A three-axis governance scoring framework built entirely on publicly verifiable regulatory filings. Designed to measure what single-score ratings structurally cannot.
Consider a company that publishes a 200-page sustainability report, holds quarterly earnings calls, and discloses every related-party transaction down to the last won. Impressive transparency. Now consider that the same company has a board stacked with insiders, no meaningful separation between chairman and CEO, and an ownership structure that makes minority shareholders irrelevant.
Most ESG ratings would give this company an acceptable governance score. Both dimensions — disclosure and power distribution — are collapsed into a single number. A company that tells you everything but lets you decide nothing receives the same credit as one that does both.
This is not a hypothetical edge case. It is a structural pattern across the Korean market, and it is the primary reason the Korea Discount persists. The measurement tool is broken: it evaluates whether governance institutions exist, not whether they function. The Apex G-Score was built to separate what existing frameworks conflate.
Does the company disclose honestly, consistently, and on time?
The T-axis measures the structural quality of disclosure — not the volume. All listed companies are legally required to file. The question is how they file. A company that provides shareholder meeting notices weeks in advance signals a different management discipline than one that meets the legal minimum by days. A company that files corrections repeatedly reveals a different internal process than one that gets it right the first time.
In the Korean market, transparency risk is not primarily about data absence. DART provides extensive public filings. The risk lies in structural opacity: information is fragmented across filings in ways that obscure cumulative patterns. A series of individually unremarkable related-party transactions, each disclosed in a separate filing, may constitute a systematic pattern of value extraction that no single filing reveals. The T-axis captures the structural accessibility of information, not merely its existence.
Is decision-making power distributed — or concentrated in the hands of a few?
The B-axis evaluates whether governance mechanisms that check management power are functioning in practice, not merely present on paper. Most Korean listed companies have audit committees, outside directors, and shareholder meetings. The institutions exist. The question is whether they operate. A board that has never recorded a dissenting vote in five years receives the same formal compliance score as one with active deliberation.
Balance carries the highest weight (40%) because Korea's governance failures most frequently manifest as power concentration. The pattern is well-documented: boards that never dissent, audit committees staffed by affiliates of the controlling shareholder, nomination processes that produce predetermined outcomes. These are not violations of law — they are legal structures that concentrate decision-making power in ways that make minority shareholders irrelevant. The B-axis measures whether checks and balances are substantive or ceremonial.
Are shareholders exposed to structural conflicts of interest that erode value over time?
The R-axis tracks structural exposure to conflicts of interest and capital leakage — what the academic literature terms "tunneling." We measure the channels through which value can be extracted: related-party transactions conducted at asymmetric terms, capital dilution through convertible instruments, treasury stock dispositions to allied parties, and historical patterns of embezzlement or breach of fiduciary duty.
Each of these mechanisms is individually legal. Collectively, when they recur across multiple fiscal years, they constitute a systematic pattern of value transfer from minority shareholders to controlling stakeholders. These risks are latent under normal conditions but catastrophic when triggered. A firm that has been quietly building related-party lending exposure for three years may show no distress in current financials. The R-axis captures the structural preconditions for value destruction — the loaded gun, not the shot.
Company B scores higher on a single-number scale. Yet its governance structure — high transparency masking concentrated power — creates a fragility invisible to conventional screens. The gap between what a company tells you and what it lets you do is where risk concentrates. This is why the Apex G-Score separates what other ratings combine. The tension between axes is the diagnostic signal.
Beyond the composite score, each firm is classified into one of five governance archetypes based on the interaction of its axis-level scores. The classification captures qualitatively distinct governance profiles that a single composite score would obscure.
All input data is sourced exclusively from DART (Data Analysis, Retrieval and Transfer System) — Korea's electronic disclosure platform operated by the Financial Supervisory Service. The data pipeline ingests annual business reports, corporate governance reports, external audit opinions, shareholder meeting records, and real-time event disclosures.
No surveys. No self-assessments. No management interviews. By restricting inputs to documents that carry legal liability for misrepresentation, the model eliminates self-reporting bias at the architectural level. Every data point used in the G-Score can be independently verified by any market participant with access to DART — which is free and publicly available.
What is proprietary is not the data. It is how the data is weighted, scored, and combined.
The model is built. Here's what it found.
See the Evidence →