On August 3 the Industrial Bank of Korea (KRX:024110) launched a ₩1tn lending programme for small and mid-sized companies with mid-to-low credit ratings. The loans are fixed-rate, which protects the borrower if rates rise. They also let the borrower switch to floating if rates fall.
Sit with that for a second. The bank has written a one-way option and handed it to its weakest borrowers. Heads the client wins, tails the client doesn't lose. A commercial lender either prices that optionality into the spread or doesn't offer it. Six weeks earlier the bank had launched a ₩500bn "AI Transformation" facility offering up to ₩3bn per company at a rate discount of as much as 1.5 percentage points.
This is why the shares trade at ₩20,900 against a book value of roughly ₩47,571 — 0.44 times equity — while the bank earns a return on that equity of about 7.7%. Plug those into a standard bank valuation and you back out a cost of equity somewhere near 17%, which is not a number any credit analysis of this balance sheet would produce. The discount isn't about credit. It's about who decides what the bank does.
American readers need this spelled out, because there is no clean US equivalent and the closest analogies mislead.
IBK is not a commercial bank chartered under Korea's Banking Act. It was created by its own piece of legislation, the Industrial Bank of Korea Act (중소기업은행법), and exists to finance small and mid-sized enterprises. The government holds 59.5% of it, and the statute requires the state to remain the majority owner — this isn't a stake that gets sold down when it suits the Treasury. The same Act provides for the government to make good losses that the bank's reserves cannot cover.
That last provision is the load-bearing one. It is why IBK funds itself heavily by issuing SME finance debentures that trade close to sovereign risk rather than by competing for retail deposits, and why a bank with ₩526,866.3bn of assets against ₩37,934.0bn of equity — roughly 14 times leverage — isn't priced as if it might fail. It won't be allowed to.
The price of that backstop is the mandate. IBK lends where policy directs. Its SME loan balance passed ₩260tn in 2025, and a ₩14.7tn surge in SME lending pushed it past Woori to fourth place in Korea's won-denominated loan market. When Korean small manufacturers are under strain, IBK is the instrument the government reaches for, and it cannot respond by shrinking the book and moving into safer assets. Fannie Mae and Freddie Mac are the nearest US comparison in structure, but they were placed into conservatorship; IBK is a listed, profitable, dividend-paying bank that has simply never had full control of its own strategy.
So a minority shareholder owns 40.5% of an institution whose lending decisions serve a public purpose first and a return second. Half a times book is the market saying it understands the arrangement.
First-half net income was ₩1,442.9bn, down 4.4% from ₩1,508.6bn. That's a mild result, and what's underneath it is more interesting than the headline.
Net interest income actually rose 7.2%, from ₩3,803.5bn to ₩4,076.3bn, and the second quarter's ₩2,084.4bn was the strongest of the last eight. The mechanism is straightforward: interest expense fell from ₩2,554.8bn in Q2 2025 to ₩2,482.0bn while interest income rose, so the funding side repriced faster than the asset side. Net fee and commission income also improved, from ₩200.5bn to ₩267.8bn.
Two things ate it. Credit loss provisions were ₩852.5bn in the half against ₩765.0bn, up 11.4%, continuing a climb that has taken annual provisions from ₩940.4bn in FY2021 to ₩1,774.2bn in FY2025. And operating income fell ₩103.9bn despite all that revenue improvement, which means something worse than ₩390bn happened in other operating lines — securities, foreign exchange and derivative valuations. The half-year report itemises that; the summary statements don't, and anyone underwriting this stock should look it up rather than assume.
The provisioning trend is the honest worry, and it's the mandate showing up in the numbers. A bank whose loan book is concentrated in Korean SMEs, and which keeps being asked to extend credit to lower-rated borrowers at discounted rates, will carry a higher through-cycle credit cost than a diversified commercial bank. Provisions now consume roughly 23% of net interest income, against about 15% in FY2021.
The structural weakness alongside it is fee income. At ₩431.6bn in FY2025 it was 5.6% of net interest income, and it has fallen from ₩631.5bn in FY2021. This bank is almost purely a spread lender with no meaningful wealth, card or asset-management business to cushion a rate cycle.
On July 13 the board declared a quarterly dividend of ₩210 a share, ₩167.5bn in total, on 797,425,867 shares. The record date was July 31 and it was paid on August 28. The filing states the quarterly rate against market price at 1.0%.
The FY2025 annual dividend was ₩1,048 a share, which at ₩20,900 is a 5.0% yield. IBK, along with Korea Development Bank and the Export-Import Bank, is paying the government a combined record ₩1.854tn in 2026, up 19.9% on the year.
And there's the crux of the bull case. The government owning 59.5% cuts both ways. It constrains the lending, but it also makes the state by far the largest recipient of the dividend, and a Korean government running a persistent budget gap has a direct fiscal interest in these banks paying more. Dividends paid have risen every year in the record here, from ₩375.6bn in FY2021 to ₩850.8bn in FY2025. The controlling shareholder wants the cash, and minority holders get it at the same rate per share.
There's a procedural detail in that filing worth flagging too. The amount was fixed on July 13, the record date fell on July 31, and payment followed on August 28 — so an investor knew the size of the dividend eighteen days before needing to own the stock to receive it. That ordering is recent in Korea. For years the record date came first, at the end of December, and the amount was only announced at the annual meeting months later, meaning buyers committed blind. A reform to the capital markets rules let companies reverse the sequence, and this filing cites the provision it relies on to set the payment date. American investors have never had to think about this because US dividends have always been declared before the record date.
The bear case doesn't need the bank to lose money. It needs the discount to be permanent, and it might be. Nothing about the statute is changing. The state is not selling down. The mandate produces exactly the lending decisions that keep the credit cost elevated, and a 0.44 times multiple can persist for a decade — it largely has.
Worse, the discount can widen. If Korean SME conditions deteriorate, IBK is contractually and politically obliged to lend into it, so provisions rise at precisely the moment a commercial bank would be pulling back. And the bank raised write-down contingent capital securities in June — instruments that absorb losses by being written off — which is a normal capital-management step but a reminder that the equity sits behind a lot of leverage.
The counter is arithmetic. A 5% yield, growing, from an institution the state is legally required to stand behind, at 44 cents on the dollar of book, does not need the multiple to re-rate in order to work. It needs the dividend to keep coming.
The year-end dividend declaration, which is the number that matters. Three quarterly payments of ₩210 would total ₩630, so the FY2026 outcome hinges entirely on the final declaration and whether it matches or exceeds the ₩1,048 paid for FY2025.
Then the Q3 provisioning line in late October. Net interest income is recovering and costs are contained; the only variable that decides whether this year's earnings fall or rise is credit. If provisions keep climbing while the government keeps announcing discounted lending programmes, that tells you the mandate is getting more expensive, and the discount is doing its job.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.