KOSPI Drops 775 Points: Citi Says Take Half Off the Table
KOSPI Plunges 775 Points in Three Days: Citi Warns 'Take Half Off the Table' as Foreign Selling Hits 100 Trillion Won
The numbers from Seoul this week are genuinely hard to process. KOSPI dropped another 3.25% on May 19, closing at 7,271.66. That makes three straight days of selling — 6.12% on Black Friday (May 15), another 4%-plus on May 18, and now this. From the 8,046 level three sessions ago, the index has shed 775 points, or 11.2%. About 680 trillion won in market value — a figure that actually exceeds Samsung Electronics' entire market cap of roughly 500 trillion won — just evaporated.
The VKOSPI volatility index has been brushing against 80 — territory last seen during the March Middle East war flare-up. Circuit breakers on the sell side triggered for the fourth straight day at the open. Nobody I talked to on the floor could remember a stretch quite like this outside of 2008 or the worst days of COVID.
The Crash by the Numbers
The damage spread across almost every sector. Samsung Electronics touched down 8.6% intraday, sliding into the 240,000 won range. SK Hynix wasn't far behind at -7.6%, falling to around 1.37 million won. The gap between analyst targets and reality right now is staggering: consensus targets for Samsung sit at 450,000 to 590,000 won — a 47.8% chasm — while SK Hynix targets of 3.1 million to 4.0 million won represent a 55.8% disconnect from the current price.
Bio got wrecked too. The KOSPI Bio index fell another 4.7%, bringing its May losses to 11.2%. Alteogen dropped 3.1%, LigaChem Bio cratered 15.2%, and HLB lost 5.8%. Robot stocks, which had been riding AI enthusiasm, surrendered early gains to end-of-quarter profit-taking. The only sector that finished green was shipbuilding: HD Hyundai Heavy Industries gained 4.2%, Hanwha Ocean rose 3.8%, and Samsung Heavy Industries added 2.9%.
Behind the scenes, forced selling is escalating. Margin calls averaged 54.8 billion won per day in May — 4.6 times the 12 billion won daily average in April and 2.1 times the 26.2 billion won in March. Credit balances sit at 36 trillion won. When the broker calls and you can't meet the margin, they sell your position at whatever price the market offers. That mechanical pressure feeds the downdraft.
Foreign Exodus and the Citi Warning
Foreign investors sold 6.2 trillion won on May 19 alone. That extends the sell streak to nine consecutive trading days and pushes the 2026 cumulative above 100 trillion won. For context: foreign investors sold about 9 trillion won in all of 2025. We've done 11 times that in five months.
The selling has been extremely concentrated. Merrill Lynch noted in a research report that the bulk of foreign outflows targeted Samsung Electronics and SK Hynix specifically. Their analysis: Korea's weight in the MSCI Emerging Markets index climbed from 8.4% at the end of 2024 to 12.8% by mid-May — a 1.5x increase. When your passive EM fund says "keep Korea at X percent," and the market pushes Korea to X+4 percent, you sell mechanically. That's what's happening.
Then Citigroup dropped a bombshell on May 19. Their note was blunt: the Korean market is "far more overheated" than the US, and investors should take profits on half their Korea exposure. Their math: Korea's 12-month forward P/E sits at 12.5x, above the MSCI EM average of 11.8x. Semiconductors trade at 18.3x versus the global semiconductor average of 15.1x. I think that premium is hard to justify given the macro headwinds, and frankly, I think Citi is right to be cautious — though "take half off" feels more like a headline grabber than actionable advice for most institutional investors.
But the Korean houses aren't buying it. KB Securities analyst Lee Eun-taek called the selling "mechanical rebalancing, not a trend reversal" and held his KOSPI target at 10,500. Korea Investment Securities analyst Kim Dae-jun pointed out that the 8.09x forward P/E for KOSPI represents a 35% discount to global markets — so Citi's warning, in his view, is about short-term overheating risk, not a structural problem. My view is that both sides have a point: Citi is right about the near-term risk, but the Korean houses are right that this isn't a structural crisis. The truth is somewhere in between — and that's what makes this market so hard to trade.
Meanwhile, pension funds are quietly stepping in. The National Pension Service and other domestic pension funds net bought 2.3 trillion won in May — triple April's 800 billion won. Add the expected NPS mid-year rebalancing of 5 to 10 trillion won, and there is genuine institutional buying power forming a floor.
Semiconductor Peak-Out Debate
The real gut punch for semiconductors came from an unexpected direction. Seagate's CEO mentioned during an earnings call that semiconductor supply chain bottlenecks could persist into the second half of the year. That single comment sent the Philadelphia Semiconductor Index (SOX) down 2.47%, and Korean chip stocks followed right off the cliff.
The "memory peak-out" narrative — the idea that the semiconductor cycle has hit its top — now has a fresh data point. Kiwoom Securities analyst Han Ji-young framed it well: if Seagate's comment turns out to be a one-off remark about temporary logistics, semiconductor stocks snap back fast. If it's a genuine signal of demand softening, the correction has room to run. She points to Nvidia's earnings on May 21 as the real watershed moment. I think a strong Nvidia print and the peak-out narrative collapses. A miss, and things get worse before they get better. My view is that Nvidia's report is genuinely the most important single data point for Korean markets this month — more than any KOSPI level or foreign selling number.
Meanwhile, financial regulators are watching the May 27 launch of single-stock leveraged ETFs for Samsung Electronics and SK Hynix with visible anxiety. More than 74,000 investors have already pre-registered for these 2x daily return products. At current daily KOSPI volatility of 4 to 6%, these ETFs could produce daily losses of 8 to 12%. Korea Exchange senior researcher Lee Hyo-seop warned flatly that leveraged ETFs are "short-term trading instruments, not long-term investment products," and pushing leverage in this kind of volatility is asking for trouble.
What This Means for Investors
The raw numbers are frightening: 775 points gone, 100 trillion won in foreign selling, margin calls spiking 4.6x month-over-month. But I think the market is also showing signs that panic may be overdone. The RSI has fallen to 29.3 — below the 30 threshold that marks oversold territory. Samsung Securities published data showing that buying KOSPI when RSI dips below 30 has produced an average 17.3% return over the subsequent six months. My view is that the 12-month forward P/E of 8.09x and EPS growth of 61% don't look like a market that's fundamentally broken — they look like one that's being mechanically repriced by passive flows and risk management algorithms. That mechanical repricing can overshoot, and when it does, the snap-back can be violent.
Three events on May 21 could flip the script: the FOMC minutes, Nvidia's earnings, and any progress on the Samsung Electronics union negotiations. If even two of those break positive, a snap-back toward 7,500-7,800 is realistic. If all three disappoint, we probably test 7,000 and maybe 6,800. The 2008 and 2020 crashes both produced far deeper drawdowns — 53% and 36% — but both were followed by powerful recoveries. An 11.2% correction in three days is fast, but it is not unprecedented, and the structural supports (NPS, pension ETF inflows, deep value P/E) haven't gone anywhere. The smart money tends to show up when everyone else is running for the exits.
My Take: Citi's Warning Was Overdue, Not Overdone
I think Citi's "take half off the table" call is getting more pushback than it deserves. The Korean houses are dismissing it as short-term thinking from a foreign desk that doesn't understand Korea's structural story. But in my view, Citi is simply saying what everyone in the FX and EM bond markets has been saying for weeks: Korea got expensive on a relative basis, and the global macro backdrop has shifted against it.
My view is that the real debate isn't about valuation — it's about time horizon. If you're investing for the next three months, I think you should absolutely reduce Korea exposure. The foreign selling is not going to stop overnight, the won is not going to reverse course without policy intervention, and the macro headwinds from US yields and oil are not going to dissipate in a week. But if you're investing for the next three years, I think the 8.09x forward P/E is screamingly cheap and you should be adding, not reducing.
I think the most honest answer is that Citi is right for the next quarter, and the Korean houses are right for the next three years. The challenge is that most institutional investors don't have the mandate to sit through a 10-15% drawdown while waiting for the long-term thesis to play out. That tension — between short-term pain and long-term value — is what makes this market cycle so difficult to navigate. My take: if you can handle the volatility, buy in tranches starting at 7,000. If you can't, follow Citi's advice and lighten up until the macro picture clarifies.
Related Keywords
- KOSPI 775 point crash 3-day selloff May 2026
- Foreign investors Korea KOSPI 100 trillion won net selling
- Citigroup Korea stock market overheated warning profit-taking
- Samsung Electronics SK Hynix single-stock leveraged ETF risk May 27
- Nvidia earnings May 21 semiconductor peak-out debate impact Korea
Comments
Post a Comment