30 Years of History: Where We Stand
Today, the wind seemed to ease. KOSPI −1.23%; KOSDAQ: −2.70%. A quiet day compared with the consecutive circuit breakers through yesterday. But raise the hot-air balloon a little higher and put 30 years into one view, and today's position becomes clear. This forest has fallen this fast only twice in this forest's history.
27 Trading Days Ago, We Were at a Record High
The high remains within sight, defining this episode. The 2000 dot-com bubble fell −88.5% over 4 years and 5 months; the 2008 financial crisis took 15 months. People had time to adapt. This time, they did not.
A Position Not Seen in 30 Years
First, KOSDAQ. Use the same ruler — the worst 60 trading days of each crisis — and the ranking is below. Every figure uses KOSDAQ closing index levels, without mixing indices.
| Rank | Episode | KOSDAQ worst 3-month fall |
|---|---|---|
| 1 | Dot-com crash (2000-05-24) | −55.21% |
| 2 | Global financial crisis (2008-10-27) | −51.50% |
| 3 | 2026 current (today) | −45.92% |
| 4 | Pandemic (2020-03-19) | −33.88% |
| 5 | Asian financial crisis (1997-12-24) | −27.26% |
| 6 | 2022 rate shock | −23.41% |
| 7 | 2018 US–China dispute | −20.70% |
| 8 | 2011 European debt crisis | −15.00% |
| 9 | 2024 emergency martial law | −13.27% |
Deeper than the pandemic, much deeper than the Asian financial crisis. Only dot-com and 2008 sit above today's KOSDAQ. One measure is clearer: cumulative 60-trading-day falls below −45% occurred over 30 years on just 23 days, concentrated in three years.
In between, not once in 25 and a half years did the index enter this zone. That is the KOSDAQ. Switch to the KOSPI and the record is worse.
| Rank | KOSPI episode | KOSPI 20-trading-day fall |
|---|---|---|
| 1 | 2008-10-24 financial crisis | −37.48% |
| 2 | 2020-03-19 pandemic | −33.61% |
| 3 | 2026-07-29 current | −33.19% |
| 4 | 1997-12-12 IMF | −30.23% |
| 5 | 1998-05-26 | −25.93% |
Deeper than the Asian financial crisis. Today's −32.64% ranks among 9,418 windows over 36 years. Only five windows were worse — bottom 0.05%.
The Asian Financial Crisis Took 15 Months; This Took 3
The Asian crisis was deeper overall. The KOSDAQ fell from 1,371.90 on 1997-08-09 to 607.00 on 1998-10-07, a decline of −55.75%. Today it is −47.4%. But the Asian crisis took 15 months to get there; this took 3.
| Episode | Fall / Duration | Speed |
|---|---|---|
| 2020 pandemic | −38.2% / 23 days | −1.661pp/day |
| 2026 current | −47.4% / 63 days | −0.752pp/day |
| 2024 emergency martial law | −16.9% / 25 days | −0.676pp/day |
| 2011 European debt crisis | −24.8% / 37 days | −0.670pp/day |
| 2008 financial crisis | −68.5% / 316 days | −0.217pp/day |
| Asian financial crisis | −55.8% / 339 days | −0.172pp/day |
| Dot-com bubble | −88.5% / 1,077 days | −0.082pp/day |
The Asian crisis saw −18.6% in 1997 and −23.6% in 1998: an attritional war eroding 20% per year. Today, 47% in one quarter. Only the pandemic was faster, and it ended in 23 days. This episode is on day 63.
Verdicts on Yesterday's Five Observations
Then last night's FOMC. Before the meeting, consensus favored a hold; attention centered on September hike odds. The result was a hold (3.50–3.75%), but three dissented in favor of a hike. Markets now weigh a hike this year rather than a cut. The hawkish side was confirmed.
Beyond the Pandemic Low
Yesterday's report gave 1.045 (estimated), using the 7/28 balance because 7/29 margin data were not yet released. Today's observation confirmed 1.029, with today added at 1.058. That is the upward turn at the green line's end.
Meanwhile, investor deposits rose instead: KRW 103.88tn (7/22) → 109.59tn(7/29), over 5 days, +5.71tn. Stock-sale proceeds have not left the market; they are sitting as cash. Debt did not fall, cash accumulated, and only the index declined.
China Was the Epicenter
#04 called the trigger the AI multiple. Today that trigger acquired a name. China's ChangXin Memory Technologies (CXMT) listed on July 27 and surged +466% on its first day. It raised CNY 57.9 billion (about KRW 11.6 trillion; #04's KRW 8.6 trillion came from pre-listing reports), Asia's largest IPO in 2026, with intraday market cap of KRW 802 trillion. On top came Chinese localization of DUV lithography equipment.
What collapsed was the competitive landscape rather than earnings, and the speed of price increases rather than prices.
Prices are still rising. What turned down was the rate of increase. The market responded to slowing, rather than falling prices. Combine that slowdown with record earnings and the phrase peaking appeared.
The two companies' consensus-estimate P/Es are now in the 4x range: Samsung Electronics 4.44x, SK hynix 4.20x. Shares are −45% and −56% from 52-week highs while earnings rose 18x. Foreign buyers' argument is that earnings like these make 4x P/E unreasonable. Sellers' argument is that those earnings are the peak, and 4x is a normal multiple for peak earnings. Today's market split over this proposition.
Korea Was Hit Harder
| Market | 1 month | 3 months | 1 year |
|---|---|---|---|
| KOSPI | −34.01% | −15.44% | — |
| KOSDAQ | −30.62% | −47.42% | — |
| Philadelphia Semiconductor (SOX) | −23.79% | +4.10% | +45.54% |
| TAIEX | −13.43% | +1.60% | +45.38% |
| Nikkei | −11.70% | +2.20% | +22.72% |
| Nasdaq | −5.33% | −0.90% | +4.13% |
| S&P 500 | −1.67% | +2.48% | +5.94% |
| Dow | −1.13% | +4.99% | +6.46% |
| Hang Seng | +13.01% | +0.70% | +0.22% |
S&P −1.67%, Dow −1.13%. The broader US market is intact. SOX alone is −23.79%, yet even SOX over 1 year remains +45.54%. It is giving back gains rather than collapsing. Taiwan, despite TSMC, is −13.43%, but Korea is −34%. The extra 7–10 percentage points of damage is Korea-specific, and its name is leverage.
Single-stock leveraged ETFs/ETNs based on Samsung and SK hynix had accumulated KRW 6.8 trillion. In the slump, these products traded KRW 5.4 trillion a day and individual products recorded −26~29% as they sold themselves. Daishin Securities expects meaningful easing of liquidation pressure in 3–4 weeks.
A Crash Without Fear
FX is the strongest evidence. USD/KRW was KRW 1,554.9 on July 1, its highest since March 2009. Today it is KRW 1,438.64. While the KOSPI collapsed −34%, the won appreciated 7.5%. In the Asian and global financial crises, the won fell sharply as capital left. Today's direction is exactly the opposite.
Crypto says the same. In July, as stocks collapsed, Bitcoin rose +10.7% and Ethereum +22.4%. It did not act as Nasdaq's high-beta counterpart as in the past. Even within crypto, ONDO, which handles tokenized government bonds and stocks, alone rose +57.8% over 3 months — even as Bitcoin fell −17.1% over the same period. Money did not flee risk; it moved toward assets that pay interest.
The market fears Interest rates . The US 30-year yield hit 5.21%, its highest since 2007. MOVE jumped +8.83% today, and three FOMC members dissented in favor of a hike. This is position liquidation, not a systemic crisis.
Oil is recovering the −12% drop that followed Sunday's reports of a halt to attacks — last night, Iran declared the ceasefire over, and Brent returned to $87.5 → $92.1. This news also accounted for half of the Dow's worst session since April. The test left by #04 — whether traffic through Yanbu and Bab el-Mandeb holds up in the next weekly observation — remains in place. Rising oil also strengthens §08's conclusion, because oil fuels the interest-rate fear troubling the market.
The Three Investor Groups, and What Comes Next
This is not "foreign versus domestic." Pension funds' No. 1 July net purchase was also SK hynix (+KRW 425.8bn), while they were net sellers of Samsung Electronics (−KRW 111.5bn). Foreign investors and pension funds bought semiconductors on the same side; retail investors stood opposite them. Pension funds fill KOSPI positions first not because they rule out KOSDAQ, but because KOSDAQ trading value is just 12% of KOSPI's: buying pushes prices up. It is a question of sequence.
① Trading money displaced by regulation. What dried out KOSDAQ was retail money cycling through single-stock leveraged products at KRW 5.4tn a day (§07). That game closes tomorrow (7/31) — minimum deposits, premium/discount controls, and financial firms' own restrictions. The government chose this "casino closure" over a rescue through a market stabilization fund. Paradoxically, it could channel that money back to its former habitat: small and mid-cap KOSDAQ stocks.
② Cash waiting on the sidelines. Investor deposits are KRW 109.6tn (+KRW 5.7tn / 5 days) — the money from sales has not left the market (§05). The trigger for re-entry is liquidation intensity turning lower through the numerator (liquidation) channel.
③ Foreign investors. 3 consecutive days of buying, totaling +KRW 643.6bn. Short balances have already shrunk −10.5% in July, using up the fuel for covering. If buying continues after balances stop falling, it is bargain buying.
Pension funds come last (see above). There are also three headwinds: the BOK's hiking stance, residual liquidation intensity of 1.058, and shattered retail sentiment. The test is whether KOSDAQ trading value (baseline KRW 7.2tn) and its share of trading value (17.5%) recover in the first two weeks of August. If water returns, returning money will come before new money.
The calendar ahead — when we can check
Takes effect · July close
Decision at midday
09:00
08:00
21:30 KST
Takes effect
Earnings
Symposium