The number of IPO shares you receive in Korea is the sum of two calculations that follow completely different rules. Equal allocation (gyundeung baejeong) divides its pool by the number of subscribers. Proportional allocation (birye baejeong) divides its pool by the deposit you put up. Because half the retail tranche goes through the equal route, someone who deposits 100,000 won and someone who deposits 100 million won receive the same single share there. The gap only opens in the other half.
Miss this structure and you will misread a headline like "1,000-to-1 subscription ratio." That ratio applies to the proportional pool only. The equal pool is decided by account count, not by the ratio.
Headcount splits one pool, deposits split the other
Per Shinhan Securities' subscription guide, "at least 50% of the total offering is allocated 1/N according to the total number of participants," with the remainder distributed in proportion to subscribed quantity and the final ratio. KB Kookmin Bank's explainer likewise treats a 50/50 equal-to-proportional split as the default, subject to the underwriter's policy.
So the math runs in two steps. First, divide the equal pool by the number of subscribers. If the quotient is under one share, nobody gets a guaranteed share and the leftover goes to a lottery. Second, divide the proportional pool by the ratio: your subscribed quantity divided by the combined ratio is your allocation, and fractions below one share are normally discarded.
The deposit itself does not buy allocation. At a 50% deposit rate you simply park half the value of your subscription in advance, and the unallocated portion comes back on refund day. Shinhan notes refunds are paid out sequentially from 7 a.m. on allocation day. The real cost is the few days your cash sits frozen. Reading it alongside how margin loan interest accrues retroactively in Korea makes it clear where those days of interest come from if the deposit is borrowed.
How much deposit buys how many shares
Fixing the assumptions makes comparison possible: a 20,000 won offer price, a 50% deposit rate (10,000 won held per subscribed share), one guaranteed share from equal allocation, and a combined proportional ratio of 1,000 to 1. Fractional proportional shares are truncated.
| Shares subscribed | Deposit (KRW) | Proportional (shares) | Equal (shares) | Total (shares) | Deposit per share (KRW) |
|---|---|---|---|---|---|
| 10 | 100,000 | 0 | 1 | 1 | 100,000 |
| 100 | 1,000,000 | 0 | 1 | 1 | 1,000,000 |
| 500 | 5,000,000 | 0 | 1 | 1 | 5,000,000 |
| 1,000 | 10,000,000 | 1 | 1 | 2 | 5,000,000 |
| 5,000 | 50,000,000 | 5 | 1 | 6 | 8,333,333 |
| 10,000 | 100,000,000 | 10 | 1 | 11 | 9,090,909 |
| 30,000 | 300,000,000 | 30 | 1 | 31 | 9,677,419 |
The rightmost column is the point of the table. Inside the proportional pool, one share costs a fixed 20,000 × 50% × 1,000 = 10 million won of deposit. Multiply the deposit thirtyfold and the per-share cost still bottoms out near 9.68 million won. A ten-share account, meanwhile, walks away at 100,000 won per share thanks to its guaranteed equal share. On capital efficiency alone that is a hundredfold difference.
The step between 500 and 1,000 shares is worth noticing too. Only when the deposit doubles from 5 million to 10 million won does allocation move from one share to two; below that step, extra deposit buys nothing at all in the proportional pool. The practical difficulty is that the location of that step is unknown until the final ratio is fixed.
In the proportional pool, the price of one share is set by the subscription ratio, not by your deposit. Double the ratio and the same share ties up twice the cash.
When equal allocation pays zero
Equal allocation rarely divides cleanly. Holding the equal pool at 500,000 shares and varying only the number of subscribers produces this:
| Subscribers | Per-person share | Guaranteed (shares) | Lottery odds on the remainder (%) |
|---|---|---|---|
| 200,000 | 2.50 | 2 | 50.0 |
| 400,000 | 1.25 | 1 | 25.0 |
| 500,000 | 1.00 | 1 | 0.0 |
| 600,000 | 0.83 | 0 | 83.3 |
| 1,000,000 | 0.50 | 0 | 50.0 |
| 2,000,000 | 0.25 | 0 | 25.0 |
At 600,000 subscribers the guaranteed allocation drops to zero and only 83% of accounts draw a single share by lottery. Past a million accounts, half go home empty. In February 2026, Axvis drew roughly 8.96 trillion won in deposits at a 2,711-to-1 ratio — exactly the kind of crowding that pushes the guaranteed equal share to zero. A high ratio and a large allocation point in opposite directions.
More accounts do not mean more shares
Subscribing to the same deal through multiple brokerages is prohibited, and opening several accounts still counts as one person in the equal headcount. What does vary is the split of allocated volume across underwriters: crowds concentrate at the lead manager, while smaller syndicate members can end up with a thinner queue and a proportionally thicker equal share.
Subscription limits matter as well. At Shinhan, limits run 300% for premium tiers, 200% for the general preferential tier, and 100% for everyone else. A low limit makes the proportional step unreachable; filling a high limit freezes cash until refund day. Limit, deposit, and refund date belong in the same table before any comparison holds up.
What to watch
- The equal allocation percentage — the prospectus states whether it is 50% or more. Above 60% favors small accounts.
- Allocated volume and subscription count per brokerage — the count at day-one close hints at whether the guaranteed equal share will fall below one.
- Deposit rate — 50% is typical, but 100% deals exist and double the cash needed for the same quantity.
- Fractional-share handling — truncation or lottery, spelled out in the notice. Around the one-share boundary that single line flips the outcome.
- Refund date and days of frozen cash — count the days from subscription to refund, and check whether that money is already spoken for.