📚 Korean Financials Capital-Buyback Compounding Series — Part 2/N. Previous: Meritz Financial Holdings — The Capital-Buyback Compounding Standard for Korean Financials, and the Landscape Beyond Its Peak

Part 1 framed the broader recognition shift in Korean financials. This is the natural follow-up — if Meritz is the static “ROE × payout-ratio” model, then Kiwoom Securities is the company that completed the recognition shift on the dynamic “ROE × trading-volume beta × capital turnover” variant of the same matrix. The April 30 sell-off after the strong 1Q26 print is not a model rejection. It is the self-stabilization a peak price naturally produces. This piece reads that signal.

Executive Summary

• The recognition shift is already complete. Kiwoom Securities PBR moved from 0.55× (2024) → 1.14× (2025) → 1.39× (2026E). The market no longer classifies Kiwoom as a “retail-#1 discount” name. It has already been reclassified as an ROE-20% capital-efficiency brokerage.
• So 1Q26’s strong print was not “discovery” — it was “confirmation.” Operating profit ₩621.2B (+90.9%), net income ₩477.4B (+102.6%), Kiwoom domestic-equity daily-average trading volume ₩27.8T (+215.9%). Clearly strong numbers. But the market had been pricing this trajectory since early April. The post-print -6.02% reaction is not “new information” pricing in. It is the start of the next verification phase after the recognition has already been priced.
• This is the interesting contrast with the Meritz model. Meritz = static capital-buyback compounding (ROE × payout ratio). Kiwoom = dynamic capital-turnover compounding (ROE × trading-volume beta). Two different mechanisms producing the same “ROE-20% brokerage” classification. After the recognition shift, the market evaluates both on the same matrix — at different points.
• The current price already embeds a self-stabilization mechanism. PBR 1.39× closes the math at ROE 20.7% × cost of equity ~14.9%. The market has already priced sustained ROE in the low-20s. From here, further upside is not discovery alpha but model-durability validation — and the May–June ₩44.8T daily-trading-volume threshold is the first check.
• The follower matrix creates the new landscape. On the same matrix, Korea Investment Holdings (071050; ROE 16.8%, PBR 1.07×), Samsung Securities (016360; ROE 15.8%, PBR 1.05×, dividend yield 5.4%), and NH Investment & Securities (005940; ROE 17.1%, PBR 1.18×, dividend yield 5.9%) each occupy different positions on the same standard. If Kiwoom sits at the peak of one variant, the others each carry distinct time-gap alpha.

1. The Starting Position — Reading the Landscape After Recognition

1.1 Picking up where Part 1 left off

Part 1 condensed the move in Korean financials into a single line: the era of “low-PBR discount asset” is over; the market now reprices Korean financials through the matrix of ROE × payout ratio × EPS growth. This piece goes one layer deeper inside the same landscape — what does that matrix look like when its key variable is retail trading-volume beta rather than capital allocation? Kiwoom Securities is the answer.

1.2 Kiwoom’s Position in One Table

Arithmetic checks:

• Market cap = ~26.23M shares × ₩398,000 ≈ ₩10.44T ✓
• 2026E PBR = 398,000 / 285,909 = 1.392× ≈ 1.39× ✓
• 2026E PER (Mirae Asset, EPS 59,426) = 398,000 / 59,426 = 6.70× ≈ 6.7× ✓
• 2026E PER (Samsung, EPS 53,228) = 398,000 / 53,228 = 7.48× ≈ 7.5× ✓
• 2026E dividend yield = 15,500 / 398,000 = 3.89% ≈ 3.9% ✓
• vs low = 398,000 / 132,100 - 1 = 201.3% ✓

The single line this table tells: PER looks low; PBR sits at historical highs. And those two facts are not contradictory — they are the natural landscape of a recognition shift fully completed.

2. Recognition Shift Already Complete — What the PBR Path Shows

2.1 The Historical PBR Trajectory

The clearest evidence sits in the PBR path itself.

Arithmetic check: 2024 → 2026E change = 1.39 / 0.55 - 1 = +152.7%. Even after accounting for BPS growth across the same window, PBR itself expanded roughly 2.5×.

This trajectory is not “still changing.” It has already changed. The single biggest jump (0.55× → 1.14×, +107%) happened in 2025, and 2026 is fine-tuning on top. The same diagnosis from Part 1 — “the recognition shift has already happened” — applies identically to Kiwoom.

2.2 The Model the Market Is Now Using

For Kiwoom to move from PBR 0.55× to 1.39×, the market’s underlying model of the company has to change. The shift looks like this:

Single line: the market has already reclassified Kiwoom into the ROE-20%-class category. That is the same kind of recognition shift Meritz received on its way to the ROE-22%-class category. The mechanism is just different.

2.3 The Meritz Model vs the Kiwoom Model — Two Variants on the Same Matrix

This is the most interesting table in the piece. Both companies have been recognized by the market as ROE-20%-class names. The difference is how they get there. Meritz makes per-share value bigger by shrinking the capital base (buyback-and-cancel). Kiwoom makes earnings bigger by turning the capital base over faster (trading volume × margin × prop). The endpoints look similar; the paths are different.

That difference shows up in the multiple. Meritz, the lower-volatility model, gets the higher PBR (1.6×). Kiwoom, the higher-volatility model, gets the slightly lower PBR (1.39×). The market is pricing both correctly as different positions on the same matrix.

3. 1Q26’s Strong Print — The Meaning of a “Confirmation Event”

3.1 The Numbers Themselves Are Clearly Strong

Arithmetic check: Annualized 1Q ROE = ₩477.4B × 4 / avg equity ~₩6.84T ≈ 27.9% (matches KB Securities). ✓

A naive 1Q × 4 annualization implies controlling-shareholder net income ~₩1.91T. At the current ₩10.44T market cap, that drops PER to:

Naive annualized PER = 10.44T / 1.91T = 5.47×

Pure arithmetic, this looks “cheap.” But that is just arithmetic.

3.2 The Sell-Side Models a Decelerating Quarterly Path

Mirae Asset Securities’s 2026 quarterly path looks like this:

Arithmetic check: 4,774 + 4,230 + 3,160 + 2,620 = ₩14,784B ≈ ₩1.48T ✓

The sell-side reads 1Q as the peak quarter. 2Q at -11% vs 1Q, 3Q at -34%, 4Q at -45%. On those assumptions, 2026E NI is ~₩1.48T — the denominator behind the PER 6.7× figure.

3.3 The Meaning of a “Confirmation Event”

Reading the two tables together unpacks the April 30 sell-off precisely:

What the market knew since early April:  1Q trading-volume surge → strong print incoming
What the April 30 release added:         Almost nothing (within expected range)
What the market wanted to learn:         "Is 1Q the peak, or not?"
What April trading-volume data implied:  April daily-avg below 1Q average

So the print itself was a confirmation event, not new information. And April trading-volume data suggested 1Q may indeed have been the peak. Those two together produced the -6.02% adjustment.

This is not a model rejection. The market still classifies Kiwoom as ROE-20%-class. PBR 1.39× is stable on top of that classification. The single thing that changed is the question — “is the ROE 20% level sustained for the full year 2026, not just 1Q?” — and that verification has moved from 1Q to 2Q.

This is the same flavor of signal Part 1 mentioned in section 4.2 on Meritz: the model’s self-stabilization mechanism. A price not running in a straight line is not a weakness — it is the natural landscape after the recognition is complete, where the price now demands quarterly verification data.

4. The Arithmetic of Self-Stabilization — How PBR 1.39× Closes

4.1 The PBR Closure Identity

For financials, the justifying-PBR identity is simple:

PBR ≈ ROE / cost of equity

Substituting Kiwoom’s PBR 1.39× and ROE 20.7%, the implied cost of equity emerges:

Implied cost of equity = ROE / PBR = 20.7% / 1.39 = 14.89% ≈ 14.9%

Arithmetic check: 20.7 / 1.39 = 14.892% ✓

That 14.9% is the cost of equity the market is applying to Kiwoom. It is ~3.4 percentage points higher than Meritz’s implied 11.5% (22.4 / 1.94). That 3.4 ppt is exactly the “retail-cycle volatility discount.”

4.2 What the Identity Says About the Price-Range Scenarios

Plug in different ROE assumptions and the price range falls out naturally.

Arithmetic check: 285,909 × 1.39 = ₩397,414 ≈ ₩397,400 (within 0.1% of the ₩398,000 close) ✓

Single-line read: the current price closes exactly with the ROE 20.7% assumption. That is not a coincidence — it is the market having priced this correctly. If the market re-anchors to ROE 22%, the price moves naturally to the ₩423K range. If it re-anchors to ROE 18%, to the ₩346K range. This is what PBR does inside a recognition-completed regime.

4.3 What the Self-Stabilization Means

The arithmetic shows two things.

First, the price is internally consistent with the model. PBR 1.39× is consistent with ROE 20.7%. The price that looks “rich” is actually the price that closes the model. Same kind of consistency as in Part 1, where Meritz PBR 1.5–1.6× closed against ROE 22.4%.

Second, the price is directly tethered to the ROE assumption. ROE going to 22% pushes PBR toward 1.48× automatically. ROE coming in at 18% pulls PBR toward 1.21×. A self-correcting mechanism is already embedded inside the price. Strong quarters lift the price; soft quarters pull it back. That is the most concrete evidence of recognition having been completed.

That self-stabilization is what the April 30 -6.02% really reflects: a market re-anchoring its ROE assumption from 20.7% to roughly 19.6%. The model didn’t break — the model is waiting for its next data point inside its own regime.

5. Followers on the Same Matrix — Kiwoom + Korea Investment + Samsung + NH

5.1 The Same ROE × PBR Matrix Applied to Securities

Reusing Part 1’s framework, the four Korean securities names map like this:

Arithmetic checks:

• Kiwoom = 20.7 / 1.39 = 14.89% ≈ 14.9% ✓
• NH = 17.1 / 1.18 = 14.49% ≈ 14.5% ✓
• Korea Investment Holdings = 16.8 / 1.07 = 15.70% ≈ 15.7% ✓
• Samsung = 15.8 / 1.05 = 15.05% ≈ 15.0% ✓

Observation. The earnings-yield leader is Korea Investment Holdings (15.7%). Kiwoom (14.9%) sits 0.8 ppt behind. In a recognition-completed market, that 0.8 ppt is not “discovery alpha” — it is the market correctly distinguishing each company’s model. Kiwoom = high-volatility dynamic model. Korea Investment Holdings = more stable capital-management + IB model. The market has priced both as different positions on the same matrix.

5.2 What Each Firm’s Variant Looks Like

The four firms carry different mechanism variants on the same standard.

The key point: all four have crossed the recognition threshold. All four have PBR ≥ 1×, all four print ROE ≥ 15%. The Korean securities sector as a whole has exited the “low-PBR discount asset” regime. Within that, each firm carries a distinct variant.

5.3 What Kiwoom Being at the Peak Means

The single-line summary of Kiwoom’s position inside this matrix:

Kiwoom Securities sits at the peak of the “ROE × trading-volume beta” variant. If Meritz is the peak of the “ROE × payout-ratio” variant, the same standard has bifurcated into two peaks of two variants on the same matrix.

The interesting thing is that the two peaks are not in conflict — they are complementary. Meritz: low-volatility capital-allocation algorithm → higher PBR. Kiwoom: high-volatility capital-turnover algorithm → higher ROE but slightly lower PBR. The market has correctly assigned both to the same “ROE-20%-class” category at different points.

Once the peaks are anchored, the next layer of the landscape comes from how Korea Investment Holdings, Samsung Securities, NH Investment & Securities each evolve their own model. If Korea Investment Holdings starts shifting capital-return form toward share-buyback-and-cancel, it tilts toward the Meritz variant. If Samsung and NH drive dividend yields higher, they consolidate around the capital-return-track identity.

6. The Next Verification Step — Signals That Track Model Durability

Not trading triggers. Observation points that show how the model carries forward into the next quarter.

6.1 The Trading-Volume Threshold — ₩44.8T

The cleanest piece of arithmetic in this entire post. 1Q26 KRX (KOSPI + KOSDAQ combined) daily-average trading volume was ~₩43.8T. April was ~₩41.9T, about -4.3% below the 1Q average.

For 2Q to exceed 1Q:
(41.9 + May + Jun) / 3 > 43.8

May–Jun average > (43.8 × 3 − 41.9) / 2
              = (131.4 − 41.9) / 2
              = 89.5 / 2
              = ₩44.75T ≈ ₩44.8T

Arithmetic check: (43.8 × 3 - 41.9) / 2 = ₩44.75T ≈ ₩44.8T ✓

Translation: if May–June daily-average trading volume averages ₩44.8T or higher, 2Q brokerage exceeds 1Q, and the sell-side “1Q peak → quarterly deceleration” assumption breaks. That re-anchors the market’s ROE assumption upward.

This single number is the fastest verification signal for Kiwoom’s model durability.

6.2 The Depth of Capital In-Market

Trading volume alone is too narrow. Capital depth has to be read together with it.

All three at near-all-time-high levels matters. Even with April daily trading volume slightly off, the depth of capital inside the market is greater. April’s dip looks more like “post-rally repositioning” than “capital outflow.”

6.3 The Tracking Set

Variables to watch as the cohort moves into the next quarters.

6.3.1 Kiwoom — Verification at the Peak

• May–June KRX daily-average trading volume ≥ ₩44.8T threshold. Most direct signal.
• Margin loan balance ≥ ₩35T sustained. Defends the interest-income line.
• Customer deposits ≥ ₩120T. Capital-depth-not-outflow confirmation.
• Brokerage market share. Whether Kiwoom maintains share even as KOSPI rallies — the structural verification of its retail-#1 position.

6.3.2 Korea Investment Holdings — Time-Gap Alpha Progression

• Buyback-and-cancel disclosures from Korea Investment Securities. Signals whether the dividend-led capital-return form is shifting toward more cancellation.
• 2026–2027 ROE durability in the 16–17% range.

6.3.3 Samsung Securities + NH Investment & Securities — Capital-Return Track Verification

• Dividend yield ~5% sustained. The accounting verification of capital-return identity.
• NH IB-revenue recovery cadence.

6.3.4 Sector-Level Meta Signals

• Frequency of “low-PBR brokerage” framing in Korean sell-side language. The recognition completion deepens as the framing fades.
• Whether the Korean brokerage cohort PBR average stabilizes above 1×. Verification of sector-level reclassification.

7. Two Honest Limits

Constructive tone shouldn’t mean overstating durability. Two real limits.

7.1 Volatility Has Not Disappeared

Kiwoom’s “ROE 20%-class” is not stable. Mirae Asset’s quarterly path alone implies 1Q ₩477.4B → 4Q ₩262B — a -45% intra-year move. The annual ROE of 20.7% is the average of a ROE-30%-class 1Q and a ROE-15%-class 4Q. That volatility is precisely why Kiwoom’s PBR sits below Meritz’s.

This is not a weakness — it is the model’s identity. Owning Kiwoom means accepting quarterly volatility in exchange for a higher annual mean ROE. That is a different exposure than owning Meritz’s lower-volatility capital-allocation algorithm.

7.2 Margin Loan ₩36T Is a Two-Sided Signal

End-April margin loan balance at ~₩36T is an all-time high. For Kiwoom, that is a near-term tailwind — interest income up, retail activity sustained. But the same number is also a two-sided signal:

• Overheated margin lending can lead brokerages to temporarily suspend new credit issuance.
• High margin-loan stocks face concurrent forced-selling and trading-volume contraction during corrections.
• If the regulator flags overheating, retail-leverage profitability gets a partial discount.

This two-sidedness is not a model defect. It is the structural feature of the “ROE × trading-volume beta” variant itself. Same type of model-identity feature as Meritz’s capital-sensitivity exposure to insurance/securities cycles.

8. The Single Closing Line

Kiwoom Securities’s recognition shift is already complete. The 2024 PBR-0.55× “cheap brokerage” era is gone, and the market has already classified Kiwoom as a verified ROE-20%-class capital-efficiency brokerage. The April 30 -6.02% post-print sell-off is not a model rejection. It is the operation of the self-stabilization mechanism that a recognition-completed price uses to demand quarterly model-verification data.

On the same matrix, Meritz holds the static peak (ROE × payout ratio); Kiwoom holds the dynamic peak (ROE × trading-volume beta). Korea Investment Holdings, Samsung Securities, and NH Investment & Securities each carry their own variant in between. The Korean securities sector as a whole has exited the “low-PBR discount asset” regime — and that fact alone is sufficient reason to keep this series tracking the cohort.

The next post in the series returns when (1) the May–June ₩44.8T daily-trading-volume threshold prints, (2) Korea Investment Holdings’s capital-return-form transition signals appear, and (3) the Korean brokerage-cohort average PBR stabilizes above 1×.

Appendix — Evidence Tier

[Fact]

• Kiwoom Securities April 30, 2026 close ₩398,000; market cap ~₩10.44T; 52-week range ₩132,100–₩517,000.
• 2026E EPS ₩59,426; 2026E BPS ₩285,909; 2026E ROE 20.7%; 2026E PBR 1.39×; 2026E DPS ₩15,500.
• Kiwoom 1Q26: operating profit ₩621.2B (+90.9% YoY), net income ₩477.4B (+102.6% YoY), equity-brokerage commission revenue ₩311.5B (+120.8% YoY), Kiwoom domestic-equity daily-average trading volume ₩27.8T (+215.9% YoY), customer AUM ₩21.8T (+43.4% YoY).
• Kiwoom historical PBR: 2024 0.55× → 2025 1.14× → 2026E 1.39×.
• April 30, 2026 post-1Q26-earnings reaction: -6.02%.
• 1Q26 KRX (KOSPI + KOSDAQ) daily-avg trading volume ~₩43.8T; April daily-avg ~₩41.9T.
• End-April 2026 customer deposits ~₩130T; margin loan balance ~₩36T (all-time high); April average margin loan ~₩33.8T.
• 2026E peer multiples: Korea Investment Holdings (071050) ROE 16.8% / PBR 1.07×; Samsung Securities (016360) ROE 15.8% / PBR 1.05% / dividend yield 5.4%; NH Investment & Securities (005940) ROE 17.1% / PBR 1.18× / dividend yield 5.9%.

[Inference]

• Kiwoom’s recognition shift from “cheap brokerage” to “ROE-20%-class capital-efficiency brokerage” is materially complete; PBR 1.39× sits inside the justified range under cost of equity ~14.9%.
• The April 30 sell-off reflects the market re-anchoring its ROE assumption (~20.7% → ~19.6%) rather than rejecting the model.
• The Meritz model and the Kiwoom model are two complementary peaks on the same standard — static capital-allocation compounding vs. dynamic capital-turnover compounding.
• The May–June ₩44.8T daily-trading-volume threshold is the most direct verification signal for whether the 2026 path is “1Q peak + decel” or “1Q is a base.”

[Speculation]

• A May–June trading volume sustained above ₩44.8T would re-anchor the market’s ROE assumption upward and pull the price back toward the ~₩423K range.
• Korea Investment Holdings shifting capital-return form toward buyback-and-cancel would narrow its valuation gap to Kiwoom on the matrix.
• Samsung Securities and NH Investment & Securities consolidating around the dividend-yield-track identity would create a clearer three-way variant landscape inside the cohort.

[Blocked]

• Per-quarter brokerage M/S, retail margin-loan stock concentration, and prop-trading P&L composition beyond what has been disclosed.
• Per-firm CET1-equivalent capital-headroom for further capital-return uplift in the brokerage cohort.
• Forward-looking capital-return-form transition timing across the four-firm cohort.

Disclaimer: This post is research commentary, not investment advice. ROE / payout / PBR / trading-volume scenarios are based on publicly available sell-side estimates (Mirae Asset Securities, Samsung Securities, Hana Securities, KB Securities, others) and company IR materials; actual results may differ. Tickers cited are illustrative for the framework, not recommendations. Do your own due diligence and consult licensed advisors before any investment decision.

Disclaimer: For research and information purposes only. Not investment advice. Names cited are for analytical illustration; readers should perform their own due diligence and consult licensed advisors before any investment decision.