The Federal Reserve is no longer actively shrinking the total securities portfolio in the same way it was during quantitative tightening. But it is still stepping away from agency mortgage-backed securities. That distinction matters for anyone trying to understand whether mortgage rates should move lower or higher.
As of the latest H.4.1 release available on April 27, 2026, the Fed reported $1.993 trillion of mortgage-backed securities held outright on April 22, 2026. That was down about $192 billion from April 23, 2025. The Fed is still a massive holder of agency MBS, but it is not acting like a marginal buyer of new mortgage production. The current instruction is to reinvest principal payments from agency securities into Treasury bills, not back into MBS.
That is mildly upward pressure on mortgage spreads, all else equal. It does not automatically mean higher mortgage rates, because mortgage rates also depend on Treasury yields, inflation expectations, recession risk, volatility, bank demand, foreign demand, REIT demand, and lender margins. But it does mean one historically powerful source of mortgage demand is still fading.
Even after overall securities runoff ended, agency MBS balances can keep falling as principal payments are redirected into Treasury bills. Source: Federal Reserve H.4.1.
What the Fed Actually Changed
The Fed began reducing its securities portfolio in June 2022 by allowing principal payments to run off, subject to monthly caps. The original plan allowed up to $35 billion per month of agency debt and agency MBS runoff after the phase-in period.
That cap often did not bind. When mortgage rates rose, borrowers stopped refinancing, mortgage prepayments slowed, and the Fed’s MBS portfolio ran down more slowly than the cap allowed. A Federal Reserve FEDS Note published in September 2024 made this point directly: from June 2022 through June 2024, monthly agency MBS redemptions averaged about $18 billion, well below the $35 billion cap.
Then the Fed changed course on overall balance sheet runoff. On October 29, 2025, the FOMC announced that it would cease the runoff of securities holdings starting December 1, 2025. But the instruction was not to rebuild the MBS book. The Desk was directed to roll over Treasury principal payments and reinvest agency principal payments into Treasury bills.
The March 17-18, 2026 FOMC minutes repeated the same operating instruction:
- Maintain the federal funds target range at 3.50% to 3.75%.
- Increase SOMA securities through Treasury bill purchases, and if needed short Treasury securities, to maintain ample reserves.
- Roll over Treasury principal payments at auction.
- Reinvest all agency principal payments into Treasury bills.
The practical result is that the Fed is supporting reserve management with short Treasuries while letting its agency MBS footprint decline over time.
Fed demand supported the mortgage basis.
Private buyers must absorb more MBS supply.
This is liquidity support through short Treasuries, not renewed mortgage-credit support. The exhibit shows policy direction, not a forecast of monthly dollar volume. Sources: Federal Reserve Policy Normalization and March 2026 FOMC minutes.
Why This Matters for Mortgage Rates
Mortgage rates are usually described as a spread over Treasuries, but the better way to think about them is as a clearing price for a prepayable credit instrument. The borrower can refinance if rates fall. The investor owns a bond whose cash flows change when rates move. That embedded refinance option makes agency MBS different from Treasury notes.
When the Fed buys MBS, it removes duration and convexity risk from the private market. It also adds a large, price-insensitive buyer to the agency mortgage market. That tends to compress MBS spreads and lower mortgage rates relative to Treasuries.
When the Fed does not buy MBS, private investors must absorb more of the risk:
- Banks must decide whether MBS yields are attractive relative to deposit costs, capital, liquidity, and unrealized loss risk.
- Money managers must compare MBS to Treasuries, investment-grade credit, CMBS, and other spread products.
- Mortgage REITs must decide whether the levered spread works after repo funding and hedging.
- Foreign investors must evaluate dollar yields after currency hedging.
- Dealers must intermediate TBA inventory and pipeline hedges without becoming the final buyer.
If those investors require more compensation, the MBS spread widens. A wider MBS spread usually feeds into a wider primary mortgage rate, unless lender margins compress enough to offset it.
Does the Fed’s Current Action Signal Lower or Higher Long-Term Rates?
It sends mixed signals, so the clean answer is split into Treasury rates and mortgage spreads.
For Treasury rates, ending overall runoff is less restrictive than continuing quantitative tightening. Maintaining ample reserves through Treasury bill purchases is not the same as a new long-duration QE program, but it removes one source of liquidity pressure. By itself, that leans modestly lower for funding stress and front-end liquidity risk.
For long-term mortgage rates, the MBS-specific signal is less friendly. Reinvesting agency MBS principal into Treasury bills means the Fed is still allowing its MBS share to decline. That does not force mortgage rates higher, but it removes a large marginal buyer from the market. If private demand is not strong enough, MBS spreads need to stay wider to attract buyers.
So the Fed’s latest balance-sheet stance is not a simple “rates down” signal. It is closer to this:
- Less pressure from aggregate balance-sheet shrinkage.
- Continued pressure on agency MBS relative value because Fed agency holdings are still being reduced.
- A preference for short Treasuries over mortgage credit.
- No sign yet that the Fed wants to use MBS purchases to lower housing finance costs.
That mix points to lower long-term rates only if inflation, growth, and term premium also move lower. It does not, by itself, point to sharply lower mortgage rates.
The Marginal Buyer Problem
The Fed still owns nearly $2 trillion of MBS. But mortgage rates are not set by the average holder. They are set by the marginal buyer.
During QE, the marginal buyer was often the Fed. The Fed did not need a mortgage REIT’s levered return target, a bank’s deposit-cost calculation, or a foreign investor’s currency-hedged yield. That changed the market-clearing price.
Today, the marginal buyer is more likely to be a private balance sheet. That buyer has alternatives. A bank can hold reserves or Treasuries. A fund can buy corporates. A REIT can reduce leverage. A foreign investor can buy another dollar asset or hedge less duration. The MBS coupon has to clear against all of those alternatives.
That is why mortgage spreads can remain wide even after the policy rate falls. The Fed can lower overnight rates and still not be buying mortgage duration. If long Treasury yields stay elevated and MBS investors demand a wider spread, the primary mortgage rate can stay near 6% even with lower fed funds.
What Would Make This Bullish for Mortgage Rates
The balance-sheet signal would become more clearly bullish for mortgage rates if one of three things happened.
First, the Fed could stop shrinking the MBS book by reinvesting agency principal back into agency MBS. That would stabilize the Fed’s footprint in the market.
Second, the Fed could restart net MBS purchases. That would be a stronger housing-finance signal, but it would also raise the credit-allocation issue the Fed has repeatedly tried to avoid. The Fed’s longer-run stated preference is to hold primarily Treasury securities.
Third, private demand could replace the Fed without requiring wide spreads. That would require a friendlier combination of lower volatility, stable funding, attractive yield levels, lower hedging costs, and stronger bank or foreign investor demand.
Until one of those conditions appears, the Fed’s MBS runoff is not a direct mortgage-rate relief valve.
What Would Make This Bearish
The current stance could become more bearish for mortgage rates if private MBS demand weakens while the Fed’s book continues to amortize.
That could happen if:
- Treasury volatility rises.
- Inflation expectations drift higher.
- Banks remain cautious because deposit costs and securities losses reduce appetite.
- Mortgage REIT funding becomes more expensive.
- Foreign buyers face high currency hedging costs.
- The Treasury curve backs up because of fiscal supply or term premium.
In that scenario, the Fed is not adding MBS demand at the margin, so the market clears through wider MBS spreads and higher borrower rates.
The Development Underwriting Takeaway
For development pricing, the Fed’s balance sheet argues against assuming mortgage rates will fall just because the Fed has cut, or may cut, the fed funds rate. The relevant question is broader:
Will the 10-year Treasury yield fall, will MBS spreads tighten, and will lenders pass that improvement through to borrowers?
The latest Fed action helps liquidity more than it helps mortgage basis. It supports an ample-reserves framework through Treasury bills while allowing the MBS portfolio to keep running down. That is not a housing-credit easing program.
For underwriting, that means base cases should not rely on a fast return to 4% or 5% mortgage rates unless the pro forma also explains why long Treasury yields and MBS spreads both compress. A better approach is to sensitize exit cap rates, buyer affordability, refinance assumptions, and absorption against mortgage rates that stay higher for longer.
The Fed is still important. But in today’s mortgage market, it is important partly because it is no longer the buyer it used to be.