In early August, a phenomenon emerged in the Japanese government bond market: the yield on 15-year Japanese government bonds approached 4.25%, significantly higher than not only the 10-year and 20-year bonds, but even exceeding some 30-year and 40-year bonds. According to traditional term structure logic, this partial inversion is unnatural in an environment where both inflation risk premiums and fiscal term premiums are rising simultaneously. More importantly, this does not entirely indicate a sudden loss of control over the real financing cost of 15-year bonds, but is closely related to the yield calculation methods used in the Japanese market, the coupon structure of existing bonds, and changes in the supply of ultra-long-term government bonds. As of August 6, the yields on 10-year Japanese government bonds were approximately 2.81%, 20-year bonds approximately 3.685%, 30-year bonds approximately 3.97%, and 40-year bonds approximately 4.01%. Yields had risen significantly in the previous trading session due to weak primary market demand, but subsequently corrected as energy prices declined. The Bank of Japan is currently guiding the unsecured overnight call rate to around 1.0%, a significant departure from the previous long-term ultra-low interest rate environment. This means that the differences between duration, coupon rate, and bond pricing methods are being amplified.

The abnormal 15-year yield is primarily due to a problem with the calculation method.

The key to understanding this anomaly is not to first discuss fiscal risks, but to first confirm what the "yield" displayed on the screen actually means. Japanese government bond benchmarks have long used simple yields, while the more common comparison method in global bond trading is yield to maturity. The core of simple yields is to linearly process coupon income with the annualized return resulting from the bond's reversion from its current price to face value, and then divide by the market price. It's convenient to calculate, but it doesn't fully consider the compounding of cash flows and the time value of cash flows at different maturities. Under normal circumstances, the difference between the two calculation methods is limited, so in the long run, it doesn't seriously interfere with market judgments. However, the problem arises with a bond with a low coupon, a price significantly below face value, and a relatively long remaining maturity. The bond currently used as the 15-year general benchmark has a coupon rate of only 0.5% and a maturity of September 2041. With market interest rates significantly higher than its coupon rate, the bond price has been in a state of significant discount for a long time. Simple yield calculations over-linearize the "reversion to face value" over a long period, thus inflating the displayed yield. This traditional calculation method needs special attention when observing the Japanese general government bond index. This means that the 15-year screen yield is close to 4.25%, which cannot be directly equated with the actual 15-year holding return calculated using the internationally commonly used yield-to-maturity metric.

Why is it that the most obvious distortion occurs specifically in the 15-year period?

The 10-year and 30-year benchmarks typically shift frequently to newly issued bonds. These new bonds have coupon rates closer to current market interest rates, and their prices are closer to face value. Therefore, even using a simple yield curve, the deviation from the yield to maturity is usually not particularly exaggerated. The 15-year maturity is different. It's not a continuously issued large volume of new bonds; instead, it's often based on older bonds with remaining maturities close to 15 years. When these older bonds were issued during periods of low interest rates and had low coupon rates, and the current interest rate environment has fundamentally changed, the distance between the price and face value widens significantly. Therefore, the "peak" on the 15-year yield curve essentially contains two parts of information. The first part is the actual term risk and changes in supply and demand. The second part is the statistical amplification created by the combination of low-coupon older bonds and the simple yield curve formula. If these two parts are not separated, and 15-, 20-, 30-, and 40-year maturities are compared solely based on the general yield curves displayed on trading terminals, it's easy to misjudge the curve shape.

Anomalies do not all stem from formulas; ultra-long-term supply is also altering the curve.

The calculation method can explain why the 15-year yield "looks particularly high," but it cannot explain all changes in the term structure. In fiscal year 2026, Japan's government bond issuance plan has proactively reduced the supply of ultra-long-term bonds. Official plans show that the issuance of 20-year, 30-year, and 40-year government bonds will decrease by 100 billion yen per month. Specifically, the annual market issuance of 20-year bonds will decrease from 10.2 trillion yen to 8.4 trillion yen, 30-year bonds from 8.7 trillion yen to 7.2 trillion yen, and 40-year bonds from 2.5 trillion yen to 1.8 trillion yen. This reduced supply means that, all else being equal, 30-year and 40-year bonds are more likely to receive a scarcity premium, resulting in relatively stronger bond prices and relatively suppressed yields. Meanwhile, long-term funds' liability duration management remains focused on ultra-long-term bonds. Some life insurance institutions are adjusting their holdings of low-coupon medium- and long-term bonds and increasing their allocation to newly issued ultra-long-term bonds. This behavior involves both asset-liability duration and book profit and loss management. Therefore, the anomaly in the curve from 15 to 30 years is the result of the combined effects of pricing formulas, issuance policies, and institutional funding behavior, rather than being explained by a single macroeconomic variable.