportional share but, since banks are the largest MBS holders, it is highly likely that any Fed sales would lead to higher bank holdings of MBS.
Figure 3: Holders of MBS
Source: Federal Reserve, author’s calculations
How would sales work?
MBS investors, including the Fed, are informed of how much of their holdings prepaid around the eighth business day of a month. Those prepayments (and regular principal payments) are made on the 25th calendar day of the month. Regardless of whether the Fed chose a portfolio reduction rate or monthly amount, it would likely set the sales amount after staff learned of prepayment rates for the Fed’s portfolio. Sales would be conducted through auctions with the New York Fed’s primary dealers, who would then resell the securities to end users. The New York Fed periodically conducts test sales, so portfolio rebalancing would largely be a scaling up in frequency and size of existing operations.
Since the Fed sells securities through auctions and auction prices are likely to be better with uniform products, the Fed would benefit from selling securities with the same yield simultaneously (evaluating the relative value of bids for different securities requires sophisticated judgment). For example, the Fed might sell Ginnie Mae 2.5 percent securities in one auction and Fannie/Freddie 2 percent securities in another auction. The Fed would also need to balance auction sizes with auction frequency. Dealers with winning bids would then look to find long-term investors to sell the securities to, with investors in the categories shown in Figure 3 providing guidance on likely candidates.
While sales in the quantities described above would be unlikely to have a large effect on the market, a so-called “announcement effect” could drive MBS prices somewhat lower. The Fed would likely follow its usual script of preparing investors for policy changes over many months through hints in speeches, testimonies and press conferences, making it difficult to relate a policy change to discrete changes in MBS rates. MBS dealers trade over $200 billion a day, so sales to bring the Fed’s portfolio reductions up to its maximum allowable amount would not be significant — unsurprising given the considerations that led to the $35 billion per month limit in the first place.
Conclusion
The Fed’s purchases of MBS, along with Treasury securities, were intended to “sustain smooth market functioning and help foster accommodative financial conditions.” While smooth market functioning was not in doubt, the purchases were accommodative — at times to the detriment of the Fed’s inflation objective. As the Fed is now in quantitative tightening mode to shrink its balance sheet, selling MBS would theoretically put marginal downward pressure on housing prices, which are still rising faster than overall inflation. Given the potentially helpful pricing effects, the inclinations of at least some Fed governors to return to an all-Treasury portfolio and the likely small market impact from Fed MBS sales of the size discussed above, it is curious that the Fed has not conducted sales already.
At this time, MBS sales by the Fed are not the most likely outcome, but the conditions for sales (high housing sector inflation, relatively high economic growth, continuing low unemployment) already exist. Monetization of the Fed’s losses from sales would not affect its ability to conduct monetary policy (just as its existing unrealized losses have not) and a smaller Fed balance may be seen as desirable. Fed officials may ultimately conclude that MBS sales, if undertaken cautiously, could support its macroeconomic goals.
