GLOSSARY DEEP DIVE

Mortgage-Backed Securities: Why the Bond That Pays You Back Early Is a Problem

A normal bond gives you a predictable schedule of payments until a fixed maturity date. A mortgage-backed security does not, because the underlying homeowners, not the bond issuer, control when the principal actually comes back, and empirically they tend to exercise that control at exactly the worst times for the investor holding the security.

Deep dive10 min readUpdated 2026

The core principle

A mortgage-backed security (MBS) pools thousands of individual home mortgages together and passes the principal and interest payments those homeowners make through to investors who hold shares of that pool. The cash flow an investor receives each month is a blend of scheduled interest, scheduled principal, and, critically, any unscheduled early principal repayment, called a prepayment, that occurs when homeowners in the pool refinance, sell their homes, or otherwise pay off their mortgage ahead of schedule.

Prepayment behavior is not random with respect to interest rates; it is systematically tied to them, and in the direction that hurts the investor. When mortgage rates fall, homeowners refinance into cheaper loans in large numbers, which returns an investor's principal early, exactly when that principal can only be reinvested at the new, lower prevailing rate. When mortgage rates rise, homeowners who locked in an old, lower rate have every incentive to stay put rather than refinance into a costlier loan, so the investor's money stays locked in the pool longer than expected, exactly when the investor would rather be free to reinvest at the new, higher rate.

This asymmetric behavior is called negative convexity. An ordinary, option-free bond's price moves in a predictable, curved relationship as interest rates change, and that curve is favorable to the holder: prices rise more for a given rate decline than they fall for an equivalent rate increase. An MBS instead shortens in expected life exactly when rates fall and lengthens exactly when rates rise, the reverse of what a fixed-income investor generally wants, and the market compensates investors for absorbing this asymmetric risk with extra yield, commonly quoted as an option-adjusted spread over a maturity-matched Treasury security.

Key idea An MBS's yield is not just compensation for credit risk the way a corporate bond's yield is; a meaningful part of it specifically compensates investors for prepayment risk, the uncertainty over when principal actually comes back, which is why comparing an MBS fund's yield directly to a Treasury fund's yield without this adjustment understates the real risk being taken.

It is worth distinguishing agency MBS, issued or guaranteed by government-sponsored entities and generally carrying an implicit or explicit government backing against homeowner default, from non-agency MBS, sometimes called private-label MBS, which carry no such backing and instead depend entirely on the credit quality of the underlying mortgage pool and any structural protections built into that specific deal, such as subordination of lower-priority tranches. Non-agency MBS played a central role in the 2008 financial crisis, when underlying mortgage credit quality proved far weaker than assumed and losses cascaded through the securities built on top of those mortgages. Agency MBS did not carry the same default exposure during that period, since the government backing addressed exactly the credit risk that impaired non-agency securities, though agency MBS still experienced the prepayment and extension behavior described above as rates moved sharply during and after that period.

How the math works

Example 1: how a rate drop shortens expected life and cuts future income. Suppose an MBS pool has an average expected life of 8 years at current mortgage rates, and pays a coupon of 5.0% on a $100,000 position, generating roughly $100,000 x 0.05 = $5,000 a year in income. If mortgage rates fall sharply and prepayments surge, the pool's expected life might shrink to 3 years as homeowners refinance in large numbers, returning much of that $100,000 principal back to the investor years earlier than planned. If new MBS at that point yield only 3.5%, reinvesting the returned principal now generates roughly $100,000 x 0.035 = $3,500 a year, a drop of $5,000 minus $3,500 = $1,500 in annual income, purely as a consequence of the rate environment that made reinvestment necessary in the first place.

Example 2: how a rate rise extends duration and delays reinvestment. Suppose the same pool, instead of prepaying quickly, sees mortgage rates rise and prepayments slow to a crawl, extending its expected life from 8 years to 13 years, since homeowners locked into their now-below-market 5.0% mortgages have little incentive to refinance. The investor continues collecting the 5.0% coupon, which sounds fine in isolation, but is now unable to redeploy that principal into newly issued MBS or bonds yielding a higher 6.5% for five additional years beyond what was originally expected. The opportunity cost over those extra five years, roughly $100,000 x (0.065 minus 0.05) = $1,500 a year in foregone extra income, is the practical cost of extension risk showing up on the other side of the same negative convexity.

How it shows up in real portfolios

The most common way individual investors hold MBS exposure is indirectly, through a bond fund or a broad-market bond index fund that includes agency mortgage-backed securities as a meaningful sleeve of its total holdings, often without the investor realizing the extent of that allocation. Checking a bond fund's factsheet for its allocation to agency MBS is a useful step for any investor who assumes their "core bond fund" behaves identically to a Treasury-only fund of similar duration; it generally does not, particularly in sharply falling rate environments.

A second scenario involves a retiree or near-retiree relying on a bond allocation specifically for predictable income, who discovers during a period of falling rates that their bond fund's income distributions are shrinking faster than expected, a pattern directly traceable to a wave of mortgage refinancing returning principal early inside the fund's MBS holdings, which the fund manager must then reinvest at the new, lower rates.

A third scenario, common among more sophisticated fixed-income investors, involves deliberately choosing between agency MBS, which carry government backing or guarantee against default but still carry full prepayment risk, and Treasury securities of similar duration, which carry neither meaningful default risk nor prepayment risk at all, accepting a lower yield in exchange for genuinely more predictable cash flow timing when that predictability matters more than the incremental yield.

A fourth scenario involves an investor who bought individual MBS pools directly rather than through a fund, perhaps as part of a laddered fixed-income strategy, and is surprised to receive a much larger than expected principal payment in a single month during a period of falling mortgage rates. Unlike a bond fund, which reinvests prepayments across the whole portfolio automatically and smooths the effect across many investors and many holdings, an individual investor holding a single pool directly experiences that reinvestment decision personally and immediately, often at a moment when reinvestment options are meaningfully less attractive than when the original pool was purchased, which is one reason most individual investors access this asset class through a diversified fund rather than direct pool ownership.

Actionable breakdown

  • Understand what MBS yield actually compensates for:
    • Prepayment risk, not just credit risk.
    • Negative convexity relative to an option-free bond.
  • Check the issuer type:
    • Agency MBS (Fannie Mae, Freddie Mac, Ginnie Mae) carry government backing or guarantee.
    • Government backing reduces default risk, not interest rate or prepayment risk.
  • Access MBS through a diversified bond fund rather than individual pools.
  • Expect MBS funds to lag plain Treasuries when rates fall sharply.
  • Check a fund's average life estimate, not just its stated maturity.
Key idea Government backing on an agency MBS protects against the homeowner defaulting, not against the homeowner refinancing; those are two entirely different risks, and agency guarantees address only the first one.

Common pitfalls

  • Treating an MBS fund's headline yield as directly comparable to a Treasury fund's yield without accounting for prepayment risk baked into that extra yield.
  • Assuming government-backed means price-stable; agency guarantees address credit risk, not interest rate or prepayment risk.
  • Ignoring how a falling-rate environment can quietly shrink a bond fund's future income well before it shows up clearly in the fund's price.
  • Overlooking a core bond fund's meaningful agency MBS allocation when assuming it behaves like a Treasury-only fund of similar duration.

For the interest rate sensitivity concept underlying negative convexity, see duration and interest rate risk. For the government debt MBS is often compared against, see Treasury bond and Treasury note. For the fund structure most investors use to access MBS, see mutual fund and ETF (exchange-traded fund). For broader context, see the guide on bonds.

The bottom line

Mortgage-backed securities pay extra yield specifically because their timing works against the investor whenever rates move, shortening when reinvestment rates are low and lengthening when reinvestment rates are high, so they belong in a bond allocation only with that trade-off clearly understood.

Back to the full glossary