Buying a Korean IPO on its first day
and holding: what happens
You applied for the IPO and got a handful of shares, or none. On listing day the price is flying. Should you just buy it in the market? I counted 1,019 listings from 2014 to 2026, from the first day to a year later.
01The rules, and Korea's IPO quirks
Two things about the Korean market first. KOSPI is the main board; KOSDAQ is the secondary board, closer to the Nasdaq, and it is where most IPOs list. The session runs 09:00–15:30 Korea Standard Time (UTC+9).
Korean IPO allocations are heavily retail-oriented and oversubscribed hundreds of times over, so most applicants receive a few shares at most. That creates a large crowd who did not get an allocation and are watching the first-day price. Until June 2023, the listing-day open was set between 90% and 200% of the offer price, and then the normal ±30% daily limit applied — so the best possible first day was the open at double the offer price followed by a limit-up close, a pattern with its own nickname, tta-sang. From June 26, 2023 the first-day band became 60% to 400% of the offer price.
This article is not about people who received shares at the offer price. They mostly did well on day one; everyone knows that. It is about people who bought in the market on the first day.
- Day-one open — buy the moment it starts trading
- Day-one close — watch the first day, buy at the bell
- Day-two open — skip the first day entirely
Holding periods are 1 day, 1 week, 1 month, 3 months, 6 months and 1 year, always selling at the open. 0.22% is deducted on every buy. The baseline is any stock on any day, same holding period.
The listing day is the first daily bar, except for 59 stocks that had pre-IPO over-the-counter prices attached; for those I used the first day with volume above 200,000 shares. SPACs are excluded. The median is the middle result when every trade is lined up from worst to best — half did better, half worse.
02Day one
| Value | |
|---|---|
| Open → close · share up | 39.1% |
| Open → close · median | −2.82% |
| Open → day's high · median | +9.79% |
| Open → day's low · median | −9.38% |
| Closed at the day's low (slid all day) | 13.7% |
| Closed at the day's high (rose all day) | 18.1% |
| Closed more than 10% below the open | 36.2% |
| Closed more than 30% above the open | 6.6% |
Buy at the open and six times out of ten you are down by the close. The day swings 10% each way, which invites the thought "I'll just sell at the top" — but a close 30% above the open happened once in fifteen listings, and a close 10% below it happened once in three.
03Holding on
| Hold | Win rate | Baseline | Average | Median | Baseline median |
|---|---|---|---|---|---|
| 1 day | 53.2% | 42.5% | +1.73% | +0.23% | −0.22% |
| 1 week | 40.1% | 44.4% | +0.10% | −2.92% | −0.41% |
| 1 month | 36.0% | 44.6% | −3.71% | −8.13% | −0.88% |
| 3 months | 38.4% | 43.5% | −2.28% | −9.04% | −2.13% |
| 6 months | 38.5% | 43.2% | +5.97% | −10.43% | −3.26% |
| 1 year | 35.2% | 42.3% | +14.79% | −19.15% | −5.85% |
Exactly one holding period works: buy at the day-one close and sell at the day-two open, 53% winners. There is a small overnight drift after the first close.
After that it only goes down. Median −8.1% after a month, −19.2% after a year. Any random stock over the same periods: −0.9% and −5.9%. New listings did 13 points worse than the market over a year.
The one-year average is +14.8%, positive. A few listings that went up several times over lift the average. But only one buyer in three was ahead after a year, and the buyer in the middle had lost 19%. Do not read this table by its average.
Buying at the open, or waiting until day two, made almost no difference.
| Entry | 1-month win rate | 1-month median | 1-year win rate | 1-year median |
|---|---|---|---|---|
| Day-one open | 35.1% | −10.80% | 33.5% | −24.17% |
| Day-one close | 36.0% | −8.13% | 35.2% | −19.15% |
| Day-two open | 34.7% | −9.53% | 34.1% | −19.73% |
| Any stock, any day | 44.6% | −0.88% | 42.3% | −5.85% |
Timing within the first two days did not matter. Entering in the first two days at all was the problem.
04Since June 2023
From June 26, 2023 the first-day price band widened to 60–400% of the offer price. The stated aim was to let the first-day price find its level faster instead of queuing up at limit-up for days.
| Count | Day-one open→close median | 1-month win rate | 1-month median | 1-year median | |
|---|---|---|---|---|---|
| Before June 26, 2023 | 749 | −0.97% | 39.5% | −4.42% | −14.62% |
| After | 269 | −10.51% | 26.0% | −19.10% | −39.49% |
Since the change, the median listing has fallen 10.5% from its open to its close on day one. The open now prints at a multiple of the offer price and sells off during the session. Buy that close and hold a month, and only one buyer in four is ahead; a year on, the median buyer is down 39%.
The price did find its level faster. The level was far below the first-day open, and it took a year to get there.
05The bigger the first-day pop, the worse
| Day-one move | Count | 1-month win rate | 1-month median | 1-year median |
|---|---|---|---|---|
| Below −10% | 369 | 31.2% | −11.91% | −22.29% |
| −10% to 0% | 251 | 41.8% | −1.24% | −1.71% |
| 0% to +10% | 170 | 41.8% | −1.39% | −12.94% |
| +10% to +30% | 161 | 36.0% | −11.04% | −24.61% |
| +30% to +60% | 62 | 27.4% | −16.98% | −40.52% |
Listings that closed more than 30% above their open — the ones everybody wants — did worst: one buyer in four ahead after a month, median −40.5% after a year.
The least bad group was the quiet one: listings that drifted slightly lower on day one (−10% to 0%). Their one-year median of −1.7% beat the market's −5.9%. Nobody chased them.
06Year by year
| Year | Count | Day-one open→close median | 1-month median | 1-year win rate | 1-year median |
|---|---|---|---|---|---|
| 2014 (from July) | 47 | 0.00% | −0.46% | 48.9% | −0.46% |
| 2015 | 98 | 0.00% | +2.57% | 39.8% | −6.34% |
| 2016 | 60 | −4.41% | −1.13% | 40.0% | −8.34% |
| 2017 | 77 | −4.38% | −1.70% | 42.9% | −8.32% |
| 2018 | 84 | −0.25% | −8.20% | 31.0% | −15.69% |
| 2019 | 91 | −0.27% | −8.36% | 38.5% | −5.18% |
| 2020 | 71 | 0.00% | −2.70% | 47.9% | −3.05% |
| 2021 | 99 | 0.00% | −12.96% | 13.1% | −40.40% |
| 2022 | 80 | −4.97% | −10.71% | 42.5% | −13.98% |
| 2023 | 107 | −0.85% | −14.17% | 26.2% | −47.78% |
| 2024 | 111 | −12.22% | −23.84% | 32.4% | −29.45% |
| 2025 | 79 | −12.50% | −13.61% | 35.7% | −39.01% |
The one-year median was never positive, not in a single year. The IPO boom of 2021 produced listings where one buyer in eight was ahead a year later. 2023–2025, overlapping the rule change, are the deepest.
07Caveats
No offer prices. The data only has post-listing prices, so I could not measure "versus the offer price" and the returns of people allocated at the offer are not here. This is about market buyers on day one only.
Delisted companies are not in the data. Listings that later disappeared are missing; including them would make every number above worse, in the same direction as the conclusion.
Spin-off relistings and SPAC mergers may be mixed in. I excluded only names containing "SPAC". Nineteen REITs are included.
Costs are the 0.22% buy-side deduction only. Spreads on listing day are wide, so reality is a little worse.
08What I take from it
- Buy at the open and 6 in 10 are down by the close (median −2.8%)
- Buy at the close and sell next morning: 53% winners, the only winning window
- After that it keeps falling: −8.1% at one month, −19.2% at one year (any stock: −5.9%)
- The bigger the first-day pop, the worse (+30% and up → −40.5% at one year)
- Since the June 2023 rule change it is twice as bad (−39.5% at one year)
- In 12 years the one-year median was never positive
The people who make money on Korean IPOs are the people allocated at the offer price. Whoever buys from them in the market on day one has, for twelve years with almost no exceptions, paid their profit for them. The first-day price is made by subscription ratios and first-day excitement; the company's value was somewhere below it.
Listing day is the sellers' day.
The buyers' day comes about a year later.
댓글 쓰기