Zero T-Weak: A Saturation Pattern in Taiwanese Governance

TSMC Is a Celestial

TSMC, read on the Apex G-Score framework, classifies as Celestial — the framework’s designation for governance profiles that clear the structural threshold for strength on every measured axis.^[1] So does MediaTek. So does Delta Electronics. Among the largest technology companies on the Taiwan Stock Exchange, the Celestial archetype is not the rare exception it is in Korea. It is close to the default.

Across the full Taiwanese listed universe of 1,962 companies, forty-nine percent of issuers — 963 companies — classify as Celestial. Another thirty-nine percent — 775 companies — classify as Chameleon. The remaining twelve percent splits across Kill Switch overrides at 7.7%, Hidden Gem at 2.4%, Poison Apple at 1.4%, and Time Bomb at zero.

Korea, measured on the same ruler, reads as 88% Chameleon. Taiwan reads as a near-even Celestial-Chameleon split. The same framework, the same three axes, the same archetype classifier. Two East Asian markets producing two structurally different distributions.

The reason is not that Taiwanese governance is categorically stronger. It is that one axis has saturated.

What 49/39 Looks Like

The full distribution under the v2 framework (T 0.30 / B 0.30 / R 0.40) is the following:^[1]

Two features of this distribution require structural explanation. The first is the empty Time Bomb classification — zero of 1,962 issuers qualifies. The Time Bomb designation requires all three axes to fall below the framework’s structural threshold simultaneously. In Taiwan, the T-axis distribution makes this mechanically unreachable: nearly the entire universe clears 70 on the T-axis, preventing the three-axis floor condition from triggering. Companies that would otherwise qualify as Time Bomb are instead caught by the Kill Switch override, which fires on specific indicator-level conditions that operate independently of the composite score.

The second is the 49/39 split itself. Korea’s 88% Chameleon concentration produces a market that the framework reads as essentially one archetype with fine-grained differentiation within it. Taiwan’s near-even split produces a market that the framework reads as structurally bifurcated — two large populations with different governance profiles, coexisting within the same regulatory regime.^[2]

The Celestial-Chameleon split partially aligns with a technology and non-technology sector divide. Technology companies — spanning semiconductor, communications, optical, and electronics industry codes — constitute 42% of the universe and produce a 59% Celestial rate. Non-technology companies produce a 42% Celestial rate and a Kill Switch prevalence 1.6 times higher.

The Saturated Axis

The structural driver of Taiwan’s distribution shape is the Transparency axis. The T-axis measures disclosure compliance, filing completeness, and information accessibility. In Taiwan, the axis has reached a saturated plateau: 96.7% of the universe scores at or above 70, and the interquartile range has compressed to an exceptionally narrow band. The distribution is not bell-shaped. It is a cliff — nearly the entire universe clusters in a tight range in the medium-high band, with very few firms either substantially below or substantially above the cluster.

The saturation is genuine market signal, not a measurement artifact. Taiwan’s disclosure regime has been shaped by two decades of cumulative reform: mandatory IFRS-equivalent adoption since 2013, English-language corporate governance report requirements for qualifying TWSE issuers, the FSC Corporate Governance Evaluation system applied to both TWSE and TPEx boards, and twenty-one years of progressive independent director mandates.^[3] Four of six T-axis indicators in the framework’s Taiwan calibration score at maximum levels for over ninety percent of the universe — a direct reflection of the disclosure infrastructure Taiwan has built.

The consequence for archetype classification is mechanical. When the T-axis cannot differentiate between issuers, the archetype classifier’s ability to sort companies into structurally distinct profiles depends entirely on the B-axis and R-axis. Taiwan’s 49/39 Celestial-Chameleon split is, fundamentally, a B-axis and R-axis split.

The contrast with Korea is informative. Korea’s T-axis has not saturated — it retains meaningful variance across the universe, and T-weak Chameleons constitute 9% of the Korean Chameleon population. In Taiwan, T-weak Chameleons constitute exactly zero percent.

Zero T-Weak

Of Taiwan’s 775 Chameleon issuers, the sub-tag distribution is the following:

The T-saturation regime is so complete that T-axis cannot be the lowest axis for any Taiwanese company. Even among non-Celestial firms — issuers that fail the structural threshold on at least one axis — the T-axis still clears the threshold. The Chameleon decomposition collapses to three sub-tags. This is not a classifier choice. It is a structural feature of the disclosure regime.

The B-weak dominance within Taiwan’s Chameleon population — 71.4% — echoes the Korean pattern, where B-weak Chameleons account for 89.7% of Korean Chameleons. Both East Asian markets exhibit board-axis weakness as the dominant Chameleon pathology. The direction is shared. The intensity is not.^[4]

Two Pathologies, Not One

Taiwan’s Chameleon population carries two governance pathologies, coexisting under a single archetype label.

The B-weak cohort — 553 companies, 71.4% of Chameleons — exhibits the pattern the framework reads across the East Asian regional cluster: board structures exist but do not generate the substantive oversight, challenge, and shareholder-rights infrastructure that the framework’s structural threshold requires.

The R-weak cohort — 203 companies, 26.2% of Chameleons — exhibits a different pattern. The R-axis measures conflict-of-interest exposure: related-party transactions, insider pledging, cross-shareholding structures, capital dilution. In Taiwan, the R-axis distribution is bimodal — one cluster at lower-medium and one at upper-medium-high — reflecting a structural split between companies with concentrated family or cross-shareholding exposure and companies without.

The Kill Switch composition confirms the reading. Two B-axis-derived triggers — KS-1 (excessive insider pledging) and KS-4 (board capture by a single institutional shareholder) — account for ninety-five percent of all Kill Switch firings in Taiwan.^[5] The dominant risk pathways in this market are balance-of-power failures, not transparency failures.

The T-saturation regime makes both pathologies visible because the T-axis can mask neither.

Where This Leads

Three features of Taiwan’s governance distribution set the structural context for what follows.

The first is that Taiwan’s governance baseline floor is materially higher than Korea’s. The T-saturation is not cosmetic — it reflects a disclosure infrastructure built across two decades that produces measurable compression of the T-axis distribution. The Celestial population at 49% indicates that nearly half of the listed universe clears the framework’s structural threshold on all three axes.

The second is that the variance that remains sits on the B-axis and R-axis, and the two axes produce different pathologies. The B-weak Chameleon pathology is a regional East Asian pattern shared with Korea. The R-weak Chameleon pathology is Taiwan-specific in scale.

The third is that the distribution as reported is aggregate — TWSE and TPEx combined. Whether the two boards exhibit the same distribution shape, or whether the 49/39 split masks a board-level bifurcation analogous to Korea’s KOSPI/KOSDAQ gap, is the question Note 2 addresses.

Apex Governance LLC · Taiwan Foundation Series · Note 1 of 6. Production data: GS-TWN v06 (scored 2026-04-27). Universe: 1,965 issuers — TWSE Mainboard 1,082 + TPEx Mainboard 883. Source: TWSE OpenAPI · MOPS public filings · doc.twse.com.tw PDF cohort. Taiwan’s annual report filing deadline falls at end of Q2; FY2025 financials were not universe-complete at the April 2026 scoring date. FY2024 is the latest fiscal year with full coverage across all 1,965 issuers. Korea reading: FY2025 cross-section. Taiwan reading: FY2024 cross-section. The one-year offset reflects FSC filing-deadline calendar differences (Taiwan Q2 vs Korea Q1). Per Apex G-Score IP guardrails: indicator weights, classifier thresholds, and per-firm exact scores (outside Sample 3-firm) are NDA.