As Japan's 10-Year JGB Yield Rose 1.3 Points in a Year, How Did WACC Move?

How has the JGB yield moved over the past year?

Pulling the 10-year yield out of the Ministry of Finance's daily JGB interest-rate release on a monthly basis, the trend looks like this.

The yield stood at 1.633% in September 2025, climbed to the 2.2% range by around February 2026, eased back slightly in March, then resumed its climb through April, May, and June before reaching 2.943% at the end of August — a 1.31-point rise over the year. Over the same period, the short-term prime rate also climbed, from 1.475% in March 2026 to 2.375% in September, suggesting that both long- and short-term rate levels have been moving up together (source: beluuga.ai).

It's the Cost of Equity (Ke), Not the Cost of Debt (Kd), That Responds to Rising Rates

It's intuitive to assume that "rising rates push up the cost of debt (Kd), which then pushes up WACC" — but that's not necessarily true.Looking at the WACC formula, it's Ke, not Kd, where a rise in Rf flows through directly and in full. At beluuga.ai, we calculate the cost of equity, Ke, under CAPM (Ke = Rf + β × ERP), and the cost of debt, Kd, on an after-tax basis (Kd × (1 − statutory effective tax rate)), then weight them by market capitalization and interest-bearing debt (E/V and D/V) to arrive at WACC. Because of the shape of the formula Ke = Rf + β × ERP, any rise in Rf is added to Ke in full regardless of the size of β (the β × ERP term itself is unaffected by Rf). So whether a company's equity risk is high (large β) or low (small β), a 1.31-point rise in Rf lifts Ke by a uniform 1.31 points across the board. So what about Kd, the channel one might expect to matter most? At beluuga.ai, Kd is primarily sourced from a company's own bond market yield (based on JSDA over-the-counter reference prices) or, failing that, an industry median — and these market-observed values do tend to move up alongside rising JGB yields. In fact, if a company were to issue bonds in the current market, it would likely need to set a higher yield to place the issue. But there is no formula for Kd analogous to Ke = Rf + β × ERP, where Rf flows through one-to-one. How much of the rate rise passes through to Kd instead depends on things like the timing of refinancing existing debt, remaining maturity, and how credit spreads move — Rf's change does not mechanically translate into Kd on a one-to-one basis. In the estimate that follows, we hold Kd fixed at its current value and move only Rf, precisely so we can isolate how much capital structure (E/V) alone changes WACC's sensitivity. How Kd actually moved over this period is shown later, with real data, in the section titled "Kd (Own Bond Yields) Rose, but the Credit Spread Actually Narrowed."

Even With the Same +1.3-Point Rise in the 10-Year JGB Yield, the WACC Increase Ranges From +1.31pt to +0.15pt

Rf's rise is added to Ke uniformly. But because this estimate holds Kd fixed, the direct effect of the Rf rise on WACC runs only through Ke. As a result, the impact on WACC is proportional to E/V: the higher a company's E/V, the more of the Rf rise flows through to WACC, and the higher its D/V, the more that effect is diluted.

Lining up the current WACC composition for four companies with different capital structures against an estimate where only Rf is reset to its level a year ago (1.633%), the picture looks like this.

Nintendo, with E/V = 100% (zero interest-bearing debt), sees the full 1.31-point rise in Rf pass straight through to WACC. Tokyo Electric Power Holdings, at the other extreme with an exceptionally high D/V of 88.3%, sees WACC move by only +0.15 points. Toyota Motor (E/V≈48%) and Tokyu Fudosan Holdings (E/V≈33%) sit in between, at +0.63 points and +0.43 points respectively. For this same 1.31-point rate move, the increase in WACC ranks as:

Nintendo +1.31pt > Toyota Motor +0.63pt > Tokyu Fudosan Holdings +0.43pt > Tokyo Electric Power Holdings +0.15pt

In other words: the calculation shows that a single variable, E/V, is almost entirely what determines how a rate rise translates into WACC.

Kd (Own Bond Yields) Rose, but the Credit Spread Actually Narrowed

beluuga.ai has been collecting the market yields on bonds issued by the target companies themselves (JSDA over-the-counter reference prices) daily since April 17, 2026. It isn't a full year of data, but for this roughly 4.5-month observation window, we can look at how Kd actually moved using real data. For the three companies above (Toyota Motor, Tokyu Fudosan Holdings, Tokyo Electric Power Holdings), we picked one bond each with a similar remaining maturity as of April 2026 (roughly 3 years in each case) and set the change in its yield, from the start of collection to the most recent date, alongside the change in the comparable 3-year JGB yield over the same period.

All three companies' own bond yields rose by +0.26 to +0.27 points over the 4.5 months. The 3-year JGB yield over the same period rose +0.35 points — meaning the rise in these three companies' bond yields was actually smaller. In other words, over this window, the rise in JGB yields itself explains most of the increase, and the credit spread over government bonds actually narrowed slightly. This observation window (Apr 17 – Aug 28) is shorter than the one-year Rf move covered elsewhere in this article (Sep 2025 – Aug 2026), and is limited to the period since beluuga.ai began collecting daily own-bond yields. What we can confirm here is only this roughly 4.5-month window of actual movement — we don't know how Kd moved between September 2025 and March 2026.

Assumptions Behind This Estimate

The estimate behind "Even With the Same +1.3-Point Rise in the 10-Year JGB Yield, the WACC Increase Ranges From +1.31pt to +0.15pt" is a sensitivity analysis that asks: "holding β, ERP, Kd, and capital structure (E/V, D/V) fixed at their current (September 1, 2026) values, what would WACC have been if only Rf were reset to its level a year ago?" It is not a reconstruction of the WACC that was actually being calculated a year ago. By moving only the single variable Rf, we isolate how capital structure (E/V) shapes the way Ke feeds through to WACC, without that effect being confounded by changes in other variables. In reality, β (reflecting share-price moves) and capital structure (changes in market cap and interest-bearing debt) have also moved over this year, so the actual path of WACC did not follow this estimate exactly. As for Kd, as shown above in "Kd (Own Bond Yields) Rose, but the Credit Spread Actually Narrowed," the credit spread for the three target companies actually narrowed over the roughly 4.5 months we could observe (Apr 17 – Aug 28, 2026), and the rise in Kd was smaller than the rise in JGB yields over the same period. We cannot say with certainty in which direction, or by how much, the actual WACC increase would deviate from this fixed-Kd estimate, since we don't hold Kd data for the full period (including September 2025 – March 2026).

Data assumptions

• Rf (10-year JGB yield) source: beluuga.ai (primary source: Ministry of Finance, "JGB Interest Rate Information")
• Ke/WACC calculation date for the four target companies: September 1, 2026 (Rf=2.94%, ERP=5.14%, statutory effective tax rate=30.62%, source: beluuga.ai)
• β (levered β): company-specific OLS regression of weekly returns against TOPIX over the trailing two years, or, where the coefficient of determination is low, the Damodaran Japan industry average re-levered using the Hamada formula
• Kd (pre-tax cost of debt): the company's own bond market yield (based on JSDA over-the-counter reference prices) where available, otherwise the industry median
• Own bond yields and 3-year JGB yield trend (in the "Kd (Own Bond Yields) Rose, but the Credit Spread Actually Narrowed" section): Apr 17, 2026 – Aug 28, 2026, source: beluuga.ai (primary sources: JSDA over-the-counter bond reference statistical prices; Ministry of Finance, "JGB Interest Rate Information")
• E/V, D/V: calculated from market capitalization and each company's latest disclosed interest-bearing debt
• The Rf sensitivity estimate (the "WACC (if Rf were at last year's level)" column in the table) is a sensitivity analysis holding β, ERP, Kd, E/V, and D/V fixed at current values (see "Assumptions Behind This Estimate" above for detail)
• Source/calculation: beluuga.ai