Korea Curbed 2x Leverage; Retail Moved to 3x in the U.S.

Daily turnover in Korea's single-stock leveraged exchange-traded funds was 12.4 trillion won on July 30. By August 11 it was 700 billion won. That is a drop of 94.35%.

From the regulator's side, that reads as success. In the first nine trading days of July, retail investors lost roughly 8.8 trillion won in these products. Access was made harder, and the product effectively disappeared.

The leverage did not disappear. It relocated.

What the Rules Actually Say

Korea's Financial Services Commission announced the package on July 16 and phased it in. The detail matters.

To buy these products, an account must hold 30 million won (about $21,500) in cash; the old threshold was 10 million won and was tripled on July 31. The minimum trading unit went from one share to twenty. Mandatory education went from two hours to three. New listings were suspended indefinitely, and brokerages and fund managers were barred from advertising the products.

The last piece took effect on August 19: before buying a single-stock leveraged product for the first time, a new investor must complete five trading days of simulated trading on the Korea Exchange website, at least one hour per session. The days need not be consecutive. The commission described the purpose plainly:

"expose investors to features such as negative compounding before they use real money"

The same day, pricing-gap rules tightened as well. The premium or discount a liquidity provider may tolerate fell from 3% to 2% on domestically listed products, and from 6% to 5% on overseas-listed ones.

The Reach of the Instrument

Every one of these rules is real and in force. The question is what they cover.

The text governs leveraged funds listed in Korea and tied to a single stock. Index- and sector-based leveraged funds were left out of the package. For overseas-listed products, only the pricing-gap threshold changed — none of the cash, simulation or education requirements apply to them.

A Korean investor already has an overseas trading account at the same brokerage. From the same screen, that investor can buy SOXL, the New York-listed fund that triples the daily move of the semiconductor index, without holding 30 million won in cash, without five days of simulation, without three hours of education.

That is what happened. On July 16, the day the package was announced, SOXL became Korean retail's largest overseas net purchase at $505.79 million. TQQQ, which triples the daily move of the Nasdaq 100, took in $31.96 million and jumped from 13th to 2nd. Before July 16, neither product was in the year's top 50.

The pace held into August. Over two sessions on August 12 and 13, SOXL took in $662.85 million of Korean net buying. As of August 12, Korean holdings in the fund reached $6.56 billion, making it the fourth-largest U.S. equity position among Korean investors; in April it ranked tenth.

As the domestic product's turnover melted, the offshore leveraged share climbed. Between August 1 and 20, the combined settlement value of the 50 U.S. securities Korean retail traded most was $16.02 billion; $6.95 billion of that — 43.4% — was in leveraged products. A year earlier the figure was 27.8%. Total settlement volume, meanwhile, fell 11.7%. Less trading overall, a larger share of it levered.

Hong Kong stayed open too. Hong Kong-listed products delivering twice the daily move of Samsung Electronics and SK Hynix account for 12.4% of the top 50 securities Korean investors trade there. The product banned in Korea is available on another exchange, from the same account.

The Mechanism: What the Step From 2x to 3x Costs

A leveraged fund's decay does not scale with leverage; it scales with the square of leverage. That is why a difference that looks small produces a result that is not.

This is precisely why Korea imposed the simulation requirement: to show investors the arithmetic before real money meets it. In products the requirement does not cover, the arithmetic is unchanged; only the lesson is missing. More: What Is Leverage? and Volatility.

This Morning in Seoul

On August 27 the Kospi opened 187.91 points, or 2.76%, higher at 6,996.12, approaching 7,000 for the first time in seven sessions. At 9:05 a.m. Seoul time (8:05 p.m. ET the previous evening), SK Hynix was up 5.45% at 1.78 million won and Samsung Electronics up 3.25% at 270,000 won. Foreign investors bought 53.9 billion won at the open and took that past 200 billion won within fifteen minutes.

The cause was singular: Nvidia's results, released overnight. Quarterly revenue reached $96.22 billion, up 106% year over year; the company guided to $108 billion for the current quarter and said memory supply would stay tight through the end of fiscal 2028. For a memory maker, tight supply means pricing power.

The chart above tracks the American producer exposed to the same memory demand, not the move in Seoul. The chart below tracks the index behind TQQQ, the product Korean retail pushed to second place; a leveraged fund traces a different curve from it.

An index rising does not mean the investor who levered into it is made whole. The Kospi is still roughly 25% below its July 19 intraday high of 9,385.59; returning there requires a 34% gain from today's level. Korean retail's average cost basis in SK Hynix was 2.28 million won at the end of July; the stock trades at 1.78 million won today. That gap is about 22%, and closing it requires a 28% rise in the shares.

For the roughly 320,000 accounts liquidated in July, none of this arithmetic applies. Once a position is closed, the recovery has no surface left to touch.

Leverage Is Rebuilding

Margin balances did not stay at their lows either. The balance set a record of 38.4786 trillion won on June 19, fell to 27.4 trillion won by August 4 in the July unwind, and stood at 30.93 trillion won on August 13 — seven consecutive daily increases, a 13% recovery off the trough.

Investor cash deposits moved the other way over the same stretch: from a June 4 peak of 139.69 trillion won to 97.93 trillion won on August 11, then 100.07 trillion won on August 13, some 28% below the peak. Cash in accounts is shrinking while borrowed positions grow.

By JPMorgan's estimate, net assets in Korea-linked leveraged ETFs fell from $50 billion at the end of June to $26 billion, and the bank reckons about 75% of the contraction toward what it considers an appropriate $18 billion is complete. The same note keeps a 12-month Kospi target of 12,500. A completed contraction does not mean leverage has fallen; it means leverage in that particular product has fallen.

The Other Side

An asset management executive quoted by the Seoul Economic Daily in July had already sketched this outcome:

"investment demand will spread to other overseas leveraged products not subject to the regulation"

Another executive quoted by the same paper argued that raising deposit requirements to make investing inconvenient was unlikely to change much in the disposition of investors already accustomed to leverage.

What Is Left

On August 2 the commission moved to give itself emergency authority to adjust leverage ratios when volatility spikes. Whether that authority extends to overseas-listed products has not been clarified. If it does not, the current picture continues as it is; if it does, how Korea would act on a product not listed on its own exchange is a separate question.

That is the concrete thing to watch: a clarification of the authority's scope, and the distance between the margin balance and its 38.48 trillion won June record.

For a restriction to protect an investor, its reach has to be as wide as the investor's. In Korea those two areas are not the same size. Risk management and ETFs begin with knowing exactly where that difference sits.

This piece draws on the Korean Financial Services Commission's July 16 and August 2 rule texts, settlement and turnover data reported by the Seoul Economic Daily, Korea JoongAng Daily, Aju Press and FinanceMagnates, leveraged-ETF asset estimates relayed from a JPMorgan client note, and Nvidia's August 26 earnings release. Kospi and share levels refer to the first hour of the August 27 morning session in Seoul; closing values may differ. JPMorgan's contraction estimate rests on the bank's own methodology and may not match other institutions' figures.