Industry Insights 11 min read

Why Is Everyone Trading Stocks in South Korea?

South Korea's KOSPI surged over 100% this year, with active trading accounts exceeding the population, driven by tax‑free arbitrage, leveraged retail enthusiasm, and a semiconductor boom centered on Samsung and SK Hynix, raising concerns about market fragility.

Model Perspective
Model Perspective
Model Perspective
Why Is Everyone Trading Stocks in South Korea?

What Is Driving the Surge?

By the end of June 2026 the KOSPI had risen more than 100% year‑to‑date, and total active trading accounts topped 108 million despite a national population of only 52 million. Although the account count exceeds the population, many individuals hold multiple accounts for different purposes, so the estimated number of actual retail investors is about 17 million.

The rapid increase in accounts—over ten million in six months—and a 272% jump in accounts opened by minors are explained by two distinct forces:

Regulatory arbitrage : Korean tax law permits parents to transfer assets to children’s accounts tax‑free up to 20 million won every ten years.

Emotion‑driven buying : Retail investors chasing price gains.

The former is a rational calculation, while the latter signals heightened risk.

This Rally’s Core Drivers

The KOSPI is nominally weighted across 800 companies, but Samsung Electronics and SK Hynix together account for roughly 50% of the index weight. Their performance has powered most of the index’s gains, while the remaining 600‑plus stocks have largely underperformed, making the index effectively a weighted reflection of two semiconductor firms.

Both companies produce high‑bandwidth memory (HBM), the primary memory solution for AI accelerators. Nvidia’s GPUs rely heavily on HBM, and SK Hynix holds about 62% of the global HBM market, with 2026 capacity already fully booked.

SK Hynix reported Q1 2026 revenue of 52.58 trillion won, a 198% YoY increase, and an operating margin above 70%, outpacing even TSMC. Valuation multiples of 5–7× forward earnings are low for the semiconductor sector, suggesting a solid industry‑driven foundation for the rally. U.S. banks have labeled 2026 a “memory super‑cycle” with projected 51% YoY growth in global DRAM revenue.

Historical Parallel: The 1995 PC Memory Cycle

The article compares the current situation to the 1993‑1997 PC memory super‑cycle. In the early 1990s, Windows GUI adoption spurred a surge in PC shipments, driving DRAM demand far beyond supply. Prices rose, Korean firms like Samsung and the predecessor of SK Hynix (Hyundai) posted record profits, and semiconductor exports reached 13.4% of Korea’s total exports.

Subsequent capacity expansion—about 50 new fabs announced in 1995‑1996—created a lagged oversupply. In 1996 DRAM spot prices fell 51% and fell another 65% in 1997, triggering a debt crisis for highly leveraged chaebols and contributing to Korea’s 1997 IMF bailout.

The key difference today is the debt holder: in the 1990s it was corporate balance sheets; in 2026 it is leveraged retail investors.

Who Is Buying and Who Is Selling?

Goldman Sachs estimates foreign investors sold roughly $62 billion of KOSPI stocks by May 2026. This was not a flight but a mechanical reduction of positions because the Korean market’s weight in global and emerging‑market benchmarks had risen sharply, forcing fund managers to stay within allocation limits.

Nomura’s Chetan Seth notes that investors and clients were compelled to sell. Domestic retail investors, however, bought the shares that foreigners sold. Each time foreign institutions reduced holdings and the index fell, Korean retail investors net‑bought, evident during June’s circuit‑breaker events where foreigners sold over 5 trillion won in a day while retail bought about 1.3 trillion won.

Retail buying relies heavily on margin financing. By mid‑June the market’s total margin debt reached 37.8 trillion won, a historic high, with Samsung and SK Hynix alone accounting for 9.1 trillion won—almost three times the amount at the start of the year. New 2× leveraged ETFs on these stocks allow retail exposure of 4–6×.

Yuanta’s Daniel Yoo describes the moves as “purely a leverage effect, unrelated to fundamentals.” A 7% earnings miss by Broadcom in the U.S. triggered a 5.5% KOSPI drop and a circuit‑breaker, followed by an 8% decline the next day, illustrating how modest news can be amplified by high leverage.

Is This a Bubble?

The rally is not a bubble in the sense of having no fundamentals: HBM demand is real, SK Hynix and Samsung’s earnings are solid, and 5–7× valuations are modest. However, the bubble‑like element is the leverage‑amplified retail sentiment that makes the system unusually fragile.

With nearly half the index’s market cap concentrated in two correlated stocks and retail investors holding multiple‑times leverage on the same positions, any adverse signal from the semiconductor supply chain can trigger moves far beyond what fundamentals justify.

The article also draws a parallel to the 2017 “kimchi premium” in Korean cryptocurrency markets, where limited economic mobility led to speculative demand and heightened vulnerability.

The true risk lies not in the semiconductor sector’s continued growth but in whether the leveraged retail structure can unwind orderly if the market reverses. Historical memory cycles ended with capacity expansions; the decisive factor for 2026‑2027 will be whether new fab capacity coming online can sustain HBM prices enough to meet the expectations built into today’s margin‑heavy positions.

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

Market AnalysissemiconductorHBMstock marketleverageSouth KoreaKOSPI
Model Perspective
Written by

Model Perspective

Insights, knowledge, and enjoyment from a mathematical modeling researcher and educator. Hosted by Haihua Wang, a modeling instructor and author of "Clever Use of Chat for Mathematical Modeling", "Modeling: The Mathematics of Thinking", "Mathematical Modeling Practice: A Hands‑On Guide to Competitions", and co‑author of "Mathematical Modeling: Teaching Design and Cases".

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.