A jeonse loan in Korea is not money the bank lends on its own. Most of it goes out behind a guarantee from an institution such as the Korea Housing Finance Corporation. With the same structure standing behind the risk, you would expect similar pricing. Put the published rates side by side and that is not what you get.
On 11 August 2026 we pulled the per-bank jeonse loan rates that the Korea Housing Finance Corporation publishes weekly. For the week of 3 to 9 August 2026, 12 banks returned a rate. The lowest was 3.79%, the highest 5.78% — a spread of 1.99 points. On a KRW 100 million loan that is KRW 1.99 million a year.
1. How wide is the gap? 3.79% to 5.78%, 1.99 points. Drop the single highest and 3.79% to 4.73%, 0.94 points still remains.
2. Who is cheaper? In this week’s publication the three internet-only banks averaged 3.99% against 4.21% for the four large commercial banks (our own calculation) — a difference of 0.22 points.
3. Is this my rate? No. This is the rate the bank applies; what you are quoted depends on the guarantee terms, your income and the product. Use it for the ordering between banks.
The 12 banks that reported this week
| Bank | Applied rate | Annual interest on KRW 100m |
|---|---|---|
| Kakao Bank | 3.79% | KRW 3.79m |
| KB Kookmin | 4.04% | KRW 4.04m |
| K Bank | 4.06% | KRW 4.06m |
| Toss Bank | 4.11% | KRW 4.11m |
| Woori | 4.13% | KRW 4.13m |
| Busan Bank | 4.20% | KRW 4.20m |
| Hana | 4.28% | KRW 4.28m |
| Shinhan | 4.38% | KRW 4.38m |
| IBK | 4.41% | KRW 4.41m |
| NH Bank | 4.42% | KRW 4.42m |
| Gyeongnam Bank | 4.73% | KRW 4.73m |
| Jeju Bank | 5.78% | KRW 5.78m |
The annual interest column is simply the rate applied to KRW 100 million (our own calculation). Jeonse loans commonly repay principal at maturity, so this stays close to reality — but on an amortising product the figure changes. To try different repayment structures, use the loan repayment calculator.
Once the list is sorted, the mean and the median part company. The plain average of the 12 is 4.36% while the median is 4.24% (our own calculation). Jeju Bank’s 5.78% alone lifts the mean by 0.12 points. Quote “an average of 4.36%” on its own and you overstate what most borrowers are looking at.
Grouped by the kind of bank
| Group | Banks | Average |
|---|---|---|
| Internet-only banks | Kakao 3.79 · K Bank 4.06 · Toss 4.11 | 3.99% |
| Four large commercial banks | Kookmin 4.04 · Woori 4.13 · Hana 4.28 · Shinhan 4.38 | 4.21% |
| Policy and special banks | IBK 4.41 · NH 4.42 | 4.42% |
| Regional banks | Busan 4.20 · Gyeongnam 4.73 · Jeju 5.78 | 4.90% |
The grouping is ours, not a category in the published data. And with only two or three banks in each group, one member moves the average a long way — take Jeju out of the regional group and 4.90% becomes 4.47%. This table is worth reading for direction and no further.
The direction is clear enough, though. Two of the three lowest rates this week belong to internet-only banks. The usual explanation is a different branch cost structure, but this dataset cannot confirm any reason.
What arrived, and what did not
This feed went dark on 3 August and came back on 11 August. Opening the revived response showed it had only half returned. Our own check caught it: on every row the arithmetic base rate plus spread equals applied rate should hold, and what actually came through was 0 + 0 = the applied rate.
| Field | State | So in this article |
|---|---|---|
| Applied rate | Normal — 12 banks | Printed as received |
| Base rate | All zero | Not printed |
| Spread | All zero | Not printed |
| Guarantee ratio | All zero | Cannot split rates by guarantee ratio |
| Banks with no value | KDB, Jeonbuk, SC First, Gwangju, Suhyup, IM — 6 | Excluded from the table |
“0%” and “no value returned” are different things. Treat the six silent banks as zero and the average comes out at a nonsensical 3.36%. So every average here is computed over the 12 banks that reported. For the same reason we will not declare which bank is cheapest — one of the six missing banks could be lower.
The same table sits on our live data page. Because publication is weekly, opening it next week will show different numbers.
| 은행 | 적용금리 |
|---|
What gets decided before the rate does
Before chasing a fraction of a point between banks, several things are settled earlier.
- The guarantor and the guarantee ratio. Jeonse loans go out behind a guarantee, and the guarantee ratio feeds into the rate. This week every ratio came back as zero, so we could not split the data along that axis.
- Whether to take jeonse at all. Loan interest and monthly rent have to be computed side by side, and the tax treatment differs on each side, so a naive comparison fails. The arithmetic is in jeonse versus monthly rent.
- How large a deposit to agree. Change the deposit-to-rent mix and the loan you need changes wholesale — the deposit-to-rent converter lets you move the split.
- Fixed or variable. Most jeonse loan products are variable rate. How to think about that choice is in the fixed versus variable guide.
If the lease is being renewed rather than newly signed, the renewal terms come before the loan. We counted how often the statutory renewal right is actually used, from the filing data itself, in the lease renewal right article.
Where readers usually get stuck
Can I apply at the rate in the table?
Not as printed. The published figure is the rate the bank applies; what an individual is quoted turns on credit standing, income, guarantee product and any preferential terms. The table is for ordering, not for quoting.
Why does it change every week?
Publication is weekly, and most products are variable rate, so they move with their benchmark. Everything here is as at the week of 3 to 9 August 2026.
Do the six silent banks not offer these loans?
Do not read it that way. Their values were simply absent from this week’s response, and whether they offer the product cannot be told from this data. Ask the bank directly.
Is jeonse loan interest deductible?
Article 52(4) of the Income Tax Act lets a household head with employment income who owns no home deduct 40% of principal and interest repaid on housing lease borrowings from employment income, capped at KRW 4 million a year shared with housing subscription savings. The commonly missed part is that it covers repayment of principal and interest, not interest alone.
Same guarantee behind them, same week of publication, and nearly two points between the banks. Even in a week when the data came back only half filled, the ordering was still visible.
Sources
- Korea Housing Finance Corporation — Korea Housing Finance Corporation. The public institution that runs jeonse loan guarantees and other housing finance guarantees; the per-bank rates in this article are the figures it publishes weekly.
- Korea Public Data Portal — Korea Housing Finance Corporation dataset — jeonse loan rate information. The 12 applied rates here were pulled from this dataset on 11 August 2026, and the reference week (3 to 9 August 2026) came with the response. The base rate, spread and guarantee ratio all arriving as zero was observed in the same response.
- Korea Ministry of Government Legislation — statute text — Income Tax Act, article 52 (special deductions). Source for “40 per cent of principal and interest repaid” in paragraph (4) and the KRW 4 million annual combined cap.
- Bank of Korea — base rate history — base rate history table. The table that governs what variable-rate products follow. 2.75% from 16 July 2026 was read directly from it.
Where to check further
- Your own rate. The published figure is what the bank applies — the Korea Federation of Banks consumer portal comparison and each bank’s own quote are where your number comes from.
- The six banks with no value. Availability cannot be told from this data — calling the bank’s customer centre and asking whether they write jeonse loans is faster.
- How the guarantee ratio changes the rate. Every ratio came back zero this week, so we could not split it — the Korea Housing Finance Corporation customer centre (1688-8114) explains terms by guarantee product.
Written as at August 2026. The rate figures were pulled directly from the Korea Housing Finance Corporation dataset on the Public Data Portal on 11 August 2026 and are printed without rounding. The reference week is 3 to 9 August 2026, and because publication is weekly the figures change. The six banks that returned no value are excluded from every average — their rate is not zero, the data simply did not arrive. This is general information, not financial advice.
