The mortgage spread is one of those concepts that's invisible until you know about it, then unavoidable once you do. Every credible forecast of where mortgage rates are heading involves a forecast of both the 10-year Treasury and the spread, because mortgage rates are essentially the sum of those two things. Treasury moves get the financial-press attention; spread moves usually don't, even though the spread is responsible for roughly a third of the difference between the all-time-low 2.65% mortgage rate in January 2021 and today's 6.37%.
What the spread actually is
The mortgage spread is the difference between the 30-year fixed mortgage rate and the 10-year Treasury yield. With the Freddie Mac PMMS 30-year at 6.37% and the 10-year Treasury at about 4.38% (week of May 8, 2026), the current spread is roughly 1.99 percentage points.
The reason mortgages price off the 10-year Treasury and not the 30-year Treasury comes down to expected duration. Despite being called "30-year" mortgages, very few actually last 30 years. People move, refinance, sell, or pay extra. The average actual life of a 30-year mortgage is closer to 7-10 years, which makes the 10-year Treasury a better duration match for the cash-flow profile mortgage investors are pricing.
The spread exists because mortgage-backed securities carry risks that U.S. Treasuries don't. Treasuries are the global benchmark for risk-free lending in dollars — backed by the U.S. government, completely predictable in their cash flows, infinitely liquid. Mortgage-backed securities are riskier on two dimensions:
- Credit risk. Borrowers default. For agency MBS (loans guaranteed by Fannie Mae, Freddie Mac, or Ginnie Mae), this risk is small — the agencies make investors whole on principal even if individual borrowers default. The agencies themselves carry an implicit U.S. government backstop. So credit risk contributes only a small portion of the spread on agency MBS.
- Prepayment risk. This is the big one. Borrowers can pay off their mortgages whenever they want — by refinancing, selling, or making extra principal payments. This optionality is one-sided: it favors the borrower, not the investor. Investors charge for it through a higher required yield, which translates directly into the spread.
Fannie Mae's published research puts the historical mortgage spread in a range of about 0.71 to 1.4 percentage points, with a long-term average around 1.7 when measured against the PMMS specifically. The current spread of ~2.0 sits well above that historical band. We'll get to why.
The prepayment problem, more carefully
Prepayment risk is worth understanding because it explains most of why the spread moves. Imagine you're an investor who buys a mortgage-backed security yielding 5% when 10-year Treasuries are at 3%. You're earning a 2-percentage-point premium for the prepayment risk.
Now consider two adverse scenarios:
Rates fall to 2%. Borrowers refinance en masse. Your MBS gets paid off early. You get your principal back at par — but you now have to reinvest it at 4% mortgages instead of the 5% you were earning. The lower-rate environment that drove the refinances also lowered your reinvestment options. You lose money.
Rates rise to 7%. Borrowers don't refinance — their 5% loans are now well below market. They also don't sell their homes (the "rate lock-in effect" — they don't want to give up their below-market mortgage). Your MBS extends in duration just as interest rates have risen, locking you into a below-market yield for longer than expected. You lose money.
This is negative convexity. In bond-investor language, MBS investors get hurt when rates fall (their bonds prepay early) and hurt when rates rise (their bonds extend in duration). Treasuries don't have this problem — their cash flows are predictable regardless of rate environment. Investors demand additional yield to take on the negative convexity, and that additional yield is most of the mortgage spread.
The key insight: prepayment risk gets worse when rate volatility is high. If rates are stable, prepayment behavior is predictable. If rates are bouncing around, investors don't know whether their MBS will prepay or extend, so they require more yield to take on the uncertainty. This is why the MOVE Index — a measure of Treasury market volatility — is one of the best predictors of the mortgage spread. When the MOVE is high, the spread is wide. When the MOVE is low, the spread tightens.
A historical tour of the spread
Looking at where the spread has been over the past 15 years tells you a lot about what's normal, what's not, and what moves it.
| Period | Approx. spread | What was happening |
|---|---|---|
| 2008-2009 | 2.5-3.0 pp | Financial crisis, MBS demand collapsed, spread blew out |
| 2010-2014 | 1.5-2.0 pp | Fed QE buying MBS heavily; spread compressed below historical norm |
| 2015-2019 | 1.5-1.7 pp | Calm rate environment, normal spread |
| 2020-2021 | 1.0-1.5 pp | Peak Fed MBS buying during pandemic; spread compressed to historic lows |
| Late 2022 - early 2023 | 2.7-3.1 pp | Fed begins QT (stops buying MBS); rate volatility spikes; spread peaks |
| 2024-2025 | 2.2-2.6 pp | Spread partially normalizes but remains elevated; volatility still high |
| May 2026 (now) | ~1.99 pp | Continued partial normalization; MOVE Index moderated to mid-70s; spread sits near long-term midpoint but above pre-2008 historical band |
Two things to notice from the history:
The pandemic period gave borrowers a double benefit. Both the 10-year Treasury yield AND the spread fell to historic lows simultaneously. The 10-year hit 0.5% in mid-2020. The spread compressed to about 1.0 percentage point. The combination produced the 2.65% mortgage rate of January 2021. Recreating those rates today would require both numbers to return to historic-low territory at the same time, which would require some combination of an economic shock severe enough to push Treasury yields back near zero plus a return of massive Fed MBS buying. Possible, but neither is currently in the cards.
The 2022 spread blowout was unusual but explainable. Three things hit simultaneously: the Fed reversed from being the largest MBS buyer to being a net seller (via runoff), Treasury market volatility spiked to historic highs as the Fed hiked aggressively, and prepayment uncertainty rose because borrowers with 3% mortgages were suddenly very unlikely to refinance regardless of where rates went. All three pushed the spread wider at the same time.
What moves the spread today
The four main inputs to the spread, ranked by current relevance:
1. Treasury market volatility (the MOVE Index)
The MOVE Index measures implied volatility on U.S. Treasury options — essentially, how much investors expect rates to move over the near term. When the MOVE is low (under 60), prepayment outcomes are more predictable, MBS investors require less compensation, and the spread tightens. When the MOVE is high (above 100), prepayment uncertainty rises and the spread widens. The MOVE peaked above 180 in October 2022 (around the same time the spread peaked). It's currently around 77, which has helped pull the spread down from its 2022 highs but is still well above the calm-market readings of 2015-2019 (often below 50). Until volatility settles further, the spread is unlikely to return all the way to its pre-2008 historical band.
2. Fed balance sheet policy
The Federal Reserve has been a major MBS holder since 2008. Peak MBS holdings were $2.7 trillion in 2022, when the Fed owned roughly a quarter of the entire agency MBS market. Since 2022, the Fed has allowed its MBS holdings to roll off as loans pay down — quantitative tightening. The withdrawal of the largest single MBS buyer from the market widens the spread by reducing demand. If the Fed eventually slows or stops MBS runoff, the spread should compress somewhat. If the Fed resumes MBS purchases (which would only happen in a serious economic downturn), the spread could compress dramatically. Until either of those things happens, the Fed's posture is a structural drag on spread tightening.
3. Bank demand for MBS
Commercial banks hold a meaningful share of agency MBS in their securities portfolios. Bank willingness to hold MBS depends on capital requirements, deposit stability, regulatory treatment, and whether competing assets (commercial loans, Treasuries) look better on a risk-adjusted basis. The 2023 regional banking stress reduced bank MBS demand as banks rebuilt liquidity and shifted away from longer-duration assets. As banks have stabilized, demand has partially recovered but remains below pre-2022 levels. A meaningful return of bank demand would narrow the spread.
4. Foreign and institutional demand
Foreign central banks, sovereign wealth funds, pension funds, and insurance companies all hold MBS as part of their fixed-income portfolios. Their willingness to buy depends on dollar exchange rates, the steepness of the yield curve, and whether MBS yields look attractive relative to corporate bonds and other high-grade fixed income. This demand is harder to forecast and less commented on than the other three factors, but it matters.
Why this matters for borrowers
The spread isn't just a curiosity. It directly affects the rate you're being quoted. Three practical implications:
- Don't expect mortgage rates to fully track Treasury yields. If the 10-year drops 0.3 percentage points, mortgage rates won't necessarily drop 0.3 points. If the spread widens by 0.1 points at the same time (which can happen if volatility spikes during the Treasury move), the actual mortgage rate decline is closer to 0.2 points. Conversely, the spread can compress on its own and pull mortgage rates lower even when Treasury yields are flat.
- The "fair value" mortgage rate may be lower than today's quote. If the spread eventually normalizes from ~2.0 to ~1.7 (the long-term historical average), mortgage rates would fall by ~0.3 percentage points independent of any Treasury move. Whether that normalization happens — and when — depends primarily on the MOVE Index and the Fed's MBS runoff policy. Neither is in your control as a borrower, but tracking them gives you a signal for whether the rate environment is gradually improving on the spread channel.
- Forecasts that focus only on the 10-year miss half the picture. Most consumer financial press coverage discusses "where the Fed thinks rates are heading" or "where the 10-year Treasury is going." Neither directly determines your mortgage rate. The complete forecast requires both Treasury direction AND spread direction, which is harder, less covered, and more uncertain.
For deeper coverage of what's behind the 10-year Treasury moves themselves, see what actually drives mortgage rates. For the headline numbers and the current PMMS averages, see the Rates Center hub.
Putting today's rates in historical context
One question this content often raises: are today's mortgage rates "high"? Relative to the 2020-2021 pandemic lows, yes. Relative to the 50-year history of mortgage rates in the United States, no — they're around the long-term median. The 30-year mortgage averaged about 7.5% in the 1990s. It averaged about 6.5% in the 2000s. The all-time peak was 18.6% in late 1981. The 2.65% trough of January 2021 was a once-in-a-generation anomaly driven by extraordinary policy interventions.
The "stuck rate" view that has gained traction among housing economists argues that the 6% range may be the new normal for the next 12-24 months, anchored by persistent service-sector inflation, geopolitical risk premia in commodity markets, and a Fed willing to accept slightly above-target inflation rather than risk a recession. This view could be wrong. But planning for it — buying or not buying based on today's rates rather than a hoped-for return to 3% — is the conservative posture. If rates drop substantially, refinancing is always available. Locking in a transaction based on rate optimism is one of the most common ways people end up overextended.