Five Industries, One B-Weak: Family Conglomerate Cluster Pathology

The Far Eastern Signature

Far Eastern New Century — industrial polymer. Heng Yang Petrochemical — petrochemical. Far EasTone — telecom. Far Eastern Department Stores — retail. Farglory Construction — construction. Five listed companies in five different industries. All five classify on the Apex G-Score framework as Chameleon, sub-tagged B-weak, Grade B.^[1]

When one ultimate controller produces an identical governance signature across five unrelated industries, the explanation is not the industries. The explanation is the controller.

The Far Eastern Group, controlled by the Hsu family, operates the purest example of family conglomerate cluster pathology in the Taiwanese listed universe. Each of the five entities satisfies the framework’s T-axis threshold. Each clears the R-axis threshold. On the B-axis — board independence, insider board concentration, ownership-control wedge — each falls short. The pattern is uniform. The pathology is inherited.

Thirteen Clusters

Taiwan does not publish a government-curated designated enterprise group list equivalent to Korea’s Fair Trade Commission system.^[2] Thirteen clusters spanning 34 listed firms represent the identifiable lower bound on family-controlled firm presence in the Taiwanese listed market.

The cluster cohort skews toward Chameleon (50% versus 39% universe) and away from Celestial (44% versus 49%). Hidden Gem and Poison Apple are entirely absent.

B-Weak, Not R-Weak

The hypothesis entering this analysis was that family conglomerate clusters would concentrate R-axis weakness. The data does not support this hypothesis.

Cluster Chameleons are 88% B-weak — seventeen percentage points above the universe baseline of 71%. The family conglomerate governance signature is board-axis weakness, not conflict-of-interest-axis weakness.

Three Patterns

The thirteen clusters do not exhibit a single pattern. The framework reads three distinct signatures across major groups.

The absorption pattern. The Formosa Plastics Group, controlled by the Wang family, lists four entities. Three classify as Celestial. Only the parent entity reads as Chameleon with an R-weak sub-tag.^[3] The cluster pathology concentrates in the parent, which absorbs the cross-firm related-party transaction signal. The implication is counterintuitive: buying a subsidiary of a family conglomerate is not necessarily buying the cluster pathology.

The mixed pattern. The Hon Hai / Foxconn Group lists five entities under the Gou family’s historical orbit. Three classify as Celestial. Two do not — one reads as R-weak Chameleon and one reads as B-weak Chameleon. The mixed pattern demonstrates that cluster membership does not determine archetype mechanically.

The pure pattern. The Far Eastern Group’s five-for-five Chameleon B-weak result is the outlier, not the norm. Among the thirteen identified clusters, only Far Eastern produces a uniform governance signature across all listed entities.

What Clusters Don’t Trigger

Of 78 KS-1 firings — excessive insider share pledging — across the entire Taiwanese universe, zero occur within the identified cluster cohort.^[4] The cluster KS rate overall (5.9%) is lower than the universe baseline (7.7%).

The pledging risk that Note 2 identified as concentrating on TWSE is not a family conglomerate phenomenon. It is a non-cluster controlling-shareholder phenomenon. This distinction matters for policy analysis. Pledging reform that targets conglomerate structures addresses a problem that the data says conglomerates do not have.

Where This Leads

Taiwan’s independent director mandate has been in force for over two decades. The audit committee migration is structurally complete. The independent director ratio meets or exceeds the regulatory minimum at 92% of the universe. And yet — the cluster cohort produces an 88% B-weak Chameleon rate, seventeen points above the universe baseline. The boards exist. The independent directors are seated. The cluster governance signature persists through all of it.

Note 4 takes the question to the universe level: what has twenty-three years of board reform actually delivered — and what has it left unmeasured?

Apex Governance LLC · Taiwan Foundation Series · Note 3 of 6. Production data: GS-TWN v06 (scored 2026-04-27). Cluster cohort: 13 groups, 34 firms (curated lower bound). Cluster identification: publicly-known related-party disclosures (公司法 §369); Taiwan does not publish a government-designated enterprise group list. 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. Per Apex G-Score IP guardrails: indicator weights, classifier thresholds, and per-firm exact scores (outside Sample 3-firm) are NDA.