Korea's 2026 market crash, explained: what broke, what it cost, and where the KOSPI stands now

For most of the first half of 2026, the Korean stock market was the best story in global equities. Then, over about five weeks, it became a cautionary one.

Both things were caused by the same feature of the market, and if you only read the headlines from either phase you will draw the wrong conclusion about what Korea is. This piece is an attempt to describe the mechanism plainly — what rose, why, what broke, and what the market actually looks like as of 12 August 2026.

The scale of the move, in two numbers

The KOSPI roughly doubled in the first six months of 2026, peaking at an all-time high of 9,385.59 on 19 June. That is not a normal year; that is not even a normal bull market year. It is one of the largest half-year gains any major national index has produced in the modern era, and it happened in a market that had spent the preceding decade as a byword for a valuation discount.

Then, from that 19 June peak into early August, the index gave back about 44%. Estimates of the market value destroyed cluster above $2tn — a drawdown deeper, in percentage terms, than either the 1997 Asian financial crisis or the 2008 global financial crisis.

Who was actually buying

The single most important fact about the 2026 rally is that foreigners did not drive it. They sold into it, heavily.

Foreign selling of that magnitude alongside a doubling index is genuinely unusual, and the explanation is largely mechanical rather than bearish. Three things were happening at once:

1. Profit-taking on the memory names. Foreign institutions had been overweight Samsung Electronics and SK hynix through the AI memory cycle. When those positions ran, they trimmed.
2. Benchmark-weight compression. As Korean stocks surged, Korea’s weight inside global and emerging-market indices rose sharply. Active managers with position limits and tracking-error budgets were forced to sell simply to stay inside their mandates. This is the least intuitive driver and probably the largest.
3. The won. A weakening currency erodes dollar-denominated returns even when local prices rise, which pushed some unhedged foreign holders towards the exit.

Meanwhile domestic retail absorbed most of it — net buying figures for the half range from ₩64.6tn to ₩97tn depending on whether you count direct equity purchases alone or include ETF flows. That is the setup that matters, because the marginal buyer determines what happens on the way down.

What actually broke: the leverage stack

The rally’s ending was not caused by a bad earnings print or a policy shock. It was caused by the structure of who owned the market and how.

Single-stock leveraged ETFs. In early 2026, Korean regulators approved leveraged ETFs tracking individual companies. These funds use derivatives to deliver a multiple of one stock’s daily move — and they rebalance daily to maintain that exposure, which means they must buy more as the stock rises and sell as it falls. That is fine when a market drifts. It is combustible when the underlying is a mega-cap with a 400%-plus annual move and the fund holders are individuals.

Margin debt. Retail buying was substantially financed. When prices turned, brokerages issued margin calls; unmet calls triggered automatic liquidation; liquidation pushed prices lower; lower prices triggered the next tranche of calls. Margin debt peaked at a record ₩38.63tn on 24 June; total forced liquidations then reached roughly ₩2.3tn in about two and a half months, with around 1.2 million retail accounts hit by margin calls.

Index concentration. Samsung Electronics and SK hynix together account for a very large share of KOSPI capitalisation, and they drove most of the index’s 2026 gain. An index that concentrated does not behave like a diversified market — it behaves like a leveraged bet on one industry’s capex cycle. On the way up that looked like genius. On the way down it removed any cushion.

Put the three together and you get the outcome: a 10.84% single-day fall, followed the next session by a record 17.91% surge on 31 July. Across 2026 the Korea Exchange has triggered seven market-wide circuit breakers and 36 trading halts — already more than half of all such events since the mechanism was introduced in 2000, with three market-wide halts inside about three weeks in July alone.

Where the market stands in August 2026

Where the floor actually sits is genuinely contested. Some accounts put it around the 6,200 level in the first week of August; others record a close of 5,593.56 as the rebound faded. Both are defensible depending on the session measured, and the honest description is a volatile 5,600–6,200 band rather than a settled bottom. Sell-side range forecasts cluster at 6,000–7,000, with foreign flows, the semiconductor cycle and CPI treated as the swing factors.

The honest summary is that published range forecasts cluster around 6,000–7,000, but individual twelve-month targets vary far more widely than that. A spread that broad is not a forecast; it is an admission that nobody knows which regime the market is in.

What this does and doesn’t tell you about Korea

It does not tell you the AI memory cycle was imaginary. Korea genuinely supplies a large share of the world’s high-bandwidth memory, and that demand is being driven by capital expenditure decisions taken far outside Korea.

It does tell you three things worth carrying forward:

• The KOSPI is a sector bet wearing an index costume. If you buy the benchmark, you are overwhelmingly buying memory semiconductors. Size the position as you would a sector fund, not a country allocation.
• Market-structure reform cuts both ways. Korea has spent years liberalising to attract capital. Leveraged single-stock products arrived as part of that modernisation and materially worsened the drawdown. Both facts belong in the same sentence.
• Retail depth is now a first-order variable. A domestic investor base large enough to absorb ₩148tn of foreign selling is a genuine structural change in this market. It cut the fall’s floor higher than it would have been in 2008 — and it also supplied the leverage that made the fall so fast.

What we’re watching next

The September launch of Korea’s revamped FX trading infrastructure is the next structural event, and it matters more for foreign flows than any single earnings print — a friendlier won market removes one of the three reasons foreigners sold in the first half. We’ll cover it when it lands.

Figures in this article are current as of 12 August 2026 and are sourced to the outlets and data portals listed below. Where estimates differ between sources, we have given the range rather than picking one.