Read any policy-loan brochure in Korea and you will find two interest rates: the one the bank charges, and the subsidy rate the city pays on the borrower’s behalf. What the borrower actually carries is the gap between them — and no published dataset states that gap. The two numbers are disclosed separately, bank by bank.
So we subtracted. On 11 August 2026 we pulled the average loan rate and average subsidy rate that Seoul Open Data Plaza publishes per bank, and across 12 banks and 16,769 loans made between 11 May and 10 August 2026, the loan-count-weighted effective rate came to 2.65%. And the first thing the ranking showed — the best bank and the worst bank — turned out to rest on a sample of one loan each.
1. What does it cost? A weighted average of 2.65% (our own calculation). The published average loan rates run 4.54% to 4.91%, and Seoul pays 1.80 to 2.50 points of that.
2. Does choosing a bank help? Barely. Among the nine banks that made 100 loans or more, the effective rate runs 2.59% to 2.74% — a spread of 0.15 points.
3. Then where does it split? Unexpectedly, on the subsidy side. Across those same nine banks the loan rates span 0.06 points while the subsidies span 0.10. What the city pays varies more than what the banks charge.
First, what kind of money this is
These are not personal loans. They are loans made under Seoul’s SME promotion fund through its commercial bank cooperation channel — that is, business lending to small and medium enterprises and small business owners located in Seoul. These figures do not belong next to mortgage or personal credit rates.
| Item | Detail |
|---|---|
| Programme | Seoul SME promotion fund — commercial bank cooperation funds |
| 2026 size | KRW 2.02 trillion |
| Rate type | Variable — the loan rate moves with its benchmark |
| Applications | Seoul Credit Guarantee Foundation (1577-6119) |
| Fund types | Economic revitalisation, start-up, inclusive finance, rapid relief, refinancing, restart and others — ten in all |
| Published data | Seoul Open Data Plaza — per bank, the min, max and mean for the last three months of lending |
The word “variable” puts an expiry date on every figure here. The 2.65% holds at the current benchmark, and if the base rate moves, the loan-rate side follows. Nothing in the data promises the subsidy will move with it. How to think about the fixed-versus-variable choice itself is set out in the fixed versus variable guide.
Twelve banks, ranked by effective rate
| Bank | Effective | Loans | Loan rate | Subsidy |
|---|---|---|---|---|
| Gyeongnam [only 1] | 2.22% | 1 | 4.72% | 2.50% |
| K Bank | 2.59% | 767 | 4.55% | 1.96% |
| NH Bank | 2.60% | 1,008 | 4.54% | 1.94% |
| Saemaul | 2.62% | 585 | 4.54% | 1.92% |
| Shinhan | 2.62% | 2,021 | 4.58% | 1.96% |
| Woori | 2.64% | 3,346 | 4.57% | 1.93% |
| Hana | 2.65% | 4,045 | 4.57% | 1.92% |
| Kakao Bank | 2.66% | 1,488 | 4.55% | 1.89% |
| KB Kookmin | 2.67% | 2,665 | 4.59% | 1.92% |
| IBK | 2.74% | 816 | 4.60% | 1.86% |
| SC First [only 26] | 2.76% | 26 | 4.69% | 1.93% |
| Suhyup [only 1] | 3.11% | 1 | 4.91% | 1.80% |
Read the table top to bottom and it looks like a 0.89-point spread, from 2.22% to 3.11%. But Gyeongnam at the top made one loan, and Suhyup at the bottom made one loan. A single borrower’s terms became each bank’s “average”, so the two ends of this ranking describe a sample, not a bank.
Keep only the nine banks with 100 loans or more and the range narrows to 2.59% (K Bank) to 2.74% (IBK) — 0.15 points. On KRW 100 million borrowed for a year that is KRW 150,000. To put your own principal through it, change the rate in the loan repayment calculator.
The variation came from the subsidy, not the bank
| Measure | Lowest | Highest | Spread |
|---|---|---|---|
| Loan rate | 4.54% (NH, Saemaul) | 4.60% (IBK) | 0.06 pts |
| Subsidy | 1.86% (IBK) | 1.96% (K Bank, Shinhan) | 0.10 pts |
| Effective | 2.59% (K Bank) | 2.74% (IBK) | 0.15 pts |
Every figure in that table is an average across the bank’s loans. This was the surprise in the data. The twelve banks price this programme almost identically. Among the nine with real volume, everything sits between 4.54% and 4.60% — six hundredths of a point. For a cooperation programme, similar terms make sense. Yet the subsidy Seoul paid ranged from 1.86% to 1.96%, a wider band than the lending itself.
Why the subsidy average differs by bank, this dataset cannot say. There are ten fund types and the subsidy terms may differ by type, so the likeliest explanation is a different mix of fund types flowing through each bank. That mix is not a published field, so we could not verify it. One thing is certain: the effective-rate ranking was not produced by banks competing on price.
“Effective” is a number we made
The “effective” column is not a published field. It is the average loan rate minus the average subsidy rate — our subtraction. Two caveats belong with it.
- It is an average minus an average. It does not pair each borrower’s loan rate with that same borrower’s subsidy. The data publishes bank-level means only, so nothing finer is possible.
- Interest subsidies usually run for a fixed period. They are not guaranteed for the life of the loan, and when the subsidy ends the burden returns to the loan rate. Read 2.65% as the cost while the subsidy is live.
Note too that the weighted and unweighted averages came out almost identical this time (2.65% against 2.66%). That is because the two single-loan banks happened to sit on opposite sides and cancelled out — not a reason to use a plain average. With only one of them present, the figure would have shifted.
Where the volume actually sits
Of 16,769 loans, Hana, Woori, Kookmin and Shinhan account for 12,077 — 72.0% (our own calculation). The two internet banks, Kakao and K Bank, follow with 2,255 loans, 13.4%. And yet the lowest effective rate belongs to K Bank at 2.59%, while Hana, the largest lender by far, sits at 2.65%.
So the busiest bank is not the cheapest one. Whether 0.15 points justifies moving your business banking is a separate question — these loans clear through a credit review at the Seoul Credit Guarantee Foundation, so a good deal is decided a step earlier than the choice of bank. What a small business owner should line up first is set out in the small business owner checklist.
Interest paid is also a business expense. How it lands in the annual return is covered in the comprehensive income tax guide, and the part people confuse with input tax credits is in the VAT guide. The same table sits on our live data page — when the dataset refreshes, that copy moves first.
| 은행 · 실행 건수 | 실부담(평균) |
|---|
Where readers usually get stuck
Will I be offered 2.65%?
No. This is an average over loans already made; an individual rate depends on the fund type, credit standing and guarantee terms. Use it to see the ordering between banks, not as your own number.
Are personal loans in this data?
They are not. Only business loans made through the commercial bank cooperation channel of Seoul’s SME promotion fund. If you want personal lending rates, this dataset is not the answer.
Is there an equivalent outside Seoul?
This dataset is Seoul’s. Other metropolitan cities and provinces run comparable promotion funds, but whether they publish per-bank rates like this varies. We did not verify any region outside Seoul.
Does a missing bank mean it does not offer these loans?
Do not read it that way. The table only lists banks with lending activity in the last three months. Zero activity means no row at all — the data does not distinguish “does not offer it” from “made none this quarter”.
The rate on a policy loan looks like something the bank sets. What actually decided the burden was how much the city paid on the borrower’s behalf — visible only because that figure is published separately.
Sources
- Seoul Open Data Plaza — public dataset — Seoul bank-by-bank loan rates, commercial bank cooperation funds (provided by the Seoul Credit Guarantee Foundation). The published fields include min, max and mean loan rate and min, max and mean subsidy rate per bank. The 12 banks and 16,769 loans here were pulled from this API on 11 August 2026, and the reference window of 11 May to 10 August 2026 came with the data.
- Seoul Metropolitan Government — SME promotion fund — 2026 SME promotion fund lending programme. The KRW 2.02 trillion size, the variable rate, applications through the Seoul Credit Guarantee Foundation (1577-6119) and the ten fund types were confirmed here. That page carries the KOGL Type 4 licence (attribution, no commercial use, no derivatives), so nothing is reproduced from it — only the confirmed facts, in our own words.
- Bank of Korea — base rate history — base rate history table. Because these loans are variable rate, this is the table that governs what the lending side follows. 2.75% from 16 July 2026 (previously 2.50% from 29 May 2025) was read directly from it.
Where to check further
- Your own rate and how long the subsidy runs. The data publishes averages only — the Seoul Credit Guarantee Foundation (1577-6119) advises on terms and subsidy periods by fund type.
- Why subsidy averages differ by bank. The fund-type mix is not a published field, so we could not verify it — the per-fund pages on the same site let you compare subsidy terms by type.
- The same data for other regions. We did not verify any — your local credit guarantee foundation and your province’s open data portal will show quickly whether a comparable dataset exists.
Written as at August 2026. The rate figures were pulled directly from the Seoul Open Data Plaza API on 11 August 2026, and the “effective” column is not a published field but our own subtraction of the average subsidy from the average loan rate. These are business loans to SMEs and small business owners in Seoul, not personal lending, and because the rate is variable the figures move when the benchmark does. This is general information, not financial or tax advice.
