The 30-Year Yield Hit 5.34%: Why Mortgage Bonds Stay Discounted
The U.S. 30-year yield printed 5.34% in mid-August, putting the bond on its worst stretch since 2006 as it entered September 2026, according to Bloomberg-compiled data. Bloomberg Yield is the annual return demanded to hold the bond. That print reset the reference rate for long-dated credit.
The selloff overlapped with a flattening Treasury curve in early to mid-September. Reuters The Bloomberg account of the August drop also described flattening into September. The sequence was inversion, with short rates above long rates, easing into a flatter curve, not a return to a steep upward slope.
In a November 2022 deep dive, Harley Bassman identified inversion as the dominant force behind mortgage rates and estimated MBS prices were discounted by roughly five to six points. Convexity Maven MBS are pools of home loans sold to investors. The paper tied those cheap dollar prices, what investors pay for $100 of loan balance, to curve shape rather than credit deterioration.
Bassman made the same link in April 2024, attributing soft mortgage bond prices to the inverted curve and to market pressure pushing those bonds lower. Financial Times Together, the two statements describe a market where curve structure forces a repricing of long loans borrowers can repay early.
The broader context here is how curve shape feeds into TBA and specified-pool pricing, the main marketplaces for mortgage bonds. An inverted front end raises funding and hedging costs, while the embedded prepayment option shifts in value with rate volatility and expected loan life. When long yields rise and dollar prices fall, discount coupons extend and last longer. That extension lifts negative convexity costs, where prices fall faster than they rise, and widens the option-adjusted spread investors demand. That is why a five to six point discount is the market clearing level for extension, volatility and carry, not a pricing error.
Looking at what this means for relative value, the shift from inversion toward flattening does not by itself normalize mortgage spreads. What is known is the August high and the September flattening. What is priced in for prepayments, turnover, and bank and REIT demand is less certain. Weak long-end performance on a scale not seen since 2006 counsels caution. A flatter curve can cut the inversion penalty but leave term premium, the extra pay for holding long debt, and supply absorption as the binding constraints. The question for managers is whether current dollar prices compensate for that duration and convexity profile, not whether headline mortgage rates look high alone.
