The Bank of Korea raised its base rate to 3.00% on 27 August 2026, the second consecutive increase and the first back-to-back move since early 2023. For a bank holding company that has spent the last two years telling investors about insurance acquisitions, that decision puts the attention back where it belongs: on the spread.

Woori Financial Group Inc. (KRX:316140) makes most of its money the plainest way a bank can. It borrows short from depositors, lends longer, and keeps the difference. That difference has been remarkably static for four years, and the mechanics that held it up through the last cycle are now working the other way.

Thirty-Four Percent More Balance Sheet Bought 3.8% More Spread Income

Start with the headline series. Net interest income was ₩8,696.6bn in FY2022, ₩8,742.5bn in FY2023, ₩8,886.3bn in FY2024 and ₩9,030.8bn in FY2025. That is 3.8% cumulative growth across three years — roughly 1.3% a year, well under Korean inflation over the period.

Now the balance sheet. Total assets went from ₩480.5tn at the end of FY2022 to ₩601.5tn at the end of FY2025, up 25%, and to ₩635.2tn by the second quarter of 2026. Take FY2021's ₩447.2tn as the base and the growth is 34% through FY2025.

Those two facts together define what Woori is. The bank grew the book by a third and got 3.8% more spread income out of it. Margin per won of assets compressed by roughly the difference. Some of the asset growth is insurance reserves that were never going to produce interest income, so the comparison overstates the compression on the lending book itself. But even confining the view to the pre-acquisition years, FY2022 to FY2024, assets grew 9.4% and net interest income grew 2.2%. The direction is the same.

The Gross Lines Show How The Trick Worked

The interesting part is underneath. Interest income — what the assets earn — ran ₩14,654.5bn in FY2022, ₩20,641.6bn in FY2023, ₩22,013.3bn in FY2024, then fell to ₩21,188.7bn in FY2025. Interest expense followed almost the same shape: ₩5,958.0bn, ₩11,899.0bn, ₩13,127.0bn, then ₩12,158.0bn.

FY2025 is the year to study. Interest income fell ₩824.6bn. Interest expense fell ₩969.0bn. Funding costs came down faster than asset yields, and that ₩144.4bn gap is the entire net interest income improvement for the year. Woori didn't earn more on its loans in 2025. It paid less for its deposits, by more.

That is what a liability-sensitive balance sheet looks like in a cutting cycle, and it is a perfectly respectable way to make money. It is also entirely dependent on rates continuing to fall.

This Year The Squeeze Reversed

They stopped falling. In the second quarter of 2026, interest income was ₩5,548.4bn against ₩5,114.6bn a year earlier, up 8.5%. Interest expense was ₩3,191.9bn against ₩2,852.8bn, up 11.9%. Expense is now growing faster than income — the mirror image of FY2025.

Net interest income still rose, to ₩2,356.6bn from ₩2,261.8bn, or 4.2%, because the base is larger on the asset side. First-half interest income for the group came to ₩4.66tn on management's reporting, up 3.2%. So the number is fine. The composition is not the same as last year's, and the trend in the gap is unfavorable.

Deposit pricing is where this shows up first. The average rate on newly taken Korean deposits was 3.21% in July 2026, up 13 basis points from June. Deposits reprice on maturity, in a stream, so a bank absorbs higher funding costs over several quarters after the market rate moves. Loan books reprice too, but Korean household lending carries a heavy fixed-rate and policy-linked component, and regulatory attention to household debt limits how fast a bank can grow into a better-priced book. The result is that a hiking cycle usually hurts a Korean commercial bank's margin before it helps.

Credit Costs Eat A Bigger Share Than They Used To

There is a second claim on the spread. Provisions for credit losses were ₩1,894.9bn in FY2023, ₩1,716.3bn in FY2024 and ₩2,102.8bn in FY2025. Measured against net interest income, credit costs took 19.3% of spread income in FY2024 and 23.3% in FY2025.

The quarterly figures have improved: ₩575.7bn in Q3 2025, ₩583.8bn in Q4, ₩526.8bn in Q1 2026 and ₩439.2bn in Q2. Two quarters at the recent pace annualizes near ₩1.9tn, below FY2025. Woori's second-quarter net income of ₩1,026.5bn, up 9.24% year on year in the 24 July disclosure, leans on that.

Higher policy rates are not obviously good for this line either. The BOK explicitly cited household debt and Seoul-area home prices in its August decision. Whatever that does for margins, it raises debt service costs for borrowers who are already stretched, and provisioning follows with a lag.

What The Whole Thing Costs

At the ₩30,900 reference close in Woori's 24 July cancellation filing and 728,089,682 shares outstanding, the market values the group near ₩22.5tn (~$16bn, using roughly ₩1,400 to the dollar as a working rate). That is about 2.5x annual net interest income. Buying the entire company costs two and a half years of the spread it earns before any expense, provision or tax.

That framing makes the point better than a P/E does. The market is not paying for growth in this line. It is paying for the line to keep existing.

The Case That Rising Rates Help

The other side is worth stating properly, because plenty of Korean bank analysts hold it. A hiking cycle steepens what a bank earns on its non-interest-bearing and low-cost deposit base, and Woori Bank has a large retail deposit franchise where a chunk of balances pay close to nothing regardless of policy rates. Those balances reprice on the asset side and not on the liability side. If the BOK's forecast path holds — board members most often cited 3.25% for six months out — the eventual effect on margin could be positive once the deposit repricing wave passes through.

Growth helps too. The BOK lifted its 2026 growth forecast to 3.3% from 2.6%, driven by a semiconductor export boom. Faster nominal growth means more corporate loan demand, and Korean press coverage of Woori's first half attributed part of the interest income increase to large-corporate lending. A bank that can grow the book at better spreads in a strong economy is not obviously worse off at 3.00% than at 2.75%.

And the compression I described has a ceiling. Net interest income has been flat, but flat at roughly ₩9tn a year through a period that included a full hiking cycle, a full cutting cycle and a household debt scare. Stability is not nothing.

What Would Settle It

Interest expense in the third quarter. Not net interest income, which nets out the story, and not the margin the company reports on a slide. The gross expense line in the Q3 consolidated statements, against Q2's ₩3,191.9bn and Q3 2025's ₩3,149.1bn. If it climbs past ₩3.3tn while interest income lags, the deposit repricing has arrived and net interest income growth stalls or reverses despite a bigger book.

The second thing is the provisioning line. Two quarters near ₩440bn is a pattern, three would be a trend. If Q3 comes in above ₩520bn while rates are rising, the credit assumption behind this year's earnings improvement needs redoing.

Both land in the third-quarter report, and both matter more to Woori's next twelve months than anything the insurers do.