Five Origins, One Exchange: The 5-Tier Architecture of HKEX Governance

Four Numbers, Five Regimes

H-share issuers carry a Balance of Power axis mean of 46.1 — the lowest in the panel. Red Chip issuers carry a Transparency axis mean of 86.7 — the highest. HK-Local and P-chip issuers carry nearly identical Transparency means (72.5 versus 71.5), but their Balance of Power means diverge by 9.7 points. Foreign secondary listings register at composite mean 47.3, on a base of ten issuers.

These four numbers describe what Note 1 only sketched. The 52.2 percent Chameleon share at the universe level, and the R-weak signature within it, average over five distinct governance regimes that share a single listing venue. HKEX hosts the five, but the disclosure floor it imposes lifts only one axis uniformly. Below the disclosure axis, origin separates into different patterns of compliance, control, and conflict-of-interest exposure. The framework's three-axis architecture surfaces those patterns; this Note describes them.

---

The Five Origin Types

The framework classifies HKEX equity-only issuers along the dimension of incorporation jurisdiction, controlling-shareholder regime, and listing pathway^[1]. Five Origin Types result.

Origin	Definition	n	% of universe
HK-Local (HKL)	HK-incorporated, HK-operating	2,325	84.0%
P-chip (PCH)	Cayman/BVI offshore-incorporated, mainland-private operating	373	13.5%
H-share (HSH)	Mainland-incorporated, direct HKEX listing	40	1.4%
Red Chip (RCH)	State-controlled, offshore-incorporated, mainland-operating	20	0.7%
Foreign (FOR)	Non-HK, non-mainland secondary listings	10	0.4%

HKL is the dominant cohort by a wide margin. PCH is the second-largest at 13.5 percent — Cayman or BVI parent vehicles whose operating activity is mainland-based but whose incorporation, listing, and equity-issuance vehicle are offshore. HSH is small in count but distinct in pattern: mainland-incorporated companies that list directly on HKEX as H-shares^[2], typically with a state-supervised mainland parent or sponsor as majority shareholder. RCH is also state-controlled but offshore-incorporated, routing mainland operating activity through a non-mainland parent. FOR is the smallest cohort: foreign issuers using HKEX as a secondary listing venue, with primary disclosure governed by a home-market regulator.

A structural overlay applies on top of the origin tier: the Variable Interest Entity (VIE) structure, n=49 in the universe (1.8 percent), all of which sit within the PCH origin (13.1 percent of the PCH cohort)^[3]. The VIE overlay is treated separately later in this Note.

---

Where Each Origin Lands on the Grade Distribution

The framework's grade bands aggregate composite scores into S/A/B/C/D bands plus a KS override for issuers triggering the Kill Switch criteria.

Origin	n	S	A	B	C	D	KS	Composite mean	Median	IQR	SD
HKL	2,325	0.0%	5.4%	32.9%	32.5%	20.9%	8.1%	65.5	67.0	60–72	9.83
PCH	373	—	2.7%	21.4%	27.9%	41.6%	6.4%	60.9	61.0	52–69	10.26
HSH	40	—	—	5.0%	42.5%	50.0%	2.5%	59.0	59.5	55–64	6.96
RCH	20	—	—	30.0%	65.0%	5.0%	—	66.6	66.0	62–70	5.70
FOR	10	—	—	—	10.0%	90.0%	—	47.3	45.0	45–45	7.27

HKL is universe-pacing — its composite mean (65.5) and median (67.0) sit close to the universe values, with grade representation spanning every band including the single S-grade issuer in the universe. As 84 percent of the panel, HKL is the cohort against which other origins are read.

PCH and HSH both carry high D-rates — 41.6 percent and 50.0 percent respectively. Their composite means (60.9 and 59.0) sit several points below HKL, and their distributions skew to the lower grade bands. The two cohorts reach the same bottom-band skew through different governance mechanisms, examined in later sections.

RCH presents an inverted pattern: composite mean 66.6 — above HKL — but with the distribution compressed into a narrow band. Ninety-five percent of RCH issuers sit in B or C, with no S, A, or KS firms and an SD of 5.70 (the tightest in the panel by a wide margin). FOR is the panel's other end — composite mean 47.3, with nine of ten firms in D-band. The FOR pattern reflects calibration mismatch rather than underlying governance and is examined directly later in this Note.

---

Where Each Origin Lands on the Archetype

The grade distribution describes how each origin scores in absolute terms. The archetype distribution describes which axis pattern each origin tends toward.

Origin	Celestial	Hidden Gem	Chameleon	Poison Apple	Time Bomb	KS
HKL	7.7%	9.0%	52.2%	23.0%	0.04%	8.1%
PCH	5.9%	2.9%	53.1%	31.6%	—	6.4%
HSH	—	5.0%	30.0%	62.5%	—	2.5%
RCH	—	—	55.0%	45.0%	—	—
FOR	—	—	100%	—	—	—

The HSH cohort is the panel's most distinctive archetype concentration. At 62.5 percent within-origin Poison Apple, H-share issuers carry the highest within-origin Poison Apple density the framework has observed across the eight Asian markets it covers. Note 3 in this series unpacks the Poison Apple distribution directly.

RCH presents a clean two-archetype distribution: Chameleon at 55 percent and Poison Apple at 45 percent, with no other archetypes. The 20-firm cohort distributes binomially.

FOR is uniformly Chameleon (100 percent on a base of ten). The cohort's distinctive archetype profile reflects the calibration-mismatch pattern referenced earlier — every Foreign secondary issuer registers as a Chameleon because every one carries the high-T-low-B-low-R profile that mismatch produces.

---

The Three Most Distinctive Origin Patterns

The framework's reading of the five origins surfaces three governance regimes that diverge most sharply from the universe baseline.

HKL versus PCH

The largest two cohorts in the universe — HKL and PCH — share a single feature on the Transparency axis (means 72.5 and 71.5 respectively, essentially identical) but split sharply on the Balance of Power axis. HKL's B-axis mean is 66.8; PCH's is 57.1. The 9.7-point gap is the panel's largest between any two adjacent-density origins.

Mechanically, HKL issuers are HK-incorporated under the Hong Kong Companies Ordinance (Cap. 622)^[4], listed under HKEX Chapter 8, and operating in HK or HK-extended jurisdictions. The HKL cohort spans family-conglomerate holdings (the Cheung Kong / CK Hutchison axis, the Henderson axis, the Wharf-Wheelock axis), property-and-financial conglomerates, regulated utilities (MTR Corporation, the Power Assets axis, China Light and Power), and SFC-regulated financial institutions. The B-axis pattern within HKL — mean 66.8 — reflects independent-director architecture under HKEX Listing Rule 3.10 plus the nine-year tenure cap effective from 2023, with controlling-shareholder structures running through transparent equity blocks rather than through contractual or offshore-vehicle arrangements.

PCH issuers are Cayman or BVI offshore-incorporated, with mainland-operating activity routed through subsidiary chains. Controlling-shareholder structure typically tracks a mainland founder-block pattern — single founders or founder families holding majority equity through offshore special-purpose vehicles, with board compositions that operate within HKEX listing-rule formal compliance but whose functional decision-making sits with the founder block. The B-axis penalty captures that founder-control depth in a way disclosure-axis indicators do not.

R-axis means converge — HKL 56.8 versus PCH 55.3. Both cohorts are exposed to Chapter 14A connected-transaction reporting; both register R-axis weakness from controlling-shareholder transaction streams. The R-axis is uniform across the two largest origins, the B-axis diverges, and the T-axis is essentially identical. The structural distinction between HK family-conglomerate governance and mainland-private founder-control governance, on this evidence, sits primarily on the board axis.

---

HSH and the State-Shareholder Pattern

The H-share cohort is small (40 firms) but it carries the panel's sharpest within-origin pattern.

HSH Balance of Power mean: 46.1. Median: 45.0. The cohort's IQR on the B-axis runs from 35.0 to 55.0 — fully below the universe median of 65.0. Where HKL's B-axis distribution centers at 67.5 and PCH's at 55.0, HSH's centers at 45.0. The cohort is uniformly board-axis-deficient — not as an outlier subset but as the entire cohort.

The structural mechanism is direct. H-share issuers are mainland-incorporated and listed directly on HKEX under Chapter 19A, typically with majority equity held by a state-controlled mainland parent or sponsor (subject to SASAC oversight on the parent side and HKEX listing-rule compliance on the subsidiary side)^[5]. The framework's B-axis indicators read controlling-shareholder concentration, board appointment patterns consistent with state-board nomination, and director-tenure structures that depart from the HKEX-mandated independent-director model. H-shares register weakness on each.

The cohort's archetype distribution follows from the axis profile. Transparency at 77.9 (well above market average, lifted by both HKEX listing-rule disclosure and CSRC-side annual reporting) paired with Balance of Power at 46.1 produces a sharply asymmetric T-B profile. The framework places high-Transparency-low-Balance profiles in the Poison Apple archetype, and HSH ends up at 62.5 percent within-origin Poison Apple — a within-origin density not matched by any other cohort in the panel. The remaining 30 percent of HSH classifies as Chameleon, with Balance of Power again dominant as the weakest sub-tag within that subset.

The HSH cohort is small enough that small-n caveats apply to specific axis values. The directional pattern is consistent and reproduces inside both the Chameleon and Poison Apple sub-cohorts of HSH. State-shareholder concentration is the cohort's defining structural feature, and the framework reads it as a B-axis depression that no amount of disclosure-axis lifting can offset.

---

RCH and SOE Double-Supervision

The Red Chip cohort is smaller still (20 firms) but exhibits the panel's clearest top-axis pattern.

RCH Transparency mean: 86.7. Median: 86.7. The cohort's T-axis SD is 5.19 — the tightest of any cohort, by a wide margin. Three out of every four RCH issuers score within five points of each other on the T-axis.

Mechanically, the RCH cohort operates under a double-supervision overlay. Red Chip issuers are state-controlled but offshore-incorporated (typically Hong Kong, Cayman, or Bermuda parent vehicles), with mainland operating presence routed through subsidiary structures. The state-control side imposes SASAC governance disclosure requirements on the parent. The HKEX listing side imposes Chapter 13, the Corporate Governance Code, and the bilingual disclosure mandate on the listed vehicle. Both regimes mandate detailed annual financial reporting, audit by an internationally recognized firm, and structured disclosure of related-party transactions. RCH issuers — subject to both — score uniformly at the top of the T-axis.

Below the T-axis ceiling, the RCH cohort's other axes track patterns adjacent to HSH but with less severity. B-axis mean 58.4 sits between HKL and PCH; state ownership produces some board-axis weakness but at a less extreme level than direct mainland-incorporation does. R-axis mean 58.5 sits at the universe baseline. The clean two-archetype distribution — 55 percent Chameleon, 45 percent Poison Apple, no other archetypes — reflects the narrow axis-profile band the cohort occupies.

RCH n=20 is a small base. The directional pattern — top-T, mid-B, mid-R — is robust across the cohort and reproduces the SOE double-supervision interpretation cleanly.

---

FOR and the VIE Overlay

The Foreign secondary cohort and the VIE structural overlay each describe a specific subset of the universe that warrants separate treatment.

FOR (n=10) is the panel's smallest origin and lowest-scoring cohort by composite mean (47.3). The pattern is consistent across the ten firms: low Transparency (45.7), low Balance of Power (41.8), mid-range Conflict-of-Interest Risk (56.0), all-Chameleon archetype assignment, all 100 percent [B-weak] sub-tag. The interpretation is structural rather than diagnostic. Foreign secondary listings on HKEX are issuers whose primary listing is elsewhere — a UK-primary listing under the FCA, a US-primary listing under the SEC, a Singapore-primary listing under the MAS. Their primary disclosure is calibrated to the home-market regulator, and the framework's HK-calibrated indicators read only the HK secondary-listing filings. The result is a systematic axis-score depression that reflects mismatch between framework calibration and dual-listing structure rather than the firms' substantive governance quality. The n=10 base limits inferential weight; the interpretation is qualitative.

The VIE overlay is structurally different from the Origin Type axis. VIE-structured issuers are not a separate origin tier but a subset of an origin — in the HK universe, all 49 VIE-overlay firms sit within PCH. The VIE structure layers a contractual control architecture on top of equity incorporation: a Cayman-incorporated parent vehicle holds equity in a wholly foreign-owned enterprise (WFOE) that operates inside mainland China, with the underlying mainland operating company controlled through a chain of contractual arrangements (consulting agreements, exclusive technology licenses, equity pledges, voting proxies) rather than through equity ownership. The structure is most associated with the Chinese internet, education, and online-services sectors, where direct foreign equity ownership has historically been restricted under mainland law.

The framework's reading of VIE versus non-VIE PCH cohorts is asymmetric.

Cohort	n	T mean	B mean	R mean	Composite mean
VIE (all PCH)	49	72.1	53.9	61.6	61.8
PCH non-VIE	324	71.4	57.5	54.4	60.7
Δ (VIE − non-VIE)		+0.7	−3.6	+7.2	+1.1

VIE-structured issuers carry slightly lower B-axis scores than non-VIE PCH peers (by 3.6 points) but materially higher R-axis scores (by 7.2 points), with composite means essentially equivalent. The VIE structure does not produce a uniform governance penalty across axes — its effect runs in opposite directions on the board axis (slight depression) and the conflict-of-interest axis (substantial improvement). The R-axis improvement reflects a feature of the architecture sometimes overlooked: contractual-control structures generate explicit, documented service agreements between the parent vehicle and the operating entity, and those agreements register in the framework's R-axis indicators as structured rather than informal related-party flows. Whether the explicit documentation produces substantively better governance is a separate question; the framework reads it as cleaner.

The VIE archetype distribution within the 49-firm cohort is Chameleon 31, Poison Apple 13, Celestial 4, KS 1. The cohort skews more toward Celestial than the PCH non-VIE cohort does, again on a small base.

---

Five Regimes, One Exchange

The HKEX equity universe is a single listing venue. Its disclosure regime, board governance architecture, and Chapter 14A connected-transaction framework apply uniformly to all listed issuers. The framework's reading is that this uniform architecture produces five distinct governance patterns when laid over five distinct incorporation, control, and listing regimes.

HK-Local issuers — the dominant cohort — sit at the universe baseline on every axis. PCH issuers replicate HKL's Transparency baseline but diverge on Balance of Power, where mainland-private founder-control structures register as a 9.7-point B-axis depression. HSH issuers sit at the panel's lowest B-axis mean (46.1) and the panel's highest within-origin Poison Apple share (62.5 percent), driven by direct mainland incorporation and state-supervised parent-shareholder structures. RCH issuers sit at the panel's highest T-axis mean (86.7) under SOE double-supervision but compress into a narrow C-and-B grade band. FOR issuers exhibit a calibration-mismatch pattern at the panel's lowest composite mean (47.3) on a base of ten secondary-listed firms whose primary disclosure happens elsewhere. The VIE overlay, sitting on top of the PCH origin in 49 cases, produces an asymmetric axis profile that runs against the common assumption of uniform VIE governance penalty.

The R-weak Chameleon signature this series identified in Note 1^[6] — the universe-level pattern that defines the framework's reading of HKEX governance — is, at the origin level, primarily an HKL pattern. The mainland-linked cohorts (PCH, HSH) carry B-weak as the dominant Chameleon sub-tag instead, and FOR carries it uniformly. What unifies the five origins is the disclosure floor; what divides them is the controlling-shareholder regime, and the regime separates by jurisdiction of incorporation rather than by sector or market capitalization. Note 3 takes up the cohort that most concentrates the divergence: the 686 Poison Apple issuers, of whom 534 are HKL, 118 are PCH, and 25 are HSH — carrying high-Transparency-low-Balance-and-Conflict profiles that the framework reads as Hong Kong's signature governance distortion.

---

The Apex G-Score framework currently covers 2,768 Hong Kong listed companies under v2.0 calibration. Origin Type classifications are derived from incorporation jurisdiction, listing pathway (Chapter 8 versus Chapter 19/19A), controlling-shareholder structure, and operating-presence geography. The VIE structural overlay is identified through prospectus and corporate-structure language consistent with contractual-control architecture; the n=49 figure reflects manual verification of the canonical VIE pattern and may understate firms using less standard VIE phrasing in their structural disclosure.