September 25, 2026 KOSPI Market Brief

★ Yields have "settled in" above 5.1%. The 10-year hit 5.17 to 5.18% intraday (highest since 2007) and closed near 5.11 to 5.12%, roughly unchanged. The 30-year touched 5.47% (highest since 2004). Japan's 10-year hit its highest since 1996, Germany's since 2009. This is a global bond sell-off.
★ Stocks erased the morning's losses on the Hormuz report. S&P 500 7,704.13 (−0.02%) · Nasdaq 26,939.37 (+0.01%) · Dow 51,349.98 (−0.31%) · Russell 2000 2,835.57 (−0.11%). Morning drawdowns of −0.5% (S&P) and −0.9% (Nasdaq) disappeared.
7-year auction at 5.085%, highest since April 1993. The tail was about 0.7bp, short of the 1bp bar we set yesterday, but bid-to-cover fell to 2.42 (from 2.50) and indirect bidders to 57.2% (from about 61%). Two weak auctions in two days.
The Fed is again signaling "one more." New York Fed President Williams: expecting another hike by year-end is "reasonable." CME odds of an October hike roughly 70 to 77%. 30-year mortgage rate 7.37%, highest since May 2024.
Oil read the day opposite to stocks. Brent $106.60 (+3.4%) · WTI $94.61 (+2.6%). The Trump–Xi summit ended with little beyond the two-month tariff-truce extension. USD/JPY near 159, with 160 as the intervention line traders are watching.

This is Mr. Can. Here is the September 25, 2026 brief ahead of the Korean market open.

TL;DR — Yields settled at 5.1%. Stocks listened to Hormuz.

The Big One — Stocks and oil heard the same headline differently

Korea Variable — Two nights of backlog: rates and FX, not stocks

Checkpoints — Four things to watch before Seoul reopens

1 / 5

Written: 2026-09-25 12:04 KST (Vancouver 9/24 20:04 PDT)
Basis: US 9/24 regular-session close + Europe 9/24 close + Tokyo, Hong Kong, Shanghai 9/24 close (Korea closed for Chuseok)

🟡 Scheduled 16:00 PDT, started 16:07 PDT. A long tool stall mid-draft pushed publication to about 12:04 Korea time on 9/25, roughly four hours late. Korean markets are closed today, so no open is missed. Tokyo, however, opened at 09:00 KST and this report does not cover its 9/25 session. Checkpoint 4 will be judged on Tokyo's close (15:30 KST).
🟡 Korean markets are closed today (9/25, Chuseok Day). Chuseok is Korea's autumn harvest holiday; the exchange is shut 9/24 to 9/27 and reopens Monday 9/28. This is the second holiday edition. The question stays the same: what is piling up while Seoul is closed?

SEO Intro

The US 10-year Treasury yield touched 5.17 to 5.18% intraday last night, its highest since 2007. Seven-year notes sold at 5.085%, the highest auction yield since April 1993.

Yet US stocks finished almost exactly where they started. S&P 500 −0.02%, Nasdaq +0.01%.

What lifted equities was not bonds but the Strait of Hormuz. Midday, Reuters reported that US and Iranian negotiators are discussing a phased deal: Tehran reopens the strait, Washington lifts its economic blockade.

Oil heard the same headline differently. Brent slipped from $108 but still settled at $106.60, up 3.4%.

Stocks bought the possibility of a deal. Oil and bonds kept pricing the reality that there is no deal yet. Seoul will absorb both when it reopens on September 28.

1. Grading the 9/24 AM Checkpoints (last night's US session)

Yesterday morning's axis: "Does last night's yield spike harden into a new level, and can chips withstand it?"

# · Criterion we set · Result · Verdict / 1 · 7-year auction tail 1bp or more · 5.085% vs 5.078% when-issued, ~0.7bp · ❌ Below the bar · still a weak auction / 2 · 10-year close above 5.10% / below 5.05% · ~5.11–5.12% · ✅ Above 5.10% / 3 · Chip index falls less than the Nasdaq · Nasdaq +0.01% · large chipmakers reported down up to 3% intraday · 🔲 SOX close not secured · 🔲 Pending (evidence leans "chips lagged") / 4 · Written deliverables from the summ...

[Interpretation] Checkpoint 1 was less "wrong" than set too high

The tail missed 1bp, but bid-to-cover dropped and the indirect share, which includes foreign buyers, fell from about 61% to 57.2%. Judged by the single line we wrote, it's a ❌; judged by direction, demand weakened for the second straight day. We keep the verdict as written. Changing criteria after the fact is exactly what this project avoids.

[Interpretation] Checkpoint 2 is the midpoint of Friday's test

The 10-year ran to 5.17–5.18% and came back to 5.11–5.12%. There was force pushing up and force pulling back, and the net was zero. Yesterday we said yields were "resting"; today "settled" fits better. The 9/23 PM checkpoint (Friday close above 5.10%) has one day left.

2. The Big One: Stocks and Oil Heard the Same Headline Differently

[Fact] The Hormuz report (9/24 ET, midday, Reuters)

US and Iranian negotiators meeting in New York are discussing a phased deal: Iran reopens the Strait of Hormuz; the US lifts its economic blockade of Iran.
The obstacle: neither side wants to give up its leverage first.
The war is now about seven months old.

[Fact] Before and after

Asset · Before report (morning) · Close / S&P 500 · ~−0.5% · −0.02% (7,704.13) / Nasdaq · ~−0.9% · +0.01% (26,939.37) / Brent · intraday $108 · $106.60 (+3.4%) / WTI · $94.61 (+2.6%) / US 10-year · intraday 5.17–5.18% · ~5.11–5.12% / VIX · 15.67 (+3.2%)

★ [Interpretation] Score "possibility" and "reality" separately

On 9/18 this project wrote a rule: when stocks and bonds are watching different variables, grade them separately. Last night was that day.

Stocks bought the possibility of a deal. Equity indices buy the future first.
Oil kept pricing the reality that the strait is still shut. Crude only gets cheaper when tankers actually pass. The headline trimmed the rally but oil still closed up more than 3%.
Bonds watched that oil price, along with hot growth data and Fed talk of another hike. So they pulled back from the highs but did not come down.

[Interpretation] For Seoul on September 28, the question is which of these three turns out right over the weekend. If the deal moves from "reported" to "announced," stocks were right. If it stays talk, oil and bonds were right.

3. Korea Variable: Two Nights of Backlog

[Observation] What has piled up since Seoul closed on 9/23 (two of four nights)

Driver · At Seoul's 9/23 close · Now · Direction for Seoul / US 10-year · ~4.95% · ~5.11–5.12% · 🔴 / Korea 10-year govt bond · closed at 4.405% (−5.6bp) · 🔴 closed · 🔴 First gap to close on 9/28 / S&P 500 · 9/22 close basis · ~−0.8% over two days · 🟡 / Brent · ~$99 (9/22) · $106.60 · 🔴 / US–China tariff truce · expires 11/10 · extended two months to 1/10 · 🟡 Shorter than hoped / Nikkei 225 (chip-heavy index) · 9/24 +0.97% · broader TOPIX down · 🟡...

★ [Interpretation] Two nights, net: stocks roughly flat, rates and oil a notch higher

The S&P 500 is down only about 0.8% across two sessions. By equities alone, Seoul has little to fear. But the two numbers Seoul locked in on 9/23, a 4.405% 10-year yield and 1,358.4 won per dollar, were both set before the US rate spike. [Interpretation] The 9/28 gap may show up more in Korean bonds and the won than in stocks. Bonds catch up first, then the currency.

[Observation] Still a narrow window on the won

We again could not secure the offshore NDF won/dollar quote (🔲 third straight edition). The neighbors give a direction: the dollar index was around 101, its highest since late July, as of yesterday morning, and USD/JPY is near 159. [Interpretation] It is hard for the won alone to strengthen against a dollar this firm. We keep yesterday's call that the won will likely open weaker than its 9/23 close on 9/28.

4. Macro Backdrop: Global Bonds, the Fed, the Summit

[Fact] Bonds

Item · Level · Note / US 2-year · ~4.91% · Near highest since May 2024 / US 10-year · 5.17–5.18% intraday → ~5.11–5.12% close · Highest since 2007 / US 30-year · 5.47% intraday · Highest since 2004 / 7-year auction · 5.085% · bid-to-cover 2.42 · indirect 57.2% · Highest since April 1993 / Japan 10-year · 🔲 level not secured · Highest since 1996 / Germany 10-year · 🔲 level not secured · Highest since 2009 / 30-year mortgage · 7.37% · Highest since May ...

[Fact] The Treasury bought back about $4 billion of 20- and 30-year bonds on the day. Yields nonetheless hit session highs around the announcement.

[Interpretation] When long yields rise even as the government buys long bonds, the sell-off is about more than a temporary shortage of buyers; the market is saying rates should be higher. With Japan and Germany also at multi-decade highs, it isn't just an American story.

[Fact] The Fed

NY Fed President John Williams: it is "reasonable" to expect another hike by year-end.
CME odds of an October 28 hike: roughly 70–77% (69.7%, 73.5% and 77.5% depending on source and time). Under our single-source rule, we adopt only the direction: above 70%.
Current policy rate: 3.75–4.00% (raised 9/16).

[Fact] Trump–Xi summit (9/24 ET, White House)

Bilateral meeting of a little over an hour, followed by a state dinner.
Confirmed deliverable: tariff truce extended two months (11/10 → 1/10/2027), announced the evening before by Treasury Secretary Bessent.
AI, fentanyl, and military-incident talks produced no confirmed written agreements. Analysts had expected a cautious summit with few deliverables.

[Fact] Commodities and more

Item · Value / Brent · $106.60 (+3.4%) · $108 intraday / WTI · $94.61 (+2.6%) / Gold · ~$4,308 (+0.2%) / Bitcoin · ~$84,400 (flat) / USD/JPY · ~158.96 (+0.4%) / US August new home sales · 684,000 (+6.4%), fastest pace this year · median price $393,700 (−5.8% y/y) / Costco (after the close) · EPS $6.75 (est. $6.55) · sales $93.9B (+11.2%) · +0.3% after hours

[Interpretation] New home sales rose because builders cut prices. Homes selling at a 7.37% mortgage rate thanks to discounts and incentives shows the rate burden has already reached consumers.

5. Next Korean Session (Mon 9/28 KST) Preview

[Observation] What's left of the four days

🟢 Tokyo today (9/25): already open. Does yesterday's narrow rally (Nikkei up, TOPIX down) continue?
🟢 US Friday (9/25 ET): weekly close; final test of the 9/23 PM checkpoint.
🟡 The weekend: does the Hormuz deal move from "reported" to "announced"?
🟡 Next week's big items: US August PCE (the Fed's preferred inflation gauge) on 9/30 and Micron earnings, both landing on 10/1 Korea time.

✅ Four Checkpoints (before Korea reopens on 9/28)

The axis: is Seoul's 9/28 gap driven by the "reality" of rates and oil, or the "possibility" of a Hormuz deal?

Does the US 10-year close Friday (9/25) above 5.10%? (Final verdict on 9/23 PM checkpoint 1.) Above means a new level; below 5.00% means one day of overheating.
Is the phased Hormuz deal officially announced by both sides? Our line is Brent's 9/25 settle: below $100 favors possibility, above $105 favors reality.
Does USD/JPY reach 160 or does Japan intervene? Intervention would soften the dollar and give the won room; a quiet break above 160 adds pressure on the won.
Does Tokyo's TOPIX fall for a second day today? If so, "chips carry the index, rates weigh on everything else" becomes the default setup for the KOSPI on 9/28.

[Interpretation] The most informative one is #2. Hormuz was the only thing that pulled stocks back last night, and it is the only driver that could lower oil and yields at the same time.

6. Study Note: What Is a Treasury "Buyback," and Why Didn't It Cap Yields?

Concept

A buyback is when the government repurchases its own older bonds before they mature. The US Treasury has run regular buybacks since 2024. The goal is not to push yields down but to keep the market working smoothly by retiring older, thinly traded issues (liquidity support).

Analogy

Think of a used bookstore owner buying back a few dusty old titles to tidy the shelves. The shelves look better, but it doesn't change the price of new books.

Why it didn't cap yields

Yesterday's buyback was about $4 billion. The 7-year auction alone that day sold $44 billion, and this week's 2-, 5- and 7-year auctions together are far larger. The government sold more than ten times what it bought. And the real reasons yields are rising (oil, hot growth, Fed signals of another hike) have nothing to do with buybacks.

Applying it to Korea

Korea's government also announces bond purchases or smaller issuance when rates get unstable. When you see that headline, put "how much will it buy" next to "how much will it sell over the same period." The comparison tells you how much it can really do.

One-line takeaway

What sets yields is not how much the government buys, but how much more it has to sell.

7. Corrections & Open Items

🟡 Note on setting 9/24 AM checkpoint 1

We set the 7-year bar at "tail of 1bp or more." The result was 0.7bp, with weaker bid-to-cover and foreign demand. The verdict stays ❌, but future auction checkpoints will use "two of three: tail, bid-to-cover, indirect share" so a single number doesn't decide it.

🟡 Index change discrepancy

TheStreet reported S&P −0.51% and Nasdaq −0.78%, which appear to be intraday figures. Yahoo Finance's closing data (S&P −0.02%, Nasdaq +0.01%) and NBC's "closed flat" agree, so we use the latter.

Source discrepancies (values adopted)

Item · Discrepancy · Adopted / US 10-year close · ~5.10% (Yahoo) / 5.12% (Trading Economics) · ~5.11–5.12% / 10-year intraday high · 5.17% (Yahoo) / 5.18% (NBC) · 5.17–5.18% / WTI · $94.61 +2.6% (NBC, CNBC settle) / $94.22 −0.41% (Yahoo; likely different contract or time) · $94.61 settle / October hike odds · 69.7% / 73.5% / 77.5% · "above 70%"

🔲 Not yet confirmed

Item · Status / SOX and Micron 9/24 closes · 🔲 third edition / Offshore won/dollar NDF · 🔲 third edition / Dollar index 9/24 close · 🔲 / Japan and Germany 10-year levels · 🔲 only "multi-decade high" descriptions / Tokyo's 9/25 session · 🔲 already open at publication

8. Event Calendar (KST)

Time (KST) · Event · Korea / Fri 9/25 09:00 · Tokyo open (in progress) · 🔴 Closed (Chuseok) / Fri 9/25 night – Sat 9/26 early · US Friday session · weekly close · checkpoints 1 & 2 · 🔴 Closed / 9/26–27 · Weekend · Hormuz talks · 🔴 / Mon 9/28 08:00 · Next brief: Korea reopening preview (four days in one) / Mon 9/28 09:00 · Korean market reopens · 🟢 / Wed 9/30 21:30 · US August PCE · Q2 GDP third estimate / Thu 10/1 early · Micron earnings (US 9/30 af...

Sources

Market Analyst · This brief does not recommend buying or selling any individual stock. It offers macro context and interpretation.

📌 More market briefs

※ This post is market observation and organization intended to support your own judgment. It is not a recommendation to buy or sell any security. Figures are as of the time of writing and may differ between sources.

If this was useful, bookmark it or pass it along. It genuinely helps me write the next one.

I can survive. We can survive.