Updated: Apr 3
March 31, 2022 | In this edition of The Institutional Risk Analyst, we return to the key topic when it comes to changes in US monetary policy, namely how will the FOMC manage the reduction in the system open market account or SOMA. Many talking heads spend time squawking about the changes to interest rate targets, but what happens to the Fed's balance sheet is far more important to investors, lenders and markets. The economists who run the central bank assume that they can actually sell the mortgage backed securities in the SOMA trove despite the change in interest rates. But maybe not.
We asked the following question on Twitter this AM. What was the duration of the Fed’s MBS portfolio a year ago? And what is the duration now? The folks at FIDO have a nice summary definition of duration:
“Duration is expressed in terms of years, but it is not the same thing as a bond's maturity date. That said, the maturity date of a bond is one of the key components in figuring duration, as is the bond's coupon rate. In the case of a zero-coupon bond, the bond's remaining time to its maturity date is equal to its duration. When a coupon is added to the bond, however, the bond's duration number will always be less than the maturity date. The larger the coupon, the shorter the duration number becomes.”
In simple terms, you can think of duration as the weighted average time required to recoup your investment in a bond or servicing asset. But the bond or modified duration of a security shows how volatile that security is given a change in benchmark interest rates. Because most of the Fed’s $2.5 trillion in MBS are 2% or 2.5% coupons, these securities are more volatile than the broad market. Today in New York, of note, UMBS 2s for delivery in April were trading 92 26-28 at midday. The screenshot from Bloomberg shows the TBA market summary with Fannie/Freddie MBS (UMBS) at the top and Ginnie Mae MBS at the bottom
Source: Bloomberg 03/31/22
TCW outlined the liquidity issue facing the FOMC in a research note:
“The Fed has served as the buyer of last resort for years and suddenly removing its role in the agency MBS market will be problematic. In addition to the possibility of destabilizing the agency MBS market, it is likely to have knock-on effects on other sectors. Currently, agency MBS is already facing multiple challenges including an YTD rise in mortgage rates of over 100bps. We do not believe the Fed will want to put further stress in the market.”
While we agree that the mortgage market is vulnerable to further Fed manipulation, unfortunately the damage is already done. When the FOMC purchased trillions of dollars in MBS from the markets, the average life of a mortgage loan was a couple of years. In QE, the Committee also removed huge amounts of duration from the markets, forcing investors to look elsewhere for returns. The purchase of MBS during QE also pushed down interest rates and pushed asset prices higher, depressing visible default rates. Now we are reverting back to normal. Buckle up.
Unlike a regular bond, the maturity of MBS vary with the rate of prepayment of the residential mortgages behind the security. Every borrower has a free option to prepay the loan w/o penalty. When rates rise, bond prices fall and prepayment rates plummet. With effective MBS maturities now in excess of 10 years and headed higher, the FOMC wants to sell several times more duration back into the markets than was removed via QE. Hello. This change in duration of MBS is called extension risk, BTW.
Not only is the dealer community unwilling to deploy capital to support these risky, low-coupon securities, but these government-insured MBS may soon be under water in terms of cash returns. As the rate of monthly loan prepayments fall down to, say, high single digits in 2022, a Ginnie Mae 2 or 2.5% MBS is not an easy asset to sell. TCW continues:
“It will also be an operationally daunting task for the Fed to sell its MBS holdings. A mortgage portfolio’s paydowns are dependent on interest rates due to prepayment speeds… Over the past QE cycles, the Fed’s purchases have been focused in purchasing production coupons with the goal of matching market liquidity. For the latest QE cycle, the bulk of purchases have been focused in 30-year UMBS 2.0s and 2.5s. As a result, 70% of Fed’s 30-year UMBS holdings are in 2.0s and 2.5s. With the recent selloff in interest rates, the 30-year mortgage rate is hovering around 4% and the production has shifted to 3.0s and 3.5s.”
If interest rates continue to sell off, TCW notes, “the liquidity on the Fed coupons may be thin.” This may be the understatement of the year to date. Many observers are concerned about a massive sale of off-the-run MBS by the Fed. Some believe that given the Fed’s stated schedule to return to a Treasury-only portfolio, the Fed needs to sell MBS at some point in the coming years. We have a different view, namely that the Fed will ultimately need to change its schedule for balance sheet reduction.
The conditional prepayment rate or "CPR" is an estimate of the percentage of a loan pool's principal that is likely to be paid off prematurely. We think that prepayments on the Fed's portfolio of 2s and 2.5s could fall down into low single-digit CPRs. With the duration of low-coupon MBS falling and prices also plummeting, the FOMC may have to accept the natural runoff rate for the MBS in the SOMA and avoid outright sales for many years to come.
Of note, most the third-party valuation firms currently use 8-10% CPR for modelling LT MSR returns, but we think those 2% MBS held by the Fed could fall down to 4% CPRs. That is, we do not think that the FOMC will be able to OK outright sales of MBS from the SOMA beyond a token amount.
The only other option for the FOMC is to convince the Bank of Japan and other supranational buyers of MBS to swap the mortgage paper on the Fed’s books for Treasury securities. The good news is that prepays are down and returns on UMBS and GNMA MBS are again positive on a nominal basis after several years of negative returns. Buyers of MBS were annihilated by QE when CPRs rose to 30-40% annual prepayment rates.
Is Fed Chairman Jerome Powell smart enough to call the BOJ and ask the question? As the Dutch like to say, if you don’t ask the answer is no.
Trying to sell a couple of trillion in low-coupon MBS on the Fed’s current schedule will cause a market crisis in government and agency securities, an eventuality that even the folks on the FOMC will hopefully have the good sense to avoid. Given that the Fed is currently playing chicken with the markets with the abortive “transition” from LIBOR to SOFR, hopefully cooler heads will prevail. Unlike inflation and unemployment, risk is never transitory.