Korea's retail leverage gauge: margin loans, deposits and forced selling
Updated September 2026 · Data through 1 September 2026
Korea has one of the most active retail investor bases in the world, and unusually, it publishes the plumbing behind that activity every day. The Korea Financial Investment Association (KOFIA) releases three series that together work as a leverage-and-sentiment gauge: the outstanding balance of margin loans, the cash sitting idle in brokerage accounts, and the amount of stock liquidated by brokers to cover unmet margin calls. This guide explains each and puts the 2026 readings in context.
1. Margin loans (credit balance)
"Margin loans" here means sinyong yungja: money brokerages lend to individual clients to buy stocks, secured against the stocks themselves. The balance is the best single measure of how much borrowed money is in the Korean equity market, because Korean retail leverage runs almost entirely through this channel rather than through derivatives.
The series is small before 2007 (under ₩1 trillion), jumps to ₩7 trillion in the 2007 fund boom, collapses to ₩1.5 trillion by the end of 2008, and then grows with the market: ₩10 trillion by 2017, ₩19.5 trillion at the 2020 peak, and a record ₩25.7 trillion on 13 September 2021 — the top of the pandemic retail cycle. It took until late 2025 to exceed that level. On 24 June 2026 the balance hit ₩38.6 trillion, a new record, before the July sell-off cut it by ₩8 trillion in a month. It stood at ₩33.4 trillion on 1 September.
One nuance: relative to the size of the market, 2026 leverage is lower than 2021. Margin loans were 0.95% of combined KOSPI and KOSDAQ market capitalisation at the 2021 peak but only 0.52% at the 2026 peak, because market value had roughly tripled. Another: the KOSDAQ share of margin debt has fallen from ₩11 trillion (2021) to ₩7 trillion, a sign that retail leverage in this cycle went into KOSPI large caps rather than small-cap speculation.
2. Investor deposits
Investor deposits are the cash balances individuals hold at brokerages: money that has been moved into a trading account but not yet invested. It is widely read as "dry powder". The series roughly doubled in the 2020 boom (₩27 trillion at the end of 2019 to ₩65 trillion a year later), stayed there through 2021, and drifted down to the mid-₩40s and ₩50s trillion in 2022–24.
2026 set records that make 2020 look modest. Deposits reached ₩139.7 trillion on 4 June 2026. They then fell 30% in three months, to ₩98.2 trillion on 1 September, the sharpest outflow of brokerage cash on record. Some of that cash bought the July dip; a good deal of it, according to Korean press reports on retail flows into US stocks, left the domestic market altogether.
3. Forced selling
When a margin-financed position falls enough that the collateral no longer covers the loan and the client does not top it up, the broker sells the stock at the next open. KOFIA publishes the daily amount of such forced sales and their share of receivables outstanding. It is a small series most of the time — a few billion won a day — which is exactly why spikes are informative.
The all-time spike came in October 2023, when a cluster of thinly traded stocks including Yeongpoong Paper fell by the daily limit for several consecutive sessions, trapping leveraged retail positions. Forced selling hit ₩550 billion on 20 October 2023, and on the previous day 69% of outstanding receivables were liquidated. In 2026, the daily average for June was ₩53.5 billion and for July ₩45.1 billion, the highest monthly averages on record and roughly seven times the 2025 norm.
2026 month by month
| Month-end | Margin loans (₩ tn) | Deposits (₩ tn) | Forced selling, daily avg (₩ bn) |
|---|---|---|---|
| Dec 2025 | 27.3 | 87.8 | 7.1 |
| Jan 2026 | 30.3 | 106.0 | 10.2 |
| Feb | 32.7 | 118.7 | 13.5 |
| Mar | 32.9 | 110.3 | 26.2 |
| Apr | 35.7 | 124.8 | 12.0 |
| May | 38.0 | 131.6 | 39.3 |
| Jun | 37.3 | 121.6 | 53.5 |
| Jul | 28.9 | 104.1 | 45.1 |
| Aug | 33.3 | 99.7 | 15.4 |
The long view
| Year | Margin loans, peak (₩ tn) | Year-end | of which KOSDAQ | Deposits, year-end (₩ tn) |
|---|---|---|---|---|
| 2007 | 7.0 | 4.5 | 0.9 | 9.5 |
| 2008 | 4.4 | 1.5 | 0.3 | 9.2 |
| 2011 | 6.9 | 4.4 | 1.5 | 17.4 |
| 2015 | 8.1 | 6.5 | 3.5 | 20.9 |
| 2017 | 10.2 | 9.9 | 5.4 | 26.5 |
| 2018 | 12.6 | 9.4 | 4.6 | 24.8 |
| 2019 | 10.8 | 9.2 | 5.2 | 27.4 |
| 2020 | 19.5 | 19.2 | 9.6 | 65.5 |
| 2021 | 25.7 | 23.1 | 11.0 | 67.5 |
| 2022 | 23.8 | 16.5 | 7.8 | 46.4 |
| 2023 | 20.6 | 17.6 | 8.5 | 52.8 |
| 2024 | 20.2 | 15.8 | 6.6 | 54.2 |
| 2025 | 27.5 | 27.3 | 10.2 | 87.8 |
| 2026* | 38.6 | 33.4 | 7.0 | 98.2 |
* 2026 as of 1 September. Source: KOFIA. Deposit and margin series are published with a two-business-day lag.
How to read the three together
- Loans rising with deposits is a healthy boom: new money is arriving and part of it is leveraged (early 2026).
- Loans high while deposits fall is the fragile state: the cushion under leveraged positions is thinning. Margin loans were 34% of deposits on 1 September 2026, up from 31% in June and comparable to the 38% at the 2021 top.
- A forced-selling spike after a sharp drop usually marks the point of maximum stress in a correction, because it removes the most fragile holders. July 2026 fits the pattern; whether it marked a durable low is a different question.
- None of this is about foreigners. These series cover domestic individuals only. Pair them with the daily retail flow rankings to see what the leveraged money is actually buying.
This article is general information, not investment advice.