DSR, LTV & Stress DSR: How Korean Loan Limits Work

If you estimate a Korean mortgage by asking only “what percent of the home’s price can I borrow?”, you will get the number badly wrong. A house expensive enough to support ₩500M under LTV lends you only ₩300M if your income and existing debts cap DSR there. Flip it around: income good for ₩700M under DSR still yields at most ₩400M on a ₩1B home at 40% LTV.

Your limit is the smallest of three, not the largest

Each rule caps a different thing; the binding one wins

LTV — how much the property supports40–70% of value, by area and property type

DSR — how much your income supports40% at banks, 50% at non-banks, across all your debt

The product's own capSet by the lender, independent of both rules above

The working rule for a real-world limit:

Actual loan limit = the smallest of the LTV limit, the DSR limit, and any institutional or policy maximum

And since July 2025, stage-3 stress DSR — extended across effectively the whole lending sector — pushes the DSR leg down further. The FSC defines DSR exactly as it sounds: all household loans’ annual principal and interest payments divided by annual income. This article is the reference for how the three mechanisms interact — the hub behind the foreigner loans pillar.

LTV: what is the collateral worth?

LTV (loan-to-value) is the loan as a share of recognized collateral value:

LTV = mortgage amount ÷ recognized home price × 100

An ₩800M home at 70% LTV supports ₩560M; at 40%, ₩320M.

As of 2026, purchase-purpose mortgages for ordinary non-owners (and disposal-conditioned single-home owners) run at up to 70% in non-regulated areas and 40% in regulated areas — and regulation moves: areas newly designated in July 2026 (Dongtan-gu, Giheung-gu, Guri) dropped from 70% to 40% at designation. First-time-buyer and policy-mortgage programs carry their own relaxed standards.

In the capital region and regulated areas, clearing LTV is still not the end, because purchase mortgages also face price-tier maximum loan caps:

Home price	Max mortgage (capital region / regulated)
≤ ₩1.5B	₩600M
₩1.5B–2.5B	₩400M
> ₩2.5B	₩200M

So a ₩2.0B home at 40% LTV computes to ₩800M — and actually caps at ₩400M. And the “home price” is not the seller’s asking number: apartments use recognized price sources such as KB market prices, with appraisals where market data is unavailable.

DTI vs DSR: does the other debt’s principal count?

Both are income-based affordability ratios; the computation scope differs.

DTI = (mortgage annual P&I + other loans’ annual interest) ÷ income

DSR = all loans’ annual principal and interest ÷ income

That difference bites. A borrower earning ₩80M with a ₩50M credit loan and a card loan sees mostly interest drag under DTI — but under DSR, the principal repayment of those debts enters too. This is why DSR became the true wall of Korean lending. (One precision note: per FSC guidance, auto installments and leases are exception items in the borrower-level DSR calculation — they do not enter the formula the way credit and card loans do, though they remain real credit-screened obligations that banks can weigh in internal reviews; see the rental-vs-lease-vs-buying comparison.)

The 40% / 50% split — and why non-banks aren’t a 25% bonus

Borrower-level DSR applies at 40% in banks and 50% in non-bank lenders for borrowers with over ₩100M of total loans, per the FSC’s supervisory framework.

On a ₩50M income, the bank-sector ceiling on annual debt service is:

₩50M × 40% = ₩20M/year (≈ ₩1.67M/month)

Non-bank at 50%: ₩25M/year (≈ ₩2.08M/month).

But “banks refused me, so the second tier lends 25% more” oversimplifies dangerously: higher non-bank rates inflate the annual P&I of the same principal, so the principal difference is smaller than the ratio gap suggests — and the new second-tier debt sits in your file at the next bank review, as covered in the loans pillar’s tier-2 warning.

Existing debt is measured in annual payments, not balances

Before a new mortgage, convert every existing debt into annual P&I. A bank borrower earning ₩70M:

DSR ceiling: ₩70M × 40% = ₩28M/year Existing credit + car loans: ₩8M/year Available for the new loan: ₩28M − ₩8M = ₩20M/year

Then reverse-engineer the principal that ₩20M/year services at the new loan’s rate and term. The sequence matters: you do not compute a limit from the ₩70M — you first compute how much DSR space existing debt already occupies.

Income means recognized income, not deposits

This is where foreigners, freelancers, and business owners diverge hardest. The strongest category is documented income (증빙소득): withholding receipts, NTS income certificates, pension statements — government-verifiable records. Where those fall short, the framework allows deemed income (인정소득) estimated from national-pension or health-insurance payments, and declared income (신고소득) from card-usage or financial-income data.

Two cautions. First, there is no universal formula like “₩200,000/month of health insurance = exactly ₩X of income” — estimation methods, caps, and eligibility vary by product and sector. Second, foreigners get no separate DSR formula; the binding constraint is verification. Earning USD 100,000 abroad means little if no Korean lender can recognize it — which is the same domestic-income problem that governs every foreigner credit decision.

Stress DSR does not raise your interest rate

The most misunderstood mechanism in the system. A 4% loan with a 3% stress rate does not cost 7%. The bank charges the contracted 4%; it merely computes your DSR limit at a higher virtual rate:

DSR calculation rate = actual loan rate + applicable stress rate

The point is pre-testing: could you still service the debt within DSR if rates rose? The FSC states explicitly that the stress rate is not charged on the actual loan. Stage 3, in force since July 2025, extends stress DSR across effectively all DSR-covered household lending — with credit loans joining once total credit-loan balances exceed ₩100M.

The 2026 stress map is regional, not uniform

• Capital region and regulated areas: the stress-rate floor was raised to 3.0% in October 2025 and carries into 2026.
• Non-regulated regional mortgages: relief holds through H2 2026 — a 1.5% stress base at a 50% application ratio (stage-2 level) from July 1 to December 31, while credit and other loans run 1.5% at the full stage-3 100% ratio.

So “everything gets 3% in 2026” is wrong in both directions.

Rate structure changes your stress — and your limit

The stress applied scales with rate-change risk: pure long-term fixed rates carry the least, floating rates the most, and mixed (fixed-then-floating) or periodic-reset loans a partial ratio by fixed period and reset cycle — ratios that stage 3 tightened for mixed and periodic loans specifically, by design, to push borrowers toward long fixed rates. The practical consequence: choosing a floating loan for a 0.1% cheaper rate can cost you not just future interest risk but loan limit today.

What stress DSR does to a ₩50M income

A bank borrower, ₩50M income, no existing debt, 30-year amortizing loan:

• DSR 40% → ₩20M/year of allowable P&I.
• At the actual 4% rate only: supportable principal ≈ ₩349M.
• Assume a capital-region floating loan where the full 3% stress applies → 7% calculation rate: the same ₩20M/year supports ≈ ₩251M.

Borrowing capacity on a ₩50M income — before and after stress DSR

Bank DSR 40% · 30-year amortizing loan · no existing debt

Actual rate 4%≈ ₩349M

Calculated at 7% (4% + 3% stress)≈ ₩251M

Nearly ₩100M of limit evaporates — while the interest actually charged stays 4%. The bank is lending only what survives a 7% world.

The reverse-engineering formulas, for your own estimates:

Allowable annual P&I = recognized income × DSR cap − existing annual P&I

Loan principal = monthly capacity × [1 − (1+r)^(−n)] ÷ r

with r the monthly rate including stress and n the term in months. Bank systems apply per-product DSR maturities, repayment types, and rate structures, so treat any online estimate — including our loan calculator — as directional, not an approval.

Which wall you hit first depends on you

A ₩1B non-regulated home at 70% LTV offers ₩700M — but a ₩50M income will hit the DSR wall far below that; the bottleneck is income. A ₩200M earner with no debt buying a ₩1B regulated-area home clears DSR easily and stops at LTV’s ₩400M; the bottleneck is collateral policy. Before any contract, compute both:

Collateral limit = recognized price × applicable LTV Income limit = principal reverse-engineered from allowable DSR P&I

— and plan on the smaller one.

“Jeonse loans never count in DSR” is now outdated

Jeonse financing was long the textbook DSR exemption. Since October 29, 2025, that is only partly true: a single-home owner taking a jeonse loan in the capital region or a regulated area has the loan’s interest portion reflected in DSR. Non-owners’ ordinary jeonse loans have not been wholesale absorbed — but the blanket exemption statement is dead. When running the jeonse-vs-wolse math, homeownership status and the property’s region now belong in the loan column.

Exemptions are not bypasses

Certain loans receive exceptional DSR treatment for policy reasons — portions of jeonse lending, interim/relocation payments on presale apartments, credit loans of ₩3M or less, some inclusive-finance products, deposit-secured loans, and auto installments and leases. Two things follow. Exempt debt is still debt — it never disappears from your finances. And business loans are not a personal-mortgage bypass: the FSC is explicitly targeting diverted business lending and real-estate workarounds in its 2026 inspections, so business borrowing must track real business purposes and repayment sources.

Sequencing your borrowing changes your limits

Take a ₩50M car installment and a ₩50M credit loan this year, and next year’s mortgage has that much less DSR space. Reverse the order — clear the big housing loan first, then evaluate consumer financing — and the same person ends up with a different capital structure.

If a major loan sits 6–12 months out: pay down high-rate credit and card loans, add no new car installments, keep withheld income stable if salaried, and — if freelance or self-employed — objectify income through proper filings, since unreported income cannot be recognized income. Foreigners add stay duration and domestic credit history to the same checklist. A high KCB/NICE score cannot break through DSR, and abundant collateral cannot break through LTV.

Korean loan limits reduce to three questions:

What is the collateral worth? How much can you repay each year? Does that capacity survive rising rates?

The first is LTV, the second is DSR, and the third is stress DSR.