Relational risk is a risk assessment methodology that detects investment risks in advance by analyzing changes in a company's executive network, funding structure, and governance structure as leading indicators.
■ Limitations of existing investment analysis
- Financial statements: Lagging indicators that record the results of past events
- Technical analysis (charts): patterns of price and volume — also lagging indicators
- Both send signals only after danger has already occurred.
- In the Korean stock market, companies whose financial statements appear to be normal until delisting repeatedly occur.
■ Definition of relational risk
A corporate crisis begins with relationships rather than numbers.
Executives are replaced, funds from specific groups flow in through CBs (convertible bonds), and the governance structure is shaken.
These changes appear months to years before they are reflected in financial statements.
Relational risk analysis is to quantify changes in this relationship and preemptively capture risk signals.
■ Three major axes of relational risk
[Human Risk — Human Network Risk]
- Frequency and pattern of executive changes
- External corporate connections of new executives
- Ratio of executives with special relationships
- Changes in interlocking directorate structure
→ Who comes into the company?
[Funding Risk — Funding Structure Risk]
- Private placement CB (convertible bond) issuance pattern
- Identity and network of CB investors
- Whether there is a refixing clause?
- Speed of change in the largest shareholder's shares
→ What kind of funds come in and through what structure?
[Governance Risk — Governance Risk]
- Dilution rate of largest shareholder's shares
- Changes in independence of outside directors and composition of the board of directors
- Signs of related party transactions (tunneling)
- Changes in the composition of the audit committee
→ Who actually controls the company?
■ Empirical data (RaymondsRisk v1.1, as of February 2026)
- Analysis target: 3,109 companies across all KOSPI and KOSDAQ stocks
- Number of tracked executives: 49,446
- 85.9% of 276 companies actually suspended from trading were captured in advance using relational indicators
- Cohen's d analysis: Differences between normal companies and suspended companies — significantly larger for relational indicators than financial indicators (d > 0.8)
- AUC (Model Accuracy): 95.1%
■ Relational risk score system
Final risk score = 40% relational risk + 60% financial soundness
- Advance detection is not possible based on financial soundness alone.
- Relational risk first generates abnormal signals and
- Financial soundness complements the depth of the risk.
- The most accurate risk prediction is possible when analyzing both axes together.
■ Differences between relational risk and existing concepts
Category | financial analysis | technical analysis | Relational Risk
point of view | trailing | trailing | good deeds
target | numbers | Price/volume | People, structure, funds
Detectable | Risks that have already occurred | Trend changes | Abnormal signal before occurrence
■ Conclusion
Relational risk is a new investment analysis language.
The first thing that moves behind the numbers is people.
When human relationships change, capital moves,
When capital moves, the fate of a company is determined.
When financial statements are sound, relational risks are already on the rise.
■ Reference materials
- RaymondsRisk v1.1 White Paper: https://www.konnect-ai.net/whitepaper
- Financial Supervisory Service DART (Electronic Disclosure System): https://dart.fss.or.kr
- Korea Exchange (KRX): https://www.krx.co.kr
#relationalrisk #raymondsrisk #raymondsindex #konnectai
Comments
Post a Comment