Banks: 0.5%. Savings banks: 5%. Loan sharks: 10%. Mutual aid: 5%. Stacking four tiers side by side reveals two rival definitions of credit.
Bank Default Rates vs. SAF Mutual Aid — Two Definitions of Credit, in Numbers
Korean finance is layered. Commercial banks sit at the top, with card issuers, savings banks, capital firms, and licensed lenders descending in turn. Each layer has its own default rate, and that rate stands as the headline indicator of the layer's "credit strength."
But alongside this stratified column sits a different system, run by an artists' community for itself: mutual aid lending. Place its default rate next to the four layers above, and something strange surfaces.
This piece lays out that comparison and dissects the two definitions of credit the numbers point to.
A Comparison Table — Default Rates, Interest Rates, Borrowers Across Five Systems
| Financial System | Default Rate | Avg. Interest | Primary Borrowers | Underwriting Basis |
|---|---|---|---|---|
| Commercial bank household loans | 0.4–0.6% | 4–6%/yr | Salaried workers, registered businesses | Credit score, income, collateral |
| Card cash advances | 2–3% | 15–18%/yr | General cardholders | Card usage history |
| Savings banks | 4–6% | 10–19%/yr | Those rejected by tier-1 banks | Eased credit thresholds |
| Licensed consumer lenders | 8–12% | 15–20%/yr (statutory cap) | Those rejected by tier-2 | Minimal screening |
| Artist mutual aid (SAF) | ~5% | 2–5%/yr | Artists previously rejected by tier-1 | Community testimony, purpose review |
Averages for 2023–2025. Drawn from Financial Supervisory Service disclosures, press reporting, and SAF internal data. Licensed-lender figures cover registered firms only; the unregistered/illegal market is excluded.
Three Strange Observations
Observation 1. Mutual aid matches the savings-bank default rate, but at one-quarter the interest
A savings-bank default rate of 4–6% lines up with mutual aid's roughly 5%. Yet interest is 10–19% versus 2–5% — three to five times lower.
Intuitively, higher rates should mean higher defaults; the data point the other way. Lower rates appear to protect borrowers' ability to repay, not erode it.
Observation 2. Mutual aid defaults at less than half the rate of consumer lenders
Mutual aid's main borrowers are "artists previously rejected by tier-1 banks." By conventional credit metrics, they sit in the same band as licensed-lender borrowers. Yet repayment outcomes are about 5% — half the 8–12% seen at licensed lenders.
What this says: there's a point at which credit scoring loses predictive power. Someone classified as "high risk" by institutional grading can act as "sound" under a different framework.
Observation 3. The qualitative gap in underwriting
Look at the rightmost column. From commercial banks down to licensed lenders, the axis is the same, only varying in precision: credit score, income, collateral, usage history. All past financial data.
Mutual aid uses a different axis: community testimony, purpose, work history. Not past data, but present relationships and intent.
Two Definitions of Credit
The table reveals that the word "credit" actually carries two simultaneous meanings.
Definition A — Credit as Financial Position
- What's measured: income, assets, debts, prior delinquencies
- How: statistics, algorithms, collateral valuation
- Premise: past financial behavior predicts future repayment
- Strengths: scale, speed, objective numbers
- Limits: weak for irregular income, new borrowers, the signal-less
Definition B — Credit as Density of Relationship
- What's measured: reputation in community, weight of commitment, purpose of funds
- How: peer testimony, conversation, in-depth review
- Premise: those who know a person best predict their willingness to repay best
- Strengths: works for the signal-less, resilient to uncertainty
- Limits: small-scale, time-intensive, hard to scale
The two are complementary, not interchangeable. A finance built only on Definition A excludes the signal-less; Definition B alone cannot move a large economy.
Why Mutual Aid Defaults So Low — Five Mechanisms
Five structural reasons sit behind the table's numbers, each illustrating how Definition B works.
(1) Money With a Face
A KRW 5 million bank loan doesn't carry the felt presence of any depositor. A KRW 5 million mutual-aid loan does: "Artist A sold a work last summer, and that's the money." When the source has a face, repayment becomes a different obligation.
(2) Communal Binding
Falter on repayment, and the next person in line is concrete. "If I don't pay, B can't make rent next month" — that fact connects in a way that can't be hidden. The bind is sometimes stronger than legal collection.
(3) The Right-Sized Loan
Banks lean toward maximum credit (interest income depends on it). Mutual aid lends "only what's truly needed." Because peers vet whether the request fits the purpose, un-repayable amounts simply don't get lent.
(4) Low Rates and Flexible Repayment
2–5% rates decisively lower the borrower's burden. Concentrated repayment after an exhibition, early repayment when income arrives, deferral by mutual agreement in a crisis — all possible. Repayment runs at a rhythm the borrower can sustain, so default is rare.
(5) Repayment Itself Means Something Different
In a bank loan, repayment is "individual financial responsibility." In mutual aid, repayment is "the practice of solidarity." The motivational structure is fundamentally different, and the same act (paying back principal) carries different psychological weight.
So Why Don't Banks Use This Approach
A fair question. If mutual aid produces such results, why don't banks adopt the model?
The Scale Constraint
Definition B underwriting takes time. Gathering peer testimony and discussing purpose takes hours to days per case. That doesn't scale to the tens of thousands of loans a commercial bank processes daily.
The Community Premise
The model only works inside a community where borrowers know one another. Extend it to large, anonymous populations and the communal binding evaporates.
Liability Structure
Mutual aid rests on small-scale voluntary trust. Banks have legal obligations to depositors for principal and interest, so subjective underwriting raises legal and regulatory issues.
The two systems aren't competitors so much as complements, with mutual aid filling the blind spots banks can't see.
Not Just Korea
Similar tables appear in other countries.
- Grameen Bank, Bangladesh (Muhammad Yunus): Uncollateralized microloans with 95–97% repayment rates — basis of the 2006 Nobel Peace Prize.
- U.S. CDFIs (Community Development Financial Institutions): Microloans to low-income borrowers default at 3–6%, comparable to or better than traditional consumer finance.
- Germany's Künstlersozialkasse (Artists' Social Insurance): Not a lending system, but a community-contribution welfare model that resolves artist exclusion.
All show the same shape: people excluded by financial-position screening repay reliably under relational systems. SAF's mutual aid is the Korean artist's version of this international current. For more, see International Artist Finance Cases.
The Citizen's Role — Growing One Row in the Table
The most direct way to grow the last row (artist mutual aid) — to make it larger, more stable — is the purchase of a single work.
The Money's Path
- A citizen buys a work at SAF
- Part of the proceeds enter the mutual-aid fund
- The fund makes a low-rate loan to an artist
- Repayment cycles back into the fund
- The next artist receives the next loan
As the loop grows, more artists can access "KRW 5 million at 2–5%," and people from the table's right column (artist borrowers in the licensed-lender market) move into the middle column (mutual aid). A single person moving from a 24%-per-year market to a 4%-per-year market converts 10–15% of annual income from interest into creative spending.
This is not natural market adjustment. It's a movement only citizen participation can produce, because the mutual-aid fund grows only when citizens grow it.
A One-Line Summary
"No credit" and "creditworthy" can be said of the same person at once. Lined up in one table, you can see which measurement is more accurate.
Frequently Asked Questions
Q. As of when are these figures? A. Averages for 2023–2025. Commercial bank, card, savings bank, and licensed-lender figures draw on Financial Supervisory Service and Bank of Korea disclosures and press reports. Mutual-aid figures come from three years of SAF internal operations data. There is quarterly and annual variance — check each institution's disclosures for the latest.
Q. Who receives the mutual-aid interest? A. As a non-profit fund, no individual takes interest income. Interest covers operations (review, management, follow-up) and grows the fund's principal.
Q. Won't default rates rise as mutual aid scales? A. A reasonable concern. International cases (Grameen, CDFIs) show defaults rise modestly during early scaling, then stabilize within 10% under regional manager systems. SAF is designing a regional and genre-distributed structure.
Q. Does mutual aid serve both deceased and living artists? A. The fund supports the creative work of living artists. Royalty trusts and similar instruments for deceased artists fall under separate frameworks.
Q. Do other Korean freelance professionals (developers, web designers, etc.) face the same problem? A. To a significant extent, yes. Platform workers and freelancers broadly read as "signal-less" under Definition A underwriting. If artist mutual aid proves out, the methodology may extend financial access for irregular workers more widely.
Related reading
If this piece helped, you may also enjoy these related articles:
- Painting on Janji — An Eunkyung and the Contemporary Voice of Korean Painting — Janji is a thick traditional Korean surface made by layering hanji. Through An Eunkyung's paintings, we read its absorption, thickness, and quiet emotional effect.
- Understanding Dansaekhwa: The Korean Monochrome — A painting that repeats the same stroke a thousand times — why is that art? The key to Dansaekhwa isn't the single color but the repetition. From Park Seo-bo to Ha Chong-hyun, the aesthetics of Korean monochrome.
- Introduction to Minjung Art Through Shin Hak-chul — Resistance in the 1980s, reappraisal in the 2000s, reinterpretation in the 2020s. Minjung art never disappeared — follow a lineage still alive as one spine of Korean art, read through Shin Hak-chul.
A single table says a lot. Set the four institutional layers next to mutual aid, and the word "credit" reveals itself as two languages bound together without translation. To take part in growing that last row, the most direct path is to bring home a work from the SAF artwork gallery.
Further Reading
- Five Numbers That Map the Financial Reality of Korean Artists
- What a 95% Repayment Rate Means — How SAF Differs from a Bank
Start Collecting
Seed Art Festival
Published May 30, 2026
