KakaoBank Corp. (KRX:323410) earned ₩328bn in the first half of 2026, a record, up 24.4% from the same period last year. The Korean press ran it as a record half on loan and fee income growth, and it is one.
Look one line higher. Operating profit for those six months was ₩351.6bn. For the first half of 2025 it was ₩353.2bn. Slightly down.
So a bank whose operating result was flat reported a quarter more net profit. The whole difference sits in the March quarter, where pretax income was ₩250.7bn against operating income of ₩157.6bn — ₩93.1bn arriving between the two lines. The June quarter behaved normally: ₩194.0bn operating, ₩190.8bn pretax, ₩140.8bn net.
That's not an accusation. Non-operating items are ordinary at banks and the figures are the company's own filed numbers. But it changes what the record means. The underlying franchise did not grow its operating earnings in the first half of 2026, and almost everyone reporting the result said the opposite.
KakaoBank's net interest margin reached 2.13% in the June quarter, against a 1.62% average across Korea's four major commercial banks. That is a real advantage, and it comes from exactly where the branchless model says it should: cheap deposits gathered through an app most Koreans already have open, funding consumer loans without a branch network underneath.
The interest lines confirm it. Interest income of ₩690.4bn less interest expense of ₩288.1bn gave ₩402.3bn of net interest income in the June quarter, up from ₩318.6bn a year earlier. Across the half, net interest income grew about 21%.
Then look at what that turns into. Trailing four-quarter net income is ₩544.7bn on equity of ₩6.68tn. Roughly 8% return on equity. Korea's large commercial banks, running half KakaoBank's margin, generally earn more than that.
The reconciliation is the balance sheet. Loans were ₩48.2tn at the end of June against total assets of ₩75.9tn, so barely 63% of the asset base is lending. The rest sits in securities and deposits with other institutions, earning market rates rather than lending spreads. And equity of ₩6.68tn against ₩75.9tn of assets is about 11 times leverage, low for a bank. KakaoBank raised a large amount of capital at its 2021 IPO and has not yet found enough loans to put behind it.
An American reader used to capital-constrained neobanks should sit with that inversion. This is the opposite problem. The constraint isn't capital, it's assets.
Two Korean rules shape this in ways that have no US equivalent, and both are worth knowing before judging the numbers.
The first is why KakaoBank exists at all. Korean banking law generally caps a non-financial company's stake in a bank at 4% of voting shares. The Internet-Only Banks Act of 2018 raised that ceiling to 34% specifically for internet banks, which is what allowed Kakao Corp. to become the controlling strategic shareholder; its stake sits around 27% after the IPO diluted it. In the United States, the Bank Holding Company Act does the reverse — it keeps commercial firms out of banking, and the fight over industrial loan companies has run for decades without a statute like Korea's. Korea simply wrote one.
The second is a quota. Internet banks in Korea must direct a minimum share of their unsecured personal lending to mid- and low-credit borrowers; the supervisory target has been at least 30%, with the three internet banks pushing toward higher ratios. No US bank operates under a rule shaped like that. It means KakaoBank's credit costs are partly a policy variable, not purely a risk-appetite choice.
Layer on Korea's household lending restraints, which have capped how fast any bank can grow mortgages and personal loans, and the strategic move becomes readable. Loans grew only ₩518bn in the June quarter — about 1% — with growth concentrated in lending to sole proprietors and small business owners. That is where a Korean consumer bank goes when the consumer side is administratively closed.
Credit costs have been improving through it. Provisions were ₩56.8bn in the June quarter against ₩58.1bn a year earlier, and full-year provisions fell from ₩270.5bn in 2024 to ₩243.8bn in 2025 even as the loan book grew. On a ₩48.2tn book, the June quarter annualizes near 0.47%.
The company's second-quarter presentation led on fee income growth of 92%. Filed operating revenue for the half tells a duller story: ₩1,648.3bn against ₩1,562.6bn, up 5.5%, while interest income rose 11.9%. Subtracting interest income leaves non-interest revenue lower than a year ago, not higher.
Those can both be true — fee income is a subset of non-interest revenue, and other components such as securities valuation could have moved the other way. The filed statements don't break it out finely enough to say. I'm flagging it rather than resolving it, because a 92% growth headline sitting on top of a 5.5% revenue line is the kind of thing worth checking in the third-quarter disclosure.
The shares closed at ₩21,900 on September 2, in a 52-week range of ₩19,750 to ₩28,700, for a market capitalization of ₩10.45tn — about $7.6bn at roughly 1,370 won per dollar, which is an approximate rate. Against equity of ₩6.68tn that is 1.57 times book. Trailing earnings of ₩544.7bn put it near 19 times.
Korea's incumbent bank holding companies have persistently traded well below book. So the market is paying about three times the sector's book multiple for a bank earning a return on equity that isn't higher than theirs.
The bull case for that gap is not this year's earnings. It is 27.63 million customers, up 920,000 since the end of 2025 in a country of about 52 million, a cost base without branches, and the option that the excess capital eventually gets deployed into lending at that 2.13% margin. If loans went from 63% of assets to something closer to a normal bank's 70-plus percent, the return on equity arithmetic changes materially without any new customer.
The bear case is that the constraint isn't temporary. Household lending curbs are policy, the mid-and-low-credit quota is policy, and neither is scheduled to disappear.
Deposits fell 3% quarter on quarter in the June quarter, with the company noting a return to net growth in July. For a bank whose entire margin advantage rests on cheap retail deposits, that line matters more than the loan line, and one quarter of outflow is not a trend but it is the first thing to watch.
The pivot into sole-proprietor lending moves the book toward a borrower class with different loss behavior than salaried consumers, and it is doing so late in a cycle where Korean small-business credit has been under pressure. Provisions have been falling; that is the assumption most exposed if this mix shift goes wrong.
Comprehensive income has been running well below net income — ₩52.7bn against ₩140.8bn in the June quarter, and ₩380.9bn against ₩480.3bn for full-year 2025. That gap is securities marks flowing through equity, and it is why book value has barely moved in eighteen months despite retained earnings rising.
And the concentration risk that doesn't appear in any financial statement: distribution runs through the Kakao messaging app, so anything that damages Kakao Corp.'s platform position damages the bank's customer acquisition cost before it damages anything else.
The third-quarter report, on three specifics. Whether operating profit grows year on year without help from below the line — the June quarter's ₩194.0bn against ₩151.1bn in September 2025 suggests it can, and one more clean quarter would settle whether the flat first half was a Q1 artifact.
Whether deposits recovered through the third quarter after June's decline.
And the loans-to-assets ratio. If it starts moving up from 63%, the return on equity story fixes itself and the 1.6 times book stops looking odd. If it doesn't, that multiple is being paid for a customer count.
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.