GF Securities has released a research report stating that reviewing this year, US nominal GDP grew 5.8% and 6.3% year-on-year in the first and second quarters, while the average 10-year US Treasury yield stood at 4.2% and 4.4% respectively, roughly below normal valuation levels.

From June to October, interest rates rose notably, with monthly central levels of 4.6%, 4.7%, 5.0%, and 5.1%.

Third-quarter US nominal growth data actually did not climb further from the second quarter, coming in at 5.9% year-on-year. If we roughly estimate using a nominal GDP growth of around 6.0% for the second half and a 1.3x valuation multiple, the 10-year US Treasury yield should theoretically be around 4.6%.

Therefore, the rate movements since the third quarter contain other pricing factors.

Looking ahead to 2027, with a higher inflation base, US nominal GDP growth will most likely be slightly lower than this year. The Federal Reserve forecasts real growth year-on-year of 2.3% and 2.4% for 2026-2027, and PCE price year-on-year of 3.7% and 2.3%. Considering that PCE is more affected by energy prices and import prices, the deflator will show stronger resilience. Assuming US nominal GDP growth slows to around 5.2%, based on the 1.3x empirical relationship, this theoretically corresponds to a 10-year US Treasury yield of around 4.0%.

What should the central level for the 10-year US Treasury yield be?

From a macro framework perspective, nominal GDP growth is an important reference anchor for long-term nominal interest rates, with both reflecting real growth plus inflation compensation from macro and micro angles respectively. Looking at US data, the average ratio of nominal GDP growth to the 10-year US Treasury yield from 2000 to 2019 was 1.3x; from 2020 to the second quarter of 2023, due to the dual influence of a special economic environment and monetary policy during a special period, the ratio had no reference value; since the second quarter of 2023, the 13-quarter average has returned to 1.3x, and the quarterly average over the past two years since 2025 is also around 1.3x.

Reviewing this year, US nominal GDP grew 5.8% and 6.3% year-on-year in the first and second quarters, while the average 10-year US Treasury yield stood at 4.2% and 4.4% respectively, roughly below normal valuation levels. From June to October, interest rates rose notably, with monthly central levels of 4.6%, 4.7%, 5.0%, and 5.1%. Third-quarter US nominal growth data actually did not climb further from the second quarter, coming in at 5.9% year-on-year. If we roughly estimate using a nominal GDP growth of around 6.0% for the second half and a 1.3x valuation multiple, the 10-year US Treasury yield should theoretically be around 4.6%. Therefore, the rate movements since the third quarter contain other pricing factors.

We suspect there are two major reasons why rate levels during this period were above the theoretical central level and gradually rose: first, it is related to inflation expectations, as geopolitical direction remained uncertain, oil prices generally trended upward from July to October, and the market at one point worried about a Fed rate hike in October; second, it is related to AI companies issuing large amounts of long-term debt, a new phenomenon not encountered before, which has had a certain impact on the supply-demand structure for long-term bonds.

Since November, with an October rate hike ruled out and oil price movements easing, the 10-year US Treasury yield has returned below 5%, reaching 4.86% on November 6, initially regressing toward the theoretical central level. Looking ahead to 2027, with a higher inflation base, US nominal GDP growth will most likely be slightly lower than this year. The Federal Reserve forecasts real growth year-on-year of 2.3% and 2.4% for 2026-2027, and PCE price year-on-year of 3.7% and 2.3%. Considering that PCE is more affected by energy prices and import prices, the deflator will show stronger resilience. Assuming US nominal GDP growth slows to around 5.2%, based on the 1.3x empirical relationship, this theoretically corresponds to a 10-year US Treasury yield of around 4.0%.

Of course, this can only serve as a macro thinking anchor, and a series of factors still exist that could cause real rate levels to exceed this standard: first, the slowdown in US nominal GDP growth alongside rate hikes is a gradual process, and the corresponding rate decline is also a gradual process; second, at least in the first half of next year, the trend of AI capital expenditure expansion will continue, and AI companies issuing long-term debt will continue to push up term premiums to some extent; third, the rate hike cycle will bring dynamic pricing, and according to current mainstream market expectations, the US may still raise rates in the first half of next year, with policy rate increases and policy uncertainty constraining declines in market rates.

Taking into account the slowdown in nominal GDP growth plus these constraining factors, it is reasonable for the 10-year US Treasury yield to settle at a central level of 4.0-4.5%, similar to the first half of 2026. There are two major themes on the macro front for 2026: first, the pricing of AI supply; second, the pricing of rising US Treasury yields. After 2027, the first thread may ease somewhat; the key is whether new changes form in the former thread, such as a new narrative on the demand side.

Global asset trading sees US midterm elections conclude, AI policy uncertainty, geopolitical premium decline, US equities oscillate, bond market recovers, domestic stocks and bonds both strengthen, crude oil geopolitical premium clears, gold rebounds from oversold, dollar oscillates stronger

The US midterm elections' "blue wave" has concluded, the Philadelphia Semiconductor Index fell 12.15% for the week, the Nasdaq led declines, and Japanese and South Korean stock markets came under deep pressure. The 10-year US Treasury yield fell 8BP to 4.86%. Domestic small-cap growth showed the advantage of a "counterattack option," with the market unfolding along AI applications, and the all-A valuation deviation at +0.08σ still offering cost-effectiveness.