The Bipolar Prime: Japan's 87% Concentration in Two Archetypes

The universe

The framework's Japan production scope is TSE Prime — 1,576 firms in the FY2025 yuho window (filings between March 2025 and April 2026).^[2] After NR-gating (foreign issuers, preferred-class duplicates, late-listed tickers), the active universe is 1,556. Standard and Growth tiers are not yet in production; the boundary reads they would generate are available through tier crosswalk but not used for the universe-level distribution below.

The universe is pre-financials-excluded by sector convention but not pre-bank-excluded — banking sits inside the distribution at 69 firms (4.4% of the universe). This matters for the sectoral readings further down.

A B-modal grade distribution

Grade distribution in Japan Prime:^[3]

Nearly half the universe — 770 firms — sits at Grade B. Add Grade A and 76% of Prime is in the upper-middle band. The tails are thin: only 5 firms reach S, only 8 sit at D, and 115 fire a Kill Switch.

This is a B-modal distribution with compressed tails. It is the shape of a market where the floor has been raised — almost every firm clears a threshold of basic disclosure adequacy — but the ceiling is also compressed. Few firms reach the elite tier; very few collapse to the rated bottom.

Korea's distribution, for contrast, is flatter and more skewed: a wider tail at C and D, fewer firms concentrated in any single grade band.^[1] Japan's is the shape of a market that has spent a decade operationalizing comply-or-explain and is now harvesting the floor effect.

A polar archetype distribution

The archetype distribution is the more striking number:^[3]

Celestial and Chameleon together account for 87% of Japan Prime. The four narrative archetypes — KS, Poison Apple, Hidden Gem, Time Bomb — collectively describe only 13%. Time Bomb fires zero times in the production universe.

Compare with Korea, where Chameleon alone accounts for 88% of the universe and Celestial sits below 2%.^[1] The Korean monoculture is single-pole; the Japanese signature is bipolar.

What does the bipolar signature mean? In the framework's archetype taxonomy:

• Celestial — every axis clears 70. No structural concern surfaces on the headline read.
• Chameleon — exactly one axis sits below 70 while the others remain above. The classification carries a sub-tag indicating which axis lags.

The bipolar reading, then, is that Japan Prime divides into two roughly equal cohorts: one where every axis clears the floor, and one where exactly one axis sinks below it. The middle archetypes — Hidden Gem (modest on T but strong on B and R), Poison Apple (high T disclosure with hidden B or R weakness), Time Bomb (multi-axis collapse short of KS) — are residual.

This is not a Korean-style story of "almost everyone is partway compromised." It is a story of "almost everyone is either clean on all three axes or compromised on exactly one."

The missing T-weak

The Chameleon sub-tag distribution sharpens the picture:^[3]

T-axis is never the lowest axis on Japan Prime. Not rare — zero. This is a structural feature, not a measurement artifact.

The mechanics: T-axis at sector-level median is 79.0 across every one of the ten largest sectors.^[3] T-02 (audit opinion) sits at maximum value for 97% of the universe. T-04 (J-SOX internal control attestation) sits at maximum for 99%. The T-axis floor is high enough that T-axis is mathematically prevented from being the weakest of the three for almost any Prime firm.

What this says about Japan: the formal disclosure layer — the yuho, the audit opinion, the J-SOX attestation, the comply-or-explain depth on the CG Code — has saturated. After a decade of CG Code 2015 / 2018 / 2021 revisions,^[4] three rounds of Stewardship Code revisions,^[5] and the 2022 TSE Prime market reform,^[6] formal disclosure quality is no longer the differentiator across firms. Where firms now differ is what they disclose about — not whether they disclose.

The Korean analog has not saturated this way.^[1] In Korea, T-axis still discriminates because filing-timeliness and disclosure-completeness gaps remain meaningful at the universe level. Japan has moved past that floor.

This is the empirical content of the often-asserted claim that Japan's governance reform has worked. It has — on the disclosure layer. The question Japan analysts now have to answer is not whether firms disclose but what the disclosures conceal.

Sub-tag inversion

Within the Chameleon cohort, the dominant lag is on the R-axis — conflict-of-interest risk — at 57.6%. B-axis lag is the runner-up at 42.4%.

This inverts the Korean pattern. In Korea, B-weak Chameleons dominate: outside-director independence is technical not substantive, audit committees lack teeth, parent-subsidiary structures concentrate power. The Korean Chameleon is fundamentally a board-architecture problem.^[1]

Japan's Chameleon is fundamentally a related-party problem. Cross-shareholding (政策保有株式), capital-efficiency profile, cash-earnings divergence, large-shareholding patterns, anti-takeover defenses — these are where the modal Japanese Chameleon lags. The board-architecture indicators have improved enough across the universe — partly through the 2015 Audit-and-Supervisory Committee company (監査等委員会設置会社) introduction,^[7] partly through the 2022 TSE Prime tightening of independent-director definitions^[6] — that they are no longer the dominant Chameleon weakness.

Two markets, two opposite stories about where the same archetype actually breaks.

Sectoral and size gradients

Two cross-cuts sharpen the universe shape further.

Banking is the universe outlier. Of 69 banks in Prime, 19 fire a Kill Switch — a 28% KS rate, nearly four times the universe-wide rate of 7.3%.^[3] Banking R-axis median (57.6) sits 18 points below the universe-wide R-axis median (75.3). The dominant trigger is cross-shareholding. Japanese regional banks remain the structural-keiretsu / structural-shareholder cluster that has not unwound, and the framework reads them as such.

Size correlates with archetype. Celestial share by TOPIX size bucket:^[3]

Celestial concentration declines steeply across the size buckets. KS concentration moves the other direction, though more gently. Poison Apple — high disclosure paired with hidden weakness — is large-cap-rare and small-cap-concentrated.

This is partly intuitive: larger firms have more disclosure resources, more institutional scrutiny, more historical pressure to comply with code revisions. But the gradient is steep enough to be analytically consequential. The archetype distribution at TOPIX Core30 is fundamentally a different distribution from the one at TOPIX Small 2, and a framework reading at the index level should not be applied uncritically across size buckets.

The Large70 cohort is worth a separate note: it outperforms Core30 on Celestial share. The "big-but-not-mega" tier carries the strongest archetype profile in the universe, suggesting that the Core30 mega-caps carry structural-keiretsu legacies their slightly smaller peers have shed.

What the distribution implies for the analyst

Three readings follow from the distribution shape.

First, Japan rewards axis-specific reading rather than composite reading. Because 87% of the universe is concentrated in two archetypes that differ on whether any single axis breaks below 70, the composite g-score by itself loses information. The diagnostically useful question is not "what is the composite?" but "which axis, if any, is the weakest?" — and within Chameleon, the sub-tag carries more signal than the headline grade.

Second, the dominant pathology in Japan is not board structure. The R-weak Chameleon majority, combined with the cross-shareholding cluster that accounts for the overwhelming share of Kill Switch firings, means the analytic center of gravity for Japan governance research is conflict-of-interest risk — related-party transactions, capital efficiency, structural shareholding, anti-takeover defenses — not board independence. Frameworks that lead with board-of-directors metrics systematically under-weight what actually moves the Japanese governance distribution.

Third, the Banking sector and the small-cap tail carry the unwind problem. The 7.3% of firms that fire a KS are heavily concentrated in regional banks and the TOPIX Small 2 segment. The structural-keiretsu cluster that has not responded to a decade of unwinding pressure sits there. Cross-shareholding reform — the headline TSE Prime Market initiative since 2022^[6] — has worked at the top of the distribution and stalled at the bottom.

The distribution map is the foundation. The pathology Notes that follow read into the cross-shareholding architecture, the parent-subsidiary unwinding wave, the shadow-governance layer of sodanyaku and komon (相談役 and 顧問), and the four-tier audit-governance model. Each is a vertical cut into the universe described above.

The horizontal map is bipolar. The vertical cuts explain why.

Apex G-Score™ Japan Foundation Series. Production cohort: TSE Prime, 1,576 issuers (1,556 rated; 20 NR). The "FY2025" label denotes a rolling-window cohort — each firm's most recent yuho filed within 2025-03-11 to 2026-04-15. Fiscal year coverage by 決算月 (fiscal year-end month): 3月 (March) ~76.5% (reporting period 2024-04 to 2025-03); 12月 (December) ~13.7% (calendar year 2025); 9月 (September) ~5.1% (2024-10 to 2025-09); 6月 (June) ~4.6% (2024-07 to 2025-06). The December-end sub-cohort shares the calendar-year reporting window of the Korea cohort. Framework: Apex G-Score v2 (T 0.30 / B 0.30 / R 0.40); archetype classifier: unified-v1. Sample firm-level scores in this article are limited to the IP-guardrail-v2.0 disclosed set (Toyota Motor; Samsung Electronics and Reliance Industries appear in cross-market comparisons only). All other firm-level scores remain unpublished. Multi-year indicators where cited are noted explicitly with source vintages. This article is research output, not investment advice; readers should consult their own advisors before making investment decisions.